Colorectal cancer (CRC) is the third cause of cancer death worldwide. CRC arises from adenomatous polyps (adenomas) which are initially benign; however, with time some of them can become malignant. Currently, the standard approach to reduce CRC-related mortality is to perform regular screening in search for polyps and colonoscopy is the screening tool of choice. During the examination, clinicians visually inspect intestinal wall in search of polyps. Once detected, they are resected and sent for histological analysis to determine their degree of malignancy and define the corresponding treatment the patient should undertake.
Though colonoscopy is still considered as the gold standard for colon screening, it presents some drawbacks being its miss-rate the most relevant of them (some polyps are still missed, specially those small/hidden behind folds). Besides, though several visual classification protocols are being tested, actual colonoscopy still needs of lesion removal to obtain an accurate diagnosis of lesions’ degree of malignancy.
Our objective is to develop computational support tools to aid clinicians in all stages of colonoscopy procedures: from lesion detection to diagnosis, going also through efficient navigation and assessment of trainee skills. Our methodologies are built in cooperation with experts from Hospital Clinic of Barcelona, Spain in order to develop solutions that are really focused on actual clinical needs. Considering our 20-year expertise on Machine Vision, we also aim at offering efficient solutions that are able to meet real-time constraints, compulsory for a given computational system in order to be clinically useful.