Clinical assessment of tumor regression grade systems in gastroesophageal adenocarcinoma following neoadjuvant chemotherapy

Background: The standard treatment for gastroesophageal cancer is neoadjuvant chemotherapy, followed by surgery, which has been shown to increase survival compared with surgery alone. Evidence is mounting that characterization of the oncologically induced tumor regression is of prognostic importance. However, no consensus regarding the optimal system for describing tumor regression exists. Thus, this study aims to explore three validated/promising tumor regression systems with a focus on their interobserver reliability and usability. Methods: We included 100 consecutive patients with gastroesophageal adenocarcinoma who had undergone neoadjuvant oncological treatment followed by surgery. The tumors underwent tumor regression grade (TRG) assessment according to the Standard Mandard-, Modified Mandard-, and Becker systems to assess the interobserver reliability between two consultant pathologists. The interobserver reliability was determined by both Fleiss kappa and weighted kappa metrics. Besides, a semi-quantitative usability questionary was completed and it was expanded with usability comments. Results: The Fleiss kappa interobserver agreement was 0.67 [95% CI, 0.55–0.79], 0.88 [95% CI, 0.73–1.00], and 0.88 [95% CI, 0.73–1.00] for Standard Mandard-, Modified Mandard-, and the Becker systems, respectively. The weighted kappa (linear) was 0.80 [95% CI, 0.72–0.89], 0.91 [95% CI, 0.84–0.98], and 0.91 [95% CI, 0.84–0.98] for the Standard Mandard-, Modified Mandard-, and the Becker systems, respectively. The usability was scored on a scale of 8–24 by both raters. The systems were scored accordingly: 47 (Modified Mandard system), 43 (Becker system), and 37 (Standard Mandard system). Conclusion: The Modified Mandard- and Becker systems had excellent interobserver reliability and usability. However, the systems could be improved by a better characterization of the different tiers and tumor morphology.