Cervical cancer remains an important public health concern in the United States with an estimated 12,200 new cases and 4,210 deaths in 2011.1 Recurrent cervical cancer can be deadly. Early detection of central recurrence may improve the chance for survival, and Pap testing may facilitate detection. However, there are no prospectively validated surveillance methods. The National Comprehensive Cancer Network guidelines for surveillance after cervical cancer treatment recommend Pap testing every 3–6 months for the first 2 years, then every 6 months for 3 years, and then annually thereafter.2 Previous retrospective studies evaluating cervicovaginal cytology have shown poor sensitivity3–7 in detection of recurrence, but all these studies were performed using conventional Pap tests. Most previous studies were undertaken before the adoption of the Bethesda system for cervicovaginal cytology reporting, which incorporates new categories including atypical squamous cells of undetermined significance (ASC-US), ASC that cannot exclude high-grade squamous intraepithelial lesions (ASC-H), and low-grade squamous intraepithelial lesions (LSIL). The significance of these abnormalities in cervical cancer survivors is poorly understood. In addition, the risk for high-grade cervical intraepithelial neoplasia (CIN) and vaginal intraepithelial neoplasia in women with high-grade intraepithelial lesions (HSIL) found at Pap testing is unclear. Thus, the appropriate management for cervical cancer survivors with abnormal cytology remains uncertain.
The objective of this study was to evaluate the utility of liquid-based cytology in detecting recurrent cervical cancer. A secondary objective was to determine the utility of colposcopy with or without biopsy done for patients with abnormal, but not frankly malignant, cytology after treatment.
MATERIALS AND METHODS
This was a retrospective multi-institution study evaluating all patients treated for cervical cancer at three institutions (Washington University in St. Louis, University of Kentucky, and University of Alabama at Birmingham) from January 1, 2000, to November 1, 2009. After obtaining Institutional Review Board approval at each institution, patients were identified through local cancer registries and patient databases. All women between the ages of 18 and 100 with a histologically confirmed diagnosis of cervical cancer treated at participating institutions during this time period were evaluated for inclusion. Patients were included who had at least one posttreatment liquid-based cytology or colposcopy. Decisions on whether to follow abnormal results with serial cytology or colposcopy were made by the managing physicians. Women with CIN and vaginal intraepithelial neoplasia were combined for analysis. Patients were excluded for lack of follow-up after cervical cancer treatment or for incomplete treatment data. Pap test results were read according to the Bethesda system for cervicovaginal cytologic diagnosis: normal, atypical squamous cells of uncertain significance (ASC-US), atypical squamous cells of uncertain significance with high-risk HPV positivity (ASC-US+HPV), low-grade squamous intraepithelial lesions (LSIL), high-grade squamous intraepithelial lesions (HSIL), atypical glandular cells of uncertain significance, atypical squamous cells favor high-grade (ASC-H), and favor neoplasia.
Distributions of the demographic and clinical characteristics were summarized with contingency tables. Although all the Pap data were tabulated in a test basis rather than a patient basis, the association between cytology result and other clinical characteristics (eg, stage, treatment) was assessed using the generalized estimating equations method to account for potential correlation among multiple Pap tests from the same patient.8 All the analyses were performed using SAS, and a P<.05 was used to indicate a significant effect.
We identified 929 women meeting inclusion criteria who contributed 4,167 cervical cytology results. The median age was 49 years (range, 25 to 100) and the mean number of Pap tests per patient was 4.5 (range, 1–18) with a median follow-up of 32 months (range, 2.5 to 118.2). At least one abnormal cytology result was found during follow-up in 312 (34%) women, representing 15% (626/4,167) of all Pap tests. Baseline data are reported in Table 1. Abnormal results included ASC-US in 296 (47%) or LSIL in 179 (29%), ASC-H in 59 (9%), HSIL in 55 (9%), atypical glandular cells in 14 (2%), and favor neoplasia in 23 (4%). The frequency of each cytologic diagnosis at each stage is shown in Table 2. As shown in Table 2, when compared with women with lower-stage cancers, patients with previously treated cervical cancers of higher stage were not statistically different (P=.32). The heterogeneity among institutions was also explored. Compared with the samples from Washington University, patients from the University of Alabama Birmingham were more likely to have abnormal Pap test results (odds ratio [OR] 2.34, 95% confidence interval [CI] 1.75–3.11, P<.001) whereas these from the University of Kentucky had similar results (OR 1.32, 95% CI 0.94–1.86, P=.11).
Abnormal Pap test reports led to 201 colposcopies in 135 women (15% of all women and 43% of those with abnormal Pap test results). As shown in Table 3, most patients with an ASC-US Pap test result, including those with positive assays for high-risk human papillomavirus (HPV), or LSIL Pap test results did not have colposcopy, whereas most women with HSIL and other higher-risk cytology results did (Table 3). Information on colposcopy was missing for 24 (3.8%) of the abnormal Pap test results. As shown in Table 4, the 201 colposcopic evaluations identified 45 women with CIN 2 or worse, 25 with CIN 3, and 12 with cancer. Only 5 of 475 (1%) women with ASC-US or LSIL had CIN 3.
We reviewed potential clinical correlates of abnormal cytology. Patients treated with radiation therapy had more abnormal Pap test results (391/2,646, 14.8%) compared with those treated with surgery alone (133/1,521, 8.7%); however, the type of treatment, surgery, or radiation or chemoradiation was not associated with colposcopy for abnormal Pap test results (P=.49). When stratified by stage and institution, patients treated with surgery were less likely to have abnormal results (OR 0.41, 95% CI 0.29–0.59, P<.001). The majority of patients with stage II or higher cancers were treated with radiation or chemoradiation and the majority of patients with stage IA1 or IA2 cancers were treated with surgery. The stage IB1 and IB2 group was more evenly divided between those patients treated with radiation and those patients treated with surgery. In this group, controlling for stage and institution, those patients treated with radiation had more abnormal Pap test results (17% compared with 9.4%, P<.001) (Table 5).
We attempted to determine the association between cancer recurrence and abnormal cytology. Cancer recurrence was documented in 147 (15.8%) women, with 12 cases (8.1% of recurrences) detected by Pap test. All the recurrences detected by Pap test were in the first 5 years after treatment. The latest recurrence detected by Pap test was 3.6 years after diagnosis. Most (8/12) were detected in the first 2 years. All except one of the women with recurrences had Pap test results of ASC-H or worse. One patient with AS-CUS+HPV had a visible lesion when called back for assessment 2 months after Pap testing. Colposcopy for cytology less than HSIL without a visible lesion on examination did not detect any recurrence or CIN 3. Of the six cases of recurrence detected on Pap test at Washington University in St. Louis, all were assessed by positron emission tomography. Five had distant disease on positron emission tomographic scan, all opted for chemotherapy, and all ultimately died of disease. One had disease limited to the pelvis, but this included positron emission tomography 5-fluoro-deoxyglucose avid pelvic lymph nodes. She was offered an exenteration but declined and underwent Syed needle radiotherapy and chemotherapy. She died of her disease 11 months later. Of the two cases of recurrence detected by Pap test at the University of Alabama Birmingham, one patient had disease limited to the pelvis and underwent a total pelvic exenteration; she remains without evidence of disease after 2 years of follow-up. The second patient had metastatic disease in the chest on positron emission tomographic scan and elected to be treated elsewhere. Four cases of recurrence were detected by Pap test at the University of Kentucky. Two patients had been treated initially with radical hysterectomy and received chemoradiation for recurrence and are without evidence of disease after 3 and 7 years of follow-up, respectively. One patient had local recurrence treated with chemotherapy, and she died of disease. The last patient had distant disease on computed tomographic scan and failed treatment with combination platinum-taxane therapy. She was alive with disease after 15 months of observation. Thus, Pap surveillance appears to have led to salvage for recurrence in 3 of 929 (0.3%) cervical cancer survivors. Assuming this was a representative sample of the population, 810 Pap tests would be required to detect at least one cancer with 90% probability.
Abnormal cervicovaginal cytology is a common problem among cervical cancer survivors, with nearly one-third of cervical cancer survivors having at least one abnormal Pap test result during the first years after treatment. This presents a significant survivorship burden. Despite the frequency of abnormality, clinicians at leading academic centers lack consensus on appropriate follow-up. Only 31% of abnormal Pap test results were evaluated by colposcopy in this group, and rates were lower for those women with lesser cytologic abnormalities. On the other hand, 20% of women with ASC-H, 25% of women with atypical glandular cells, 43% of women with HSIL, and 67% of women with Pap test results read as favor neoplasia who underwent colposcopy had CIN 2 or worse. For women without distant disease or limiting comorbidity, colposcopy is indicated for cervical cancer survivors with these grades of cytologic abnormality.
On the other hand, most abnormalities in our series did not reflect a cervical or vaginal recurrence of their cancer, or a significant new premalignant change. In the absence of symptoms or visible lesions, cervical cancer survivors with ASC-US and LSIL cytology can be followed without colposcopy.
Our data build on those of others. Larson et al3 evaluated 249 patients treated with radical hysterectomy and pelvic lymphadenectomy followed with a standard protocol of clinical history, physical examination, and vaginal cytology every 3 months for the first postoperative year, every 4 months for the second postoperative year, and then every 6 months for 3 years. In this study, five women (18% of recurrences) were diagnosed with vaginal cytology, and all except one patient had either symptoms or physical examination findings. This study reported vaginal cytology as either malignant or negative, so comparisons to our study are limited.
Likewise, Soisson et al4 reviewed the records of 203 patients treated for cervical cancer with radical hysterectomy and lymphadenectomy who had 2 years of follow up or more. Patients had the same surveillance protocol as the Larson et al study, with clinical history, physical examination, and Pap testing as previously described. In this study, 119 patients were treated with postoperative radiation for high-risk disease. Vaginal cytology was considered positive only if malignant cells were identified. Four patients with recurrence were noted to have malignant cytology. Two of these patients had no symptoms or examination findings and Pap testing was the only abnormality in their screening. Subsequently, both received radiation for central recurrence and were cured. The authors noted that in their population, 2,639 Pap tests were required to detect two cancers amenable to treatment. In our study, 812 Pap tests were required to detect one cancer, and of those cancers detected only three were a central recurrence treated with apparent cure. In contrast, Bodurka-Bevers et al6 reviewed 1,026 cervical cancer patient records from 1983 to 1993 and found that traditional Pap testing did not detect a single asymptomatic recurrence. Without data on colposcopy from previous studies, we cannot make a direct comparison in detection rates between traditional Pap testing and liquid-based cytology; however, given the poor detection rate of liquid-based cytology in our study, it seems there was no obvious improvement.
The strengths of this review are its large number of patients from multiple institutions, the completeness of the data, and the length of follow-up. It is limited, however, by its retrospective nature and by limited use of colposcopy and long interval of data collection, during which time practices may have changed. In particular, we cannot exclude the possibility that women with ASC-US or LSIL could have developed local recurrence over further surveillance. For this reason, colposcopy may be indicated for cervical cancer survivors with persistent ASC-US or LSIL.
Persistence of carcinogenic HPV in women treated for cervical cancer is a sensitive indicator of risk for recurrence.10 Our data show that carcinogenic HPV is common in women with cervical cancer, as reflected by HPV DNA positivity in women with ASC-US and by LSIL results. Prospective studies are needed to explore the specificity, negative predictive value, and cost-effectiveness of HPV testing as a surveillance strategy for cervical cancer before it can be adopted clinically.
Our data suggest that Pap testing has limited utility in the detection of recurrences, with only 8% of recurrences detected by Pap and only three patients salvaged. The cost of cytologic surveillance, including colposcopy when necessary, may outweigh the benefit of detection. The reality of limited resources and need for cost containment suggests that continued cervical cytology surveillance should be questioned. Cost-effectiveness analysis comparing surveillance strategies including liquid-based cytology, clinical examination, patient education regarding symptoms, and positron emission or computed tomographic imaging or both, is needed. Prospective surveillance trials are also needed to examine new strategies for earlier detection of recurrence.
Despite the lack of data on this subject, cytology should remain a component of surveillance.2 In patients treated for cervical cancer, regardless of stage or treatment modality, the low yield of colposcopy and biopsy for cytology less than ASC-H in our data suggests that in the absence of a visible lesion, patients with ASC-US or LSIL can be followed without colposcopy unless abnormalities persist. Women with ASC-H, HSIL, and other severe abnormalities deserve further evaluation with colposcopy and biopsy.
1. Jemal A, Siegal R, Xu J, Ward E. Cancer statistics, 2010. CA Cancer J Clin 2009;60:227–300.
2. National Comprehensive Cancer Network. NCCN clinical practice guidelines v1.2010, Cervical cancer. 2009: CERV-8. Available at: http://www.nccn.org/professionals/physician_gls/f_guidelines.asp
. Retrieved March 3, 2011.
3. Larson DM, Copeland LJ, Malone JM Jr, Stringer CA, Gershenson DM, Edwards CL. Diagnosis of recurrent cervical carcinoma after radical hysterectomy. Obstet Gynecol 1988;71:6–9.
4. Soisson AP, Geszler G, Soper JT, Berchuck A, Clarke-Pearson DL. A Comparison of symptomatology, physical examination, and vaginal cytology in the detection of recurrent cervical carcinoma after radical hysterectomy. Obstet Gynecol 1990;76:106–9.
5. Massad LS, Simples P. Cervicovaginal cytology in the posttreatment surveillance of women with cervical cancer. J Low Genit Tract Dis 1999;3:245–9.
6. Bodurka-Bevers D, Morris M, Eifel PJ, Levenback C, Bevers MW, Lucas KR, et al.. Posttherapy surveillance of women with cervical cancer: an outcomes analysis. Gynecol Oncol 2000;78:187–93.
7. Morice P, Deyrolle C, Rey A, Atallah D, Pautier P, Camatte S, et al.. Value of routine follow-up procedures for patients with stage I/II cervical cancer treated with combined surgery-radiation therapy. Ann Oncol 2004;15:218–23.
8. Liang KY, Zeger SL. Longitudinal data analysis using generalized linear models. Biometrika 1986;73:13–22.
9. Chien CR, Ting LL, Hsieh CY, Lai MS. Post-radiation Pap smear for Chinese patients with cervical cancer: a ten-year follow-up. Eur J Gynaecol Oncol 2005;26:619–22.
10. Song YJ, Kim JY, Lee SK, Lim HS, Lim MC, Seo SS, et al.. Persistent human papilloma virus DNA is associated with local recurrence after radiotherapy of uterine cervical cancer. Int J Cancer 2010 [Epub ahead of print].