Ovarian cancer is the leading cause of death from gynecologic cancer, but it occurs less often than endometrial cancers.
Ovarian cancers in women older than 50 years are diagnosed at a more advanced stage, leading to a worse prognosis than for younger women.
The risk of ovarian cancer is decreased by oral contraceptive use. Tubal ligation and hysterectomy also appear to decrease the risk.
The normal postmenopausal ovary is approximately 1.5 to 2 cm in diameter.
A cystic adnexal mass smaller than 8 cm in diameter in a menstruating female is most commonly functional.
A vaginal ultrasound finding of a unilocular cyst of 5 cm or smaller in a perimenopausal woman can usually be followed without surgical intervention.
Ovarian cancer risk rises from approximately 1.4% in general to 5% to 7% if the woman has one or two first- or second-degree relatives with ovarian cancer.
There is no known effective screening for ovarian cancer.
Risk-reducing surgery is an option for women at high risk for ovarian cancer.
Patients with ovarian cancer are at increased risk of developing breast cancer and endometrial cancer.
Genetic testing and tissue-based testing are recommended for all patients diagnosed with epithelial ovarian cancer. Approximately 15% to 18% of patients with high-grade serous carcinoma will have a germline BRCA mutation. In addition, of those who test negative for a germline BRCA mutation, 5% to 7% will be positive for a somatic BRCA mutation on tissue-based testing.
It is important that the follow-up of ovarian cancer patients includes monitoring for breast cancer.
Epithelial tumors are the most common ovarian neoplasm. They account for two-thirds of all ovarian neoplasms and 85% of ovarian cancers.
The major ovarian epithelial ovarian cancer subtypes are high-grade serous (the most common), mucinous, clear cell, endometrioid, and low-grade serous.
The risk of an ovarian tumor being malignant is approximately 33% in a woman older than 45, whereas it is less than 1 in 15 for those aged 20 to 45. More than 50% of ovarian cancers occur in women older than 50.
Most ovarian carcinomas start from small microscopic foci and spread throughout the peritoneum before becoming clinically evident (de novo origin), especially serous and poorly differentiated tumors.
Ovarian carcinomas having a cystic origin are primary mucinous or endometrioid and are more likely to be discovered at an early stage.
Most ovarian carcinomas are diagnosed in stage III or IV.
The primary distribution spread of epithelial carcinoma is transcoelomic to the visceral and parietal peritoneum, diaphragm, and retroperitoneal nodes.
The risk of retroperitoneal node spread of epithelial carcinoma in apparent stage I cases is greatest for poorly differentiated tumors, for which the risk can reach 10% to 20%. The risk of retroperitoneal node spread increases in higher-stage cases.
The prognosis of a patient with ovarian epithelial carcinoma is related primarily to tumor stage and tumor grade and to the amount of residual tumor remaining after primary resection.
Laparoscopic or robotic staging of early ovarian cancers appears to be feasible without compromising survival, based on retrospective data.
For women with apparent metastatic ovarian cancer, diagnostic laparoscopy can be used to determine those patients who benefit most from primary cytoreductive surgery versus neoadjuvant chemotherapy followed by interval cytoreductive surgery.
Optimal surgical debulking (R0-microscopic residual) appears to confer a survival advantage in cases of stages III and IV ovarian carcinoma.
Systematic pelvic and paraaortic lymph node dissection in advanced ovarian cancer is not beneficial based on randomized clinical trials.
Interval cytoreduction has little additional effect on overall survival if a maximal attempt is made at primary surgery.
Neoadjuvant chemotherapy can reduce surgical morbidity; randomized trials have indicated that this strategy may be equivalent to standard treatment for advanced-stage patients with surgery followed by chemotherapy.
Minimally invasive surgery for interval cytoreductive surgery should not be performed outside a clinical trial.
Hyperthermic intraperitoneal chemotherapy (HIPEC) should not be recommended outside a clinical trial for primary or recurrence treatment.
The addition of HIPEC to interval debulking surgery is associated with longer recurrence-free and overall survival compared with surgery plus conventional chemotherapy.
Assessing the ovarian CA-125 level is useful to help monitor patients with ovarian carcinoma. Reaction to the antigen is positive in approximately 80% of cases.
A rapid decrease in CA-125 values after treatment indicates a more favorable prognosis.
More than 80% of patients with epithelial ovarian cancer will be disease free at the completion of primary chemotherapy. Initial treatment usually includes a combination of platinum and taxane agents.
PARP inhibitors have demonstrated efficacy in epithelial ovarian cancer patients in both the treatment and maintenance settings.
Recurrent ovarian cancer is difficult to cure.
Factors determining response to recurrent chemotherapy regimens include time to treatment progression, distribution and volume of disease, and performance status.
Bevacizumab tends to augment the activity of platinum-based chemotherapy in the frontline and recurrent (both platinum-sensitive and platinum-resistant) settings.
Secondary cytoreductive surgery may be of benefit only in select patients with prolonged disease-free interval, single site of disease, no evidence of carcinomatosis, and optimal cytoreduction at initial primary surgery.
The 5-year survival rate for patients with borderline epithelial ovarian carcinoma (grade 0) is close to 100% for stage I and more than 90% for all stages.
Germ cell tumors are the second most common type of ovarian neoplasms and account for approximately 20% to 25% of all ovarian tumors.
In women younger than 30 years, the most common ovarian neoplasm is a germ cell tumor; approximately one-third of these germ cell tumors are malignant in those younger than 21. For women younger than 30 years, the most common ovarian neoplasm is the dermoid.
The most common germ cell tumor is the benign cystic teratoma (dermoid). It is bilateral in 10% to 15% of the cases. Approximately 30% are calcified.
Malignant germ cell tumors are usually unilateral except dysgerminomas, which are bilateral in approximately 10% to 15% of patients.
Dysgerminomas are the most common malignant germ cell tumors and account for 1% to 2% of ovarian cancers.
The prognosis for a patient with an immature teratoma is related to tumor grade and tumor stage. These tumors are the second most common type of malignant germ cell tumor.
The 5-year survival rate of stage IA pure dysgerminoma treated by unilateral salpingo-oophorectomy is more than 90%.
Pure dysgerminomas are radiocurable; however, multiagent chemotherapy, particularly with etoposide and platinum, with or without bleomycin, will often result in complete remission. Approximately two-thirds of cases present as stage IA.
Most patients with malignant ovarian germ cell tumors can be treated successfully with fertility-sparing surgery followed by bleomycin, etoposide, and cisplatin (BEP) chemotherapy. Patients who do not require postoperative chemotherapy include those with stage IA dysgerminoma and stage IA, grade 1, immature teratoma; however, there has been a trend toward surveillance rather than chemotherapy for patients with stage I tumors of any histologic subtype.
Multiagent chemotherapy has improved survival in patients with malignant germ cell tumors, preserving childbearing function in most cases. Standard chemotherapy consists of the BEP regimen.
Gonadoblastomas are sex cord–stromal germ cell tumors that usually arise in dysgenetic gonads in patients with a Y chromosome; these are cured by removal.
Granulosa cell tumors and Sertoli-Leydig tumors usually behave as low-grade malignancies, but there may be late recurrences.
For patients with primary metastatic or recurrent sex cord–stromal tumors of the ovary, platinum-based chemotherapy is the treatment of choice. Commonly used regimens include BEP and paclitaxel-carboplatin.
Some metastatic granulosa cell tumors may respond to hormone therapy, such as leuprolide acetate, tamoxifen, or aromatase inhibitors.
Fibroma is the most common benign solid ovarian tumor.
The most common sites of origin of tumors metastatic to the ovary are the lower reproductive tract, gastrointestinal tract, and breast.
There is increasing evidence that many cases of ovarian and peritoneal carcinoma may actually arise from the fallopian tube, thereby underestimating the incidence of fallopian tube carcinoma.
Fallopian tube and peritoneal cancers have similar clinical characteristics, patterns of spread, and response to treatment compared with ovarian cancer.
The primary risk factor for fallopian tube and peritoneal cancer is an inherited mutation in the BRCA1 or BRCA2 tumor suppressor gene.
The most common histologic subtype of fallopian tube and peritoneal carcinoma is high-grade serous carcinoma.
The treatment of fallopian tube and peritoneal cancer is identical to that for ovarian cancer and typically includes a combination of surgery and chemotherapy.
Ovarian cancer is the second most common malignancy of the lower part of the female genital tract, occurring less often than cancers of the endometrium but more often than cancers of the cervix; however, it is the most common cause of death from gynecologic neoplasms in the United States. Cancer Statistics 2019 has reported that approximately 22,530 new cases of ovarian cancer will be diagnosed yearly in the United States, and there will be 13,980 deaths ( ). A major contributing factor to the high death rate from the relatively few cases stems from the frequent detection of the disease after metastatic spread, when symptoms direct clinical investigation or raise clinical concern. Surprisingly, most women diagnosed with ovarian cancer do report symptoms for months before diagnosis. As detailed later, only the severity and duration of symptoms differentially segregate cancer patients from noncancer patients. The incidence of ovarian cancer ( Fig. 33.1 ) increases with age, becoming most marked beyond 50 years, with a gradual increase continuing to age 70 years followed by a decrease for those older than 80. Previous studies have noted that those older than 65 are more likely to have their cancers diagnosed at an advanced stage, leading to a worse prognosis and poorer survival compared with those younger than 65 years. Although much of the demographic information regarding this disease has been static over the past several decades, a remarkable trend of increasing prevalence has been noted. Surveillance, Epidemiology, and End Results (SEER) estimates of disease prevalence for 2015 suggest more than 225,000 women are currently alive with ovarian cancer in the United States ( ). Successful incorporation of new treatment and specialized care are believed to be driving this metric.
Despite numerous epidemiologic investigations, a clear-cut cause of ovarian cancer has not been defined. A number of theories, particularly those associated with ovulation (e.g., parity, breastfeeding, hormonal contraception, infertility), have been previously supported. Because certain forms of hormonal manipulation of ovulation have been associated with a substantially reduced lifetime risk of ovarian cancer, even among those carrying a germline mutation in BRCA1 or BRCA2, an ovarian epithelial origin has been suggested; however, salpingectomy and tubal ligation are also associated with a reduced risk of ovarian cancer, and direct transitions from benign to malignant states have been difficult to demonstrate in ovarian surface epithelia and inclusion cysts. More recently, focus has been placed on transitional states in the fallopian tube, particularly for the most common histologic type, high-grade serous ovarian cancer (HGSOC). Next-generation sequencing of primary HGSOCs reported in the Cancer Genome Atlas (TCGA) demonstrate that nearly all these tumors have a mutation in TP53. Similar findings were reported in detailed analyses of serous tubal intraepithelial carcinomas (STICs) found in the normal-appearing fallopian tubes of unaffected women undergoing risk-reducing salpingo-oophorectomy. In these studies microinvasive tubal lesions were also identified, providing a serial link of an acquired p53 signature in normal noninvasive tissue to transformed cells migrating to the ovarian surface and peritoneal cavity with invasion. In further investigation from Labidi-Galy and colleagues, they performed whole-exome sequencing and copy number analyses of laser-captured microdissected STICs, invasive fallopian tube and ovarian cancers, and metastases from nine patients. Interestingly, tumor-specific alterations in TP53, PTEN, and BRCA1/2 were present in most of the samples, and functional evaluation of the sample supported the precursor status of these lesions and the development of invasive disease. It was estimated from this work that an evolutionary window of about 7 years separated the events. Even more recently, though, primary ovarian epithelial cells were found to produce STIC-like lesions in the ovary primarily under controlled gene expression ( ). This field continues to evolve as multiomics investigations continue to describe ovarian carcinogenesis. Other traditional risk factors, including smoking, benign ovarian diseases such as polycystic ovarian syndrome, pelvic inflammatory disease, and endometriosis, continue to provide clues to carcinogenesis, including the most strongly linked, family or personal history of ovarian/breast cancer or known germline mutation of BRCA1/2. Table 33.1 lists the various epidemiologic factors that are thought to alter ovarian cancer risk.
|Industrialized country||Tubal ligation and hysterectomy, with ovarian conservation|
Familial ovarian cancer
In a case-control study, Hartge and colleagues showed that a familial history of breast cancer and a personal history of breast cancer are ovarian cancer risk factors. Lynch and associates reported on families with these hereditary ovarian cancers and noted that they tend to occur at a younger age than in the general population. It appears that germline mutations of the BRCA tumor suppressor gene on chromosome 17q are responsible for a large proportion of hereditary cancers (discussed later); however, these are a small proportion of all ovarian carcinomas. Risk alteration in these patients through oral contraceptive use is of uncertain impact. Narod and coworkers suggested that it might be possible to reduce incident risk by their administration; however, Modan and colleagues conducted a case-control study of Jewish women in whom BRCA founder mutational analysis was performed; they evaluated the risk of cancer development based by parity and oral contraceptive use. They were able to establish a protective effect by oral contraceptive use in the cohort, but subanalysis by carrier status demonstrated no effect in those harboring a BRCA founder mutation (odds ratio [OR], 1.07; 95% confidence interval [CI], 0.63 to 1.83). Further studies are needed.
Hereditary ovarian cancers are uncommon, accounting for approximately 10% to 15% of all incident cases; however, identification of affected or unaffected women with significant familial risk is important, given their accelerated risk of ovarian and other cancers. In addition, these patients are commonly diagnosed at a younger age (median, 50 years), and unaffected individuals are able to consider prophylactic procedures that can affect their lifetime risk. The term familial ovarian cancer denotes an inherited trait that predisposes to ovarian cancer development. It has been widely studied, and two definitions are important:
A first-degree relative is a mother, sister, or daughter of an affected individual.
A second-degree relative is a maternal or paternal aunt or grandmother.
As noted in the review by Kerlikowske and colleagues, previous studies suggested an increase from approximately 1.5% to 5% in the lifetime risk of ovarian cancer with one first-degree relative; with two or more, the risk reaches may exceed 7% to 12%. Because contemporary family size may preclude disease penetrance, careful attention to BRCA -associated cancers in either gender should be performed in a multigenerational history. Furthermore, the high association of fallopian tube cancer and underlying BRCA mutation has prompted some to call for routine BRCA testing in all such cases.
Most ovarian cancers develop sporadically. For the woman with a familial history of ovarian cancer (not the dominant genetic hereditary type), periodic surveillance with transvaginal ultrasonography 6 months after the age of 35 years has been suggested (see the ultrasonography discussion presented later in this chapter). Unfortunately, such a strategy has not been shown to be worthwhile or cost effective in disease prevention and may, on occasion, lead to additional tests or unnecessary procedures when a questionable ultrasound result is. The use of prophylactic oophorectomy in patients whose mothers had ovarian cancer has been a controversial topic. Kerlikowske and colleagues and Herbst have provided reasons opposing the widespread use of this practice; however, there is a clear benefit to the use of risk-reducing salpingo-oophorectomy in the population with mutations in BRCA1/2. In a meta-analysis performed by Rebbeck and colleagues, this surgery was found to yield a 79% reduction in the risk of ovarian or fallopian tube cancer among patients with BRCA1 or BRCA2 mutations (hazard ratio [HR], 0.21; 95% CI, 0.12 to 0.39). Importantly, there was also a significant reduction in the risk of breast cancer with this procedure (HR, 0.49; 95% CI, 0.37 to 0.65). The utility of risk-reducing surgery has been validated in several prospective studies, confirming the reduction in breast and ovarian cancer risk, as well as all-cause mortality, breast cancer–specific mortality, and ovarian cancer–specific mortality ( ; ). For patients with a significant family history in whom an operation such as hysterectomy is required, consideration of removal of both tubes and ovaries at the time of operation is appropriate. Similarly, mutation carriers who have finished childbearing may reduce their subsequent cancer risk by undergoing salpingo-oophorectomy. The recommendation for hysterectomy at this time is controversial but has been advocated by some to ensure complete removal of the fallopian tube (cornual segment). The woman must be aware that peritoneal carcinomatosis, a process resembling serous carcinoma of the ovary, can develop (rarely) despite the removal of both ovaries.
Because evidence has suggested a significant proportion of ovarian carcinomas arise from preinvasive intraepithelial carcinomas in the fallopian tube, interest in a two-step approach to removal of the adnexa has arisen. Preserving ovarian function in high-risk women while mitigating risk via removal of the fallopian tubes has appeal for maintaining quality of life and avoidance of the risks from premature castration. Nevertheless, much is unknown regarding the efficacy and safety of this approach, including the strength of the risk mitigation and the degree to which retaining the ovary increases the risk of breast cancer in these at-risk women. Finally, the timing for oophorectomy is unknown. Fortunately, several prospective clinical trials are under way to address this novel approach to risk reduction surgery.
The following section describes the classification and histologic characteristics of the major ovarian neoplasms. Pertinent microscopic findings, clinical behavior, and appropriate therapy are presented.
Classification of ovarian neoplasms
The most widely used classification of ovarian neoplasms is that of the World Health Organization (WHO), which was last updated in 2014 ( ). This modified classification is shown in Table 33.2 .
|Class||Approximate Frequency (%)|
|Epithelial stromal (common epithelial) tumors||65|
|Germ cell tumors||20-25|
|Sex cord–stromal tumors||6|
|Lipid (lipoid) cell tumors||<0.1|
|Soft tissue tumors (not specific to ovary)|
|Secondary (metastatic) tumors|
|Tumorlike conditions (not true neoplasm)|
Epithelial tumors (common epithelial tumors) are the most common ovarian neoplasms. They are believed to arise from the surface (coelomic) epithelium. Germ cell tumors are the second most common and are the most common among young women. Histologically they may be composed of extraembryonic elements or may have features that resemble any or all of the three embryonic layers (ectoderm, mesoderm, or endoderm). Germ cell tumors are the main cause of ovarian malignancy in young women, particularly those in their teens and early 20s. Sex cord–stromal tumors are the third most common and contain elements that recapitulate the constituents of the ovary or testis. These tumors may secrete sex steroid hormones or may be hormonally inactive.
Epithelial ovarian neoplasms
According to , two-thirds of ovarian neoplasms are epithelial tumors; malignant epithelial tumors account for approximately 85% of ovarian cancers. Table 33.3 summarizes the five cell types that most commonly constitute epithelial ovarian tumors, indicating their relative frequency.
|Sub Type||Mutations||Clinical Prognosis||Frequency|
|High-grade serous||TP53, BRCA1, BRCA2, CDK12||Often diagnosed at late stage and chromosomally unstable.||∽65%|
|Low-grade serous||BRAF, KRAS, NRAS, ERBB2||Often diagnosed in younger patients, less aggressive, gnomically stable.||∽5%|
|Endometrioid||PTEN, CTNNB1, PPP2R1a, MMR deficient||Favorable prognosis and response to chemotherapy.||∽20%|
|Clear cell carcinoma||PIK3CA, KRAS, PTEN, ARID1A||Low response to chemotherapy and intermediate prognosis.||∽5%|
|Mucinous||KRAS, HER-2 amplification||Low response to chemotherapy.||∽5%|
Epithelial tumors can be categorized as benign (adenoma), malignant (adenocarcinoma), or of an intermediate form, known as borderline tumor or tumor of low malignant potential. The thinking regarding their origin has undergone considerable modification and debate ( ; ; ; ).
Serous tumors are the most common ovarian epithelial tumors. There are two types of serous borderline tumors—atypical proliferative serous tumor and the micropapillary variant (noninvasive low-grade serous carcinoma) ( Fig. 33.2 , A and B ). The current classification for invasive serous carcinoma divides them into low-grade serous and high-grade serous subtypes based on microscopic criteria (nuclear atypia and mitotic count) and molecular signatures ( Fig. 33.2 , C and D ). Although the origin of low-grade serous carcinoma is unclear, various theories suggest that they may arise from serous borderline tumors, ovarian surface epithelium, or endometriosis. The mitogen-activated protein (MAP) kinase pathway appears to play a prominent role in their pathogenesis. On the other hand, high-grade serous carcinomas have a p53 signature, and an undetermined proportion are thought to arise from the fimbriated end of the fallopian tube.
Mucinous tumors ( Fig. 33.3 , A and B ) consist of epithelial cells filled with mucin; most are benign. These cells resemble cells of the endocervix or may mimic intestinal cells, which can pose a problem in the differential diagnosis of tumors that appear to originate from the ovary or intestine. Benign mucinous tumors are found primarily during the reproductive years, and mucinous carcinomas ( Fig. 33.3 , C ) usually occur in those in the 30- to 60-year age range. Overall they can account for approximately 25% of ovarian tumors and as many as 10% of ovarian cancers. Molecular studies have revealed KRAS mutations in approximately 30% to 40% of mucinous carcinomas and HER-2 amplification in 18% to 20%.
Endometrioid tumors ( Fig. 33.4 ), as the name implies, consist of epithelial cells resembling those of the endometrium. In the ovary these neoplasms are less common (approximately 5% to 10%) than serous or mucinous tumors, but the malignant variety accounts for approximately 20% of ovarian carcinomas. Endometrioid carcinomas usually occur in women in their 40s and 50s. They may be seen in conjunction with endometriosis and ovarian endometriomas and are most commonly diagnosed in the early stages. ARID1A mutations occur in 30%, and PTEN mutations are also common.
Clear cell tumors contain cells with abundant glycogen ( Fig. 33.5 , A ) and so-called hobnail cells ( Fig. 33.5 , B ), in which the nuclei of the cells protrude into the glandular lumen. Tumors with identical histologic features are found in the endometrium, cervix, and vagina, the latter two often associated with intrauterine diethylstilbestrol (DES) exposure. Molecular evaluation of these tumors suggests a homology to similar pathologic changes occurring in the kidney, which may have therapeutic implications. Clear cell ovarian tumors account for approximately 5% of ovarian cancers, and most are unilateral and early stage. They occur primarily in women aged 40 to 70 years and are highly aggressive. ARID1A mutations occur in 50%, and hepatocyte nuclear factor 1beta is upregulated.
The major cell types of ovarian epithelial tumors recapitulate the müllerian duct–derived epithelium of the female reproductive system (serous-endosalpinx, mucinous-endocervix, endometrioid-endometrium). This differentiation occurs even though the ovary is not derived directly from the müllerian ducts (see Chapter 2 ). The clear cell tumors also mimic this müllerian tendency, commonly being admixed with endometrioid carcinomas and with ovarian endometriomas.
Brenner tumors ( Fig. 33.6 ) consist of cells that resemble the transitional epithelium of the bladder and Walthard nests of the ovary. There is abundant stroma. These tumors constitute only 2% to 3% of all ovarian tumors.
In addition to the cell types shown in Table 33.3 , epithelial tumors may be classified as undifferentiated if the tumor consists of poorly differentiated epithelial cells that are not characteristic of any particular cell type. They may be considered unclassifiable if they cannot be placed in any of the categories shown in this table.
Many epithelial ovarian tumors can be bilateral, and the risk of bilaterality is an important consideration in therapy, particularly when an ovarian tumor is discovered in a young woman of reproductive age. Widely varying percentages have been reported for bilaterality in ovarian tumors; the most widely quoted are summarized in Table 33.4 . Compared with benign epithelial tumors, malignant epithelial tumors more often tend to involve both ovaries. Serous tumors also tend to be bilateral more commonly than mucinous tumors.
|Type of Tumor||Occurrence (%)|
|E PITHELIAL T UMORS|
|Benign Brenner tumor||6|
|G ERM C ELL T UMORS|
|Benign cystic teratoma (dermoid)||12|
|Immature teratoma (malignant)||2-5|
|Other malignant germ cell tumors||Rare|
|S EX C ORD –S TROMAL T UMORS|
|Sertoli–Leydig cell tumor||Rare|
|Granulosa-theca cell tumor||Rare|
Adnexal mass and ovarian cancer
CA-125 was described by Bast and colleagues in the 1980s. It is expressed by approximately 80% of ovarian epithelial carcinomas but less often by mucinous tumors. The marker is increased in endometrial and tubal carcinoma, in addition to ovarian carcinoma, and in other malignancies, including those originating in the lung, breast, and pancreas ( Box 33.1 ). A level higher than 35 U/mL is generally considered to be increased. Box 33.2 lists some of the benign conditions for which the CA-125 level also has often been found to be increased.
Peritoneal inflammation, including pelvic inflammatory disease
Hemorrhagic ovarian cysts
As can be seen, many of these are commonly found in women of childbearing age. This lack of specificity must be remembered when one is interpreting increased CA-125 levels in younger women with adnexal masses or when screening is being considered (discussed later). In addition, there are rare individuals who have no disease but are found to have levels of CA-125 as high as 200 to 300 U/mL as a consequence of developing idiopathic antibodies to mouse immunoglobulin (Ig) G.
One must also be cautious in the interpretation of an increased CA-125 level, particularly in a premenopausal woman with an adnexal mass. The specificity appears to be better for increased values in the postmenopausal woman. In a study of 3511 patients, Van Calster and colleagues noted that among patients with benign masses, CA-125 level was increased in a higher proportion of premenopausal patients compared with postmenopausal patients ( ). Conversely, in the setting of malignancy, postmenopausal patients had higher levels of CA-125. Similar findings have been observed in a number of different prospective studies best described in the meta-analysis by .
In addition to CA-125 and transvaginal ultrasound (TVUS), other biomarkers have been evaluated for their ability to preoperatively discriminate between benign disease and cancer. One test, OVA1, has been approved to aid this decision by providing a probability estimate of cancer based on a proprietary mathematical algorithm of five independent biomarkers: CA-125, beta 2 -microglobulin, transferrin, apolipoprotein A1, and transthyretin. OVA1 was the first blood test cleared by the U.S. Food and Drug Administration (FDA) to help evaluate the likelihood that a woman’s ovarian mass is malignant or benign before a planned surgery. The OVA1 test, when performed in 516 women with adnexal masses deemed appropriate for surgical excision, improved the sensitivity of cancer versus noncancer discrimination in a double-blind clinical study from 72% to 92% when using the biomarker panel. When administered by gynecologic oncologists, the sensitivity increased from 86% to 99%. Although not approved for surveillance or diagnosis, the test may complement clinical decision making, particularly if a gynecologic oncologist is not available to perform appropriate staging should cancer be identified.
Evaluation of the adnexal mass
Ultrasound has helped define criteria to allow conservative follow-up and the risk of malignancy of some adnexal masses. Goldstein and associates studied 42 postmenopausal patients whose ultrasound scans showed unilocular cysts smaller than 5 cm in diameter; 28 were explored, none had malignancy, and 14 were followed for as long as 6 years with no change in ultrasound appearance. Finkler and colleagues noted that the addition of a CA-125 serum assay to their ultrasound criteria in postmenopausal women increases the accuracy of preoperative evaluation. In a study to evaluate the performance of the Society of Radiologists in Ultrasound consensus guidelines for risk stratification of adnexal masses, Maturen and colleagues evaluated adnexal cysts in 500 women. Of 219 with simple unilocular cysts, only one was malignant. In contrast, malignancy rates were 12.5% (12 of 84) for complex cysts and all four predominantly solid masses were malignant. Among the 132 cysts in postmenopausal women, 14 (10.6%) were malignant. Conversely, among 438 cysts in premenopausal women, 4 (0.9%) were malignant. In an ultrasound study of cystic ovarian masses in women older than 50 years, Bailey and coworkers noted that unilocular cysts smaller than 10 cm in diameter are rarely malignant, whereas complex cysts or those with solid areas are at high risk of malignancy.
Several scoring systems have been proposed to try to determine the risk of an ovarian mass being malignant. They usually include the following:
Is the finding a simple (unilocular) or complex (multicystic or multilocular with solid components) cyst?
Are there papillary projections?
Are the cystic walls or septa regular and smooth?
What is the echogenicity (tissue characterization)?
These terms, definitions, and measurements have been standardized under a consensus opinion from the International Ovarian Tumor Analysis (IOTA) group and help refine the likelihood of malignancy. Shalev and coworkers combined transvaginal ultrasonography and normal CA-125 values in 55 postmenopausal women with simple cystic or septate cystic ovarian masses; all 55 had benign disease. Although this was a small study, it suggests the potential of applying stringent ultrasound criteria with CA-125 evaluation of ovarian masses in postmenopausal women.
Others have advocated using transvaginal pulsed Doppler color-enhanced flow studies to differentiate benign from malignant masses. The resistance index, which measures resistance to flow in the vessels, has been used and presumably is low in the presence of the neovascularization that is seen with malignant tumors. The vessels of neoangiogenesis are abnormal in their distribution, with disorganized branching and a loss of the muscularis layer, all of which contribute to the decreased resistance to flow. A resistance of 0.40 or less was found useful by Kurjak and coworkers in a study of 254 women. In contrast, Bromley and colleagues, in a study of 33 postmenopausal women, used a cutoff of 0.6, which did not greatly add to their specificities; they relied on morphologic criteria (e.g., solid elements, papillary projections) to diagnose malignancy.
Valentin and associates evaluated the characteristics of 1066 adnexal masses; 266 were malignant (55 borderline ovarian tumors, 144 primary invasive epithelial cancers, 25 nonepithelial ovarian cancers, 42 metastatic cancers). A scoring system was used as well as information from color Doppler studies. They reported that borderline and stage I ovarian cancers shared similar morphology but had different characteristics from more advanced-stage tumors. They were larger, contained more papillary projections, and were more often multilocular, without solid components, but were less often purely solid and less likely to be associated with ascites. Significant variation was noted, however. Similarly, Twickler and colleagues described a scoring model to create an ovarian tumor index for women with adnexal disease. Of 244 women with follow-up, 214 had nonmalignant findings and 30 had cancer. In addition to age, TVUS variables, including ovarian volume, the Sassone morphology scale, and Doppler determination of angle-corrected systole, diastole, and time-averaged velocity, were evaluated. An ovarian tumor index was created from discriminant variables (continuous and weighted) correctly classifying the two cohorts. The area under the receiver operator characteristic curve (AUC) was highly significant (AUC, 0.91). Unfortunately, scoring systems such as these, developed from data produced by highly skilled and proficient sonographers, are difficult to generalize and, although promising, are highly operator dependent. The IOTA group has temporally and externally validated the diagnostic performance of two logistic regression models containing clinical and ultrasound variables for malignancy identification. In this study the prevalence of invasive cancer was 3%. The likelihood of cancer from a positive screen exceeded 6 for both models; a negative likelihood ratio was less than 0.1, suggesting that the criteria may be of use for evaluating women with an adnexal masses.
It should be noted that there is a difference in using ultrasonography to screen for ovarian cancer compared with using different modalities of ultrasonography to characterize an ovarian mass as benign or malignant. For example, the addition of color Doppler sonography, which measures blood flow and direction of flow, and power Doppler sonography, which can detect slow flow in small vessels, can add useful information. These permit visualization of flow location (peripheral, central, or within a septum). Most malignant tumors have a central flow (75% to 100%) compared with only 5% to 40% of benign ovarian tumors. Schelling and colleagues studied transvaginal B-mode and color Doppler sonography for the diagnosis of malignancy in 257 adnexal masses with unclear malignant status. They achieved 92% sensitivity and 94% specificity. The development of three-dimensional (3D) ultrasonography may allow more accurate volume assessments. In addition, color Doppler with 3D ultrasonography may permit better detection of vessel irregularity, coiling, and branching. Another possibility is the use of contrast media to quantify and permit earlier detection of abnormal angiogenesis, as noted by Abramowicz. Contrast-enhanced (microbubble) power 3D Doppler sonography has been investigated to evaluate the efficacy of antiangiogenic biologic agents in serial scanning.
Ovarian cancer screening
Although ovarian cancer is characterized by advanced-stage disease at diagnosis and high mortality, early-stage disease is often curable. The greatest impact on these statistics, other than prevention, would be screening to identify early-stage disease. Four modalities, used individually or in combination, have been the common theme of this effort: physical examination, biomarkers (e.g., CA-125), proteomics-genomics (experimental), and sonography. For a disease to be amenable to screening, it should be sufficiently severe (high mortality); it should have a natural history that is well characterized from latency to overt disease; and there should be a successful outcome if early disease is treated. The screening modality should have high positive and negative predictive values; have high sensitivity and specificity; and be acceptable to the population, cost-effective, and widely available. The screening population should be identifiable, and effective therapy should be available for those in whom early disease is identified. Although ovarian cancer satisfies many of these mandates, it is rare in the general population (1/2500 postmenopausal women) and not readily characterized by an identifiable precursor, thus producing a high bar for any modality.
Of the three most commonly used modalities, the least sensitive and specific is physical examination. It is estimated that just one early ovarian cancer will be identified in 10,000 physical examinations. Although the easiest to implement, poor sensitivity limits this intervention as an effective strategy.
Biomarkers such as CA-125 are of great interest because they are easy to obtain and serial evaluation can be tracked. CA-125 has been used most consistently since being discovered as a reliable biomarker to monitor epithelial nonmucinous ovarian cancer, particularly during treatment; however, large, population-based studies highlighted its limitation as a sole strategy for ovarian cancer screening. Einhorn and associates screened 5550 women and in 1992 reported that only two stage I cancers in 175 women with elevated CA-125 values were identified. As noted, a differential effect would be expected between pre- and postmenopausal women. Using the modality in women with a pelvic mass (in whom prevalence is increased) has substantial effects on assay performance but overlooks the obvious need for cancer identification before gross ovarian enlargement. This has led to the development of combined evaluation (sonography) as described here.
Ultrasonography as an isolated modality has also been advocated for screening. Although more expensive and less amenable to population-wide screening, it has become increasingly accurate in identifying early changes within the ovary, as noted. Campbell and coworkers screened 5479 patients and obtained 338 abnormal scans. Five early-stage ovarian cancers were identified. The positive predictive value was only 1.5%. Similarly, van Nagell screened 1300 patients and obtained 33 abnormal scans. Two early-stage ovarian cancers were identified. As with single-modality testing, sonography is too insensitive to be widely used for screening.
Population-based ovarian cancer screening programs have been difficult to recommend and implement because poor sensitivity and positive predictive value characteristics accompany expensive and inefficient testing methodology and triage algorithms. Menon and colleagues approached this problem by evaluating a prospectively based algorithm in a population-based screening program in the United Kingdom. The population cohort used to evaluate this screening strategy involved 13,582 menopausal women 50 years or older with at least one ovary, of whom 6532 randomly allocated women completed a first screen; the remainder served as controls. The screening strategy was a staged process in which each CA-125 sample drawn underwent a calculation using the Risk of Ovarian Cancer Algorithm (ROCA) based on the woman’s age and CA-125 value relative to her personal baseline. In this trial an estimated risk of less than 1 in 2000 was considered normal, whereas a risk of more than 1 in 500 was considered increased; those in between were considered intermediate and required repeat testing. Those not considered normal were referred for a second stage of screening that incorporated a TVUS scan and repeat CA-125 testing. A TVUS scan was considered normal, abnormal, or equivocal based on ovarian volume and morphology. From the combination of CA-125 risk estimation and transvaginal ultrasound scan, a follow-up recommendation was made that could be a gynecologic oncology referral, repeat CA-125 testing, or TVUS scan or annual screening. In the screened group, almost 80% continued with annual screening; 91 (1.4%) were considered at increased risk. Among the intermediate group, repeat testing was normal in 92%, leaving 188 (2.9% of the initial population) to undergo second-stage evaluation. Of the 144 who stayed in the program, 95 were returned to annual screening based on CA-125 and TVUS scan findings; 6 were found to have nongynecologic malignancies; and 43 were referred to a gynecologic oncologist, of whom 27 women were returned to annual screening and 16 underwent surgery. From this group, 5 cases of ovarian cancer were identified (4 malignant epithelial and 1 borderline); the 11 remaining women had benign ovarian neoplasms. Compared with the authors’ previous algorithm based on flat CA-125 values (normal = 30 U/mL), the new process referred less than 50% to secondary screening. It was concluded that this algorithm increased screening precision.
Two other prospective trials of general population screening deserve mention: the Prostate, Lung, Colorectal and Ovarian (PLCO) study and the UK Collaborative Trial of Ovarian Cancer Screening (UKCTOCS) ( ; ). The primary objective of the PLCO study is to evaluate the impact of annual screening with TVUS and CA-125 on ovarian cancer mortality. The study is prospectively following a cohort of more than 78,000 largely postmenopausal women with intact ovaries using an algorithm indicating that an abnormal CA-125 level (≥35 U/mL) or an abnormality on TVUS is considered a positive screen. Follow-up procedures for a positive screen are not prespecified but have been tracked, along with any surgical interventions resulting from these findings. Depending on when patients joined the study, between four and six rounds of screening could be experienced. Compliance of screening decreased slightly over the 6 years of postbaseline evaluation but remained greater than 75%; in years 5 and 6 TVUS was not administered. CA-125 screen positives ranged from 1.4% to 1.8% during six screening rounds and from 2.9% to 4.6% during the four screening rounds with TVUS. Ovarian cancer was diagnosed in 212 women in the intervention group and 176 in the usual care, control group. The relative risk (RR) for incident disease was 1.21 (95% CI, 0.99 to 1.48). Accordingly, 118 and 100 deaths were recorded in the intervention and usual care groups, respectively, over a median follow-up of 12.4 years. The RR for ovarian cancer mortality was 1.18 (95% CI, 0.82 to 1.71). In addition, of the cases diagnosed, no stage migration was experienced. Collectively, this suggests that annual screening with combined CA-125 and TVUS is insufficiently sensitive to affect ovarian cancer natural history.
The UKCTOCS was designed to definitively assess the effect of ovarian cancer screening on mortality, as well as comprehensively address the cost, acceptance, physical and psychosocial morbidity, and performance characteristics of multimodal screening and ultrasound-based screening. Between 2001 and 2005, a total of 202,638 postmenopausal women 50 to 74 years of age were randomly assigned to control (no screening), annual CA-125 screening (based on the ROCA), and second-line ultrasound testing (multimodal screening [MMS]) or annual TVUS screening (USS) alone in a 2:1:1 ratio. In the prevalence screen, 50,078 women (98.9%) underwent MMS and 48,230 (95.2%) underwent USS. Overall, 9% of the MMS cohort and 12% of the USS cohort required repeat testing. Surgery was undertaken in a small proportion of both cohorts but was significantly more likely after USS. Ovarian neoplasms, benign and malignant, were identified in both screening cohorts. The proportion of stage I and II cases was 48.3% and was balanced between the two screening algorithms; however, specificity was significantly higher for the MMS cohort (99.8%) relative to USS (98.2%). The primary endpoint was reported in 2016 with 50,639 patients randomly allocated to USS and 50,640 patients to MMS (with 101,359 controls). Over a median follow-up of 11 years, 1282 women were diagnosed with ovarian cancer: 338 in the MMS group, 314 in the USS group, and 630 in the control group. The percentage of screened patients dying from ovarian cancer was 0.29%, 0.30%, and 0.34% for the MMS, USS, and control groups, respectively. The reduction in mortality relative to control over years 0 to 14 was 15% (95% CI, −3 to 30; P = .10) with MMS and 11% (−7 to 27; P = .21) with USS. In detailed analysis of time-dependent risk reduction, the strongest effect was late in the screening interval (years 7 to 14). In a sensitivity analysis, removing the prevalent cases of ovarian cancer (those diagnosed in the first year of screening) raised the hypothesis that MMS could, in fact, reduce mortality (averaging 20% over 14 years of screening). These two studies establish the feasibility of screening; determining efficacy will required further work to firmly establish a change in clinical practice ( Table 33.5 ).
|Multimodal Screening||Transvaginal Ultrasound|
|Number of ovarian cancers||42||45|
|Stage I and II cancers (48.3% of cancers identified)||16||12|
|Positive predictive value||35.1%||2.8%|
One strategy to improve the predictive index is to address a population in which prevalence is increased. A number of studies have been undertaken using transvaginal ultrasonography to screen for ovarian malignancy in higher-risk women. Bourne and colleagues screened 775 women who had at least one first-degree (n = 677) or second-degree (n = 98) relative with ovarian cancer. Overall, 43 women were referred for surgery with abnormal-appearing ovaries and 39 underwent surgery, with three stage IA ovarian carcinomas discovered (3.9 of 1000 screened); one of these was a borderline tumor. One screened patient was found to have peritoneal carcinomatosis 11 months after a normal screening study. The remainder had nonmalignant findings. The UK Familial Ovarian Cancer Screening Study (UKCTOCS) evaluated the Bayesian ROCA every 4 months with annual TVUS, if normal; if ROCA values were elevated, TVUS was to be performed within 2 months. A total of 4348 women were enrolled and followed for a median 4.8 years; 19 cases of ovarian or fallopian tube cancer were diagnosed, with 10 (52.6%) at stage I/II. Mortality data are immature. Similarly, Skates and colleagues evaluated ROCA screening with every 3 months of CA-125 testing in 3692 high-risk women. At the time of the report approximately half of incident cancers were stage I/II.
Given the potential value of an effective screening algorithm, investigation into more sensitive modalities continue. For instance, based on a deeper understanding of tissue and bloodborne factors associated with ovarian cancer, molecular-based and enhanced standard imaging are entering the clinical domain. In addition, multiomic assays, such as proteomics, microRNAs, circulating tumor DNA, exosomes, and TP53 autoantibodies, are under investigation ( ).
Diagnosis, staging, spread, preoperative evaluation, and prognostic factors
Ovarian carcinomas are usually diagnosed by detection of an adnexal mass on pelvic or abdominal examination. Occasionally the diagnosis is made from a radiographic survey carried out for the evaluation of nonspecific gastrointestinal symptoms. Unfortunately, the diagnosis is commonly made only after the disease has spread beyond the ovary, as noted earlier when we described the de novo origin of these tumors. Scully has estimated that the risk of malignancy in a primary ovarian tumor increases to approximately 33% in a woman older than 45, whereas it is less than 1 in 15 for women aged 20 to 45. In general, more than 50% of ovarian carcinomas occur in women older than 50. In a hospital-based study of ovarian neoplasms in 861 women, Koonings and associates noted that the risk of malignancy was 13% in premenopausal women but rose to 45% in postmenopausal women. More than 90% of women diagnosed with ovarian cancer report symptoms before diagnosis. Unfortunately, these symptoms are vague and not specific for early-stage disease or even ovarian cancer. Goff and colleagues conducted a prospective survey of women seeking medical care. The case patients were those about to undergo surgery for a known or suspected pelvic or ovarian mass; the controls were women presenting to one of two primary care clinics, in which approximately two-thirds were being seen for a specific problem. The voluntary questionnaire instrument administered to both cohorts asked the respondents to score the severity, frequency, and duration of 20 symptoms generally reported by ovarian cancer patients. In both groups, recurring symptoms were common and nonspecific. Symptoms in control patients were related to the purpose of their visit (general checkup vs. specific complaint), underlying disease comorbidities, and menopausal status. Not surprisingly, women with the final diagnosis of ovarian cancer generally reported numerically more symptoms and symptoms of greater severity but of shorter duration of onset compared with the clinic controls and patients with benign ovarian tumors. Patients with ovarian cancer were also statistically more likely to report increased abdominal size, bloating, urinary urgency, and pelvic pain. Because the combination of increased abdominal size, bloating, and urinary urgency was reported five times more often and had greater severity in cancer patients than in controls, the authors recommended further clinical investigation when identified. The diagnosis is established by histologic examination of tumor tissue removed at operation, although occasionally the initial diagnosis is suggested by malignant cells found in ascitic or pleural fluid obtained at paracentesis or thoracentesis, respectively.
On abdominal pelvic examination, patients may present with a suspicious or palpable pelvic mass, ascites, abdominal distention, and/or other symptoms that may not indicate an obvious source of malignancy. These may include bloating, pelvic or abdominal pain, difficulty eating or feeling full quickly, and urinary symptoms such as urgency or frequency. The initial workup for such patients, as suggested by the 2019 National Comprehensive Cancer Network (NCCN) guidelines, should include an abdominal/pelvic examination, an ultrasound, and/or abdominal/pelvic computed tomography (CT) or magnetic resonance imaging (MRI) as indicated. A chest CT is generally obtained when an ovarian malignancy is suspected. Tumor markers such as CA-125 should be obtained; other tumor markers may include inhibin, beta human chorionic gonadotropin (β-HCG), alpha-fetoprotein (α-fetoprotein), lactate dehydrogenase (LDH), carcinoembryonic antigen (CEA), and CA-19-9. A thorough family history assessment should be performed; if the patient is not being evaluated by a gynecologic oncologist, when findings are indicative of a primary ovarian malignancy, a referral to a gynecologic oncologist should be made ( ).
The staging of ovarian cancer ( Table 33.6 ) is designed according to the criteria of the International Federation of Gynecology and Obstetrics (FIGO), which are based on the results of operative exploration. Involvement of certain sites in the abdominal or pelvic cavity would lead to disease being considered inoperable to achieve an optimal cytoreduction (no gross residual disease). These include retroperitoneal suprarenal lymph node enlargement, mesenteric disease, portal hepatis disease, and bilateral parenchymal liver metastases.
|I||Tumor confined to the ovaries|
|IA||Growth limited to one ovary (capsule intact); no tumor on ovarian surface; no malignant cells in the ascites or peritoneal washings|
|IB||Tumor limited to both ovaries (capsule intact); no tumor on ovarian surface; no malignant cells in the ascites or peritoneal washings|
|IC||Tumor limited to one or both ovaries|
|1C2||Capsule ruptured before surgery or tumor on ovarian surface|
|1C3||Malignant cells in the ascites or peritoneal washings|
|II||Tumor involves one or both ovaries with pelvic extension|
|IIA||Extension or metastases to the uterus or fallopian tubes|
|IIB||Extension to other pelvic intraperitoneal tissues|
|III||Tumor involving one or both ovaries with cytologically or histologically confirmed spread to the peritoneum outside the pelvis or metastasis to the retroperitoneal lymph nodes|
|IIIA1||Positive retroperitoneal lymph nodes only (cytologically or histologically proved)|
|IIIA1 (i)||Metastasis up to 10 mm in greatest dimension|
|IIIA1 (ii)||Metastasis more than 10 mm in greatest dimension|
|IIIA2||Microscopic extrapelvic (above the pelvic brim) peritoneal involvement with or without positive retroperitoneal lymph nodes|
|IIIB||Macroscopic peritoneal metastasis beyond the pelvis up to 2 cm in greatest dimension, with or without metastasis to the retroperitoneal lymph nodes|
|IIIC||Macroscopic peritoneal metastasis beyond the pelvis more than 2 cm in greatest dimension, with or without metastasis to the retroperitoneal lymph nodes (includes extension of tumor to capsule of liver and spleen without parenchymal involvement of either organ)|
|IV||Distant metastases excluding peritoneal metastases|
|IVA||Pleural effusion with positive cytology|
|IVB||Parenchymal metastases and metastases to extraabdominal organs (including inguinal lymph nodes and lymph nodes outside of the abdominal cavity)|
* According to the International Federation of Gynecology and Obstetrics (FIGO), 2014.
Radiologic imaging has an important role in the management of ovarian cancer. Ultrasonography may be used to confirm the presence of an ovarian mass and to help distinguish between a benign or malignant lesion. CT is more commonly used to determine extent of disease and treatment response, although one of the limitations of such imaging is that it may not detect small tumor implants (<2 cm). 18 F-2-fluoro-2-deoxy-D-glucose positron emission tomography/CT (FDG-PET/CT) is a hybrid metabolic and anatomic imaging technique. In the setting of ovarian cancer, it is not routinely performed in the initial evaluation but rather more so in patients with suspected recurrence based on elevated serum CA-125 or when findings on traditional CT scans are nonspecific ( ).
Occasionally, a diatrizoate (Gastrografin) enema or colonoscopy is performed to evaluate pelvic or gastrointestinal symptoms. Consideration of gastrointestinal pathologic changes is of importance for the potential of a primary colon carcinoma, which may present initially as an adnexal mass in the older woman. Approximately 4% of colon cancers have metastatic involvement of the ovary at diagnosis. Determination of the serum CEA level may be useful in this setting and is recommended as part of the preoperative evaluation of a pelvic mass when there is suspicion of metastatic disease to the ovary. An endoscopic or gastrointestinal radiographic examination is performed if there is evidence of gastrointestinal bleeding or the suggestion of any gastrointestinal pathologic condition.
Preoperatively the use of routine bowel preparation before major abdominal surgery for advanced ovarian cancer is no longer considered the standard of care. In a Cochrane review of 260 trials including 43,451 participants with colorectal cancer, the authors aimed to determine whether rates of postoperative surgical wound infection were affected by prophylactic antibiotics compared with no treatment. The main findings from that study were that antibiotics delivered orally or intravenously before elective colorectal surgery reduced the risk of surgical wound infection. The risk of postoperative surgical wound infection was reduced by as much as 75%. ( ). In another study by Koller et al., the investigators aimed to determine the relationship between bowel preparation and surgical site infection. The analysis included 32,359 patients who underwent elective colorectal resections in the American College of Surgeons National Surgery Quality Improvement Program. Mechanical bowel preparation was not associated with decreased risk of surgical site infection compared with no bowel preparation. In contrast, both oral antibiotics or oral antibiotics and mechanical bowel preparation were associated with decreased risk of any surgical site infection compared with no bowel preparation. Venous thromboembolism prophylaxis is of particular importance in patients with ovarian cancer because a large tumor burden is associated with venous stasis and prolonged operation times. Treatment with variable-compression leg support stockings and heparin (fractionated and unfractionated) appears to reduce the risk of thromboembolism in gynecologic oncology patients undergoing surgical tumor extirpation.
Ovarian carcinomas infiltrate the peritoneal surfaces of the parietal and intestinal areas, as well as the undersurface of the diaphragm, particularly on the right side ( Fig. 33.7 ). This is particularly important because tumors that at operation appear to be confined to the ovary may have small areas of diaphragmatic involvement as the sole site of extraovarian spread. As noted earlier, most ovarian carcinomas, particularly the serous type, appear to arise from microscopic ovarian sites and do not become clinically evident until there is widespread metastatic disease. Lymphatic dissemination is also a prominent part of disease spread ( Fig. 33.8 ), and it is particularly important to note that the paraaortic nodes are at risk through lymphatics that run parallel to the ovarian vessels. In a study of 180 patients, Burghardt and coworkers observed that the proportion of positive nodes increases with higher-stage tumors: 24% in stage I, 50% in stage II, and 73.5% in stages III and IV. A study conducted by Schmeler and colleagues has evaluated the prevalence of lymph node involvement in women with primary mucinous ovarian carcinomas. A total of 107 patients were identified. Of the patients with tumor grossly confined to the ovary at surgical exploration who underwent lymphadenectomy, none had metastatic disease to the pelvic or paraaortic lymph nodes. In addition, the authors found no significant differences in progression-free survival or overall survival between patients who underwent lymphadenectomy and those who did not.
The prognosis for patients with ovarian carcinoma is related to tumor stage, tumor grade, cell type, and amount of residual tumor after resection. Information from the SEER database is presented in Table 33.7 .
|Stage||5-Year Survival (%)|
* Data from the American Cancer Society. Information is based on patients diagnosed from 2004 to 2010 and obtained from the National Cancer Institute and the Surveillance, Epidemiology, and End Results (SEER) Database.
Cell type has been reported to be an important factor in prognosis, as shown in Fig. 33.9 , which summarizes the 10-year survival rate of a group of patients. The most common invasive epithelial cancers, serous carcinomas, have the worst prognosis; prognosis may be better for mucinous and endometrioid tumors. A variant of papillary serous carcinoma termed transitional cell carcinoma is thought by some to be a rare but more chemosensitive tumor; however, this has not been established in multiinstitutional studies. Endometrioid carcinoma may be associated with endometriosis; such cases more commonly occur in younger women and have a better prognosis than typical endometrioid carcinomas of the ovary. Clear cell cancers have a worse prognosis, but Kennedy and associates noted that mitotic activity and tumor stage are important prognostic features of this tumor in their series. Serous tumors tend to be more poorly differentiated and discovered at a higher stage than mucinous tumors.
In addition to stage, the grade of the tumor is a major determinant of patient prognosis. Fig. 33.10 demonstrates the survival of 442 patients with ovarian carcinoma by grade, with a markedly worse prognosis for poorly differentiated tumors (grade 3). The relationship between grade and survival also exists when the results are examined separately for each stage of disease.
The development of gene expression profiling has enabled a more precise evaluation of clinical behavior in some tumors. Bonome and coworkers studied the gene expression of low malignant potential (LMP) serous neoplasms and invasive low-grade and high-grade serous tumors. A distinct and separate clustering was observed between LMP tumors and high-grade cancers. Low-grade serous tumors generally clustered with LMP neoplasms. High-grade tumors differentially expressed genes linked to cell proliferation, chromosomal instability, and epigenetic silencing. Based on these findings, high-grade epithelial cancers appear to have a distinct profile relative to LMP neoplasms. Low-grade serous tumors are remarkably similar to LMP serous neoplasms. These observations have ushered in the reclassifying invasive malignant cancers into two categories, low grade and high grade.
Borderline ovarian tumors: Ovarian carcinomas of low malignant potential
Approximately 20% of ovarian epithelial cancers are tumors of LMP and usually have an excellent prognosis, regardless of stage. Most studies have been confined to borderline tumors of the serous (see Fig. 33.2 , B ) and mucinous (see Fig. 33.3 , B ) varieties, which are the most common histologic types; however, other epithelial types (see Table 33.3 ) can occur. Serous borderline tumors include two patterns—atypical proliferating tumor and micropapillary/cribriform (also known as noninvasive low-grade serous carcinoma) ( Fig. 33.11 ). The cells of these epithelial tumors do not invade the stroma of the ovary. It is extremely important that the ovarian tumor be thoroughly sampled by the pathologist to ensure that a borderline tumor is not mixed with invasive elements. Numerous studies have confirmed that borderline tumors have a slower growth rate than invasive ovarian carcinomas, manifested by prolonged survival
Surgery is the primary treatment for women with borderline ovarian tumors. The principal objectives of surgery are as follows: (1) diagnosis, (2) fertility-sparing surgery for patients who have not completed childbearing or who are young and have only unilateral ovarian involvement, (3) surgical staging for apparent early-stage disease, and (4) cytoreductive surgery for the minority of patients who have obvious advanced-stage disease.
The typical scenario is surgery for an adnexal mass of unknown type. One of the initial considerations in planning surgery for a pelvic mass is the surgical approach—minimally invasive or open technique. Factors to be considered in the selection of minimally invasive surgery (laparoscopic or robotic) include size of the ovarian mass(es), extent of tumor metastasis, number and type of previous operations, and body habitus. Several reports have documented the feasibility and safety of the minimally invasive approach when appropriately used ( ; ).
Once the mass is excised, frozen-section examination is a key element in ensuring appropriate decision making. If the frozen section suggests a borderline ovarian tumor, considerations for fertility-sparing surgery include the woman’s age, her desire for future childbearing, and the degree of involvement of the ovaries—unilateral versus bilateral disease. Options for fertility-sparing surgery include ovarian cystectomy and unilateral adnexectomy. Even with bilateral borderline ovarian tumors, bilateral ovarian cystectomies or ovarian cystectomy plus unilateral adnexectomy may be performed, depending on the extent of ovarian disease. Lim-Tan and associates reported on 33 cases of stage I serous borderline tumors initially treated by cystectomy. Only 3 of 33 patients undergoing cystectomy had recurrence or persistence of borderline tumor, and these patients had positive resection margins or multiple cysts present in the ovary, emphasizing the effectiveness of conservative operation; however, for most stage IA cases, unilateral adnexectomy is performed and, if the opposite ovary looks normal, no biopsy or wedge resection is done.
Surgical staging for borderline ovarian tumors remains somewhat controversial. The most compelling reason for surgical staging in a woman with borderline tumor on frozen-section examination is the risk of invasive carcinoma on final pathologic diagnosis. For women with pathologically confirmed borderline tumors, the incidence of lymph node involvement is only approximately 5%. Therefore most investigators have recommended against routine pelvic and paraaortic lymphadenectomy; however, omental and peritoneal biopsies are recommended because peritoneal implants are usually small or microscopic ( ).
As noted earlier, there are two types of serous borderline tumors: atypical proliferating tumor and micropapillary pattern. The former is more common, and the latter is more often associated with relapse. For stage I serous borderline tumors, surgery alone is the standard of care because the cure rate associated with this treatment approaches 100%. On the other hand, approximately 30% of women will have peritoneal implants, which are classified as noninvasive (see Fig. 33.11 ) or invasive (now classified as invasive low-grade serous carcinoma based on the 2014 WHO classification) ( Fig. 33.12 ). For these women, the recurrence rate associated with noninvasive implants is 20% to 50% and the recurrence rate with invasive implants ( Fig. 33.13 ) (invasive low-grade serous carcinoma) is 50% to 70%. No studies to date have convincingly demonstrated any benefit to postoperative therapy for women with stages II to IV disease with noninvasive implants; however, for women with invasive implants (invasive low-grade serous carcinoma), postoperative treatment—paclitaxel-carboplatin chemotherapy or hormonal therapy—is recommended.
Outcomes are influenced by pathologic, clinical, and other factors. Pathologic factors that may be associated with an increased risk of relapse include the presence of the micropapillary/cribriform pattern or microinvasion in the primary ovarian serous LMP tumor and the presence of peritoneal implants ( ). Clinical factors include age at diagnosis, baseline serum CA-125, conservative surgery, and residual disease.
Mucinous borderline tumors also are associated with an excellent prognosis. Hart and Norris reviewed 97 patients aged 9 to 70 years (median, 35 years) with stage I tumors. More than 10% of the tumors were discovered during pregnancy or in the immediate postpartum period. Follow-up data were available on 87 of the patients, and there were only three tumor-related deaths during the 5- to 10-year follow-up. The actuarial survival was 98% at 5 years and 96% at 10 years. This was also noted by Bostwick and colleagues, who reported on 109 borderline tumors, 33 of which were mucinous and all of which were stage I, contributing to the good prognosis.
Mucinous borderline tumors include two distinct subtypes, gastrointestinal and seromucinous or endocervical. In the seromucinous type, the association of endometriosis is high (40%). They also may have associated microinvasion and lymph node involvement. The prognosis associated with the seromucinous type is excellent. Conversely, the gastrointestinal type may rarely be associated with the condition known as pseudomyxoma peritonei, consisting of widespread growth of mucin-producing cells in the peritoneum. The result may be the accumulation of large amounts of mucinous material, which is sometimes associated with recurrent episodes of bowel obstruction. Studies by Young and coworkers suggested that pseudomyxoma peritonei usually arises in the appendix. The review of Ronnett and coworkers supports a primary appendiceal origin for these tumors and suggest that therefore appendectomy may be indicated for women with an intraoperative diagnosis of a mucinous ovarian tumor. On the other hand, more recent studies have recommended appendectomy only if the appendix appears abnormal ( ). The disease tends to recur and is commonly characterized by repeated laparotomy to relieve bowel obstruction. Chemotherapy and mucolytic agents have been tried but are usually not successful.
Invasive epithelial carcinomas
The primary treatment of ovarian epithelial carcinoma is removal of all resectable gross disease. The woman’s abdomen is explored through a vertical incision. If ascitic fluid is present, it is sent for cytologic evaluation; if ascites is not present, 200 to 400 mL of normal saline solution is used to obtain cytologic samples from the peritoneum by irrigating at least the pelvis, upper abdomen, and right and left paracolic gutters before any resection is done. The diaphragm can be cytologically sampled by scraping the undersurface with a sterile tongue depressor and the sample placed on a glass slide and sprayed with a fixative. Biopsy or, preferably, excision of any suspicious nodules is performed. A total abdominal hysterectomy, bilateral salpingo-oophorectomy, and infracolic omentectomy are performed if technically possible. When there is no gross disease outside the pelvis, in the setting of early-stage disease, paraaortic and pelvic lymph node sampling is recommended. Evidence suggests that if all gross disease can be resected, the duration of patient survival is enhanced. Although randomized clinical trials have not been performed to document this effect, a meta-analysis of 6885 patients gathered from 81 cohorts has suggested a linear relationship between the degree of cytoreduction and overall survival. In this report, Bristow and colleagues noted that for each 10% increase in cytoreduction, a 5.5% increase in survival was observed. The surgical procedures required to achieve maximal cytoreduction may be extensive and involve splenectomy, diaphragmatic stripping resection, and posterior exenteration. It may occasionally be necessary to resect bowel to relieve impending obstruction or remove a tumor nodule, thereby eliminating all gross disease from the peritoneal cavity. Heintz and coworkers noted that prognosis is improved in patients younger than 50 years, those with good initial performance status (Karnofsky score > 80), and those whose disease could be cytoreduced to less than 1.5 cm. In a small collaborative Gynecologic Oncology Group (GOG) study, Hoskins and colleagues found that those who started with large-volume disease did worse than those who initially had small-volume disease; no survival advantage could be demonstrated for the debulking operation in the large-volume disease group. Chi and associates noted that those with advanced disease and a preoperative CA-125 level greater than 500 U/mL had less than a 20% chance of an optimal surgical debulking (discussed later).
One exception to the required removal of the uterus and opposite ovary occurs in the case of well-differentiated (grade 1) ovarian tumors confined to one ovary (stage IA). DiSaia and coworkers outlined criteria for preserving childbearing function in a young woman with stage IA, grade 1 ovarian epithelial carcinoma, as follows:
Tumor confined to one ovary
Tumor well differentiated (grade 1), with no invasion of capsule, lymphatics, or mesovarium
Peritoneal washings negative
Omental biopsy specimen negative
Young woman of childbearing years with a strong desire to preserve reproductive function
These criteria can be applied to all types of epithelial ovarian tumors but are more likely to be satisfied in the case of mucinous tumors, which are more commonly well differentiated and unilateral than serous carcinomas. Wedge resection of a normal-appearing contralateral ovary is unlikely to uncover an occult tumor. In these cases it is reasonable to follow the woman closely with vaginal ultrasonography for any evidence of future ovarian enlargement.
Mutations in BRCA1/2 are clearly associated with increased risk of development of ovarian cancer ( ; ). Thus genetic testing with or without genetic counseling is recommended during the workup of any patient with high-grade, nonmucinous, epithelial ovarian cancer ( ). This has been clearly supported by a number of national organizations, including the NCCN, the American Society of Clinical Oncology, and the Society of Gynecologic Oncology. Genetic testing should be performed even in the absence of traditional risk factors for hereditary breast and ovarian cancer syndromes. Testing can be focused on the BRCA1/2 genes alone or can be accomplished with panel testing, which may include more than 20 genes from the homologous recombination pathway that are relevant to ovarian cancer risk ( ). Results are not only important for patient and family risk of cancer but may also affect choice of therapy (discussed later). In addition to germline testing for BRCA, the use of next-generation sequencing to test for tumor tissue–based (somatic) BRCA1/2 mutations is also indicated in ovarian cancer. As detailed later, PARP (poly [ADP-ribose] polymerase) inhibitors have equivalent activity in patients with germline or somatic mutations in BRCA, and thus the performance of somatic testing is critical to guide treatment decisions.
Early-stage ovarian carcinomas
The standard therapy for all patients with early-stage ovarian cancer (EOC) includes hysterectomy (usually performed via celiotomy) with bilateral salpingo-oophorectomy. In patients interested in preserving fertility, unilateral salpingo-oophorectomy with preservation of the contralateral ovary and uterus is often feasible. In addition, all patients with early-stage disease should undergo pelvic washings, omental biopsy, cytologic sampling of the surface of the diaphragm, complete bilateral pelvic and paraaortic lymphadenectomy or lymph node sampling, and bilateral biopsies of the paracolic gutters and pelvic peritoneal surfaces. It is important to emphasize that careful assessment of the subdiaphragmatic areas and inspection of the entire peritoneum and the retroperitoneal paraaortic and pelvic nodes are important, particularly in view of the risk of diaphragmatic and nodal spread in higher-grade tumors that initially appear to be at stage I, especially those on frozen section that appear to be less well differentiated than grade 1.
Le and colleagues compared a group of patients who underwent minimal staging performed by a general gynecologist with a group of patients who underwent comprehensive staging performed by a gynecologic oncologist. They found the risk of recurrence to be increased for patients operated on by the general gynecologist. Another study by Mayer and associates showed that patients operated on by a gynecologic oncologist had a 24% improvement in 5-year overall survival.
Rupture of ovary
Occasionally, during removal, a stage I ovarian carcinoma is inadvertently ruptured (stage IC; see Table 33.6 ). There are conflicting opinions as to the potential adverse effects on prognosis. In an analysis of 394 patients, Sjövall and associates found that rupture during surgery did not affect survival, whereas there was marked reduction in survival in that study among those whose ovarian rupture occurred before the operation. In general, the spilled fluid and all residual tumor should be removed promptly from the operative field after a rupture (discussed later). Presumably, higher-grade and larger tumors are most at risk for rupture.
A study by Dembo and colleagues of 519 stage I patients found that adverse factors were grade of tumor, dense pelvic adherence (no invasion but adhesion), or more than 250 mL of ascites. Patients without these features had a 98% 5-year survival rate. It appears that patients with stage I, grade 3 tumors should have postoperative therapy, but data are unclear for stage I, grade 2 patients.
Minimally invasive surgical staging
An estimated 15% of women with EOC have early-stage disease at diagnosis. In these patients, comprehensive surgical staging is required to provide accurate prognostic information and plan treatment options. Commonly the diagnosis of EOC is made incidentally during adnexal surgery for other indications. Because minimally invasive surgical procedures are becoming more common in gynecology and more in demand by the public at large, determining the feasibility and safety of these procedures for the staging of presumed early-stage EOC has become a necessity.
Laparoscopy often leads to a shorter hospital stay, less intraoperative blood loss, and a shorter recovery period than laparotomy; however, surgical staging for EOC requires meticulous inspection of the peritoneal cavity and careful dissection of lymph nodes, vessels, and other abdominal and pelvic structures. Thus to assess the feasibility of minimally invasive surgical staging, several issues must be considered, such as the frequency of complications, frequency of conversion to laparotomy, and recurrence rate after laparoscopic staging.
In a study by Gallota and colleagues, the authors evaluated the safety and perioperative outcomes of laparoscopic staging of patients with apparent EOC ( ). A total of 300 patients were selected: 150 had been submitted to immediate laparoscopic staging (group 1), and 150 had undergone delayed laparoscopic staging (group 2). No significant differences in postoperative complications were observed between the two groups. Histologic data revealed more serous tumors (0.06), grade 3 ( P = .0007), and final upstaging ( P = .001) in group 1. Recurrence and death of disease were documented in 25 patients (8.3%) and 10 patients (3.3%), respectively. The 3-year disease-free survival (DFS) and overall survival (OS) rates were 85.1%, and 93.6%, respectively, in the whole series. There was no difference between group 1 and group 2 in terms of DFS ( P = .39) and OS ( P = .27). The authors concluded that laparoscopic management of early ovarian cancer is safe and feasible.
In summary, it appears that minimally invasive surgical staging of presumed EOC is safe and effective when performed by a trained gynecologic oncologist.
Adequacy of minimally invasive surgery compared with laparotomy for staging of ovarian neoplasms
Important concerns have been raised about laparoscopic staging for ovarian neoplasms, including concerns about the adequacy of the lymph node dissection, differences between laparoscopy and laparotomy in operative time, postoperative complications, and postoperative recovery. In one meta-analysis of laparoscopic staging surgery in patients with presumed EOC, Park and colleagues identified 11 observational studies ( ). The combined results of three retrospective studies showed that the estimated blood loss in laparoscopy was significantly lower than that for laparotomy ( P < .001). The overall upstaging rate after laparoscopic surgery was 22.6% (95% CI, 18.1 to 27.9) without significant heterogeneity among all study results. The overall incidence of conversion from laparoscopy to laparotomy was 3.7% (95% CI, 2.0 to 6.9). The overall rate of recurrence in studies with a median follow-up period of 19 months or more was 9.9% (95% CI, 6.7 to 14.4). The authors concluded that the operative outcomes of a laparoscopic approach in patients with EOC could be compatible with those of laparotomy.
Data on the application of robotic surgery for ovarian cancer staging are scant. In a study by Brown and associates, the authors evaluated the safety and feasibility of robotic-assisted systematic lymph node staging in the management of EOC. A total of 26 patients with EOC were identified. The mean operating time was 2.90 hours, and the estimated blood loss was 63 mL; there were no intraoperative complications, although one patient’s surgery was significantly prolonged because of pelvic adhesions. The mean number of pelvic and paraaortic lymph nodes removed was 14.6 (2.3% incidence of pelvic lymph node metastases) and 5.8 (3.3% incidence of paraaortic lymph node metastases), respectively. The patients’ mean duration of hospital stay was 18.4 hours, and 2 patients were readmitted for either a postoperative wound infection or vaginal dehiscence. The authors concluded that robotic-assisted surgical staging was feasible and safe; however, there was a low incidence of lymph node metastasis (2.3%), and the authors stressed the value of a systematic lymph node dissection.
Stage II ovarian cancer is initially treated by removal of all gross disease, in addition to removal of the uterus, tubes, ovaries, and omentum. The pelvic and paraaortic nodes are sampled.
Postoperative management for stages I and II
Recommendations for postoperative or adjuvant therapy generally have evolved around the identification of patients in whom a sufficient risk of recurrence is observed. The precision to make this assessment is low and has been based on morphologic features such as grade, histology, and presence of rupture and residuum. Historically, several modalities have been evaluated alone and in combination by prospective studies of unselected patients, including chemotherapy, radiation therapy, and immunotherapy. Guthrie and coworkers evaluated 656 patients treated for epithelial ovarian carcinomas that had been totally excised. Most carcinomas were stage I or II, and patients were randomly assigned to receive postoperative treatment of radiation therapy alone, chemotherapy alone, radiation therapy and chemotherapy, or no postoperative therapy. Perhaps surprisingly, the lowest frequency of death or recurrence was noted in the group receiving no postoperative therapy (2%), whereas in the other groups, the death or recurrence incidence was 14% to 17%. Thus this study found no benefit for adjuvant therapy and highlights the need for careful case selection and pathologic review. In addition, because survivorship of low-stage cancer is better, long-term follow-up is necessary to tease out the merits of intervention.
Two large multicenter trials have been conducted and were combined for analytic purposes to address this issue. The International Collaborative Ovarian Neoplasm 1 (ICON1) and Adjuvant ChemoTherapy in Ovarian Neoplasm (ACTION) trials compared platinum-based chemotherapy with observation in patients after surgery with EOC. The two trials differed somewhat in patient eligibility, with ICON1 predominantly enrolling postoperative stages I and II patients with limited staging and ACTION enrolling patients who were postoperative stage IA and IB, grades 2 and 3, stages IC to IIA, all grades, and clear cell tumor. Overall, 925 patients were collectively enrolled (477 in ICON1 and 448 in ACTION) and followed for a median 4 years. The OS rate at 5 years was 82% in the chemotherapy arm and 74% in the observation arm (HR, 0.67; 95% CI, 0.5 to 0.9; P = .001). Recurrence-free survival at 5 years was also significantly higher in the chemotherapy arm compared with observation (HR, 0.64; 95% CI, 0.5 to 0.82). In select patients, platinum-based therapy appears to improve survival and lower recurrence at 5 years compared with observation. In all, three randomized clinical trials have addressed platinum-based chemotherapy versus observation in EOC. These three trials were evaluated via meta-analysis to assess the role of adjuvant chemotherapy. As expected, the combined data mirror the results of the two larger studies with regard to the impact of adjuvant chemotherapy (HR, 0.71; 95% CI, 0.53 to 0.93) on 5-year survival; however, when subcategorized by surgical staging, this benefit retained significance only in the group in whom nonoptimal staging was performed. These data highlight the importance of accurate surgical staging information when devising an appropriate postoperative treatment plan.
The ideal regimen and number of courses of chemotherapy needed for EOC are still debated. Bell and associates reported the results of a randomized study comparing three and six cycles of adjuvant paclitaxel and carboplatin for women with stages IA and IB, grade 3, all stage IC, clear cell tumor, and completely resected stage II epithelial ovarian cancer. A total of 457 patients were recruited; 344 were alive a median of 6.8 years since entry. The overall treatment effect was a nonsignificant 24% reduction in recurrence for the six-cycle arm (HR, 0.76; 95% CI, 0.51 to 1.13; P = .18). The improved impact on estimated recurrence at 5 years was 5%, and there was no difference in OS between the arms. A post hoc analysis of the survival results by histologic type raises the hypothesis that six cycles of therapy may benefit patients with serous histologic type to a different degree than the others included in the trial.
A follow-up study, GOG-175, addressed the role of maintenance therapy in this setting. Women with the same eligibility as GOG-157 were randomly allocated to three cycles of paclitaxel and carboplatin followed by normal surveillance or 24 infusions of low-dose (40 mg/m 2 ) weekly paclitaxel ( ). Eighty percent of patients randomly allocated to the maintenance arm received all assigned therapy. Toxicity was similar between the two arms. The HRs for progression-free survival (PFS) and OS were 0.81 (95% CI, 0.56 to 1.15) and 0.78 (95% CI, 0.52 to 1.17), respectively. Subgroup analysis demonstrated no effect by stage, histologic type, or grade. There was remarkable consistency between the two GOG studies in the 5-year survival recorded by the three-cycle cohort. Although the optimal treatment is still not known, high-risk, early-stage patients clearly benefit from therapy.
Advanced-stage ovarian cancer
Role of imaging studies and serum markers in determining ideal surgical candidate
To determine which patients would be less likely to benefit from primary surgery, a number of attempts have been made to predict outcomes of primary cytoreductive surgery by using imaging modalities, tumor markers, and laparoscopic scores. The use of preoperative CT scan imaging has shown inconsistent results. The same issue has been encountered when exploring the use of serum CA-125. One should note that studies attempting to identify preoperative predictors have been limited by their retrospective design, sample size, broad inclusion criteria, and heterogeneous rates of optimal cytoreduction.
Suidan and colleagues reported on a prospective, nonrandomized, multicenter trial of patients who underwent primary cytoreduction for stage III-IV ovarian, fallopian tube, and peritoneal cancer. A CT scan of the abdomen and pelvis and a serum CA-125 level were obtained within 35 and 14 days before surgery; respectively. Four clinical and 20 radiologic criteria were assessed. A total of 669 patients were enrolled and 350 patients met eligibility criteria. The optimal (<1 cm) debulking rate was 75%. On multivariate analysis, three clinical and six radiologic criteria were associated with suboptimal debulking: age 60 years or older; serum CA-125 more than 500 U/mL; American Society of Anesthesiology (ASA) 3 to 4; suprarenal retroperitoneal lymph nodes greater than 1 cm; diffuse small bowel adhesions or thickening; and lesions greater than 1 cm in the small bowel mesentery, root of the superior mesenteric artery, perisplenic area, and lesser sac. The authors identified nine criteria associated with suboptimal cytoreduction and developed a predictive model in which the suboptimal rate was directly proportional to a predictive value score.
Role of laparoscopy in assessing resectability to R0 in advanced ovarian cancer
Laparoscopy has been proposed as a reliable predictor of R0 resection. Here we review the existing literature on the proposed criteria to predict the outcome of cytoreductive surgery and the role of laparoscopy-based scores in the management of advanced ovarian cancer. This principle was initially introduced by Fagotti and colleagues. In a study evaluating the prognostic impact of routinely using laparoscopy in patients with advanced ovarian cancer, the authors submitted all patients to laparoscopy before undergoing either primary debulking surgery or neoadjuvant chemotherapy. Among 300 consecutive patients, there were no complications related to the surgery and the laparoscopic evaluation showed that almost half of the patients (46.3%) had a high tumor load. A total of 148 patients (49.3%) were considered suitable for primary debulking surgery and the remaining 152 (50.7%) were treated with neoadjuvant chemotherapy. The authors concluded from that study that the inclusion of laparoscopy in this setting did not appear to have a negative impact on survival and that it might be helpful to individualize treatment by avoiding unnecessary laparotomies and surgical complications. The same group of investigators subsequently demonstrated that the laparoscopic assessment of peritoneal spread was a concept that could be reproduced among different institutions in a multicentric trial (Multicentre Italian Trials in Ovarian Cancer [MITO]-13) ( ).
Gómez-Hidalgo and colleagues published a comprehensive review of the evolution of laparoscopy as a tool to help identify ideal patients with advanced ovarian cancer for optimal cytoreduction (R0). The authors concluded that existing studies point to a highly valuable role for laparoscopy for objectively assessing the feasibility of optimal primary and interval cytoreductive surgery for patients with advanced-stage ovarian cancer (FIGO stages III and IV). They went on to suggest that the Fagotti laparoscopy-based score is a useful predictor of optimal cytoreduction.
Nick and colleagues published an algorithm that identifies patients in whom complete gross resection at primary surgery is likely to be achieved. Such an algorithm is currently being used to ensure that the rate of optimal cytoreduction (R0) increases and the rate of patients unnecessarily undergoing neoadjuvant chemotherapy decreases. In addition, the algorithm allows surgeon to obtain tissue before the initiation of therapy, thus targeting molecular pathways in a much more precise and personalized strategy.
In a 2018 study by Fleming and colleagues, the authors evaluated the impact of the laparoscopic scoring algorithm to triage patients with advanced ovarian cancer to immediate or delayed debulking to improve complete gross surgical resection rates and evaluate clinical outcomes. A total of 488 patients met inclusion criteria; 215 patients underwent laparoscopic scoring. The patients were stratified according to scoring of less than 8 (58.1%) or 8 or more (39%). The authors also evaluated the concordance between two surgeons in predicting the laparoscopic score. This resulted in a bivariate concordance of 98%. The implementation of this laparoscopic algorithm lead to no gross residual disease (R0) in 88% of patients in the primary surgery group and 74% in the neoadjuvant chemotherapy group.
Lastly, a 2017 multicenter, prospective randomized trial was performed within eight gynecologic cancer centers in the Netherlands ( ). The goal of the study was to investigate whether diagnostic laparoscopy can prevent futile primary cytoreduction by identifying patients with advanced-stage ovarian cancer in whom more than 1 cm of residual disease will be left after primary surgery. A total of 201 participants were included, of whom 102 were assigned to diagnostic laparoscopy and 99 to primary surgery. Results showed that futile laparotomy occurred in 10% in the laparoscopy group versus 39% of patients in the primary surgery group. These data suggested that diagnostic laparoscopy may help avoid futile laparotomy when considering primary surgery.
Primary cytoreductive surgery
Most patients with ovarian cancer present with disease that has spread beyond the pelvis and into the upper abdomen. The routine recommendation for patients with advanced disease who are surgical candidates is to perform a total abdominal hysterectomy, bilateral salpingo-oophorectomy, complete omentectomy, and resection of all visible tumor. Bristow and colleagues performed a retrospective population-based study of consecutive patients diagnosed with epithelial ovarian cancer ( ). A total of 9933 patients were identified (stage I, 22.8%; stage II, 7.9%; stage III, 45.1%; stage IV, 24.2%), and 8.1% of patients were treated at comprehensive cancer centers (National Cancer Institute Comprehensive Cancer Center [NCI-CCC]). Overall, 35.7% of patients received NCCN guideline–adherent care, and NCI-CCC status (OR, 1.00) was an independent predictor of adherence to treatment guidelines compared with high-volume hospitals (HVHs) (OR, 0.83; 95% CI, 0.70 to 0.99) and low-volume hospitals (LVHs) (OR, 0.56; 95% CI, 0.47 to 0.67). The median ovarian cancer–specific survivals according to hospital type were NCI-CCC 77.9 months (95% CI, 61.4 to 92.9), HVH 51.9 months (95% CI, 49.2 to 55.7), and LVH 43.4 months (95% CI, 39.9 to 47.2) ( P < .0001). NCI-CCC status (HR, 1.00) was a statistically significant and independent predictor of improved survival compared with HVH (HR, 1.18; 95% CI, 1.04 to 1.33) and LVH (HR, 1.30; 95% CI, 1.15 to 1.47).
Aletti and associates sought to estimate the effect of aggressive surgical resection on the survival of epithelial ovarian cancer patients. They found that the 5-year disease-specific survival rate was markedly better for patients operated on by surgeons who were most likely to use radical procedures than for patients operated on by surgeons who were least likely to use radical procedures (44% vs. 17%; P < .001). Also, the rate of optimal resection was 84% for the surgeons most likely to use radical procedures compared with 51% for the surgeons least likely to use radical procedures, highlighting the value of extensive surgical effort.
Zivanovic and colleagues evaluated the impact of upper abdominal disease (UAD) cephalad to the greater omentum on surgical outcomes for 490 patients with stage IIIC ovarian, fallopian tube, and primary peritoneal cancers. Patients were divided into three groups according to the amount of disease in the upper abdomen. Group 1 was defined as no disease in the upper abdomen, group 2 as having tumors smaller than 1 cm, and group 3 as having bulky disease, larger than 1 cm. The authors found that optimal cytoreduction was achieved in 81%, 63%, and 39% of patients in groups 1, 2, and 3, respectively. In the largest study of postoperative tumor residuum and outcome, resection to no visible intraperitoneal disease was substantially related to PFS and OS. The study population (n = 3126) was generated from three randomized phase III trials assessing primary chemotherapy regimens in patients with advanced-stage disease. Median OS was 99.1 months for patients with no postoperative tumor residua compared with 36.2 months for those with visible disease 1 cm or smaller and 29.6 months in those with more than 1 cm of tumor residua.
Chang and associates sought to quantify the impact of complete cytoreduction to no gross residual disease on OS among patients with advanced-stage ovarian cancer treated during the platinum-taxane era ( ). A total of 18 relevant studies (13,257 patients) were identified for analysis. After controlling for other factors on multiple linear regression analysis, each 10% increase in the proportion of patients undergoing complete cytoreduction to no gross residual disease was associated with a significant and independent 2.3-month increase (95% CI, 0.6 to 4.0, P = .011) in cohort median survival compared with a 1.8-month increase (95% CI, 0.6 to 3.0, P = .004) in cohort median survival for optimal cytoreduction (residual disease ≤1 cm). Each 10% increase in the proportion of patients receiving intraperitoneal chemotherapy was associated with a significant and independent 3.9-month increase (95% CI, 1.1 to 6.8, P = .008) in median cohort survival time. The authors found that the proportions of patients left with no gross residual disease and receiving intraperitoneal chemotherapy are independently significant factors associated with the most favorable cohort survival time.
The role of surgery in the upfront setting in patients with advanced ovarian cancer remains a topic of controversy and debate. There are many who argue that rather than residual disease as the primary determinant of oncologic outcomes, it may be other elements such as tumor biology that impart the most influence on such outcomes. Two ancillary data analyses have explored this question. The first was a subanalysis of the SCOTROC-1 (Scottish Randomized Trial in Ovarian Cancer 1) trial data by Crawford and colleagues. The authors reviewed the data on 889 patients with FIGO stage IC to IV ovarian cancer. A prognostic scoring system that reflected each patient’s preoperative biologic characteristics based on FIGO stage, tumor histologic type, preoperative CA-125 levels, and omental cake was established using a multivariate Cox model. In that study the authors concluded that the survival benefit associated with optimal surgery was limited to patients with less aggressive disease and that tumor biology was the primary survival determinant.
A subsequent study by Horowitz and colleagues retrospectively reviewed the GOG-182 trial data on 2655 patients with FIGO stage III or IV ovarian cancer. PFS and OS were analyzed based on three indices: preoperative disease score, surgical complexity score, and residual disease. Disease score was defined as low, with pelvic and retroperitoneal spread; moderate, with additional spread to the abdomen but sparing the upper abdomen; or high, with the presence of UAD affecting the diaphragm, spleen, liver, or pancreas. In that study the authors found that PFS and OS decreased with increasing disease score, and patients with high disease score had the worst PFS and OS. Interestingly, in patients with no gross residual disease, the high disease score still had a worse influence on PFS and OS compared with those with low-moderate disease scores. The authors concluded that although complete cytoreduction to no gross residual disease may be achieved, initial tumor burden is a primary determinant in survival and aggressive surgery alone does not seem to have a strong impact on outcomes. Both these studies have drawn some criticism for including patients with less advanced disease, having a nonobjective definition of surgical aggressiveness, defining optimal cytoreduction in the SCOTROC trial analysis as less than 2 cm, and use of complexity score as a surrogate of aggressive surgery in the GOG-182 subanalysis.
Most retrospective evidence suggests that patients with advanced ovarian cancer treated at HVHs by high-volume surgeons have better outcomes than those treated at the JVHs and by low-volume surgeons ( ).
Utility of video-assisted thoracoscopy
Unfortunately, no tools are available that allow surgeons to predict with high confidence whether there is disease in the pleural cavity. Video-assisted thoracoscopic surgery (VATS) allows surgeons, through a minimally invasive approach, to not only drain the pleural cavity of fluid but also evaluate whether pleural disease is present.
In a review by Di Guilmi and colleagues the authors summarized the literature on VATS and its applicability in patients with advanced ovarian cancer. A total of 187 patients with suspected ovarian cancer who underwent VATS procedure were identified for the analysis. The median patient age was 59.4 years (range, 20.3 to 83) and the median operative time was 32 minutes (range, 5 to 65). In 89 patients (48%), VATS revealed macroscopic disease in the pleural cavity. After VATS, 44 patients underwent neoadjuvant chemotherapy, and the remaining 143 patients underwent primary cytoreductive surgery. VATS led to a change in disease stage or management in 76 patients (41%). Among patients with pleural effusion, VATS revealed pleural disease in 57% of patients, and 73% of patients with positive pleural cytologic results had evidence of pleural disease. Interestingly, 23.5% of patients with negative pleural cytologic results had evidence of pleural disease. Whether VATS should be routinely performed in the setting of advanced ovarian cancer remains a topic of debate.
Diaphragmatic stripping or resection
At initial surgical exploration, diaphragmatic disease may be the largest-volume metastatic disease. Unfortunately, the presence of diaphragmatic disease is one of the most common factors precluding optimal tumor reduction surgery. Aletti and coworkers evaluated the therapeutic value of diaphragmatic surgery in patients with advanced ovarian cancer and found that patients who underwent diaphragmatic surgery (stripping of the diaphragm peritoneum, full- or partial-thickness diaphragm resection, or excision of nodules) had an improved 5-year OS rate relative to patients who did not undergo diaphragmatic surgery (53% vs. 15%, P < .0001).
A study by Pathiraja and colleagues compared the surgical morbidity of diaphragmatic peritonectomy versus full-thickness diaphragmatic resection with pleurectomy at radical debulking. A total of 42 patients were eligible for the study; 21 underwent diaphragmatic peritonectomy (DP, group 1) and 21 underwent diaphragmatic full-thickness resection (DR, group 2). Forty patients out of 42 (93%) had complete tumor resection with no residual disease. Histologic examination confirmed the presence of cancer in the diaphragmatic peritoneum of 19 patients out of 21 in group 1 and all 21 patients of group 2. The overall complications rate was 19% in group 1 versus 33% in group 2. The pleural effusion rate was 9.5% versus 14.5%, and the pneumothorax rate was 14.5% only in group 2. Two patients in each group required postoperative chest drains (9.5%). The authors concluded that patients in the pleurectomy group experienced pneumothorax and a higher rate of pleural effusion, but none had long-term morbidity or additional surgical interventions.
In a 2016 study by Muallem and colleagues, the authors reported on 268 patients who underwent diaphragmatic interventions. The comparison group was another 268 patients who did not undergo any diaphragmatic procedure. The surgical interventions varied between diaphragmatic partial resection (44.8%), stripping (53%), and only infrared coagulation (2.2%). The postoperative complication rate was higher in the diaphragm-intervention group compared with the group without any diaphragmatic intervention (49.6% vs. 38.8%); however, the authors recognized that most of the postoperative complications were not directly related to the diaphragmatic intervention itself. Pleural effusion was the only increased complication with a direct correlation with diaphragmatic surgery (25.4% vs. 14.2%).
Given the high risk for potential complications after diaphragmatic surgery, it would be of value to the surgeon to predict which patients would require a diaphragmatic resection to achieve optimal cytoreduction. A 2018 study evaluated the positive predictive value of preoperative imaging and the proportion of CT scans with false-negative results in patients who underwent cytoreductive surgery ( ). A total of 536 patients were analyzed in the study; diaphragmatic disease was found intraoperatively in 40.1%, and 16% underwent a diaphragmatic procedure. The positive predictive value for preoperative radiologic identification of diaphragmatic disease was 78.6%. CT imaging failed to detect diaphragmatic involvement despite obvious diaphragm disease during surgery in 29.4% of cases, thus leading to a low negative predictive value of 64.8%. The sensitivity and specificity for CT imaging in detecting diaphragm disease was 44.3% and 93.8%, respectively. The authors concluded that preoperative assessment with CT imaging is not reliable in accurately detecting diaphragmatic involvement.
For optimal cytoreductive surgery, a splenectomy may be required if there is disease involving the hilum, capsule, or parenchyma of the spleen. Magtibay and colleagues evaluated 112 patients who underwent splenectomy as part of primary or secondary cytoreductive surgery. They found that the most common indications for splenectomy were direct metastatic involvement (46%), facilitation of an en bloc resection of perisplenic disease (41%), and intraoperative trauma (13%). In that same study, the authors found that 65% of patients had hilar involvement, 52% capsular involvement, and 16% parenchymal metastases. Interestingly, patients with disease directly involving the splenic parenchyma did not have a worse prognosis than patients with disease involving the splenic hilum or capsule.
A 2018 large study by Sun and colleagues evaluated 2882 patients who underwent ovarian cancer cytoreductive surgery. Of these, a total of 38 patients (1.3%) underwent spleen resection. One patient underwent splenectomy because of trauma, but the remaining 37 patients had splenic metastases. Of these, 27 patients underwent splenectomy because of direct tumor spread in the spleen and 10 patients underwent resection because of hematogenous metastases. The authors concluded that splenectomy should be attempted, when feasible, to achieve optimal cytoreduction.
The clinical significance of hepatic parenchymal metastasis on survival in patients with advanced ovarian cancer has been studied. Lim and coworkers reported on a series of patients with hepatic parenchymal metastases. In this series, patients underwent wedge resection, segmentectomy, or hemihepatectomy as part of their tumor-reductive surgery. The 5-year PFS and OS rates for patients with stage IIIC disease and patients with stage IV disease and hepatic parenchymal metastasis from peritoneal seeding were 25% and 23% and 55% and 51%, respectively. The authors advocated that complete hepatic resection should be attempted for patients with hepatic parenchymal metastasis.
Because ovarian cancer often presents with confluent tumor in the cul de sac, rectosigmoid resection—along with or en bloc with hysterectomy and bilateral salpingo-oophorectomy—is often necessary to achieve complete tumor resection in the pelvis. This results in high rates of optimal cytoreduction, with acceptable morbidity. An average of 26% of women with ovarian cancer undergo colon resection as part of their primary cytoreductive operation according to a study by Aletti and colleagues. Peiretti and associates aimed to determine the impact of rectosigmoid resection, at the time of primary cytoreductive surgery, on morbidity and survival of patients with advanced ovarian cancer. A total of 238 patients were identified; 180 (75%) had stages IIC to IIIC and 58 (25%) had stage IV. Complete cytoreduction was achieved in 41% of the cases. Stapled coloproctostomy was performed in 98%, whereas hand-sewn coloproctostomy was performed in only 2%; a protective ileostomy and colostomy were necessary (constructed) in 2 (0.8%) and 5 (2%) cases, respectively. The complications associated to rectosigmoid resection were anastomotic leakage in 7 patients (3%) and pelvic abscess in 9 patients (3.7%) . Fifty percent of patients recurred during the study period, but only 5% of them showed a relapse at the level of the pelvis, whereas 8% presented with abdominal recurrence associated with pelvic disease as well. The median OS time among patients with complete cytoreduction was 72 months compared with 42 months among the rest of patients ( P = .002). The authors concluded that rectosigmoid colectomy may significantly contribute to a complete primary cytoreduction for advanced stage ovarian, tubal, and peritoneal cancers and that pelvic complete debulking accomplished by rectosigmoid resection could be associated with a lower rate of pelvic recurrence as well.
One of the main concerns after bowel resection is the risk of anastomotic leak, and this is particularly important in patients who have undergone multiple bowel resections. In a study by Grimm and colleagues, the authors aimed to identify risk factors for anastomotic leakage in patients undergoing primary surgery for advanced ovarian cancer. In that study, 30.1% of patients had multiple bowel resections. In a multivariate model after correcting for comorbidities, duration of surgery, patient age, and blood loss during pregnancy, the authors were not able to identify an independent predictor of anastomotic leak. The authors concluded that rate of anastomotic leak was mainly influenced by rectosigmoid resection and only marginally increased by additional bowel resections.
Whether systematic removal of retroperitoneal lymph nodes should be part of optimal cytoreductive surgery had been a topic of debate for many years. We now have evidence that systematic lymphadenectomy in patients with grossly uninvolved lymph nodes provides no benefit to the woman. A prospective randomized trial by Benedetti-Panici and colleagues showed that systematic lymphadenectomy improves PFS but not OS in women with optimally debulked advanced ovarian cancer. In addition, the median operating time was longer (300 vs. 210 minutes; P < .001) and the percentage of patients requiring blood transfusions was higher (72% vs. 59%; P = .006) in the systematic lymphadenectomy arm. A 2019 prospective randomized trial (LIONS [Lymphadenectomy in Ovarian Neoplasms]) assigned patients with newly diagnosed advanced ovarian cancer (FIGO stage IIB to IV) who underwent macroscopically complete resection and had normal lymph nodes both before and during surgery to either undergo or not undergo lymphadenectomy. A total of 647 patients underwent randomization; the results showed that systematic pelvic and paraaortic lymphadenectomy in patients with advanced ovarian cancer was not associated with longer overall or PFS compared with no lymphadenectomy and was associated with a higher incidence of postoperative complications ( ).
Postoperative therapy for advanced epithelial carcinomas (stages III and IV)
For historical interest, early adjuvant therapy attempts in advanced disease included single-agent and combination chemotherapy regimens based on the alkylating agents. A limited number of responses were observed, and treatment often continued for 1 to 3 years. With the discovery of cisplatin (and carboplatin, subsequently), several randomized trials were conducted comparing platinum and platinum combinations with nonplatinum regimens. These pivotal trials secured platinum as the agent of choice in primary adjuvant therapy, which continues to this day. In addition, several clinical trials have established that little additional benefit to treatment is observed beyond four cycles of therapy. Most recently, the development of the taxanes has documented the importance of this agent (discussed later). By convention, six to eight cycles of combination platinum- and taxane-based therapy are now recommended as adjuvant therapy for most patients with advanced disease.
The pivotal trial establishing the importance of paclitaxel in primary ovarian cancer management was reported by McGuire and associates on behalf of the GOG. They conducted a randomized trial comparing cisplatin, 75 mg/m 2 , with cyclophosphamide, 750 mg/m 2 , or paclitaxel, 135 mg/m 2 , over 24 hours and demonstrated a survival advantage in the paclitaxel arm. All patients had residual tumors larger than 1 cm after the primary operation. Response rates improved with paclitaxel relative to control in patients with measurable disease (73% vs. 60%). The median PFS was 18 months in the paclitaxel arm compared with 13 months in the platinum arm ( P < .001). OS was similarly improved (38 vs. 24 months; HR = 0.6; 95% CI, 0.5 to 0.8; P < .001).
The results of this study were confirmed in similar randomized clinical trials conducted worldwide. The taxane-platinum combination was generally considered to be the recommended first-line therapy for ovarian cancer. The platinum analog carboplatin was found to be less nephrotoxic and neurotoxic and easily administered without prehydration, thus shortening the time of infusion. After several randomized clinical studies demonstrating the equivalence of this agent to cisplatin in ovarian cancer, carboplatin was substituted for cisplatin in taxane-based regimens. In addition, paclitaxel infused over 3 hours was found likely to be equivalent to paclitaxel infused over 24 hours and, in combination with carboplatin, enabled the combination to be given on an outpatient basis. Phase III studies by Ozols and coworkers showed that paclitaxel-carboplatin is a feasible outpatient regimen with less toxicity than paclitaxel-cisplatin and is associated with equivalent survival.
It should be noted that carboplatin is quantitatively excreted by the kidney and its effective serum concentration can be calculated from a formula based on the woman’s glomerular filtration rate (GFR). This can be determined by various methods but is generally estimated by calculating the creatinine clearance. The Calvert formula is most commonly used and determines a total dose by this formula:
AUC-based dosing is preferred for carboplatin because the AUC most accurately reflects observed dose-specific toxicity and is more reliable across patients than dosing based on the body mass index. A usual dose for carboplatin is calculated for AUC values of 5 to 7.5. Both paclitaxel and platinum compounds are neurotoxic, as noted by Warner, and this is often the dose-limiting toxicity. The taxane, docetaxel, was found to be potentially less neurotoxic than paclitaxel. Vasey and colleagues reported on a large phase III study comparing docetaxel and carboplatin with paclitaxel and carboplatin in patients with stages IC to IV ovarian cancer. Almost identical survival parameters were observed between the two agents. The docetaxel arm was significantly less neurotoxic; however, it was associated with more myelosuppression. Neurotoxicity, as evaluated by several objective measures, returned to parity several months after treatment. Granulocyte colony-stimulating factor is occasionally needed to reduce the duration of significant neutropenia in these regimens. A commonly used regimen is paclitaxel, 175 mg/m 2 over 3 hours, or docetaxel, 75 mg/m 2 over 1 hour, and carboplatin (AUC, 5 to 6) given as a 1-hour infusion every 3 weeks. Premedication is required for both taxanes to combat hypersensitivity reactions, which have been attributed to the taxane itself and the carrier vehicle required to make these agents water soluble. In addition, steroid administration is necessary after treatment for docetaxel to combat fluid retention and effusion, a complication that may occur in as many as 25% of patients without prophylaxis.
Alterations in frontline treatment strategies
Although the preferred sequence in primary advanced ovarian cancer management is surgery followed by chemotherapy, several authors have attempted to take advantage of the disease’s intrinsic chemosensitivity to improve outcomes in patients with extensive disease. Two avenues have been pursued:
Neoadjuvant chemotherapy, in which, after biopsy or limited surgery, chemotherapy is administered for a reduced number of cycles (usually three to four) and an operation is planned for removal of the primary tumor (if present) and residual metastases
Interval cytoreduction, when an unsuccessful maximal attempt at cytoreduction is followed by a reduced number of chemotherapy cycles (usually three to four), followed by a second cytoreduction attempt
Both strategies are followed by three to four cycles of chemotherapy after surgery. This latter strategy has been evaluated in randomized clinical trials with conflicting results; the former was evaluated in a randomized evaluation.
Neoadjuvant chemotherapy is practiced as an alternative for patients thought to have substantial operative risk or preoperative disease distribution that could preclude optimal cytoreduction. Several authors have noted the potential benefits to this strategy, including the opportunity to allow for an improvement in performance status, decreasing operative morbidity through less extensive surgery, and increasing the opportunity to achieve an optimal result. Each of these goals has been demonstrated in several single-institution retrospective and prospective trials. In this approach, selected patients are given three to four cycles of combination chemotherapy before a planned surgical attempt. Adjuvant therapy to complete six to eight cycles is commonly administered. Because primary cytoreduction is deferred, concerns have been raised against a potential trade-off in survival benefit among patients who were not too infirm or harboring unresectable disease. Given these concerns, calls were made for prospective randomized controlled trials to examine the safety and efficacy. To date, three trials have been conducted with remarkably similar outcomes.
The first trial reported was the European Organization for Research and Treatment of Cancer (EORTC) 55971, a 670-patient randomized, noninferiority designed clinical trial; OS was the primary endpoint ( ). Of these patients, 632 (94.3%) were eligible and started the treatment. The majority had extensive stage IIIC or IV disease at primary debulking surgery (metastatic lesions that were larger than 5 cm in diameter in 74.5% of patients and larger than 10 cm in 61.6%). The largest residual tumor was 1 cm or less in diameter in 41.6% of patients after primary debulking and in 80.6% of patients after interval debulking. Postoperative rates of adverse effects and mortality tended to be higher after primary debulking than after interval debulking. The HR for death (intention/treat analysis) in the group assigned to neoadjuvant chemotherapy followed by interval debulking, compared with the group assigned to primary debulking surgery followed by chemotherapy, was 0.98 (90% CI, 0.84 to 1.13, P = .01 for noninferiority), and the HR for disease progression was 1.01 (90% CI, 0.89 to 1.15). Complete resection of all macroscopic disease (at primary or interval surgery) was the strongest independent variable in predicting OS.
In the second trial, Chemotherapy or Upfront Surgery (CHORUS), Kehoe and colleagues randomized 552 evaluable patients to primary cytoreduction followed by six cycles of platinum-taxane chemotherapy or three cycles of neoadjuvant platinum-based therapy followed by interval cytoreduction and three additional cycles of therapy ( ). The trial was designed to follow the eligibility and testing procedures of the EORTC trial with the intent of combining the databases for a future meta-analysis. In this trial 552 women with stage III or IV ovarian cancer were randomly allocated 1:1 to either primary surgery followed by six 3-week cycles of a carboplatin regimen (usually in combination with paclitaxel) or to three induction cycles of chemotherapy followed by surgery followed by three adjuvant cycles of chemotherapy. The primary endpoint was OS, which was similar (primary debulking surgery [PDS] 22.6 months vs. neoadjuvant chemotherapy (NACT) 24.1 months, upper bound of the one-sided 90% CI, 0.98), favored the NACT arm, unable to reject the noninferiority null hypothesis. High-grade adverse events related to surgery, including mortality, were higher in the PDS arm; however, like the EORTC trial, OS in the trial (both cohorts) was much lower than expected, reflecting low rates of complete cytoreduction (PDS, 17%; NACT, 39%) and heterogeneity in treatment care (76% of both arms receiving paclitaxel and carboplatin). Criticisms for both trials lie in patient selection and surgical effort potentially confounding the interpretation of contemporary management.
The third trial was conducted by the Japanese Clinical Oncology Group (JCOG-0602; ). Similar to the EORTC and CHORUS trials, this trial was designed to assess OS in a noninferiority design. In all, 301 patients were randomly allocated; unlike the previous two trials, patients who had a suboptimal cytoreduction at initial tumor reduction were allowed to undergo a second attempt. Of interest, this strategy was the focus of another randomized phase III trial conducted by the GOG (GOG-152), which failed to demonstrate any benefit on PFS or OS with a second surgical attempt in such patients ( ). Nevertheless, median OS was 44.3 months versus 49.0 months for chemotherapy versus PDS, respectively, corresponding to a HR for survival of 1.05 (90.8% CI, 0.83 to 1.33; P = .24 for noninferiority) favoring the latter. Heterogeneity in outcomes was seen in low-volume centers and patients with poor medical or performance characteristics. Compared with the EORTC and CHORUS trials, the absolute OS values are substantially longer in the JCOG trial, which is not completely clear from patient or operative characteristics among the three trials. Because the absolute PFS values are more aligned, the differential improvement in OS may reflect the recency of the latter trial and the availability of newer therapeutic agents.
Cytoreductive surgery performed after an initial failed attempt or in patients who were initially not considered candidates for cytoreductive surgery is referred to as interval cytoreductive surgery. This approach is interictally linked to the neoadjuvant chemotherapy approach described earlier, which has been largely the result of celiotomy. More recently, there has been an increased interest in the implementation of minimally invasive surgery when performing interval debulking surgery. In a study by Aletti and colleagues, the authors assessed the feasibility and early complication rates of minimally invasive surgery in 52 patients with stage III/IV ovarian cancer after undergoing neoadjuvant chemotherapy. The authors found that the median operative time was 285 minutes (range, 124 to 418) and the median estimated blood loss was 100 mL (range, 50 to 200). A microscopic residual was obtained in 96.6% of patients. No early postoperative complications were registered. Median time to restart of chemotherapy was 20 days (range, 10 to 30). The authors concluded that minimally invasive interval debulking surgery after neoadjuvant chemotherapy was safe and feasible. Subsequently, Fagotti and colleagues evaluated the role of minimally invasive surgery at the time of interval debulking surgery in different gynecologic cancer centers. A total of 127 patients were included. Among them, 96.1% of patients had no residual tumor. The rate of intraoperative complications was 4.7%, and 4.7% of patients experienced a postoperative short-term complication. The conversion rate to laparotomy was 3.9%. The authors concluded that minimally invasive surgery may be considered for the management of patients with advanced ovarian cancer when surgery is limited to low-complexity procedures.
One critical element when introducing a novel surgical modality or approach is ensuring that it does not compromise cancer-related outcomes. Although minimally invasive surgery has been shown to be safe and feasible, there is a paucity of prospective randomized data showing equivalence or noninferiority of minimally invasive interval debulking compared with open surgery. A study by Melamed and colleagues used the National Cancer Database to identify a cohort of patients diagnosed with stage IIIC and IV epithelial ovarian cancer who underwent neoadjuvant chemotherapy and interval debulking surgery by laparoscopy or open surgery and determined the 3-year survival rates, length of hospitalization, perioperative mortality, risk of readmission, and residual disease. The authors identified 3071 patients, of whom 450 (15%) underwent surgery initiated by laparoscopy. The authors found no difference in 3-year survival between patients undergoing laparoscopy (47.5%) and laparotomy (52.6%), P = .12). Postoperative hospitalization was shorter in the laparoscopy group (median 4 compared with 5 days, P < .001). Readmission (5.3% vs. 3.7%; P = .26), death within 90 days of surgery (2.8% vs. 2.9%, P = .93), and suboptimal debulking (20.6% vs. 22.6%, P = .29), did not differ between patients undergoing laparoscopy versus laparotomy.
Hyperthermic intraperitoneal chemotherapy
The value of hyperthermic intraperitoneal chemotherapy (HIPEC) in epithelial ovarian cancer appears to be inconclusive, as evidenced by contradictory and inconsistent results. Although it has been used routinely in the setting of metastasizing appendiceal cancer and later in colon cancer, the integration of HIPEC in the management of patients with ovarian cancer, particularly after upfront cytoreductive surgery, should remain experimental; however, there is evidence from prospective randomized data that HIPEC may have a role in treating patients undergoing interval cytoreductive surgery. In a 2018 multicenter, phase III randomized trial, van Driel and colleagues evaluated whether the addition of HIPEC to interval cytoreductive surgery would improve outcomes among patients who had undergone neoadjuvant chemotherapy for stage III epithelial ovarian cancer. A total of 245 patients with at least stable disease after neoadjuvant chemotherapy were randomly assigned to undergo interval cytoreductive surgery either with or without HIPEC. The authors found that the median survival was 33.9 months in the surgery group and 45.7 months in the surgery-plus-HIPEC group. Interestingly, they found that the percentage of patients who had adverse events was similar in the two groups (25% in the surgery group and 27% in the surgery-plus-HIPEC group). The authors concluded that the addition of HIPEC to interval cytoreductive surgery resulted in longer recurrence-free survival and OS than surgery alone and that it did not result in higher rates of side effects.
Additions to the paclitaxel and carboplatin backbone
It has been postulated that agents with nonoverlapping cross-resistance mechanisms or alternative mechanisms of action may be complementary opportunities in primary ovarian cancer patients to improve the therapeutic index. Several trials have been completed, with mixed results. The prevalent strategy has been to add to platinum and taxane therapy or substitute another agent for paclitaxel. The largest trial reported to date is GOG-182, which randomly allocated 4312 patients to one of four experimental arms against paclitaxel and carboplatin ( ). Two of the experimental arms involved a three-drug strategy (adding gemcitabine or pegylated liposomal doxorubicin to paclitaxel and carboplatin, with the latter triplet given every other course) and two others substituted topotecan or gemcitabine for paclitaxel for four of the eight planned cycles in a sequential administration design ( Fig. 33.14 ).