Extensive lesions and a positive cone margin are strong predictors of residual disease in subsequent hysterectomy following conization for squamous intraepithelial lesion grade 2 or 3 study design

The present study aimed to determine the predictive factors for post-conization residual lesion in subsequent hysterectomy for CIN 2/3, and we confirmed that margin involvement and extensive lesions were predictors of residual lesions in hysterectomy specimens. As reported by previous studies, the probability of residual lesions ranges 30–90% if the resection margin of a cone specimen is positive [3]. Moreover, women with margin involvement have a five times higher relative risk of residual or recurrent CIN 2/3 than that of women with negative margins [3]. Thus, margin status is generally considered as a predictive factor for residual lesions [5, 11, 15‐18]. Specifically, the endocervical margin status has been recognized to be significant, whereas the exocervical margin status remains controversial [18‐23]. The majority of our patients were of reproductive age and perimenopausal, and 210(78.7%) patients had either type I or II TZ. The most involved margin in our patients was the ectocervical margin. Our analysis indicated that the risk of residue increased with cervical margin involvement, irrespective of whether the endocervical or ectocervical margin was involved. Moreover, our study showed that involvement of both endo- and ectocervical margins was more likely to result in residual lesions than the involvement of endocervical or ectocervical margins alone. However, the specificity of our results might be related to subject bias, choice of surgical approach, excision without colposcopy assistance, and short interval between conization and hysterectomy.

In this study, the rate of positive margin after conization was higher compared with that reported in some previous studies [3, 5], especially the ectocervical margin. This may be related to the following factors. First, we performed excision without colposcopy assistance in our study. Currently, colposcopy-guided excision is the standard of care for women with CIN 2/3 in developed countries. However, where resources do not permit, some patients with CIN 2/3 are still being treated with conization without colposcopy assistance. In such circumstances, identification of high-risk patients with extensive lesions by colposcopic examination before conization is indispensable. Second, in our institution, CKC is performed as the definitive treatment of CIN 2/3. Patients treated with CKC have been shown to have lower rates of positive endocervical margin than those treated with large loop excision of the TZ [24]. Third, residual lesions at the edges of the cervix after conization may be eliminated by rapid cell turnover during cervical healing and by vaginal acidity [25]. However, in our study, the short time interval between the two operations limited the elimination of the lesions by rapid cell turnover. The results obtained by Cejtin et al. are consistent with our findings [22].

Margin involvement is regarded as an important predictor of residual disease [5, 14, 26]. Repeated cervical conization is considered an acceptable alternative for women with a positive cone margin who desire fertility preservation, and hysterectomy is the definitive therapy for women with no reproductive requirements [27, 28]. If the choice of treatment is new conization or hysterectomy, most women will unnecessarily undergo these procedures because they have no residual lesions. Such unnecessary surgery increases the risks for complications and affects these women’s gestational future. However, if surgery is not performed, there is a risk of insufficient treatment in a large number of women with CIN 2/3, as well as a risk of malignancy. This presents a problem for both patients and gynecologists when planning follow-up and further therapy. Therefore, it is necessary to distinguish the subset of patients with residual lesions from those with positive margins and identify risk factors associated with surgical margins in order to reduce or avoid positive margins.

Previous studies showed that HPV testing is an effective tool in predicting residual disease after conization [3, 14, 29]. The predictive value of resection margin for predicting residual disease improved when used in combination with the HPV test [3]. Positive excision margins and high-risk HPV infection at follow-up, appeared to be strong risk factors for residual/recurrent CIN 2/3 after conization [5, 13]. The combination was thus suggested for use in risk stratification for residual/recurrent disease [30]. Women with positive margins and high-risk HPV infection during follow-up should be considered for prompt re-treatment. However, the impact of HPV status on cervical glandular lesions is controversial [31‐33]. Therefore, further research is needed to assess the role of the combined margin and post-excision HPV status in stratifying the risk of treatment failure and follow-up management. In this study, data on HPV status after conization was scarce because most patients underwent hysterectomy shortly after conization. Surprisingly, three patients with positive margins in our cohort were found to have a previously undiagnosed cervical cancer following hysterectomy. Two of the three had HPV-related adenocarcinoma. Women with incompletely excised CIN 2/3 are at risk for cervical cancer. One study found an incidence of micro-invasive carcinoma of 10.38% in the final histopathological analysis of hysterectomies performed for CIN 3 [34]. Patients with positive margins may be considered for repeated conization or hysterectomy [35].

Currently, hysterectomy is not advisable for treating CIN 2/3 [2]. However, the procedure is acceptable, after obtaining informed consent from the patients, only if it is not possible to carry out or repeat a diagnostic excision or if adequate follow-up is not feasible [2]. Despite this, 267 women with CIN 2/3 were treated with hysterectomy in our study, suggesting that the procedure is still commonly performed. Based on their records, in addition to the aforementioned factors, the following several factors led our patients to choose hysterectomy. First, some patients had other indications of hysterectomy, such as fibromatosis, adenomyosis, dysmenorrhea. Second, the CIN diagnosis and treatment caused anxiety and fear of cancer [36], leading them to seek a permanent solution through hysterectomy. Thus, it is important to address the patients’ anxiety and fears [37], which often stem from a lack knowledge regarding their conditions. Therefore, health care managers in primary and specialized care levels should create opportunities to meet patients’ informational needs. Third, limited healthcare access and financial concerns limited patients’ ability to perform self-care. Finally, the coronavirus disease pandemic radically changed China’s healthcare, also impacting screening and colposcopy programs. Hysterectomy is unacceptable as primary therapy solely for the treatment of CIN because it can lead to complications and risks of vaginal lesions onset and overtreatment [38]. Therefore, the choice of hysterectomy for CIN 2/3 should be carefully evaluated and considered only in selected cases.

Previous studies evaluated independent pre-conization variables and concluded that some pretreatment predictors might help in planning cervical conization [28, 39, 40]. In our study, logistic regression analysis revealed that involvement of the four quadrants was the only significant independent predictor of residual lesion. A univariate logistic regression analysis also showed that postmenopausal status, age ≥ 50 years and parity ≥ 2 were associated with positive margin in patients with CON 2/3 prior to therapeutic cervical conization. This finding is in agreement with previous results [40, 41].

Another significant finding of this study was that extensive involvement of CIN 2/3 at the cone margin (3–4 quadrants) was associated with residual lesions in subsequent hysterectomy specimens, which is consistent with previous results [24]. A previous study observed residual disease in 80% of patients with involvement of three or four cervical quadrants [42]. Women with extensive cone margin involvement (3–4 quadrants) were approximately 14 times more likely to have residual lesions on subsequent surgical evaluations [43]. Extensive lesions increase the incidence of residual lesions during conization because an increase in the range of lesions may affect observations, interfere with the judgment regarding surgical margins during the operation, and increase the surgical difficulty. Some researchers have suggested that the number of disease quadrants in a conization specimen can be used as an important factor in guiding subsequent treatment [44]. Repeat conization or hysterectomy is advised for women with the involvement of 3–4 quadrants. Nevertheless, post-conization surveillance without hysterectomy may be an alternative for women with involvement of 1–2 quadrants because they have a lower risk of residual disease. Correspondingly, the number of quadrants involved on colposcopic examination may serve as an assistive tool for assessing the size and shape of the excision [45]. For instance, Kawano et al. showed that cone lengths of 15 and 20 mm were the best cut-off values for the complete excision of cervical lesions involving a single quadrant and two or more cervical quadrants, respectively [46]. For the resection of large, scattered, and multifocal lesions, colposcopy-guided CKC may be more suitable because optimally excising large multifocal lesions using a round-loop electrode is difficult, as the latter results in circular excision, and the entire volume of a lesion may not be included in the excised circle.

In addition to positive margins and extensive cervical cone margin involvement, we suspected some variables to be associated with residual lesions; however, this suspicions were not proven. In particular, we found no significant differences in menopausal status, type of TZ, gravidity, parity, and conization method. Interestingly, age ≥ 50 years and EGI were associated with residual lesions on the univariate analysis, whereas a significant correlation for both was not observed on the multivariate analysis. The relatively small sample size and the heterogeneity of the patient population might have influenced our results. Age and menopausal status have been shown to be significant predictors of residual disease in several studies. This finding is anatomically plausible; atrophy of the genital tract and deep inversion of the TZ make the complete resection of the abnormal cervical epithelium challenging after menopause. However, neither of these factors was a risk factor for residual lesions in present study. This finding may be attributed to the patient heterogeneity. The majority of our patients were reproductive premenopausal women, among whom CIN was the most prevalent. In our study, 78.7% of patients had type I or II TZ, and the majority of patients had a positive ectocervical margin, which is anatomically plausible. Therefore, these factors should be evaluated further in future studies, and more accurate prediction of residual disease after conization is indispensable, especially in women of child-bearing age.