Follow-up routines after gynaecological cancer vary. The optimal approach is unknown, and no randomised-controlled trials comparing surveillance protocols have been published. In this chapter, we summarise the diagnostic performance of ultrasound, computed tomography, and magnetic resonance imaging in the follow up of women treated for ovarian or uterine cancers. Computed tomography is today the standard imaging method for the follow up of women treated for endometrial, cervical, or ovarian cancer. Six-monthly or annual follow-up examinations have not been shown to positively affect survival. Instead, a combination of transvaginal and transabdominal ultrasound examination with clinical examination might be a more cost-effective strategy for early detection of recurrence. Positron-emission tomography might play a role in women with clinical or serological suspicion of recurrence but without evidence of disease at conventional diagnostic imaging. To create guidelines, more studies, preferably randomised-controlled trials, on follow-up strategies are needed.
Introduction
Despite continuing advances in surgical and non-surgical therapeutic strategies, gynaecological malignancies have a high probability (30–75%) of developing relapse and distant metastases after initial treatment . In view of the low survival rate of women with recurrent disease, surveillance programmes mainly aim at early detection of recurrence, the rationale for this being that earlier diagnosis of relapse could be associated with lower morbidity and mortality rates. Other objectives of routine follow up are the identification of treatment complications and detection of second tumours associated with primary gynaecological cancer. How to best follow up women with gynaecological cancers is still unclear . Published data indicate that 41–100% of all recurrences are detected through symptoms alone (i.e. the recurrence would have been detected regardless of follow-up strategy) .
The methods usually used for postoperative surveillance of women with gynaecological cancer include clinical history, pelvic examination, visual vaginal inspection with vaginal cytology, and serum tumour markers. Computed tomography, magnetic resonance imaging (MRI) and positron emission tomography (PET) are also important tools in the follow up of gynaecological relapses . Tumour markers have a limited role, because both benign gynaecological conditions and malignant non-gynaecological conditions may be associated with elevated levels of tumour markers. Moreover, elevated levels of tumour marker do not provide any information about the location of recurrence . Computed tomography, MRI, and positron emission tomography combined with computed tomography are too expensive to be proposed as first-line methods to detect recurrence, and they have a poor ability to detect small lesions . Surprisingly, scant data are available on the utility of transabdominal and transvaginal ultrasound examination in the follow up of women with gynaecological cancer. The wide availability and the low cost of ultrasound, as well as the rapidity of the examination procedure and the possibility of bed-side use of ultrasound, would justify its use for surveillance of women treated for gynaecological cancer, provided that adequate diagnostic performance could be demonstrated.
The most appropriate follow-up strategy for each type of gynaecological cancer is likely to depend on the natural history of the disease, the risk of recurrence, the most frequent site of relapse, and the appearance of the recurrent tumour at imaging (i.e. discrete solid or cystic lesions or diffuse carcinomatosis).
Ovarian cancer
Epithelial ovarian cancer is the most common cause of mortality among gynaecologic malignancies . At diagnosis, most (75%) epithelial ovarian cancers have progressed to stage III or IV , and in women with this type of cancer, the 5-year survival rate is 15–20% . Many factors are associated with poor prognosis (e.g. age at diagnosis younger than 50 years or 50 years or older), International Federation of Gynaecology and Obstetrics (FIGO) stage, grade, histotype (serous or not serous), tumour size, and CA125 levels at diagnosis . The major determinants of outcome of these women are residual tumour after the first cytoreductive surgery and platinum sensitivity .
Recurrent ovarian malignancy usually appears as a pelvic mass at the site of surgery or as diffuse peritoneal carcinomatosis, pleuropulmonary lesions, lymph node, or liver metastases. Frequently, relapses of ovarian cancer are associated with ascites or pleural effusion, and sometimes these are the first manifestations of recurrence. Pelvic relapse may involve the vaginal cuff, the parametria, the bladder, the ureters, bowel loops, or the rectum . Ferrandina et al. reported that diffuse abdominal carcinomatosis was the most common presentation of recurrence (62.1%). A single tumour nodule or multiple tumour nodules were the first manifestation of recurrence in 9.9% and 26.7% of women, respectively . Ureteral obstruction may be caused by direct encasement of the ureter or by tumour infiltration of the bladder wall, which results in ureteral obstruction and hydronephrosis. Pelvic and para-aortic lymph nodes are also common sites of recurrence. Unusual sites of recurrence are the spleen, gastrointestinal tract, skeleton musculo-cutaneous tissue, and central nervous system.
Cervical cancer
Cervical cancer is the second most common cancer worldwide in women, resulting in about 275,000 deaths annually . Primary treatment for cervical cancer (surgery or radiotherapy with or without chemotherapy) has a cure rate of 80–95% in women diagnosed with early stage disease (stage I or II) and about 60% in women with stage III disease .
Surgico-pathologic factors that predict disease recurrence are FIGO stage, tumour volume, depth of stromal invasion, lymph-vascular space invasion, parametrial extension at surgery or on histology, histologic type (e.g. small-cell cancer has poor prognosis), lymph-node involvement at surgery or on histology, and status of the resection margins.
After radical surgery for early stage disease, metastases, when they occur, are typically distributed locally (e.g. in the vaginal apex), regionally (e.g. in the pelvic sidewall), or distantly (e.g. in the lung). If primary treatment involves pelvic radiation, the location of recurrences depends on the stage of the disease at diagnosis, with a greater risk of persistent or recurrent pelvic disease in those with more advanced tumours . The most common sites of distant recurrence are lung, para-aortic lymph nodes, abdominal cavity, and supraclavicular nodes . About 90% of women who have distant recurrences will die of their disease within 5 years .
Several studies have focused on prognostic factors of relapse rather than on strategies for detecting recurrence. The largest recently published study has shown that a surveillance programme based on routine Pap smear and chest X-ray does not permit earlier detection of recurrence and does not improves survival .
Endometrial cancer
Endometrial carcinoma is the most common invasive malignancy of the female genital tract . Most women with the disease are diagnosed at FIGO stage I or II , have a good prognosis, and can be cured with surgery. Women with advanced or recurrent disease have a worse prognosis, and results obtained with chemotherapy are far from impressive . The risk of recurrence varies according to well-defined risk factors (e.g. grading, histotype, metastases to the ovaries, lymph-node metastasis, and degree of myometrial stroma infiltration), and ranges from 13–17% in large studies including all women with endometrial cancer .
The anatomic sites of endometrial cancer relapse are equally distributed between local (pelvic) and distant (abdominal and chest) sites. Counselling on the symptoms of recurrence is extremely important because most women with recurrences are symptomatic (i.e. vaginal bleeding, abdominal or pelvic pain, swelling of the legs, fever, and bowel obstruction) . Most endometrial cancer recurrences occur within 3 years after the diagnosis of the primary tumour.
No rationale (e.g. examination accuracy, cost-effectiveness, or patient’s survival benefit) is available today for any particular follow-up protocol in women with endometrial cancer . Only physical examination, including pelvic-rectal examination, has shown some utility in detecting recurrence .
Uterine sarcoma
Leiomyosarcoma is the most frequent uterine sarcoma. It is an aggressive tumour with a tendency towards early relapse. Survival for women with recurrent disease is poor. Endometrial stromal sarcoma, the second most common uterine sarcoma, is a more indolent malignancy, with a tendency towards recurrence after a long latency period.
Because of the relative infrequency of both diseases, the study and standardisation of follow up and treatment for recurrent disease is still under debate. In properly selected women, surgery and even repeated surgery for recurrent disease should be considered, because it improves survival . A prolonged disease-free interval and complete resection of disease are associated with better prognosis after surgical resection of the recurrence .
Relapse of uterine sarcomas may be local or distant. In a study of 250 women with recurrent uterine sarcoma, the relapse was pelvic in 14% of cases, distant in 33%, and both pelvic and distant in 53% . Most distant relapses of uterine sarcomas involve the lungs and upper abdomen, whereas brain metastases are less common .
Ultrasound for follow up of women treated for gynaecological malignancy
In spite of major advances in the ultrasound technique (including colour Doppler and three-dimensional ultrasound), ultrasound is still not a standard method in the follow up of gynaecological malignancies. Only a few studies have investigated the ability of ultrasound to detect recurrent disease. This is surprising considering the wide availability of ultrasound in gynaecologic departments where ultrasound could be routinely added to a clinical pelvic examination when following up women treated for neoplasms in the genital tract.
Ultrasound and detection of recurrence of gynaecological cancer
Pioneering studies have shown a high accuracy for both transrectal (accuracy 90.5%) and transvaginal (accuracy 91.6%, sensitivity 89.4%, specificity 100%) ultrasound for the detection of pelvic recurrence of malignant disease. These results, however, are based on a small number of cases and a mixture of ovarian and uterine neoplasms.
Transabdominal and transvaginal ultrasound examination was used to detect relapses in a prospective study by Testa et al. , who included women with previous ovarian, endometrial, cervical cancer, or uterine sarcoma. This study included 385 women with recurrent gynaecological cancer submitted to surgery. All the women enrolled were subjected to clinical examination, serum tumour marker assay, transvaginal and transabdominal sonography, computed tomography scan, MRI, or both. The positive predictive value (PPV) of ultrasound was 100%, the negative predictive value (NPV) was 92.7%. In the subgroup of women with positive clinical examination or abnormal tumour markers, the PPV of ultrasound was 100%, and the NPV was 22.2%. In women without clinical or serological signs of tumour recurrence, the PPV of ultrasound was 100% and the NPV was 99.6%. This prospective study showed that routine ultrasound examination may play a role in the follow up of gynaecologic malignancies, especially among asymptomatic women. In women with clinical or serological signs of recurrence computed tomography might be better to localise the recurrence.
Ultrasound-guided biopsy of a lesion suspicious for malignancy is of great help if a recurrence of cancer is suspected (Video 1).
The following is the Supplementary data related to this article: Video 1
Transvaginal ultrasound of a pelvic recurrence of cervical cancer infiltrating the distal part of the ureter. Ureteral stent is in situ .
Ultrasound appearance of recurrent gynaecological malignancy
The ultrasound appearance of abdominal and pelvic intraperitoneal recurrences of gynaecological malignancies has been described as solid nodules vascularised at colour Doppler examination ( Fig. 1 ) or as carcinomatosis with miliariform or sheet-like patterns . In a selected group of 27 women treated for gynaecological malignancy with small masses suspicious of being recurrent tumour located in the central pelvis, Doppler results differed significantly between malignant and benign lesions. Higher colour score (i.e. higher colour content at colour Doppler examination), lower resistance index and higher peak systolic and time average maximum velocities in tumour vessels were found in malignant compared with benign lesions . In a study of 45 women with previous endometrial and cervical cancer, pelvic relapses were detected as a central pelvic lesion located on the vaginal apex in 80% of cases; in 13% of cases the recurrence was located at the pelvic side-wall and in 7% it occupied the whole pelvis . The lesions had diameters ranging from 1.0 to 6.8 cm. They appeared as a solid nodular mass in 98% of cases and as cystic mass in 2% of cases ( Figs. 2 and 3 ). In 32 (71%) women, the mass surrounding the bowel was hypoechogenic, whereas in 13 (29%), it showed inhomogeneous echotexture ( Fig. 2 ). At colour and power Doppler examination, randomly dispersed blood vessels were detected in all relapses ( Fig. 2 ).
Ascites or a septated fluid collection is often found in women with tumour recurrence, but may also be the only sonographic sign of recurrent disease, especially in women with recurrence of ovarian cancer .
In a prospective study by Testa et al. which included women with previous ovarian, endometrial, cervical cancer or uterine sarcoma, the recurrences appeared at ultrasound as a solitary lesion in 50.6% of women, and in 37% they were located centrally in the pelvis. Eighty-two per cent of the recurrences of ovarian cancer appeared as solid lesions. Recurrences of other gynaecological tumours appeared as solid masses in 94% of cases.
Tumour nodules in the liver or spleen, paraaortic or visceral metastatic lymph nodes can be detected with ultrasound as signs of gynaecological cancer recurrence . Hydronephrosis can be an indirect sign of pelvic or retroperitoneal recurrent disease (Video 2).
The following is the Supplementary data related to this article: Video 2
Transvaginal ultrasound-guided biopsy of a pelvic recurrence of ovarian cancer.
Ultrasound examination can provide ‘dynamic’ information related to elasticity, mobility and tenderness of lesions suspected to be cancer recurrences. This helps to discriminate between benign and malignant lesions. A description of the lesions’ margins (smooth or irregularly infiltrating into surrounding tissues) ( Figs. 4 and 5 ) is important. It helps to plan optimal management for referral to chemotherapy, radiotherapy or surgery .
Ultrasound for follow up of women treated for gynaecological malignancy
In spite of major advances in the ultrasound technique (including colour Doppler and three-dimensional ultrasound), ultrasound is still not a standard method in the follow up of gynaecological malignancies. Only a few studies have investigated the ability of ultrasound to detect recurrent disease. This is surprising considering the wide availability of ultrasound in gynaecologic departments where ultrasound could be routinely added to a clinical pelvic examination when following up women treated for neoplasms in the genital tract.
Ultrasound and detection of recurrence of gynaecological cancer
Pioneering studies have shown a high accuracy for both transrectal (accuracy 90.5%) and transvaginal (accuracy 91.6%, sensitivity 89.4%, specificity 100%) ultrasound for the detection of pelvic recurrence of malignant disease. These results, however, are based on a small number of cases and a mixture of ovarian and uterine neoplasms.
Transabdominal and transvaginal ultrasound examination was used to detect relapses in a prospective study by Testa et al. , who included women with previous ovarian, endometrial, cervical cancer, or uterine sarcoma. This study included 385 women with recurrent gynaecological cancer submitted to surgery. All the women enrolled were subjected to clinical examination, serum tumour marker assay, transvaginal and transabdominal sonography, computed tomography scan, MRI, or both. The positive predictive value (PPV) of ultrasound was 100%, the negative predictive value (NPV) was 92.7%. In the subgroup of women with positive clinical examination or abnormal tumour markers, the PPV of ultrasound was 100%, and the NPV was 22.2%. In women without clinical or serological signs of tumour recurrence, the PPV of ultrasound was 100% and the NPV was 99.6%. This prospective study showed that routine ultrasound examination may play a role in the follow up of gynaecologic malignancies, especially among asymptomatic women. In women with clinical or serological signs of recurrence computed tomography might be better to localise the recurrence.
Ultrasound-guided biopsy of a lesion suspicious for malignancy is of great help if a recurrence of cancer is suspected (Video 1).
The following is the Supplementary data related to this article: Video 1
Transvaginal ultrasound of a pelvic recurrence of cervical cancer infiltrating the distal part of the ureter. Ureteral stent is in situ .
Ultrasound appearance of recurrent gynaecological malignancy
The ultrasound appearance of abdominal and pelvic intraperitoneal recurrences of gynaecological malignancies has been described as solid nodules vascularised at colour Doppler examination ( Fig. 1 ) or as carcinomatosis with miliariform or sheet-like patterns . In a selected group of 27 women treated for gynaecological malignancy with small masses suspicious of being recurrent tumour located in the central pelvis, Doppler results differed significantly between malignant and benign lesions. Higher colour score (i.e. higher colour content at colour Doppler examination), lower resistance index and higher peak systolic and time average maximum velocities in tumour vessels were found in malignant compared with benign lesions . In a study of 45 women with previous endometrial and cervical cancer, pelvic relapses were detected as a central pelvic lesion located on the vaginal apex in 80% of cases; in 13% of cases the recurrence was located at the pelvic side-wall and in 7% it occupied the whole pelvis . The lesions had diameters ranging from 1.0 to 6.8 cm. They appeared as a solid nodular mass in 98% of cases and as cystic mass in 2% of cases ( Figs. 2 and 3 ). In 32 (71%) women, the mass surrounding the bowel was hypoechogenic, whereas in 13 (29%), it showed inhomogeneous echotexture ( Fig. 2 ). At colour and power Doppler examination, randomly dispersed blood vessels were detected in all relapses ( Fig. 2 ).
Ascites or a septated fluid collection is often found in women with tumour recurrence, but may also be the only sonographic sign of recurrent disease, especially in women with recurrence of ovarian cancer .
In a prospective study by Testa et al. which included women with previous ovarian, endometrial, cervical cancer or uterine sarcoma, the recurrences appeared at ultrasound as a solitary lesion in 50.6% of women, and in 37% they were located centrally in the pelvis. Eighty-two per cent of the recurrences of ovarian cancer appeared as solid lesions. Recurrences of other gynaecological tumours appeared as solid masses in 94% of cases.
Tumour nodules in the liver or spleen, paraaortic or visceral metastatic lymph nodes can be detected with ultrasound as signs of gynaecological cancer recurrence . Hydronephrosis can be an indirect sign of pelvic or retroperitoneal recurrent disease (Video 2).
The following is the Supplementary data related to this article: Video 2
Transvaginal ultrasound-guided biopsy of a pelvic recurrence of ovarian cancer.
Ultrasound examination can provide ‘dynamic’ information related to elasticity, mobility and tenderness of lesions suspected to be cancer recurrences. This helps to discriminate between benign and malignant lesions. A description of the lesions’ margins (smooth or irregularly infiltrating into surrounding tissues) ( Figs. 4 and 5 ) is important. It helps to plan optimal management for referral to chemotherapy, radiotherapy or surgery .
Computed tomography for follow up of women treated for gynaecological malignancy
Thanks to its wide availability, its time-effectiveness and its high spatial resolution, computed tomography is the standard method for following up women who have been treated for gynaecological malignancies. The ability of computed tomography to assess accurately the extent of tumour in women with malignancies has increased after the introduction of thin section multi-detector scanners . Computed tomography can be used to investigate the woman’s whole body in less than 5 minutes: brain parenchyma, lung parenchyma, mediastinal structures, parenchymatous abdominal organs, the peritoneum, retroperitoneal structures, and the skeleton can be examined for the presence of metastases. The main disadvantages of computed tomography are ionising radiation and the necessity to administer intravenous iodinated contrast. Intravenous contrast is needed to increase tissue contrast resolution. Contrast, however, should be avoided in case of renal impairment, and may cause allergic reactions. Computed tomography is particularly useful for detecting distant recurrence of gynaecological malignancies. It has limited ability to detect local recurrence.
Computed tomography and detection of recurrence
Computed tomography is the most frequently used imaging method for surveillance of women after primary treatment for ovarian cancer . A large meta-analysis showed a pooled sensitivity of 79% and a pooled specificity of 84% for computed tomography for detection of recurrent ovarian cancer, and the area under the receiver operating characteristic curve was 0.88. Another study , compared the computed tomography findings in 35 women with clinical, laboratory signs of recurrent ovarian cancer, or both, with subsequent surgical data. It showed that computed tomography had a sensitivity of 58% and a specificity of 82% for identifying the precise abdominal sites of recurrence. The results of another study suggested that computed tomography can reliably predict tumour non-resectability in recurrent ovarian cancer . The main limitation of computed tomography for identifying recurrent ovarian cancer is its poor ability to detect small peritoneal implants. The reported sensitivity of computed tomography for detecting peritoneal implants smaller than 1 cm varies from 25–50% . It can be improved, however, by administration of a negative oral contrast medium, in addition to the intravenous one, to distend bowel loops and so increase the conspicuousness of solid implants .
Computed tomography is often considered the imaging modality of choice for the follow up of women treated for cervical and endometrial cancer . No reliable data on its sensitivity and specificity in detecting recurrent cervical or endometrial cancer have recently been published. Encouraging results about the use of computed tomography for the detection of recurrent uterine cancer were published in 1989 (e.g. accuracy, specificity and sensitivity 92%, 80% and 92%, respectively) . The main limits of computed tomography in the surveillance of women treated for endometrial and cervical cancer is its poor performance in identifying local recurrence. Use of computed tomography makes it difficult to discriminate between postoperative and post-radiotherapy fibrosis and local tumour recurrence.
Computed tomography and morphological appearance of recurrence
Brain metastases are rare in gynaecologic malignancies, and are usually clinically suspected because of neurological symptoms . Brain metastases are usually located at the white–gray matter interface. In most cases, they are isodense to brain parenchyma in scans carried out before contrast injection. After contrast injection, they show variable patterns of enhancement as a consequence of blood–brain barrier damage. Hypodensity of the adjacent white matter, representing hypoxic oedema, is usually present.
Pleural effusion is the most common manifestation of thoracic involvement by ovarian cancer relapse, and can be easily identified at computed tomography as the presence of hypodense (about 0 Hounsfield Units) material occupying the dependent portion of the pleural cavity ( Fig. 6 ). Solid pleural nodules or pleural thickenings may also be present. Pulmonary parenchyma involvement is seen as rounded solid nodules, variable in size, and usually located at the tip of vascular structures in the dependent pulmonary regions. Calcifications may be present if the primary tumour is a serous ovarian carcinoma. Enlarged round lymph nodes with loss of their fatty hilum indicates mediastinal involvement ( Fig. 6 ). A less common pattern of thoracic involvement is lymphangitic carcinomatosis, resulting in irregular, nodular thickening of pulmonary parenchyma interstitial septae.
Liver metastases are barely recognisable at computed tomography before contrast injection. In the portal venous phase after contrast administration, they appear hypodense compared with the liver parenchyma. Metastases from mucinous ovarian tumours usually appear markedly hypodense before contrast injection, and show a progressive peripheral rim enhancement after contrast injection. Calcifications may be present in liver metastases from serous ovarian carcinomas. Spleen, kidney, pancreas, and adrenal metastases are rare but can be detected at computed tomography.
Lymph-node metastases must be suspected in cases of enlarged rounded lymph nodes (more than 1 cm in minimum diameter) with a loss of their fatty hilum ( Fig. 7 ). A hypodense area (about 0–20 Hounsfield Units) representing necrosis in a lymph node has a positive predictive value of 100% for metastasis. Computed tomography cannot detect micrometastases in lymph nodes.
