OBJECTIVE: To estimate the effectiveness of the human papillomavirus (HPV) test performed after conization in predicting residual disease in patients who subsequently underwent hysterectomy.
METHODS: A total of 115 patients who underwent hysterectomy after conization caused by cervical intraepithelial neoplasia grade 3 (CIN 3) and microinvasive cervical cancer (IA1 cancer) were included in this prospective study. All patients underwent HPV testing with a liquid hybridization assay immediately before hysterectomy. Differences in sensitivity, specificity, and accuracy between resection margin and the HPV test in predicting residual disease in subsequent hysterectomy samples were estimated using the McNemar exact test.
RESULTS: Univariable analysis showed that age, parity, menopausal status, glandular extension, and severity of disease were not predictive for residual disease, but positive resection margin and positive HPV tests were significant factors for predicting residual disease. These factors were also significant in a multivariable analysis (positive resection margin 45.5%, odds ratio [OR] 3.09, 95% confidence interval [CI] 1.19–8.03, P=.021; positive HPV test 57.6%, OR 11.05, 95% CI 4.01–30.49, P<.001). With resection margin, the sensitivity, specificity, and accuracy in predicting residual disease were 75%, 53%, and 61%, respectively, whereas, with the HPV test, these values were 85%, 67%, and 73%, respectively (P=.454, .080, and .044, respectively). Of patients with positive resection margins, 79% of HPV-negative patients had no residual disease. Of patients with negative resection margins, no HPV-negative patient had residual disease.
CONCLUSION: The HPV test after conization was significantly more accurate than resection margin for predicting residual disease. The predictive value of resection margin for predicting residual disease was much improved when used in combination with the HPV test. Use of the HPV test is recommended for identifying patients for subsequent hysterectomy after conization for CIN 3 and IA1 cancer.
LEVEL OF EVIDENCE: III
The human papillomavirus test is significantly more accurate in predicting residual cervical intraepithelial neoplasia after conization than assessment of resection-margin status.
From the Department of Obstetrics and Gynecology, University of Ulsan College of Medicine, Asan Medical Center, Seoul, Korea.
Corresponding author: Jong-Hyeok Kim, MD, PhD, Obstetrics and Gynecology, College of Medicine, University of Ulsan, Asan Medical Center, #388-1 Poongnap-2 dong, Songpa-gu, Seoul, 138-736, Korea; e-mail: firstname.lastname@example.org.
Financial Disclosure The authors did not report any potential conflicts of interest.
Conization of the uterine cervix by procedures such as cold knife conization and loop electrosurgical excision procedure (LEEP) is considered an appropriate treatment for cervical intraepithelial neoplasia grade 3 (CIN 3) and microinvasive cervical cancer (IA1 cancer). However, residual disease after conization due to CIN 3 and IA1 cancer is found in 23–34% of patients who subsequently undergo hysterectomy.1 Therefore, accurate prediction of residual disease after conization is important for the conservative treatment and counseling of patients with CIN 3 and IA1 cancer, both for the physician and patient.
Although several demographic and clinicopathologic factors, including age, parity, menopausal status, severity of lesion, glandular extension, and resection margin, have been reported to be predictive for residual disease after conization,2 resection margin remains the gold-standard technique for prediction of residual disease after conization. However, residual disease can be found subsequently in up to 2–31% of patients with negative resection margins.3–10 This may be due to multiple lesions that were not resected during conization; by contrast, residual disease is not found in up to 10–60% of patients with positive resection margins.3–10 This may be because residual disease at the resection margin of the cervix after conization is eliminated by vaginal acidity and rapid cell turnover during cervical healing and because of frequent use of fulguration to produce hemostasis at the base of conization crater margins, which can destroy residual tumor cells.4 Therefore, resection margin is not sufficient for the prediction of residual disease after conization in a large proportion of patients, and a more accurate predictive factor is required.
Recently, the preconization human papillomavirus (HPV) test has been evaluated as a predictor of residual disease or recurrence of disease after conization in several studies,10,11 and a prehysterectomy HPV test has been proposed as a possible predictor of residual disease in some studies.12,13 High-risk HPV is known to cause up to 99.7% of cervical cancers and high-grade precursor lesions and is found in most of these lesions14,15; therefore, the presence of high-risk HPV after conization may be an accurate indicator of residual disease. The aim of this study was to estimate the role of the HPV test performed after conization (immediately before a hysterectomy) in predicting residual disease in subsequent hysterectomy samples.
MATERIALS AND METHODS
A total of 120 consecutive patients who underwent hysterectomy after conization for CIN 3 or IA1 cancer were enrolled in this prospective study from March 2007 to November 2008 at the Asan Medical Center (Seoul, Korea). Only those patients with positive HPV test results before conization were eligible for this study. All patients underwent the HPV test using the Hybrid Capture II system (Digene Diagnostics Inc., Valencia, CA) after conization (1 day before hysterectomy). Demographic data (including age, menopausal status, body mass index, and parity) and clinicopathologic data (including CIN degree, glandular extension, size and resection margin-status of conization specimen, HPV test results, and residual disease in subsequent hysterectomy samples) were obtained. The study protocol was approved by the institutional review board of the Asan Medical Center.
In all patients, LEEP was used for conization. Briefly, the procedure was as follows. The cervix was swabbed with an acetic acid solution to assist in locating the ectocervical margins of the lesion, and a 1 mL solution of local anesthetic was injected into the cervix at the 5 and 7 o’clock positions. A loop was selected according to the size of the area to be excised. The goal was to excise the complete cervical lesion via a single excision for better orientation and margin-status interpretation. The base of the resulting crater and resection margin was coagulated and cauterized using a ball diathermy. A suture was placed at the 12 o’clock position of the LEEP specimen for orientation, the inner surface was inked, and the specimens were fixed in 10% formalin for pathologic examination. Cone specimens were sectioned. Paraffin blocks were cut at 5-micrometer intervals and stained with hematoxylin and eosin. The specimens were assessed for severity of lesion, margin status (exocervical or endocervical, clear or involved), and glandular involvement (present or absent).
Cervical samples for the Hybrid Capture II test were obtained using a cytobrush (Digene Cervical Sampler, Digene Diagnostics, Inc., Valencia, CA), transferred to a vial containing Digene Specimen Transport Medium (Digene Diagnostics, Inc.) and analyzed according to the manufacturer’s instructions. Light intensity was measured using a luminometer and expressed by comparing the relative light units of clinical samples with the positive control, a 1.0 pg/mL HPV 16 cutoff standard. A relative light unit:positive control ratio of 1 or more was considered a positive result. Of several HPV tests, the commercially available Hybrid Capture II is the only one approved by the U.S. Food and Drug Administration for HPV DNA detection and involves a liquid hybridization assay designed to detect 13 high-risk HPV types (HPV types 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, and 68). This test is widely used owing to its high sensitivity and predictive value (greater than 90–95%), objectivity, ease of use, and accessibility for use in routine clinical practice.16,17
After hysterectomy, the cervix was cut at 2-mm intervals perpendicular to the long axis of the cervical canal for pathologic evaluation. Residual disease was defined as any degree of CIN or invasive cancer.
Several factors, including age, parity, menopausal status, body mass index, the severity of disease (CIN 3 compared with IA1 cancer), glandular extension, resection margin of conization specimen, and HPV test results immediately before hysterectomy, were associated with residual disease in subsequent hysterectomy samples. Frequency distributions were compared using the χ2 and Fisher exact tests, and mean or median values were compared using the Student’s t- and Mann–Whitney U-tests. A logistic regression model was used to analyze the relationship between covariates and the probability of residual disease in subsequent hysterectomy samples. Differences in sensitivity, specificity, and accuracy between resection margin and the HPV test in predicting residual disease in subsequent hysterectomy samples were estimated using the McNemar exact test. P-values (from two-sided tests) less than .05 were considered significant. Data were analyzed using SPSS 10.0 software for Windows (SPSS Inc., Chicago, IL).
During the study period, 115 out of 120 enrolled patients met the inclusion criteria and were eligible for the study. Five patients were excluded because they had negative HPV test results before conization. Of them, four patients were diagnosed with CIN 3 and one patient was diagnosed with IA1 cancer after conization. Two of four patients with CIN 3 and one patient with IA1 cancer had positive resection margins on conization specimen. On hysterectomy specimen, only one patient with CIN 3 who had a positive resection margin on conization specimen had residual disease.
The characteristics of the 115 eligible patients are listed in Table 1. The mean age of patients was 47 years (range, 26–75), 39 patients (33.9%) were postmenopausal, and 81 (70.4%) had none, one, or two children. A total of 99 patients (86.1%) were diagnosed with CIN 3, and 16 (13.9%) were diagnosed with IA1 cancer. The mean (±standard deviation) size of conization specimens was 2.6 (±0.7) cm×1.6 (±0.6) cm×0.7 (±0.3) cm (width×length×height). Analysis of conization specimens showed glandular extension in 39 patients (33.9%), endocervical resection margins positive for CIN 3 or IA1 cancer in 53 patients (46.1%), and exocervical resection margins positive for CIN 3 or IA1 cancer in 29 patients (25.2%); 66 patients (57.4%) had resection margins positive for CIN 3 or IA1 cancer. The mean time interval from conization to hysterectomy was 37 days (range 9–95 days). Prehysterectomy HPV tests were positive in 59 patients (51.3%). Analysis of hysterectomy samples revealed residual disease in 40 patients (34 patients with CIN 3 and six patients with IA1 cancer) (34.8%). The mean size of residual disease was 6.8 mm (range 1–27 mm). Of 34 patients with CIN 3, residual disease was CIN 1 in three patients, CIN 2 in five patients, CIN 3 in 23 patients, and IA1 cancer in three patients. Of six patients with IA1 cancer, residual disease was CIN 3 in four patients and IA1 cancer in two patients.
Univariable analysis showed that age (younger than 50 years compared with 50 years or older), parity (two or fewer compared with more than two children), menopausal status, glandular extension, and severity of disease (CIN 3 compared with IA1 cancer) were not predictive for residual disease in subsequent hysterectomy samples. In women with positive glandular extension, the odds ratio (OR) to have residual disease after hysterectomy was 1.57 compared with women with negative glandular extension. However, it was not statistically significant (95% confidence interval [CI] for OR 0.68–3.63, P=.290). However, positive resection margins (OR 3.25, 95% CI 1.39–7.58, P=.006) and positive prehysterectomy HPV test results (OR 11.33, 95% CI 4.20–30.56, P<.001) were significant predictive factors for residual disease (Table 2). Positive resection margins (OR 3.09, 95% CI 1.19–8.03, P=.021) and positive HPV test results (OR 11.05, 95% CI 4.01–30.49, P<.001) were also significant predictive factors by multivariable analyses (Table 2). The sensitivity, specificity, and accuracy for prediction of residual disease were 75% (95% CI 58–87%), 53% (95% CI 40–64%), and 61% (95% CI 51–69%), respectively, with resection margin and 85% (95% CI 69–94%), 67% (95% CI 55–77%), and 73% (95% CI 65–81%), respectively, with the HPV test (P=.454, .080, and .044, respectively). Of resection margin–positive patients, 78.6% (95% CI 60.1–90.1%) of patients with negative HPV test results had no residual disease, but 63.2% (95% CI 47.2–76.7%) of those with positive HPV test results had residual disease (Table 3). Of resection margin–negative patients, no patients with negative HPV test results had residual disease, but 47.6% (95% CI 28.3–67.6%) of those with positive HPV test results had residual disease (Table 3).
In our study, multivariable analysis showed that resection margin and prehysterectomy HPV test results were significant predictive factors for residual disease after conization. The diagnostic accuracy of the prehysterectomy HPV test was significantly greater than that of resection margin. In addition, when used in combination with the HPV test, the predictive value of resection margin for residual disease was much improved.
Patients with CIN 3 and selected IA1 cancers often undergo conservative treatment involving conization, such as cold knife conization or LEEP.3,4,18 However, it is important to avoid any residual disease in the remaining cervix after conization. The resection-margin status of conization specimens has been proposed as an accurate predictive factor for residual disease after conization. However, residual disease can be found in up to 2–31% of resection margin–negative patients3–10 and is not found in up to 10–60% of resection margin–positive patients.3–10 Therefore, the identification of patients for hysterectomy based on the resection-margin status alone likely would result in overtreatment of many women and undertreatment of a small but significant proportion of women. In our series, 54.5% of patients were overtreated and 20.4% were undertreated based on the resection-margin status. The sensitivity, specificity, and accuracy of resection margins in predicting residual disease were 75%, 53%, and 61%, respectively. These figures are similar to those in previous reports.3–10 Therefore, more accurate predictive factors are required.
High-risk HPV is known to cause up to 99.7% of cervical cancers and high-grade precursor lesions and is found in most of these lesions.14,15 The HPV test has been approved as an additional cervical cytologic test in primary screening and as a follow-up test after conservative management of CIN and cervical cancer. Therefore, it is reasonable to assume that use of the high-risk HPV test after conization might be an accurate predictor of residual disease. This hypothesis is supported further by reports that effective conization can eliminate HPV DNA19 and that HPV DNA is rarely present in normal squamous epithelium adjacent to CIN.20 However, to our knowledge, only two studies have investigated the role of the prehysterectomy HPV test in predicting residual disease.12,13 Jain et al investigated the use of the Hybrid Capture II high-risk HPV test immediately before hysterectomy in 79 patients who underwent conization owing to CIN 3, and they correlated the resection-margin status and HPV test results with the presence of residual disease in subsequent hysterectomy specimens.12 In their series, no residual lesions were found in HPV-negative cases; hence, they report that the HPV test was associated with a negative predictive value of 100% for predicting residual disease.12 Lin et al investigated the use of the Hybrid Capture II high-risk HPV test immediately before hysterectomy in 75 patients who underwent conization owing to CIN 3 and had cone margins or endocervical curettage specimens showing disease, and they correlated the HPV test results with the presence of residual disease in subsequent hysterectomy specimens.13 In their series, both the sensitivity and negative predictive value of the HPV test were shown to be 100%.13 The potential role of prehysterectomy HPV testing in predicting residual disease was confirmed further in our series. The sensitivity, specificity, and accuracy of the HPV test (85%, 67%, and 73 %, respectively) were higher than those of resection margin (75%, 53%, and 61%, respectively). In resection margin–positive patients, the HPV test indicated that 78.6% did not have residual disease, and in resection margin–negative patients, the HPV test indicated that 47.6% had residual disease. No patient with a negative resection margin and a negative HPV test result was shown to have residual disease. When used in combination with resection margin, the diagnostic accuracy of the HPV test was increased.
Unlike previous reports,12,13 the sensitivity and negative predictive value of the HPV test were not 100% in our study. There are several reasons that the HPV test may not accurately detect the presence of residual disease in some patients. First, in rare situations, latent HPV infection can persist in a histologically normal cervix after conization. This phenomenon has been reported by Kanamori et al21 and is supported by reports that the HPV genotype detected in residual or recurrent disease after successful conization is the same as that detected before conization in most cases.22 Second, it is possible for a new HPV infection to occur after eradication of HPV DNA by conization but before hysterectomy; this is likely if patients have different HPV genotypes. Third, the timing of the HPV test may affect the results. The 2001 American Society for Colposcopy and Cervical Pathology guidelines recommend that the HPV test be performed at least 6 months after conization to provide sufficient time for clearance of the HPV infection.23 However, some studies have reported that the predictive value of the HPV test is not affected by the time after conization.13,24 For a more accurate evaluation of the role of the prehysterectomy HPV test in predicting residual disease, future studies should investigate the high-risk HPV genotypes and the HPV test should be performed at least 6 months after conization. However, care should be taken not to delay diagnosis and appropriate treatment of occult or residual invasive carcinoma.
In conclusion, the prehysterectomy HPV test is associated with significantly greater diagnostic accuracy in predicting residual disease after conization compared with resection margin. When used in combination with the HPV test, the predictive value of resection margin in predicting residual disease was increased. Therefore, use of the HPV test is recommended when selecting patients for hysterectomy after conization for CIN 3 and IA1 cancer.
1. Buxton EJ, Luesley DM, Wade-Evans T, Jordan JA. Residual disease after cone biopsy: completeness of excision and follow-up cytology as predictive factors. Obstet Gynecol 1987;70:529–32.
2. Lu CH, Liu FS, Tseng JJ, Ho ES. Predictive factors for residual disease in subsequent hysterectomy following conization for CIN III. Gynecol Oncol 2000;79:284–8.
3. Mohamed-Noor K, Quinn MA, Tan J. Outcomes after cervical cold knife conization with complete and incomplete excision of abnormal epithelium: a review of 699 cases. Gynecol Oncol 1997;67:34–8.
4. Moore BC, Higgins RV, Laurent SL, Marroum MC, Bellitt P. Predictive factors from cold knife conization for residual cervical intraepithelial neoplasia in subsequent hysterectomy. Am J Obstet Gynecol 1995;173:361–6.
5. Kalogirou D, Antoniou G, Karakitsos P, Botsis D, Kalogirou O, Giannikos L. Predictive factors used to justify hysterectomy after loop conization: increasing age and severity of disease. Eur J Gynaecol Oncol 1997;18:113–6.
6. Husseinzadeh N, Shbaro I, Wesseler T. Predictive value of cone margins and post-cone endocervical curettage with residual disease in subsequent hysterectomy. Gynecol Oncol 1989;33:198–200.
7. Lapaquette TK, Dinh TV, Hannigan EV, Doherty MG, Yandell RB, Buchanan VS. Management of patients with positive margins after cervical conization. Obstet Gynecol 1993;82:440–3.
8. Livasy CA, Maygarden SJ, Rajaratnam CT, Novotny DB. Predictors of recurrent dysplasia after a cervical loop electrocautery excision procedure for CIN-3: a study of margin, endocervical gland, and quadrant involvement. Mod Pathol 1999;12:233–8.
9. Phelps JY 3rd, Ward JA, Szigeti J 2nd, Bowland CH, Mayer AR. Cervical cone margins as a predictor for residual dysplasia in post-cone hysterectomy specimens. Obstet Gynecol 1994;84:128–30.
10. Park JY, Lee SM, Yoo CW, Kang S, Park SY, Seo SS. Risk factors predicting residual disease in subsequent hysterectomy following conization for cervical intraepithelial neoplasia (CIN) III and microinvasive cervical cancer. Gynecol Oncol 2007;107:39–44.
11. Park JY, Lee KH, Dong SM, Kang S, Park SY, Seo SS. The association of pre-conization high-risk HPV load and the persistence of HPV infection and persistence/recurrence of cervical intraepithelial neoplasia after conization. Gynecol Oncol 2008;108:549–54.
12. Jain S, Tseng CJ, Horng SG, Soong YK, Pao CC. Negative predictive value of human papillomavirus test following conization of the cervix uteri. Gynecol Oncol 2001;82:177–80.
13. Lin CT, Tseng CJ, Lai CH, Hsueh S, Huang KG, Huang HJ, et al. Value of human papillomavirus deoxyribonucleic acid testing after conization in the prediction of residual disease in the subsequent hysterectomy specimen. Am J Obstet Gynecol 2001;184:940–5.
14. Walboomers JM, Jacobs MV, Manos MM, Bosch FX, Kummer JA, Shah KV, et al. Human papillomavirus is a necessary cause of invasive cervical cancer worldwide. J Pathol 1999;189:12–9.
15. Schiffman MH, Bauer HM, Hoover RN, Glass AG, Cadell DM, Rush BB, et al. Epidemiologic evidence showing that human papillomavirus infection causes most cervical intraepithelial neoplasia. J Natl Cancer Inst 1993;85:958–64.
16. Clavel C, Cucherousset J, Lorenzato M, Caudroy S, Nou JM, Nazeyrollas P, et al. Negative human papillomavirus testing in normal smears selects a population at low risk for developing high-grade cervical lesions. Br J Cancer 2004;90:1803–8.
17. Nobbenhuis MA, Walboomers JM, Helmerhorst TJ, Rozendaal L, Remmink AJ, Risse EK, et al. Relation of human papillomavirus status to cervical lesions and consequences for cervical-cancer screening: a prospective study. Lancet 1999;354:20–5.
18. Felix JC, Muderspach LI, Duggan BD, Roman LD. The significance of positive margins in loop electrosurgical cone biopsies. Obstet Gynecol 1994;84:996–1000.
19. Elfgren K, Bistoletti P, Dillner L, Walboomers JM, Meijer CJ, Dillner J. Conization for cervical intraepithelial neoplasia is followed by disappearance of human papillomavirus deoxyribonucleic acid and a decline in serum and cervical mucus antibodies against human papillomavirus antigens. Am J Obstet Gynecol 1996;174:937–42.
20. Tate JE, Resnick M, Sheets EE, Crum CP. Absence of papillomavirus DNA in normal tissue adjacent to most cervical intraepithelial neoplasms. Obstet Gynecol 1996;88:257–60.
21. Kanamori Y, Kigawa J, Minagawa Y, Irie T, Oishi T, Itamochi H, et al. Residual disease and presence of human papillomavirus after conization. Oncology 1998;55:517–20.
22. Nuovo G, Moritz J, Kowalik A, Chalas E, Kaplan B, Mann W. Human papillomavirus types and cervical squamous intraepithelial lesions that recur after cold-knife conization. Gynecol Oncol 1992;46:304–8.
23. Wright TC Jr, Cox JT, Massad LS, Carlson J, Twiggs LB, Wilkinson EJ, et al 2001 consensus guidelines for the management of women with cervical intraepithelial neoplasia. Am J Obstet Gynecol 2003;189:295–304.
© 2009 by The American College of Obstetricians and Gynecologists.
24. Nobbenhuis MA, Meijer CJ, van den Brule AJ, Rozendaal L, Voorhorst FJ, Risse EK, et al. Addition of high-risk HPV testing improves the current guidelines on follow-up after treatment for cervical intraepithelial neoplasia. Br J Cancer 2001;84:796–801.