A scoring system to predict complex transanal endoscopic surgery

Background Since the introduction of screening for colorectal cancer, the use of transanal endoscopic surgery (TEM) has become increasingly popular. However, the technical difficulty of this surgery varies widely. The few studies of learning curve in TEM have produced very disparate results. The aim of this study is to distinguish between straightforward and complex procedures, in order to refer more difficult cases to centers with greater experience. Method Observational study with prospective data collection and retrospective analysis was carried out between June 2004 and January 2019. All TEMs performed on rectal tumors were included. The complexity of the procedure was defined according to the weighted mean surgical time for each surgeon. A predictive model of complexity was established, with a score higher than 5 indicating a complex lesion. Results During the study period, 773 TEMs were performed, 708 of which met the study's inclusion criteria. One hundred and three tumors were defined as complex. Predictors of complexity were as follows: male sex (OR: 1.78, 95% CI 1.1-2.9, score: 1), tumor size > 5 cm (OR: 5.1, 95% CI 3.2-8.2, score: 4), TEM for recurrence (OR: 6.3, 95% CI 2.3-16.7, score: 5), and distance from the upper margin of the tumor to the anal verge > 15 cm (OR: 1.6, 95% CI 0.96-2.7, score: 1). Conclusions Rather than establishing the learning curve merely in terms of the number of TEM procedures performed, it is important to consider the surgical difficulty of the interventions. To this end, it is essential to differentiate simple TEMs from the complex ones.