Introduction
Cervical cancer is still one of the most common female reproductive tract cancers in developing countries. A total of 485 000 new cases and 236 000 deaths occurred in 2013 [
]. Cervical high-grade squamous intraepithelial lesion (HSIL) is classified as a cervical precancerous lesion, and conization is preferred because the probability of HSIL persistence and progression to invasive cervical cancer is significantly higher than that of cervical low-grade squamous intraepithelial lesion (LSIL) [
]. Cervical cold knife conization (CKC) and loop electrosurgical excision procedure (LEEP) are the two prevalent techniques for HSIL treatment. The status of cervical margin by conization is closely related with postoperative recurrence. Generally, a positive margin by conization implies that part of the lesion still remains in the cervix, which leads to lesion recurrence if the residual lesion is not removed [
–
]. A meta-analysis showed that the recurrence rate of patients with positive margin by conization was 20.38% (1649/8091), which was significantly higher than that of patients with negative margin (3.86%) [
]. However, only 23 patients (52.3%) were reported to have residual diseases in subsequent surgical samples among a total of 44 patients with LEEP endocervical positive margin [
]. That is, nearly half of patients with positive margin probably present disease-free in subsequent surgical specimens. Accordingly, re-assessment, including analysis of positive margin status, repeated HPV testing, and cytology, seems reasonable, instead of immediate re-conization [
]. Similar to the recommendation of the American Society for Colposcopy and Cervical Pathology (ASCCP), if HSIL is identified at the margins of a diagnostic excisional procedure in an endocervical sample obtained immediately after the procedure, reassessment using cytology with endocervical sampling at 4 to 6 months after treatment is preferred. Performing a repeat diagnostic excisional procedure is acceptable. Hysterectomy is acceptable if a repeat diagnostic procedure is not feasible [
]. However, the ASCCP guideline does not clearly define who are suitable for re-assessment or for immediate operation. Obviously, a triage, by confirming the indication of re-assessment or operation, will facilitate a suitable management for HSIL patients with positive margin by conization.
In the study, we retrospectively analyzed clinico-pathological data from HSIL patients with positive margin by initial conization and who underwent secondary conization or hysterectomy, calculated the residual lesion rate, and analyzed risk factors associated with residual diseases. The aim of the study is to provide a guide for the suitable management for HSIL patients with positive margin.
Clinical data and methods
Patients
A total of 119 cervical HSIL patients with positive cone margins, who underwent secondary conization or hysterectomy within 6 months, were recruited from January 2009 to December 2014 in Women’s Hospital, Zhejiang University School of Medicine. Positive margin was defined as involvement of HSIL in the margin of the tissue by initial conization. Residual lesion was defined as the detection of HSIL or cervical cancer in the tissue by secondary conization or hysterectomy.
Clinico-pathological data
Clinico-pathological data of all patients were collected, including patient age, menopausal status, previous pregnancies, pre-cone cytology, endocervical curettage (ECC) pathology, initial cone method, and pathological diagnosis. All the patients were divided into non-residual lesion group and residual lesion group, according to the pathological diagnosis in secondary operative specimens. Non-residue lesion group included LSIL or less, and residual lesion group consisted of HSIL and cervical carcinoma.
Cervical liquid-based cytology
All patients underwent cervical liquid-based cytology examination (ThinPrep) 1 month prior to initial conization. Cytological diagnoses were performed by professional pathologists and diagnostic criteria followed along the 2001 cervical cytology Bethesda reporting system.
LEEP
The LEEP knife model used was a WALLACH QUANTUM SERIES BIOVAC. The electrode entry departed from 0.5 cm of the outer edge of non-iodine staining or 1 cm of outer edge of cervical external aperture, and the electrode continuously moved to cut tissue. Cervical resection depth was approximately 0.8 cm to 1 cm. Endocervical curettage (ECC) was processed after conization, and routine pathological examination was performed on all specimens.
Cervical CKC
Entry of surgical blade departed from 0.5 cm of the outer edge of non-iodine staining or 1 cm of the outer edge of the cervical external aperture, and all lesions were removed one at a time. The depth was up to the cervical stroma and the length was approximately 2 cm to 2.5 cm. ECC and pathological examination were performed as with LEEP.
Pathological examination
All cervical tissue samples were routinely embedded, sliced, and stained with hematoxylin–eosin. Histological diagnoses were performed by professional pathologists and the diagnostic criteria followed the WHO classification of tumors of female reproductive organs. The lesions were diagnosed as cervical intraepithelial neoplasia (CIN 1, 2, or 3) before 2014, whereas they were diagnosed as LSIL instead of CIN 1 and HSIL instead of CIN 2, 3, or 2–3 after 2014.
For HSIL with positive margin, the positive margin site (endocervical site, ectocervical site, and basal site) and involved quadrants of positive margin (the whole cervix was divided into four parts) were recorded. Two or more involved quadrants were defined as multi-quadrant involvement.
Statistical methods
Univariate and multivariate analysis were performed using Statistical Package for the Social Sciences version 20.0 software, which was used to compare the clinico-pathological data between the residual lesion group and non-residual lesion group. Univariate analysis used Chi square test (χ2 test), and multivariate analysis used binary logistic regression. Statistical significance was defined as P<0.05.
Results
Residual lesions in HSIL patients with positive margin by conization
All 119 HSIL patients with positive margin underwent secondary conization or hysterectomy. Among them, 63 patients (52.94%) were found to have no residual lesion, including 58 cases (48.73%) of chronic cervicitis and 5 cases (4.21%) of LSIL. By contrast, 56 patients (47.06%) were found to have residual lesion, including 52 cases (43.69%) of HSIL and 4 cases (3.36%) of invasive cervical squamous cell carcinomas (three stage IA1 and one stage IA2 patients). The general characteristics of 119 patients are shown in Table 1, and the detailed characteristics of four residual cervical carcinoma patients are shown in Table 2. Some cases included patients with HPV-persistent infection and contact bleeding, but vaginal biopsy taken during pre-cone cytology resulted in negative for intraepithelial lesion or malignancy (NILM) or atypical squamous cells of undetermined significance (ASC-US). Biopsy pathology showed the presence of a lesion, which is confirmed by conization.
Factors related to residual lesion in HSIL patients with positive margin
Univariate analysis showed that patients aged>35 years, with post-menopausal period>5 years, and multiple-quadrant involvement were significantly related with residual lesion (All P<0.05). Logistic regression analysis revealed that multiple-quadrant involvement was an independent risk factor for residual lesion (P = 0.001; OR, 3.701; 95% CI, 1.496–9.154), as shown in Table 3.
Discussion
Residual lesion resulting from positive margin is the main cause of HSIL recurrence after conization. Previous reports revealed that residual lesion rate ranges from 11.3% to 54.8% in HSIL with positive margin [
,
–
]. Such broad rates are associated with the different definitions of positive margin and residual lesion. The residual rate is lower when residual lesions were defined with HSIL involvement or worse. If they were defined with LSIL involvement, the residual rate would be correspondingly much higher. In addition, the interval between initial conization and secondary operation also possibly influences the residual rate. Therefore, we recruited only those patients whose interval between initial conization and secondary operation was 6 months or shorter. Among 119 HSIL women with positive margin, 58 patients had residual lesions, with a residual rate of 48.73%. Similar to previous reports, only approximately 50% were found to have residual lesions in HSIL patients with positive margin; therefore, a routine secondary conization might not be necessary for nearly half the patients. Conization has been reported to cause bleeding, local or pelvic infection, cervical adhesions, and other complications [
–
]. For a woman who desires to have children, she may be faced with conization-associated premature birth, abortion, low infant birthweight, premature rupture of membranes, prolonged pregnancy, and other complications of pregnancy [
–
]. Thus, a risk factor analysis related to residual lesions would help to guide a triage for the management of HSIL patients with positive margin and to avoid unnecessary operation for some of them, especially for young women who desire to have children. Ayhan and his colleagues [
] found that lesions surrounding>50% of the cervical circumference during initial conization are associated with recurrent lesions. In the study, using univariate and multivariate analyses, we found that cervical multiple-quadrant involvement was an independent risk factor for residual lesions, suggesting that involved quadrants of positive margin may be used as a triage for the management of HSIL patients with positive margin. That is, an HSIL patient with positive margins beyond two or more quadrants during initial conization can be referred for re-conization or for reassessment using cytology, HPV testing, and colposcopy with endocervical sampling within 4 months to 6 months.
Zhu and his colleagues [
] found that age>35 years was an independent risk factor of lesion residue. Our univariate analysis showed that age>35 years was significantly related with HSIL residual lesion, but multivariate analysis only exhibited borderline significance (
P = 0.004; OR, 0.04; 95%CI, 0.876–8.422) because of insufficient samples in our study. In theory, an atrophic cervix, caused by low estrogen level after menopause, may increase the difficulty of conization and the rate of residual lesion. We found that menopausal period>5 years was significantly related with residual lesion in univariate analysis, but not in multivariate analysis, suggesting that further study with larger samples is needed.
Hysterectomy is an alternative option for those whose cervix is not suitable for re-conization or who have other indications for hysterectomy. In such situation, the probability of residual invasive cancer should be evaluated because a simple total hysterectomy is not adequate. An invasive residual rate of 3.45% to 10.7% has been reported in HSIL patients with positive margin [
,
,
,
]. In our study, out of 119 HSIL patients with positive margin, 4 patients were found to have invasive carcinoma (3.36%). As shown in Table 2, 3 patients had stage IA1 and 1 had stage IA2. All of them presented endo-cervical and multiple-quadrant involvement of positive margin and 3 of them were post-menopausal, suggesting that the probability of invasive carcinoma may be higher in post-menopausal women, especially those with endo-cervix and multiple-quadrant involvement. A diagnostic conization or big-size biopsy prior to hysterectomy was recommended to be performed for such patients to avoid inadequate hysterectomy.
In summary, a total of 56 (47.06%) patients, including 4 cases of invasive cervical carcinoma, were found to have residual lesion in subsequent operative specimens among 119 patients with positive margin by conization. Multi-quadrant involvement is an independent risk factor for residual disease. Those with multiple positive margins may consider re-conization or re-assessment.
Higher Education Press and Springer-Verlag Berlin Heidelberg
PDF
(108KB)
