Sweet-spot training for early esophageal cancer detection

Over the past decade, the imaging tools for endoscopists have improved drastically. This has enabled physicians to visually inspect the intestinal tissue for early signs of malignant lesions. Furthermore, it has paved the way for image analysis algorithms, to support the gastroenterologist in finding these early signs of developing cancer. Recent studies show the feasibility of such systems, where the problem is typically approached as a segmentation task and a binary ground truth is employed. In this study, we show that the detection of early cancerous tissue in the gastrointestinal tract cannot be approached as a binary segmentation problem and it is crucial and clinically relevant to involve multiple experts for annotating early lesions. By employing the so-called sweet spot for training purposes, a much better detection performance can be achieved. Furthermore, a multi-expert-based ground truth, i.e. a golden standard, enables an improved validation of the resulting delineations. For this purpose, we propose two novel metrics that can handle multiple ground-truth annotations. Our experiments involving these metrics and based on the golden standard show that the performance of a detection algorithm of early neoplastic lesions in Barrett's esophagus can be increased significantly, demonstrating a 10 percent point increase in the resulting F1 detection score.