Validated performance metrics for colonoscopic polypectomy remain limited. The ESGE Training Curriculum Taskforce has recently introduced the Global Polypectomy Assessment Tool (GPAT). We aimed to evaluate GPAT in real-life endoscopic training by assessing a GI trainee’s learning curve and interobserver agreement among raters with different levels of experience.

We conducted a prospective single-center study at Isola Tiberina–Gemelli Isola Hospital from January to October 2025. Thirty colorectal endoscopic resections (polypectomy and endoscopic mucosal resection, EMR) were included: 10 performed by a GI trainee under supervision and 10 performed by each of two expert endoscopists. All procedures were video-recorded and independently evaluated using GPAT by all three raters. The trainee’s learning curve was assessed by comparing mean GPAT scores in the first 5 versus the last 5 trainee cases (t-test). Interobserver agreement was measured using intraclass correlation coefficients (ICC), with p<0.05 considered statistically significant.

Thirty procedures in 23 patients were analyzed (median lesion size 25 mm; 90% conventional EMR; 57% piecemeal resection). Adverse events included 3 delayed bleedings (10%) and no perforations. GPAT demonstrated progressive improvement in the trainee’s performance, with convergence of scores after the 6th case (Figure 1). Overall interobserver agreement for GPAT was weak (ICC=0.33, 95%CI 0.09–0.57). Agreement between the two experts was poor (ICC=0.21, 95%CI −0.08 to 0.50; p=0.086). Agreement between the trainee and expert 1 was also poor (ICC=0.38, 95%CI 0.03–0.64; p=0.017), while agreement between the trainee and expert 2 was fair (ICC=0.44, 95%CI 0.01–0.72; p=0.022). When analyzing only the trainee-performed videos, agreement among the three raters remained poor (ICC=0.24). Interobserver agreement was weak for expert 1-performed cases (ICC=0.39) and fair for expert 2-performed cases (ICC=0.41).

Figure 1.  Description of the evaluations of the 10 polypectomies performed by the traineer. 

GPAT offers a structured and practical method for assessing polypectomy competency during endoscopic training. Nevertheless, the low inter-rater reliability observed in real-world conditions highlights the need for improved standardization in the interpretation and scoring of the tool. Structured training in score application, together with refinement of specific scoring criteria, may be required to enhance consistency and validity. Larger multicenter studies are warranted to establish GPAT as a reliable standard for competency assessment in colonoscopic polypectomy.