What is the role and purpose of follow-up for cervical cancer? Is the objective to identify early recurrence? If so, does this lead to a survival advantage? In the treatment of cervical cancer, after primary chemoradiation, few cases of recurrence are salvageable. Only those with central pelvic recurrences are candidates for potentially curative pelvic exenteration. Recurrences elsewhere are rarely amenable to curative therapy. Patients routinely undergo frequent follow-up, spanning years, when, to date, there have been no prospective studies to evaluate the benefits (or risks) of follow-up. History, physical examination (including pelvic examinations), smears of the cervix or vagina, in addition to various imaging studies, are commonly used. These are not only time consuming to the practitioner but potentially costly to the healthcare system with possible adverse psychological consequences for patients. Without proven benefits to follow-up, the use of this practice, applied universally to all patients, should be questioned.
In this issue of Cancer, Siva et al1 report a provocative study, challenging current standards in the post-therapy management of cervical cancer patients. This study, aimed at detecting relapse and determining survival outcomes, prospectively enrolled 105 women during a period of 5 years, with fluorine-18 (F-18) fluorodeoxyglucose (FDG) positron-emission tomography (PET) scans performed after chemoradiation for curative intent. The authors found that a complete metabolic response (CMR), noted in 73 patients, was associated with significantly decreased rates of distant failure and prolonged overall survival compared with those with only partial metabolic response (PMR). The overall survival of all patients at 3 and 5 years was 77% and 69%, respectively. In those who achieved a CMR on post-therapy FDG-PET, the overall survival rates were 95% and 91% at 3 and 5 years, respectively, demonstrating the prognostic importance of post-therapy FDG-PET.
The stated hypothesis was that the asymptomatic relapse rate at 5 years would be less than 5% in patients who achieved a CMR, making the clinical utility of pelvic examinations low. Indeed, this is what they found, with 6 patients of 73 experiencing relapse and only 1 patient relapsing without symptoms, discovered at routine follow-up, for an asymptomatic relapse rate of 1.6%. Due to the low rate of relapse after CMR, and the unlikelihood that routine physical examination would detect recurrence, the current standard of follow-up may be of less value in these patients, a finding particularly relevant to those patients who find pelvic examinations uncomfortable and anxiety-provoking.
The merits of this study include its generalizability, prospective nature, duration of follow-up, and the analysis of factors affecting prognosis. Patients were included with stage Ib to III disease, increasing the external validity of the study. Data points were prospectively collected, although no comparison group was available. The median follow-up was 3 years, during which time the majority of recurrences and disease-related deaths would be expected. Of greatest interest is the analysis revealing prognostic factors for recurrence and overall survival. FDG-PET was found to prognosticate for distant failure, nodal failure rate, and overall survival.
Although it may be possible to prognosticate with post-treatment FDG-PET, it is still uncertain whether its use may influence survival. Whether patients were converted to cure who otherwise might not have been is unclear. Furthermore, overall and cause-specific survival were measured from the time of FDG-PET, performed at the discretion of the clinician, rather than from completion of chemoradiotherapy, possibly introducing bias, as FDG-PET studies were performed anywhere between 1 and 12 months post-treatment.
At present, there is a paucity of data to guide physicians in the post-treatment follow-up of cervical malignancies. There are no prospective studies with cost-benefit analyses. Survival benefits have been suggested by only a few retrospective studies,2, 3 whereas others demonstrate equivalent outcomes between symptomatic and asymptomatic patients with recurrent disease.4, 5
Bodurka-Bevers et al2 reviewed records of 1096 patients treated for stage Ib cervical cancer. Of the 993 patients with recorded follow-up and disease remission, 461 underwent primary surgery, and 532 underwent primary radiotherapy. There were 133 recurrences (40 and 93 in the surgery and radiotherapy groups, respectively.) Of the recurrences, 7 and 12 were asymptomatic, respectively. All asymptomatic pelvic recurrences were detected by pelvic examination, and all asymptomatic pulmonary recurrences were detected by chest radiographs. Although the disease-free intervals were similar between symptomatic and asymptomatic patients (17 and 16 months, respectively), the median survival from time of recurrence and from diagnosis were significantly different (11 months vs 42 months [P < .001] and 31 months vs 83 months [P < .001], respectively). Symptom status at the time of recurrence was found to be a significant predictor of survival on multivariate analysis (P = .01). The authors suggest that the detection of recurrence in asymptomatic patients may impact survival, supporting the usefulness of post-treatment surveillance. However, only a small subset of patients may benefit, for which a cost-benefit analysis is lacking. Furthermore, as a retrospective study, although it may be possible to account for lead-time bias (where earlier diagnosis through screening artificially extends survival time), it is impossible to eliminate length-time bias (where patients with slowly progressive disease are more likely to be diagnosed).
A survival benefit from asymptomatic detection of recurrence was also noted by Zola et al3 in a study of 327 patients with recurrent cervical cancer. Patients had undergone primary radiotherapy, surgery ± adjuvant therapy, neoadjuvant chemotherapy/surgery ± adjuvant therapy, or chemoradiotherapy (11, 175, 119, and 22 patients, respectively), with 164 asymptomatic recurrences diagnosed at the time of routine follow-up. The difference in overall survival between those who recurred with and without symptoms was 37 months versus 109 months, respectively (P = .00001). Again, without a prospective analysis, the possibility of lead-time and length-time biases are significant, particularly when a difference in disease-free interval is noted (24 months for asymptomatic patients vs 36 months for symptomatic patients, P = .03).
A lack of benefit to post-treatment surveillance has conversely been reported in several retrospective studies.4, 5 Duyn et al4 performed an analysis of 277 cervical cancer patients to determine the percentage of recurrences detected during routine follow-up. Primary treatment was surgery, surgery/radiotherapy, radiotherapy, and chemotherapy in 153, 54, 66, and 4 patients, respectively. Of the 47 patients who developed recurrent disease, only 32% were detected during follow-up. Survival was not different between those who presented with recurrence by self-referral versus those diagnosed at routine follow-up. Similarly, Look et al5 reviewed 96 patients with stage Ib cervical carcinoma treated with either radiation, surgery, or a combination (55, 30, and 11 patients, respectively). There were 21 recurrences, 7 of which were asymptomatic. Time to recurrence and median survival after recurrence were similar for both symptomatic and asymptomatic patients.
If there is limited proven benefit to survival from patient follow-up at the completion of treatment, then why is it performed? Does it single out candidates for clinical trials? Is the purpose to identify and manage treatment-related toxicities? Or is it patient-driven, to provide reassurance that there is no evidence of recurrence until the next scheduled visit? With continually rising healthcare costs, current strategies and standards of care must be re-evaluated, and cost-effectiveness considered. The perspective and desires of the patient need to be taken into account as well.
To explore patient views and understanding of follow-up, Kew et al6 surveyed 104 consecutive gynecologic oncology patients. Women believed that the clinical examination was the most important component of their follow-up visit, with detection of recurrence ranked as the most important reason for attending (P < .0001). Anxiety was increased in 54% of women before their visit, with levels of anxiety unaffected by duration of follow-up. Two-thirds of the patients were followed either more frequently or for a longer duration than recommended by departmental guidelines; however, 91% believed that the frequency of their follow-up was appropriate.
Bradley et al7 reviewed 54 patients who had undergone successful treatment for early stage gynecologic malignancies. An attempt was made to identify factors predictive of follow-up preferences. Preferences were found to be highly individualized and unrelated to observable clinical details or disease factors. These findings advocate the use of informed decisions with communication between patients and specialists to tailor post-treatment follow-up. Given the choice, Howells et al8 found that 40% of patients would opt for an “open access” policy, allowing discharge from hospital follow-up with the option for self-rereferral as needed.
When patients express the need for continued follow-up, fear of recurrence is often the overriding reason, and medical reassurance is sought.9 However, in cases where recurrence is unlikely, fears tend to persist despite reassurance of low-risk disease. Follow-up is, therefore, often fulfilling a psychological rather than a medical need. The feeling of vulnerability to cancer recurrence frequently persists throughout the duration of follow-up, and the oncologist is often the only medical professional deemed appropriate to provide reassurance.9 If women could be prepared for the anticipated emotional reactions that may follow treatment, their feelings of anxiety and uncertainty might be reduced. Armed with a better understanding of their diagnosis, women could then be separated into high-risk or low-risk streams of follow-up depending on their individualized likelihood of disease recurrence and survival.
The results of Siva et al1 are supported by other works, where survival outcomes may be predicted by 3-month post-therapy FDG-PET. In Schwarz et al,10 the 3-year progression-free survival rates were 78%, 33%, and 0% for patients with complete metabolic response (n = 65), partial metabolic response (n = 15), and progressive disease (n = 12), respectively. This prospective study confirmed the retrospective analysis of Grigsby et al,11 who evaluated 152 patients after radiotherapy ± chemotherapy. Five-year cause-specific survival was 80% in 114 patients with no abnormal FDG uptake at any site compared with 32% for the 20 patients with persistent abnormal uptake. In the latter study, when FDG uptake was identified in previously unirradiated regions (n = 18), there were no survivors at 5 years.
Whereas FDG-PET scanning post-therapy may be able to predict prognosis (patients at risk for recurrence or decreased overall survival), this imaging modality could also aid in the detection of recurrent/persistent disease with the potential for a survival advantage. Unger et al12 found the sensitivity, specificity, and positive and negative predictive values of FDG-PET in asymptomatic patients to be 80%, 100%, 100%, and 88.9%, respectively, in the detection of recurrent disease. In a retrospective analysis of 249 women without evidence of recurrence, Ryu et al13 found these same values to be 90%, 76%, 35%, and 98%, respectively. In the detection of recurrent disease, Chung et al14 found the overall sensitivity and specificity of FDG-PET to be 96.1% and 84.4%, respectively. FDG-avid lesions were detected in 17 of 20 asymptomatic patients, allowing for curative intent treatment in 8, with complete response achieved in 5. FDG-PET imaging may prove to be a valuable adjunct to surveillance, when asymptomatic recurrent disease is discovered for which salvage is possible, or for use in refining the selection criteria for patients with recurrent disease who could benefit from salvage therapy.
The work of Siva et al1 brings to the forefront 2 timely questions in the management of post-treatment cervical cancer patients. The first relates to the role of follow-up, and the second to the role of FDG-PET within such follow-up. Low-risk follow-up may be possible for a subset of women identified by post-therapy FDG-PET, such as open-access policies or telephone interviews. FDG-PET may also identify a group of women better served by high-risk follow-up, which could include increased frequency of surveillance, enrollment in clinical trials, and the addition of supportive services to address psychological needs.
Patients value post-therapy surveillance. To them, the greatest value is in the early diagnosis of recurrent disease in hopes of lifesaving therapy. When a procedure or modality exists, such as FDG-PET, which allows for such early detection, and early treatment of recurrent disease is successful with improved survival outcomes, then follow-up will have achieved its purpose, both objectively (eg, overall survival, disease-free survival) and subjectively (eg, quality of life). Prospective, randomized studies will be vital to an understanding of the objective and subjective outcomes of minimalist versus intensive programs of follow-up with the addition of cost-effectiveness analyses when considering the addition of advanced and emerging technologies.