Although gynecologic cancers account for only 10% of all new cancer cases in women, these cancers account for 20% of all female cancer survivors. Improvements in cancer care have resulted in almost 10 million cancer survivors, and this number is expected to grow. Therefore, determining the most cost-effective clinical surveillance for detection of recurrence is critical. Unfortunately, there has been a paucity of research in what are the most cost-effective strategies for surveillance once patients have achieved a complete response. Currently, most recommendations are based on retrospective studies and expert opinion. Taking a thorough history, performing a thorough examination, and educating cancer survivors about concerning symptoms is the most effective method for the detection of most gynecologic cancer recurrences. There is very little evidence that routine cytologic procedures or imaging improves the ability to detect gynecologic cancer recurrence at a stage that will impact cure or response rates to salvage therapy. This article will review the most recent data on surveillance for gynecologic cancer recurrence in women who have had a complete response to primary cancer therapy.
In 2010, gynecologic malignancies were expected to afflict approximately 80,000 women within the United States. Advances within the field of gynecologic oncology have resulted in long-term survivals and a high rate of survivors. Because long-term survival is becoming more common in this patient population, insights into cancer surveillance and detection of recurrence and addressing side-effects from treatment are of utmost importance.
Currently, posttreatment guidelines call for frequent visits immediately after treatment, followed by increasing intervals over time. Typically, after the first 2-3 years, patients are transitioned back to their primary care providers. However, primary care physicians may not be comfortable with guidelines or surveillance for each specific cancer type. This is in part due to a lack of training and in part to unclear expectations for the primary care provider by the oncologist. As survivorship continues to grow, coordination of care between gynecologic oncologists, primary care providers, other healthcare providers (such as radiation oncologists), and patients ideally will allow for compliance with cancer follow-up care and routine health maintenance. The provision of a clear understanding of recommendations and responsibilities of appropriate surveillance will reduce unnecessary tests and ultimately result in cost savings.
The role of surveillance is to provide clinical and cost-effective practices that detect recurrence and impact survival outcomes. Acceptance of surveillance should be considered if there is utility of treatment for recurrence and decreased morbidity from both monitoring for disease recurrence and treatment. One should also consider the costs and the use of resources for conducting these tests. Last, patients should be counseled on the benefits and pitfalls of disease monitoring, which should include the psychologic impact of surveillance programs. Unfortunately, most studies across all cancer sites are based predominantly on retrospective studies and provide limited insight into the true benefit of recommended guidelines for posttreatment surveillance. There is a real need for prospective studies to establish the most cost-effective methods for the detection of recurrent disease. In addition, surveillance tests should be directed at detecting recurrences that are amenable to curative or significant palliative treatment. Therefore, the primary objective of this review is to provide the most recent data on surveillance for cancer recurrence in women who have had a complete response to primary cancer therapy for gynecologic malignancies. Additionally, we have included routine health screening guidelines to allow for enhanced communication between oncologists and primary care providers.
Endometrial cancer
Endometrial cancer is the most common gynecologic cancer and the fourth most common cancer in women. Yearly, there are approximately 44,000 new endometrial cancer diagnoses and 8000 deaths in the United States. Commonly, patients experience symptoms such as abnormal or postmenopausal bleeding, which warrant further investigation with ultrasound scanning and/or endometrial sampling. The combination of symptoms and diagnostic testing results in 83% of patients being diagnosed in the early stages of the disease. As a result of localized disease, 5-year survival rates exceed 95% for stage I and approach 83% overall. However, recurrence rates for patients with early-stage disease range from 2–15% and reach as high as 50% in advanced stages or in patients with aggressive histologic condition. Many local recurrences from endometrial cancer are curable; therefore, the determination of the ideal time interval and diagnostic tools for surveillance of recurrent endometrial cancer that can impact survival outcomes is critical.
Typically, surveillance guidelines are more intensive the first few years after diagnosis because many studies have shown that most (70-100%) recurrences occur within 3 years after primary treatment. Current guidelines of the National Comprehensive Cancer Network (NCCN) and the American Congress of Obstetricians and Gynecologists recommend physical examination every 3-6 months for 2 years, then every 6 months or annually. Further evaluation with vaginal cytologic evidence is recommended every 6 months for 2 years and annually thereafter. To date, there are no prospective studies that have evaluated the role of surveillance in endometrial cancer follow-up evaluation. Based on recommended guidelines and institutional practices, retrospective research and literature reviews comprise the best evidence that is available.
The most consistently used method for surveillance is the physical examination. This alone accounts for a high rate of detection that ranges from 35-68% of cases. Even more striking is that the combination of physical examination and symptoms has resulted in rates of detection that exceed 80%. In a recent literature review, Sartori et al report that only physical examination has shown utility in the detection of endometrial cancer recurrence. Therefore, physical examination, which includes a thorough speculum, pelvic, and rectovaginal examination, should be conducted during each follow-up assessment.
The role of surveillance is based on the concept that detection of recurrences in the asymptomatic stage results in better therapeutic options and outcomes. Interestingly, even in spite of intensive surveillance, many recurrences are detected based on the presence of symptoms, which occurs in 41-83% of patients. A common symptom, vaginal bleeding, is indicative of a local recurrence that is often curable if it is an isolated site of disease. However, other common symptoms include abdominal and/or pelvic pain, lethargy, and weight loss. Even in the face of monitoring for recurrence, patients who experience a distant recurrence are symptomatic in 70% of cases, such as coughing or headaches. Therefore, patient education about the signs and symptoms is a critical component of posttreatment care and may lead to the detection of recurrent disease.
Survival outcomes have been evaluated on the basis of the presence or absence of symptoms at the time of recurrence. In a report by Sartori et al, 52% of patients were diagnosed with recurrence after they had symptoms; these patients had a median postrecurrence survival of 7 months. This was significantly less than the 20-month survival that patients experienced if they were diagnosed with recurrence in an asymptomatic state that was based on examination or imaging. Several other series have evaluated the role of routine surveillance for the follow up evaluation of patients with stage I endometrial cancer and reported no difference in survival based on the presence or absence of symptoms. Of note, even patients who had symptoms were undergoing the recommended follow-up evaluations, which provided an argument against the use of routine surveillance. Although all of these studies were retrospective, they reiterate the importance of prospective trials to determine the true role and regimen for surveillance.
Because most recurrences occur at the vaginal cuff, the use of cytologic evaluation has been advocated. However, many gynecologic oncologists challenge this recommendation. Rates of recurrence detection on vaginal cytologic evidence range from 0–6.8%, even in asymptomatic patients. Although Berchuck et al and Owen and Duncan report that cytologic evaluation detected 25% of all recurrences and that cytologic evaluation alone detected only 3 of the 44 (7%) recurrences. Furthermore, in addition to a low yield of detection, Agboola et al reported that the use of vaginal cytologic evaluation at each visit resulted in a cost of $27,000 per case detected. Because most recurrences at the vaginal cuff can be found on examination, vaginal cytologic evaluation adds only significant healthcare costs without added benefit.
Similarly to ovarian cancer, the use of cancer antigen 125 (CA125) level has been investigated as a marker for recurrence. In asymptomatic patients with endometrial cancer, the use of CA125 levels accounted for 15% of detections. Rose et al reported that CA125 levels were elevated in more than one-half of the patients with advanced stage and/or high-grade histologic evidence and that of these patients most had an elevated pretreatment level. However, one must be aware of elevated CA125 levels because of other conditions or even previous radiotherapy. In addition, the role of CA125 levels for the detection of recurrence was negligible in patients with low-risk disease. At present, the use of CA125 levels should not be used routinely in patients with endometrial cancer but may be appropriate in select patients with advanced disease, serous histologic condition, or a CA125 level that is elevated before treatment.
The use of radiographic imaging has been suggested for the detection of recurrent disease. Because of low costs, chest radiographs have been advocated for the detection of asymptomatic recurrences, often on a semiannual or annual basis. The rate of detection for asymptomatic chest recurrences that are found on chest radiographs ranges from 0–20%. In another series, chest radiograph detected 7 asymptomatic pulmonary recurrences and accounted for 0.34% of all chest radiographs that were performed for surveillance, which indicates low utility for this tool. Although reports of isolated pulmonary recurrences, albeit rare, may be amenable to therapies that allow for long-term survival outcomes, the routine use of chest radiographs is not recommended.
In further evaluation of radiographic imaging for endometrial cancer surveillance, Fung Kee Fung et al conducted a review of the literature and found that only 5-21% of asymptomatic recurrences were found by computed tomography (CT) scans. Other studies have agreed that the role of CT scanning for asymptomatic patients is not warranted, because survival of patients with disease that is detected on CT scan, compared with clinical examination, did not differ significantly. To increase the detection of local recurrence, the use of pelvic ultrasound scans has also been reported. Although detection rates for local recurrence range from 4–31%, many of these recurrences were also detected on other diagnostic methods, which included physical examination. Therefore, the use of routine pelvic ultrasound and CT scanning is not advocated; however, these modalities may play a role in the evaluation of patients with symptoms, because the rates of detection approach 50% of cases.
More recently, attention has been focused on positron emission test (PET) ± CT scans for endometrial cancer recurrence. Park et al reported 100% sensitivity and 83% specificity when PET-CT scanning was used for suspected recurrence and 100% diagnostic accuracy in 64 asymptomatic patients. However, its use for routine screening has not been well studied, and larger prospective studies will determine whether PET/CT will have a role in endometrial cancer surveillance. In addition, the high cost of PET/CT may limit its use in routine surveillance ( Table 1 ).
| Method of detection | Type of cancer, % | ||
|---|---|---|---|
| Endometrial | Ovarian | Cervical | |
| Symptoms | 41-83 | — | 46-95 |
| Physical examination | 35-68 | 15-78 | 29-75 |
| Cytologic evidence | 0-7 | — | 0-17 |
| Chest radiograph | 0-20 | — | 20-47 |
| Cancer antigen 125 level | 15 | 62-74 | — |
| Computed tomography scan | 0-20 | 40-93 | 0-45 a |
| Positron emission test–computed tomography scan | 100 a | 45-100 | 86 |
In conclusion, most patients with endometrial cancer will be a low risk for recurrence, and more than one-half of all recurrences will be detected through symptoms alone. With the exception of local disease, recurrent endometrial cancer is associated with a poor prognosis, regardless of the time of detection. On the basis of the data, we recommend a surveillance regimen to include a thorough history and physical examination, which would include a speculum and pelvic examination, at scheduled intervals with further testing indicated to evaluate symptoms and abnormalities that are detected on examination. This approach may save valuable healthcare dollars. Cytologic evaluation and chest radiographs in asymptomatic women are not clearly beneficial. If patients do have a suspected recurrence, generally a CT scan of chest, abdomen, and pelvis or PET/CT scans may be performed to assess the extent of the disease ( Table 2 ) .
| Variable | Months | Years | |||
|---|---|---|---|---|---|
| 0-12 | 12-24 | 24-36 | 3-5 | >5 | |
| Review of symptoms and physical examination | |||||
| Low risk (stage IA grade 1 or 2) | Every 6 mo | Yearly | Yearly a | Yearly a | Yearly a |
| Intermediate risk (stage IB-II) | Every 3 mo | Every 6 mo | Every 6 mo b | Every 6 mo b | Yearly a |
| High risk (stage III/IV, serous or clear cell) | Every 3 mo | Every 3 mo | Every 6 mo | Every 6 mo | Yearly a |
| Papanicolaou test/cytologic evidence | Not indicated | Not indicated | Not indicated | Not indicated | Not indicated |
| Cancer antigen 125 | Insufficient data to support routine use | Insufficient data to support routine use | Insufficient data to support routine use | Insufficient data to support routine use | Insufficient data to support routine use |
| Radiographic imaging (chest x-ray, positron emission tomography/computed tomography, magnetic resonance imaging) | Insufficient data to support routine use | Insufficient data to support routine use | Insufficient data to support routine use | Insufficient data to support routine use | Insufficient data to support routine use |
| Recurrence suspected |
|
|
|
|
|
a May be followed by a generalist or gynecologic oncologist;
b Consider alternating visits with a generalist and gynecologic oncologist.
Ovarian cancer
Ovarian cancer affects almost 22,000 women each year in the United States and results in >13,000 deaths yearly. Although responsible for <30% of all gynecologic malignancies, ovarian cancer accounts for >50% of deaths. These results stem from a lack of accurate screening tools and symptoms that are vague and often not specific, which result in approximately 75% of patients being diagnosed with advanced disease. Since the 1970s, the median overall survival of patients with advanced ovarian cancer has increased from 20 months up to 65 months because of advances in surgery and chemotherapy.
Despite the achievement of a complete clinical response, recurrence rates remain high, occurring in 25% of patients with early-stage disease and >80% of patients with advanced disease. Although patients with recurrent ovarian cancer rarely are cured, patients can have significant responses to salvage treatments.
To detect recurrences, the NCCN guidelines for epithelial ovarian cancer, fallopian tube cancer, and primary peritoneal cancer recommend follow-up visits every 2-4 months for the first 2 years, followed by 6-month intervals for the next 3 years. At each visit, physical examination and identification of the CA125 level or corresponding tumor marker are recommended. Additionally, these guidelines advocate the use of radiographic imaging and laboratory testing, as clinically indicated. However, the impact of surveillance and guidelines are based predominantly on retrospective studies and expert opinions.
Because 26-50% of recurrences occur within the pelvis, a thorough physical examination is an important part of a patient’s follow-up care and should include a bimanual pelvic and rectovaginal examination. However, the rates of detection by physical examination vary significantly from 15-78%. Although physical examination is one of the most commonly used tools and is associated with low cost, the reproducibility is low and may not detect other common sites of disease recurrence, such as the retroperitoneal lymph nodes, upper abdominal organs, or lungs. Thus, in a patient with symptoms or tests that are concerning for recurrence, physical examination alone may not be sufficient.
Historically, second-look surgeries have been used to assess disease response to primary treatment. Despite negative findings, recurrence rates that range from 35-50% have been reported, and no benefit in overall survival was noted. Thus, this procedure fell out of favor and is used rarely today.
Since its discovery in 1981, the use of CA125 level for tumor recurrence has been evaluated extensively. Approximately 80% of epithelial tumors will have an elevated CA125 level at the time of diagnosis. Studies have shown that CA125 level correlates with disease status in most cases and is often elevated 2-5 months before clinical detection of relapse. Generally, the sensitivity and specificity for CA125 level and disease recurrence ranges from 62–94% and 91–100%, respectively. In 255 patients who had completed primary therapy, a CA125 level twice the upper limits of normal was consistent with disease progression in almost all patients who were evaluated. Santillan et al reported that CA125 levels with a persistently low level of increase, even within normal values of the test, were often consistent with tumor recurrence. However, other reports found that the detection of recurrent disease by CA125 level alone yielded no prognostic benefit and advocate the use of CA125 level for surveillance only after a discussion that would explain the interpretation of the test. Furthermore, in a recently completed prospective randomized trial, the European Organization for Research and Treatment of Cancer assessed the outcome of 527 patients who were treated for recurrent ovarian cancer based on CA125 level alone vs clinically evident recurrence. The overall survival outcome did not differ for either group, and the investigators concluded that routine measurement of CA125 level is not warranted for disease surveillance.
To improve early detection of recurrent disease, the role of radiographic imaging modalities has been investigated. In a retrospective analysis, surveillance with CT scans every 6 months for the first 2 years, followed by yearly intervals, demonstrated the ability to detect asymptomatic disease. The authors reported a higher rate of optimal secondary cytoreductive surgery and an improved overall survival in the group with recurrence detected asymptomatically, compared with the symptomatic recurrence. Other studies that have evaluated methods of surveillance for ovarian cancer have reported the sensitivity of CT scans to be 40-93% and the specificity to be 50-98% for recurrent disease. On the contrary, in a study of 412 patients, the use of surveillance techniques detected recurrence in 80% of patients with the following evaluations: examination (15%), imaging (27%), CA125 level (23%), and CA125 level and imaging in (35%). However, the authors did not find a difference in survival, regardless of the modality that was used. Ideally, prospective studies will help to determine the true role of interval CT scans in ovarian cancer surveillance.
Because CT scans may lack the ability to detect a small volume of disease, other imaging modalities have been reviewed. The use of magnetic resonance imaging has also been evaluated for its role in ovarian cancer surveillance. Although sensitivity ranges from 62-91% and specificity ranges from 40-100%, comparable detection rates to CT scans and increased costs have limited its generalized acceptance. Ultrasound scanning has also been investigated for ovarian cancer surveillance. Studies have shown sensitivity that ranged from 45-85% and specificity that ranged from 60-100%. However, because of user variability and limited visibility, this modality typically is not used for the evaluation of recurrent disease.
More recently, the use of PET-CT scans has been reported. Sensitivity varies from 45-100% and specificity ranges from 40-100%, although diagnostic accuracy rates approach as high as 95%. In patients with normal CA125 levels and clinical suspicion of disease (based on symptoms or surveillance CT scans), PET-CT was slightly more sensitive than CT scans for the detection of recurrent disease. Studies have shown that PET-CT will alter treatment in approximately 60% of patients with recurrent disease and many recommend PET-CT before secondary cytoreduction. However, the potential use of this modality for surveillance is limited, and currently the role of radiographic imaging is best reserved as a supplement to abnormalities in physical examination, CA125 levels, or symptoms.
Although improvements in primary treatment of ovarian cancer have occurred, outcomes after recurrence remain disappointing. Many physicians hypothesize that the detection of recurrence early potentially may improve the benefit of available treatments, especially surgery. Second-line therapies are rarely curative and often result in short-term progression-free survival. However, some patients, especially those who are good candidates for secondary surgical cytoreduction and/or those who remain platinum sensitive will have high response rates to salvage treatments. Until the ideal surveillance is determined, individualized patient plans that consist of a thorough assessment of symptoms and physical examination, which includes a pelvic examination, should be undertaken. The role for CA125 level monitoring should be discussed with patients. The pros and cons of imaging should be discussed with the patients who do not have an elevated CA125 level at the time of diagnosis. When a recurrence is suspected based on symptoms, examination, or CA125 level, a CT scan of the chest, abdomen, and pelvis should be obtained to determine the extent of the disease. PET scans are a useful adjunct when CT scans are indeterminate ( Table 3 ).
| Variable | Months | Years | |||
|---|---|---|---|---|---|
| 0-12 | 12-24 | 24-36 | 3-5 | >5 | |
| Review of symptoms and physical examination | Every 3 mo | Every 3 mo | Every 4-6 mo | Every 6 mo | Yearly a |
| Papanicolaou test/cytologic evidence | Not indicated | Not indicated | Not indicated | Not indicated | Not indicated |
| Cancer antigen 125 | Optional | Optional | Optional | Optional | Optional |
| Radiographic imaging (chest x-ray, positron emission tomography/computed tomography, magnetic resonance imaging) | Insufficient data to support routine use | Insufficient data to support routine use | Insufficient data to support routine use | Insufficient data to support routine use | Insufficient data to support routine use |
| Recurrence suspected | Computed tomography and/or positron emission tomography scan | Computed tomography and/or positron emission tomography scan | Computed tomography and/or positron emission tomography scan | Computed tomography and/or positron emission tomography scan | Computed tomography and/or positron emission tomography scan |
| Cancer antigen 125 | Cancer antigen 125 | Cancer antigen 125 | Cancer antigen 125 | Cancer antigen 125 | |
a May be followed by a generalist or gynecologic oncologist.
Low malignant potential (LMP) tumors
Tumors of LMP, also called borderline tumors, account for 10-20% of epithelial ovarian tumors; approximately 4000 cases are diagnosed annually. The average age of a woman at the time of diagnosis is 40-60 years, but a significant proportion of these tumors occur in women in their child-bearing years. In general, the prognosis for women with LMP tumors is quite good, and most women (especially those with stage I disease) are at a very low risk of recurrence. Recurrences tend to occur late, and, even in advanced stages, 70% of recurrences will be after 5 years, and 30% will be after 10 years. Many patients with recurrent LMP tumors can be salvaged with additional surgery, and <5% eventually progress to invasive cancers.
Current NCCN guidelines recommend physical examination, including pelvic examinations, CA125 level (if initially elevated), every 3-6 months and pelvic ultrasound scans for those women with fertility-sparing surgery. Complete hysterectomy with bilateral salpingo-oophorectomy is recommended once fertility is completed. However, there are no studies that suggest that this aggressive surveillance improves prognosis for women with LMP tumors.
Retrospective studies suggest that, in women who have undergone a complete hysterectomy with bilateral salpingo-oophorectomy and resection of all gross disease, surveillance should be similar that used for those women with invasive ovarian cancer. For patients who have undergone fertility-sparing surgery, either a unilateral salpingo-oophorectomy or a cystectomy, the risk of recurrence ranges from 7–30%. Current surveillance recommendations for women who have undergone fertility-preserving surgery are to undergo serial pelvic sonography because this is the most sensitive method of detection of recurrent disease in residual ovary. Ultrasound scanning with or without tumor markers is recommended on an every 6-month basis.
When recurrent disease is suspected, a CT scan of abdomen and pelvis is recommended to assess the extent of the disease. Because most women with LMP tumors can be salvaged with additional surgery, prompt attention to symptoms or physical examination abnormalities is important; however, there is no evidence that routine radiographic surveillance with CT scans is at all beneficial.
Germ cell and sex-cord stromal tumors of the ovary
Malignant germ cell tumors of the ovary are rare and account for 2.6% of all ovarian cancers. Most patients have abdominal pain and a palpable mass. Malignant germ cell tumors can produce serum tumor markers that can prove helpful in the diagnosis and posttreatment surveillance if they are elevated at the time of diagnosis. Alpha-fetoprotein can be produced by yolk sac tumors, embryonal carcinomas, polyembryomas, and immature teratomas. Human chorionic gonadotropin can be produced by choriocarcinomas, embryonal carcinomas, polyembryomas, and, in low levels, in some dysgerminomas. Lactate dehydrogenase can be a marker for dysgerminoma. Because these tumors tend to occur in young women and most are unilateral, fertility-sparing surgery has been used to include pelvic washings, unilateral salpingo-oophorectomy, peritoneal biopsies, omentectomy, and pelvic and paraaortic lymph node dissection. NCCN guidelines recommend observation for low-risk tumors such as stage I dysgerminomas and stage IA, grade 1 immature teratomas. All other malignant ovarian germ cell tumors in this country receive postoperative chemotherapy with bleomycin, etoposide, and platinum with excellent survival rates. However, in Europe some healthcare providers advocate observation of all stage I germ cell tumors.
Sex cord stromal tumors are rare and account for 1.2% of ovarian cancers. Sex cord stromal tumors of the ovary can also produce serum tumor markers such as estradiol, inhibin, Müllerian inhibitory substance, and testosterone. Granulosa cell tumors also have the possibility of late recurrence of disease, with a reported median time to recurrence of 4-6 years. Pelvic recurrence accounts for 30-45% of cases. Surveillance should include a thorough physical examination and serum tumor markers for an extended period of time because of reports of recurrence even 20 years after the initial diagnosis. The utility of imaging in sex-cord stromal tumors has not been proven, so imaging should be limited to patients with symptoms or concerning findings on physical examination.
Studies that have evaluated surveillance strategies for ovarian germ cell tumors and sex cord stromal tumors have not been performed; therefore, recommendations are based on expert opinion. NCCN guidelines for surveillance recommend tumor markers every 2-4 months for 2 years if the markers were elevated originally. Physical examination that includes bimanual examination may be less helpful than serum tumor markers, especially in adolescent patients. Although recurrences are rare and data about them in the gynecologic oncology literature are small in number, they typically occur in the first 2 years after treatment. Although prognosis for recurrent germ cell tumors is usually poor, there are potentially curative treatment options that are available with multiagent chemotherapy regimens and high-dose chemotherapy with autologous stem cell support. Recently, the American Society of Clinical Oncology issued guidelines for surveillance using serum tumor markers for men with testicular cancer. The recommendations were similar to the current NCCN guidelines for germ cell tumors of the ovary in the first 2 years, with the exception that the surveillance continues for 10 years after treatment because of a reported incidence of 50% of the recurrences occurring 5 years after treatment in men. The timing of surveillance imaging in ovarian germ cell tumors is less well-characterized. NCCN guidelines for germ cell testicular tumors recommend CT scans every 3-6 months for the first 2 years then every 6-12 months until 6 years after treatment for those who received chemotherapy alone. Because germ cell tumors of the ovary occur in young women, because serum tumor markers are very sensitive for the presence of disease, and because repeated CT scans can lead to significant radiation exposure over time, the argument could be made that imaging is not indicated without evidence of the elevation of serum tumor markers, clinical symptoms, or concerning findings on physical examination. In addition, in those patients without elevated tumor markers, radiologic assessment in the first 2 years can be helpful ( Table 4 ).
