Influence of training level on cervical cone size and resection margin status at conization: a retrospective study

This study is a retrospective analysis of a prospectively generated database including all consecutive women who had undergone LLETZ for cervical dysplasia between 2004 and 2008 at the Department of Gynecology and Obstetrics of the Medical University of Vienna. Exclusion criteria were re-conization, treatment by techniques other than LLETZ and missing information about cone dimensions. The study was approved by the Ethics Committee of the Medical University of Vienna (EK number 1712/2015, approved September 2015). According to the University Ethics Committee, a formal consent was not required for this type of study, and therefore, the need for informed consent for participation was waived.

The following parameters were abstracted from the database: woman’s age, level of training of the operating surgeon (i.e., resident or staff gynecologist with an additional subdivision of residents into those with up to 12 months vs those with more than 12 months of gynecologic rotations during residency), cone dimensions (height, width and depth), resection margin status for overall, endocervical and ectocervical margins, preoperative histology of colposcopy guided cervical biopsies, preoperative Pap smear, preoperative human papillomavirus (HPV) status, Body Mass Index (BMI) and smoking status. All residents performed the conizations under supervision of a staff gynecologist who was present at the time of surgery. The same technique was used for all conizations. Loop size was chosen according to the preoperative colposcopy report and after intraoperative application of diluted Lugol solution to the cervix. The smallest loop to allow a complete resection was chosen. Conization was performed under direct visualization without the use of a colposcope.

Assuming the surgical specimen obtained by LLETZ to resemble a hemiellipsoid, the cone volume was calculated using the formula (1/2) × (4/3) × π × (length/2) × (width/2) × depth as previously described [17]. In cases where two cones were resected, or the cone was resected in several pieces, the volumes were added. The primary outcome measures were conization depth and volume as well as resection margin status of surgical specimens obtained by gynecology residents compared to those taken by staff gynecologists. In addition, a subgroup analysis was performed to assess whether conization depth and volume as well as resection margin status differed between residents at the beginning of their residency compared to more experienced residents.

Data were analysed by descriptive statistics. Woman’s age, BMI and volume, and depth of the excised cones were compared between residents and staff gynecologists as well as between residents having already had more or less than 12 months of gynecologic rotation during residency using the Mann–Whitney U test. Cone volumes were divided using the cut-off value of 2500 mm³ and compared between groups by the Chi-squared test. Depths of cones were also divided into groups (with cut-off values of 10, 15, or 20 mm, respectively), and compared between residents and staff gynecologists (and between the two resident subgroups, respectively) by the Chi-squared test. Possible associations between the surgeon’s level of training and other characteristics, i.e., resection margin status, preoperative histology of colposcopy guided cervical biopsies, preoperative Pap smear, presence of high risk HPV and smoking status, were analysed by Chi-squared tests. Two-sided p values < 0.05 were considered statistically significant. A binary logistic multivariable regression model was fit to evaluate the association between a gynecologist’s level of training and an excised cone volume greater than 2500 mm³ together with the woman’s age. IBM SPSS version 21.0 (IBM Corp., Armonk, NY, USA) was used for data analysis.

In this retrospective, single center study, the level of training of the operating surgeon represented an independent risk factor for a greater amount of cervical tissue removed at the time of LLETZ. In contrast, treatment by residents did not compromise women’s safety with respect to the course of the disease, as there were no differences in the rate of incomplete resections (i.e., positive resection margins) compared to staff gynecologists. In univariate analysis, the cone volume excised by residents compared to staff gynecologists was observed to be significantly larger. In addition, the percentage of surgical specimens larger than 2500 mm³ was higher in the resident group compared to the staff gynecologist group. When level of training was evaluated in a binary logistic multivariable regression analysis together with the age of the woman, it remained an independent risk factor for cone volumes greater than 2500 mm³. In contrast, comparison of the cone depths showed no significant differences between the groups, neither concerning the depth nor the percentage of depths greater than 10, 15, or 20 mm, respectively. In a subgroup analysis of the resident cohort, there was no statistically significant difference of the volume and depth of the excised cones between more experienced compared to less experienced residents.

To our knowledge, this is the first study to explore whether the surgeon’s level of training has an effect on the rate of incomplete resections and/or the amount of cervical tissue resected during LLETZ. A considerable strength of this study is the large sample size and the prospective data collection over a 4 year period. Furthermore, the number of residents and staff gynecologists performing conization was high, increasing the external validity of our results.

The main limitation of this study is the secondary, retrospective data analysis, with all its inherent risks of bias. For example, data on preoperative histology were missing in 32% of cases. The vast majority of these women had a biopsy performed outside of the hospital and had to be labelled “unknown” because the histologic report was not available for this retrospective study due to limitations of the medical health care documentation system used during the study period. Furthermore, we were not able to assess the proportion of the excised amount of tissue in relation with the total cervical volume of the respective woman, since no data regarding total pre- and postoperative cervical volume were available. In a prospective study comparing pre- and postoperative cervical lengths using magnetic resonance tomography and transvaginal ultrasonography, the authors hypothesized that the proportion of the total resection volume could have a greater influence on the risk of premature birth than the depth of the excision cone [7]. However, as mentioned above, various studies found a link between the depth and volume of the excised cones with the risk of premature delivery irrespective of the total cervical volume. Therefore, comparison of these measures may nevertheless provide a good estimation of possible clinically relevant differences concerning treatment by residents compared to staff gynecologists.

Over the last decades, the incidence of CIN in young women has been increasing, as has the average age at which women give birth [18]. As a result, a growing number of women has CIN prior to their first pregnancy. Cervical conization is the standard treatment for high-grade CIN [6‐8]. The risk of adverse pregnancy outcome increases with the amount of cervical tissue removed. A recent study found both a cone volume larger than 2500 mm³ or a depth of the excised surgical specimen greater than 20 mm to be associated with a higher rate of preterm birth [11]. In the previous studies, a depth of the cone greater than 10 mm was shown to be linked to a significant increase in the risk of premature rupture of the membranes as well as in the risk of preterm delivery [6, 9]. Another group obtained similar findings regarding the depth of the removed cone, with women having received a medium (10–14 mm), large (15–19 mm), or very large (≥ 20 mm) excision showing a higher risk of preterm delivery than those with a small (< 10 mm) excision [12]. As for the cone size, in the same study, a total volume greater than 2660 mm³ was found to double the risk of preterm and very preterm delivery [12]. Other factors increasing the risk of adverse pregnancy outcome in women with treatment for CIN, such as a defective cervical antimicrobial barrier or decreased mechanical stability of the regenerated cervix, have been proposed [10]. However, reliable data are only available for the amount and dimensions of cervical tissue removed during conization.

As the volume and depth of the cone are ideally only as large as needed for a resection of the lesion with clear resection margins, it seems likely that there is a learning curve for the performance of this procedure.

In this study, LLETZ for the treatment of high-grade CIN performed by residents supervised by a staff physician resulted in the removal of a greater amount of cervical tissue compared to this procedure carried out by staff gynecologists and, therefore, may expose women to an additional risk for adverse obstetrical outcomes. An additional subgroup analysis revealed no statistically significant difference between procedures performed by more experienced compared to less experienced residents, which might be due to the limited case load performed during residency. The depth of the excised cone, another putative risk factor for adverse obstetrical outcomes, was not different in specimens obtained by residents compared to those obtained by staff gynecologists. Nevertheless, these findings underscore the importance of letting residents train on simulation models before practicing on real patients, as it is already recommended for many other surgical procedures [19]. Studies have consistently shown that skills acquired in the simulation-environment transfer to the operating room and that simulation-training programs significantly decrease the clinical learning curve of various operative procedures [20, 21]. Several inexpensive, easily constructed simulation models for conization are available [14, 22] that may help to get more confident with the technique and achieve competency through deliberate practice, thereby extending the limited case load performed during residency training.

Next to cone size and volume, the training level of the operating surgeon may also affect the rate of positive resection margins, thereby influencing the course of the disease and possibly making further treatment and/or repeat surgery necessary [2‐5]. Only one retrospective study has addressed this question, showing that high volume surgeons achieved a higher rate of clear resection margins compared to residents and low volume staff members; however, there was no significant difference between residents and staff members [13]. In line with these findings, in our study no significant difference was found in the rate of positive resection margins between supervised residents and staff gynecologists, neither regarding the overall resection margin status nor ecto- or endocervical margins evaluated separately. Furthermore, the overall rate of incomplete resection in this study was < 20%, which is acceptably low and comparable to other studies [23]. Thus, performance of cervical conization by residents under supervision of a staff gynecologist seems to be safe for the patient with regard to the course of disease.