Predictors and clinical significance of the positive cone margin in cervical intraepithelial neoplasia III patients

Transcription

1 Chinese Medical Journal 2009;122(4): Original article Predictors and clinical significance of the positive cone margin in cervical intraepithelial neoplasia III patients SUN Xiao-guang, MA Shui-qing, ZHANG Jin-xia and WU Ming Keywords: conization; cervical intraepithelial neoplasia; factor analysis I Background Conization is being widely accepted for diagnosis and treatment of cervical intraepithelial neoplasia (CIN). There is controversy as to which factors are most predictive of a positive cone margin and the clinical significance of it. We conducted this study to identify the predictive factors and to evaluate the clinical significance of a positive cone margin in CIN III patients. Methods A retrospective review was conducted of 207 patients who had undergone conization due to CIN III from January 2003 to December 2005 at Peking Union Medical College Hospital. Of these, 67 had a subsequent hysterectomy. Univariate and multivariate analysis were utilized to define the predictive factors for a positive cone margin, and to compare the pathologic results of conization with subsequent hysterectomy. Results One hundred and fifty-one (72.9%) were margin free of CIN I or worse, 37 (17.9%) had CIN lesions close to the margin and 19 (9.2%) had margin involvement. A total of 56 cases (27.1%) had positive cone margins (defined as the presence of CIN at or close to the edge of a cone specimen). Univariate analysis showed that the parity, cytological grade, multi-quadrants of CIN III by punch biopsy, gland involvement, as well as the depth of conization were significant factors correlated with a positive cone margin (P <0.05). However the age, gravidity, grade of dysplasia in punch biopsy, as well as the cone methods were not significantly correlated (P >0.05). Multivariate analysis revealed that the cytological grade (OR=1.92), depth of conization (OR=2.03), parity (OR=3.02) and multi-quadrants of CIN III (OR=4.60) were significant predictors with increased risk for positive margin. The frequency of residual CIN I or worse in hysterectomy specimens was found to be 55.6% (20/36) in patients who were margin free, 71.4% (15/21) in patients with CIN occurring close to margin, and 80.0% (8/10) in patients with margin involvement. The frequency of residual CIN III or worse was found to be 13.9% (5/36), 23.8% (5/21) and 50.0% (5/10) respectively in different groups. Conclusions Cytological grade, depth of conization, parity and multi-quadrants of CIN III in punch biopsy were significant factors with increased risk in predicting a positive cone margin. Margin status of conization did not mean the presence or absence of CIN, but rather the varied frequency of residual CIN in specimens of subsequent hysterectomy. In view of this fact, it is suggested that the margin status of conization be a valuable surrogate marker for clinical management of CIN III. n recent years, an increased prevalence of cervical intraepithelial neoplasia (CIN) has been observed worldwide. Cervical conization, as both a diagnostic and therapeutic procedure, has been used for more than one hundred years and was popularized recently in the management of patients with CIN. 1 Conization shows special value with younger patients because it preserves reproductive function. Great attention had been focused on the possibility of complete excision of a CIN lesion by conization. In this study, we reviewed patients undergoing cervical conization for CIN III or worse with the objective to: identify the pre-operative predictors of a positive cone margin and evaluate the significance of positive cone margin in the management of CIN by analysis of residual disease present in specimens obtained upon subsequent hysterectomy. METHODS Patient enrollment Between January 2003 and December 2005, 240 patients received conization in the Department of Obstetrics and Gynecology, Peking Union Medical College Hospital. Of those, 33 were excluded in this study because of incomplete medical records (17 cases) and preoperative diagnosis of CIN II (16 cases). The medical records of the remaining 207 patients with CIN III or worse were retrospectively reviewed for analysis. Preoperative diagnosis was based on the histological result of colposcopically directed biopsy (punch biopsy). The median age, gravidity and parity of 207 patients were 37 (range ) years, 3 (range 0 9) times and 1 (range 0 5) time, respectively. DOI: /cma.j.issn Department of Obstetrics and Gynecology, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences & Peking Union Medical College, Beijing , China (Sun XG, Ma SQ, Zhang JX and Wu M) Correspondence to: Dr. SUN Xiao-guang, Department of Obstetrics and Gynecology, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences & Peking Union Medical College, Beijing ,China ( xgsqc@vip.sina. com)

2 368 Diagnostic procedure before conization Preoperative examination followed a three-step technique, e.g. cytological test, colposcopy with punch biopsy and histological diagnosis. 2 In our study, 195 cases underwent thinprep cytological test (TCT), and the results were reported according to the 2001 Bethesda System. 3 Nine received a Papanicolaou smear and 3 were without cytological examination. The indication for colposcopy followed the 2001 consensus guidelines for the management of women with cervical cytological abnormalities 4 with small modification: (1) TCT showed ASC-US with positive high risk types of HPV by Hybrid Capture 2 test. (2) 2 repeat ASC-US at 3 6 month intervals. (3) TCT showed LSIL or higher than LSIL. (4) Grade III by Pap smear. In our study, 204 patients received colposcopy with punch biopsy (98.5%), of those 96 (47.1%) had colposcopy in our hospital. The remaining 108 patients (52.9%) were examined at other hospitals. Three cases received an immediate biopsy without colposcopy because of cervical polyps, two of them were given a pathological diagnosis of CIN III, and the third was diagnosed as carcinoma in situ (CIS). All 207 patients had histological results prior to conization. The pathologic slides from other hospitals were reviewed by two senior pathologists in our hospital. The cervical quadrants were divided by two straight lines through the orifice of the cervix, one line was from 1:30 to 7:30 of the cervix, and the other line was drawn perpendicular to the first line. Three or more quadrants involved by CIN III or by CINIII in the gland were defined as multi-quadrant involvement. Conization and pathology Two types of conization were performed, 142 patients (68.6%) underwent electrosurgical needle conization, whereas 65 (31.4%) received cold knife conization. For each method, the ectocervical edge of the cone was located 0.5 cm outside of the iodine negative area. The depth of the cone varied according to patients age, reproductive desire, the location of the transformational zone and purpose of the cone (diagnostic or therapeutic). The depth of cone with 1.5 cm was seen in 46 patients (22.2%), cm in 89 patients (43.0%), cm in 61 patients (29.5%), 2.6 cm in all other patients (5.3%). A fresh specimen was marked by a thread at 12 o clock for orientation. The specimen was prepared and evaluated at the department of pathology, Peking Union Medical College Hospital. After measurement of cone width and depth, the specimen was fixed in 13% formalin for 5 to 12 hours and sectioned radially from 1 12 o clock. The pathologic report was made within 72 hours and included the following information: the grade of dysplasia (CINI CINIII, CIS, microinvasive and invasive carcinoma), the location of lesions and the numbers of cervical quadrants involved, the endocervical gland involvement, the endocervical and ectocervical marginal status, as well as the width and depth of the cone specimen. Management after conization Clinical management following the conization procedure depended on the grade of dysplasia, cone margin status, reproductive desire, patients age, as well as the compliance with follow-up. The options of postcone management included: follow-up, re-conization, hysterectomy and radical hysterectomy. The indications for hysterectomy were: (1) post-menopausal women with CIN III or CIS. (2) patients with CIN III or worse with large myoma or other benign uterine disease, without reproductive desire. (3) patients with positive cone margin and without reproductive desire. (4) patients with cervical adenocarcinoma in situ or microinvasive squamous carcinoma. 5 In our study group, a total of 67 patients received hysterectomy or radical hysterectomy. Of those, 48 patients received laparoscopical total hysterectomy, 5 laparoscopically-assisted vaginal hysterectomy, 10 total abdominal hysterectomy and 4 radical hysterectomy. The median time of hysterectomy was 72 hours post- conization (range days), 94.0% (63/67) of patients received hysterectomy within 60d after conization. Among 140 patients who did not undergo hysterectomy after conization, reconization was undertaken in two patients who had a positive cone margin but with reproductive desire. The remaining 138 patients received close follow-up (cytology, HPV test and colposcopy) with a schedule of 3 6 months intervals in the initial two years and 6-month intervals in the third year after conization. Statistical analysis The demographic data (patients age, gravidity, parity), cytological results, pathologic parameters on punch biopsy (grade of disease, quadrants involved by CIN III, endocervical glands involvement ) and operative factors (methods and depth of cone) for all patients with CIN III were analyzed retrospectively. A negative margin was defined as the absence of CIN at the edge of the cone and a positive margin was judged by the presence of CIN at or close to ( 0.1 cm) the edge of the cone. Initially, we used chi-square test for univariable analysis. After excluding insignificant factors, multivariable logistic regression analysis were utilized for testing remaining factors, positive and negative cone margins were dependent variables. Finally, four factors were selected as significant in predicting the positive margin. A regression model was constructed by these four factors. The pathologic result of a cone was compared to the subsequent hysterectomy. The chi-square test was utilized to compare the residual CIN in uterine specimens among three groups of patients with different cone margin status: 36 patients with margins free of CIN, 21 patients with CIN occurring close to margins and 10 patients with margin involvement. A two-tailed P value <0.05 was considered statistically significant. For all statistical analysis, the statistical package for the social sciences for windows 11.5 (SPSS11.5) was used.

3 Chinese Medical Journal 2009;122(4): RESULTS Histological study of 207 cone specimens revealed that there were 5 cases (2.4%) with CINI, 30 cases (14.5%) with CINII, 120 cases (58.0%) with CINIII, 37 cases (17.9%) with cervical carcinoma in situ, 12 cases (5.8%) with microinvasive carcinoma and 3 cases (1.4%) with invasive carcinoma. Of 207 patients, 151 (72.9%) had cone margins free of CIN, 37 (17.9%) had CIN lesions close to the margins ( 0.1cm to margin) and 19 (9.2%) had margin involvement by CINI or a worse condition. The frequency of a positive margin was 27.1% (56/207). A positive endocervical margin was present in 43 cases (76.8%, 43/56), of which 14 had margin involvement and 29 had CIN close to the margin. A positive ectocervical margin was present in 10 cases (17.9%, 10/56), of which 2 had margin involvement and 8 had CIN occurring close to the margin. In addition, there were 3 cases (5.4%, 3/56) in which both endocervical and ectocervical margins were involved by CIN. In Table 1, we tested whether there was a correlation between positive cone margin status and various factors with univariable chi-square analysis. We found that age, gravidity, grade of disease by punch biopsy and the conization method were not significantly correlated with a positive cone margin (P >0.05), whereas parity, cytological grade, number of quadrants involving CIN III, multi-quadrant gland involvement and depth of cone had a significant correlation with a positive cone margin (P <0.05). In Table 1, 6 cases of ASC-H were classified into the HSIL group, of which 3 cases had a positive cone margin. The 12 cases who had no TCT result were excluded from the analysis of correlation between cytological result and positive cone margin. We further analyzed the correlation of operative factors to a positive cone margin, and found that the depth of the cone was a significant predictor for a positive endocervical margin, but the width of the cone was not statistically significant (data not shown) in predicting a positive ectocervical margin (P >0.05). The latter conclusion was based on the fact that the positive ectocervical margins were seen in 9 out of 143 cases (6.3%) with a cone width 2.5 cm, and 4 out of 64 cases (6.3%) with cone width 2.6 cm. Although five factors were well-correlated with a positive cone margin by univariable analysis, an integrated effect of these factors may be more valuable in predicting a positive cone margin. Table 2 outlines the result of multivariable logistic regression analysis of these five factors. Four factors, e.g. parity, cytological result, multi-quadrant CIN III and depth of cone had significant predictive values. The multi-quadrant endocervical gland involvement was also valuable, but this factor was excluded in the multivariable logistic regression model because of its high correlation with multi-quadrant CIN III (r=0.91, P <0.01). In Table 2, each significant factor Table 1. Correlation of clinical and histological factors with positive cone margin for CIN III patients Clinical and histological factors n No. of positive margin (%) P values * Age (years) (17.4) (27.7) (29.6) (25.0) Gravidity 3 times (24.7) times (34.7) Parity 1 time (24.9) times 18 9 (50.0) Cytology LSIL (18.4) HSIL (30.3) Pap smear grade III and immediate biopsy 12 6 (50.0) CIN III quadrants by punch biopsy 1 quadrant 55 9 (16.4) quadrants (25.0) 3 quadrants (35.4) 4 quadrants 19 9 (47.4) No record 17 4 (23.5) Grade of disease by punch biopsy CIN III/CIS (26.6) Micro-invasive carcinoma 15 5 (33.3) Quadrants of gland involvement by punch biopsy (18.4) (24.6) (54.5) (61.5) No record 17 4 (23.5) Methods of conization Electrosurgical conization (30.3) Cold-knife conization (20.0) Depth of conization (cm) (32.6) (19.3) * : Univariate, chi-square test. : Percentage of positive endocervial margin. : Statistically significant. Table 2. Predictive factors for positive cone margin by multivariate logistic regression analysis Factors B (Partial regression P OR coefficient) values (95% CI) Standard Parity 1 time * 2 times ( ) Cytology LSIL * HSIL ( ) CIN III quadrants by punch biopsy 1 * ( ) ( ) ( ) Depth of conization (cm) 1.6 * ( ) * : As a reference category. was subdivided into 2 4 categories, the category which was least correlated to a positive margin was chosen as the reference. Every other category had its own partial regression coefficient B and odds ratio, reflecting the

4 370 Table 3. Pathology of cone margin vs hysterectomy specimens (n (%)) Margin status Margin free (n=151) Margin closed by CIN (n=37) Margin involvement (n=19) Total (n=207) No. of hysterectomy Pathology of hysterectomy specimen No disease 16 (44.4) 6 (28.6) 2 (20.0) 24 (35.8) CIN I 20 (55.6) 15 (71.4) 8 (80.0) 43 (64.2) CIN I 8 (22.2) 3 (14.3) 1 (1/10) 12 (17.9) CIN II 7 (19.4) 7 (33.3) 2 (2/10) 16 (23.9) CIN III/CIS 4 (11.1) 4 (19.0) 3 (3/10) 11 (16.4) Micro-invasive and invasive carcinoma 1 (1/36) 1 (1/21) 2 (2/10) 4 (5.9) extent of the causal relation between the category and positive margin. The OR for cytological grade, depth of cone, parity and multi-quadrant CIN III were 1.92, 2.03, 3.02 and 4.60, respectively, thereby demonstrating an increased risk for a positive margin. Of 207 patients who underwent conization, 67 had a subsequent hysterectomy, including 36 cases with the margin free, 21 cases with the margin close to CIN and 10 cases with margin involvement. Table 3 shows a histological comparison of different cone margin status to hysterectomy specimens. The incidence of residual CIN I or worse in hysterectomy specimens for 67 cases was 64.2% (43/67), residual CIN III or worse was 22.4% (15/67). Furthermore, we found that the incidence of residual CIN I or worse was 55.6% (20/36) in the margin free group, 71.4% (15/21) in the group with CIN close to the margin and 80.0% (8/10) in the group with margin involvement. The incidence of residual CIN III or worse was found to be 13.9% (5/36), 23.8% (5/21) and 50% (5/10), respectively. The frequency of residual CIN III or worse in the margin free group was statistically lower than that of the two other groups (P <0.05). However, a significant difference was not found between the latter two groups (P >0.05). Five patients in the margin free group were found to have CIN III or worse in subsequent hysterectomy specimen, of which 3 were shown to have 3 quadrants of CIN III in punch biopsy or cone specimen. One patient had a cervical polyp with histological diagnosis of squamous carcinoma in situ and another patient was suspected of carcinoma in situ by TCT and diagnosed as microinvasive carcinoma by pathology of the cone specimen; in the latter case, the depth of the cone was <1 cm. DISCUSSION According to 2006 consensus guidelines for the management of women with cervical intraepithelial neoplasia or adenocarcinoma in situ, 6 excisional treatment, such as conization, is an acceptable method for the histological diagnosis of CIN III, under the conditions of both satisfactory or unsatisfactory colposcopy. As both a diagnostic and therapeutic procedure, conization provides a reasonable approach for clinical treatment, particularly for women who are young or wish to preserve fertility function. Although the indication and value of conization has been described previously, it was largely unknown whether it could be used for complete excision of CIN, and what factors known prior to surgery would be used in predicting the possibility of complete excision. The predictive factors we identify in this paper may be helpful for estimating the possibility of complete excision of CIN prior to conization and may serve as a guide during the operation in judging the width and depth of the cone. These factors also provide an important reference for preparing a suitable follow-up plan. Excisional methods that provide tissue specimens for pathology include loop electrosurgical excisional procedures (LEEP), electrosurgical needle conization, laser conization or cold-knife cone (CKC), all modalities having equivalent efficacy in eliminating CIN and reducing the risk of future cervical cancer. 6,7 Preoperative predictors for positive margins of different types of conization were discussed in several papers recently and are the subject of some controversy. For example, evidence of invasive carcinoma on LEEP specimen or on cytology, less than 1 cm of cone depth, as well as post-menopause were significant predictors for any cone margin involvement. 8 Parity was thought important as a predictor 7 and visibility of the squamous-columar junction was inversely related to the positive endocevical margin. 9 Our results revealed that cytological grade, depth of cone, parity, as well as multi-quadrant CIN III on punch biopsy were significant predictors with an increased risk for positive margins. Whereas age, gravity, severity of disease of specimens from punch biopsy and the method of cone excision were not significant determinants. These findings are therefore at variance with other studies. A CIN lesion is characterized by a multi-focal and multi-step development with a slow rate of progress. The different lesions on one cervix may show varied extents of severity, and may be caused by different types of HPV infection having different periods of duration. 10 Multiquadrant CIN and multi-quadrant gland involvement are generally the result of a prolonged period and severe HPV infection, and therefore show a high frequency of positive margins. Women with increased parity are more likely to have a positive margin, which was inferred as rapid progress during pregnancy and labor. 7 A high grade cytological result paralleled generally a high grade histological lesion and therefore was valuable in predicting positive margins. As for the depth of conization, some authors suggested that 1.5 cm is acceptable, 11 while others thought that 2.0 cm to 2.5 cm may be more appropriate. 5 We recommend that the depth of conization be tailored to the patient s condition. An

5 Chinese Medical Journal 2009;122(4): increased depth of conization will produce a satisfactory excision for menopausal women or women with unsatisfactory colposcopy because CIN lesion may be located deeper in the cervical canal under these conditions. In our study, the grade of dysplasia on punch biopsy was not confirmed to be a statistically significant factor, although the frequency of a positive cone margin increased as the grade of dysplasia increased. This correlation may have resulted because the sample pool of microinvasive carcinoma used in our study was not large enough to achieve a statistically significant result. A recent report found that the positive cone margin was also related to a surgeon s training. 12 Another study showed that the load of high risk HPV was a risk factor for predicting residual or recurrent dysplasia. 13 The value of high risk HPV load in predicting a positive cone margin cannot be confirmed by this study because our study was retrospective and the data was limited. The relationship between high risk HPV load and a positive cone margin needs to be confirmed by a more randomized and prospective study. The relationship of a positive margin status and residual or recurrent CIN has been of great concern. A number of studies revealed a cure rate of 42% 68% post-conization despite margin involvement. 14 A previous study noted that 5 out of 19 cases (26.3%) with cone margin involvement had no CIN in specimens obtained following subsequent hysterectomy. 12 Moore et al 15 found residual CIN in 32% of hysterectomy specimens having a negative cone margin, and an occasional invasive cancer was identified in such patients. 16 The facts above indicate that margin status does not always correlate with the presence or absence of CIN in specimens of subsequent hysterectomy. In the present study, the frequency of residual CIN was gradually increased in groups with different margin status: 55.6% in patients who were margin-free, compared with 71.4% in patients with CIN occurring close to the margin and 80.0% in patients with margin involvement. Our data demonstrate that the cone margin status reflects a variable frequency of residual CIN in patients who undergo subsequent hysterectomy. This outcome is perhaps the true value of cone margin status as a surrogate indicator for the management of CIN. Although the status of CIN close to the margin is frequently met by clinicians, its significance as an indicator of residual CIN in subsequent hysterectomy specimen is controversial. Our study demonstrate that the frequency of residual CIN is intermediate between the status of margin free and margin involvement, and did not show a significant difference with that of margin involvement status by chi-square test (P >0.05). This result explains why patients with CIN close to the margin were put into the positive margin group in our analysis. Because a different cone margin status represented a varied frequency of residual CIN, a close follow-up is necessary for patients without subsequent hysterectomy after conization regardless of the cone margin status. Reich et al 17 evaluated the outcome of a mean of 19 years of follow-up for 390 patients with involved margins after cold knife conization, 84 (21.5%) had persistent or recurrent CIN III or invasive carcinoma, of which 53 (63.1%) were diagnosed within 1 year after conization. This finding reinforced the importance of a careful follow-up, particularly during the first year. 17 The present study led us to identify four risk factors that were statistically valuable in predicting a positive margin prior to conization for CIN III patients. This conclusion will enable clinicians to estimate the surgical outcome before conization, and to address the question of complete excision of the CIN lesions by conization. We conclude that the margin status of the cone does not mean the presence or absence of CIN, but rather may suggest the varied frequency of residual CIN in specimens of subsequent hysterectomy. In view of this fact, we maintain that the margin status of a cone is a valuable surrogate factor for clinical management of CIN III. It is inferred therefore that a combined consideration of four predictive factors, the margin status of the cone, as well as high risk HPV load would be valuable in predicting the residual, persistent and recurrent dysplasia following conization. Additional clinical trials are needed to support this inference. Acknowledgement: The authors thank Dr. Susan A. Rotenberg at Queens College, City University of New York for her help in the preparation of this manuscript. REFERENCES 1. Shen K, Lang JH, Huang HF, Wu M, Shi M, Pan LY, et al. Evaluation of cervical conization in diagnosis and management of cervical intraepithelial neoplasm. Chin J Gynecol Obstet (Chin) 2001; 36: Song XH. Application of three-stage diagnosis and treatment process in screening, diagnosis and treatment of cervical precancerous lesions. Chin J Pract Gynecol Obstet (Chin) 2007; 23: Solomon D, Davey D, Kurman R, Moriarty A, O Connor D, Prev M, et al. The 2001 Bethesda System: terminology for reporting results of cervical cytology. JAMA 2002; 287: Wright TC Jr, Cox JT, Massad LS, Twiggs LB, Wilkinson EJ consensus guidelines for the management of women with cervical cytological abnormalities. JAMA 2002; 287: Shen K. Choices among the iatrotechnics for cervical intraepithelial neoplasia. Natl Med J China (Chin) 2006; 86: Wright TC Jr, Massad LS, Dunton CJ, Spitzer M, Wilkinson EJ, Soloman D consensus guidelines for the management of women with cervical intraepithelial neoplasia or adenocarcinoma in situ. Am J Obstet Gynecol 2007; 197: Tillmanns TD, Falkner CA, Engle DB, Wan JY, Mannel RS,

6 372 Walker JL, et al. Preoperative predictors of positive margins after loop electrosurgical excisional procedure-cone. Gynecologic Oncology 2006; 100: Kietpeerakool C, Srisomboon J, Ratchusiri K. Clinicopathologic predictors of incomplete excision after loop electrosurgical excision for cervical preneoplasia. Asia Pac J Cancer Prev 2005; 6: Costa S, De Nuzzo M, Terzano P, Santini D, De Simone P, Bovicelli A, et al. Factors associated with cone margin involvement in CIN patients undergoing conization-equivalent electrosurgical procedure. Acta Obstet Gynecol Scand 2000; 79: Insinga RP, Dasbach EJ, Elbasha EH, Liaw KL, Barr E. Progression and regression of incident cervical HPV 6, 11, 16 and 18 infections in young women. Infect Agent Cancer 2007; 2: Micheal S Baggish. Treatment of Cervical Intra-epithelial Neoplasia. In: Micheal S Baggish. Colposcopy of the cervix, vagina, and vulva: a comprehensive textbook. Pennsylvania: Mosby, an affiliate of Elsevier Science (USA) 2003; Dai ZQ, Pan LY, Huang HF, Lang JH. Evaluation of cervical intraepithelial neoplasia positive cutting edge after conization. Chin J Oncol (Chin) 2007; 29: Alonso I, Torné A, Puig-Tintoré LM, Esteve R, Quinto L, Campo E, et al. Pre- and post-conization high-risk HPV testing predicts residual/recurrent disease in patients treated for CIN 2-3. Gynecol Oncol 2006; 103: Lu CH, Liu FS, Kuo CJ, Chang CC, Ho ES. Prediction of persistence or recurrence after conization for cervical intraepithelial neoplasia III. Obstet Gynecol 2006; 107: 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 1996; 174: Husseinzadeh N, Shbaro L, Wesseler T. Predictive value of cone margins and post-cone endocervical curettage with residual disease in subsequent hysterectomy. Gynecol Oncol 1989; 33: Reich O, Lahousen M, Pickel H, Tamussino K, Winter R. Cervical intraepithelial neoplasia III: long-term follow-up after cold-knife conization with involved margins. Obstet Gynecol 2002; 99: (Received June 20, 2008) Edited by CHEN Li-min