{"paper":{"title":"Prediction of Rectal Cancer Regrowth from Longitudinal Endoscopy","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"A longitudinal deep learning model detects rectal cancer regrowth from paired endoscopy images with 97 percent sensitivity.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Aneesh Rangnekar, Christina Lee, Despoina Kanata, Francisco Sanchez-Vega, Hannah Thompson, Hannah Williams, Harini Veeraraghavan, J. Joshua Smith, Jorge Tapias Gomez, Julio Garcia-Aguilar, Mert R. Sabuncu","submitted_at":"2026-05-13T01:02:58Z","abstract_excerpt":"Clinical trial studies indicate benefit of watch-and-wait (WW) surveillance for patients with rectal cancer showing a complete or near clinical response (CR) directly after treatment (restaging). However, there are no objectively accurate methods to early detect local tumor regrowth (LR) in patients undergoing WW from follow-up exams. Hence, we developed Temporal Rectal Endoscopy Cross-attention (TREX), a longitudinal deep learning approach that combines pairs of images acquired at restaging and follow-up to distinguish CR from LR. TREX uses pretrained Swin Transformers in a siamese setting to"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"TREX achieved the highest accuracy in detecting LR with a high sensitivity of 97% ± 6% and a balanced accuracy of 90% ± 3%, and outperformed all baselines in early detection at both 3--6 (74% ± 1%) and 6--12 months (62% ± 4%) prior to clinical detection.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The clinical trial dataset used for training and testing is representative of broader patient populations and imaging conditions, and that performance on held-out data will translate to prospective real-world use without significant domain shift.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"TREX detects rectal cancer local regrowth from longitudinal endoscopy image pairs with 97% sensitivity and enables early prediction 3-12 months before clinical confirmation, outperforming baselines.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A longitudinal deep learning model detects rectal cancer regrowth from paired endoscopy images with 97 percent sensitivity.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"9dfdd70db4891ff3899680b91decfa91796b91aa3c0dbc505c83a3b517a41f7e"},"source":{"id":"2605.12855","kind":"arxiv","version":1},"verdict":{"id":"3f4737d8-9895-47e3-bef4-b610c9f94a1b","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T20:27:11.791585Z","strongest_claim":"TREX achieved the highest accuracy in detecting LR with a high sensitivity of 97% ± 6% and a balanced accuracy of 90% ± 3%, and outperformed all baselines in early detection at both 3--6 (74% ± 1%) and 6--12 months (62% ± 4%) prior to clinical detection.","one_line_summary":"TREX detects rectal cancer local regrowth from longitudinal endoscopy image pairs with 97% sensitivity and enables early prediction 3-12 months before clinical confirmation, outperforming baselines.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The clinical trial dataset used for training and testing is representative of broader patient populations and imaging conditions, and that performance on held-out data will translate to prospective real-world use without significant domain shift.","pith_extraction_headline":"A longitudinal deep learning model detects rectal cancer regrowth from paired endoscopy images with 97 percent sensitivity."},"references":{"count":55,"sample":[{"doi":"10.1097/dcr","year":2023,"title":"Diseases of the Colon and Rectum 67(1), 18–31 (2024)https://doi.org/10.1097/DCR","work_id":"76efcf99-1b0a-4c82-8abe-18f6618a345a","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1200/jco.22.00032","year":2022,"title":"Journal of Clinical Oncology40(23), 2546–2556 (2022)https://doi.org/10.1200/JCO.22.00032 https://ascopubs.org/doi/pdf/10.1200/JCO.22.00032","work_id":"db72d788-2002-4a36-bb4a-3ecbceb8253a","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1200/jco.23.01208","year":2024,"title":"Journal of Clinical Oncology42(5), 500–506 (2024)https://doi.org/10.1200/JCO.23.01208 https://ascopubs.org/doi/pdf/10.1200/JCO.23.01208","work_id":"2d6ca64b-cb8c-4102-9197-23cb6f584caf","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2018,"title":"An- nals of surgery268(6), 955–967 (2018)","work_id":"393ed7c8-c0a7-4b2b-8ea3-2ecb2ab22dc8","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2018,"title":"The Lancet Gastroenterology & Hep- atology3(12), 825–836 (2018)https://doi.org/10","work_id":"f9083db1-5184-4b86-92f4-e0ab8f3db209","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":55,"snapshot_sha256":"88921062008d48dbdb99999f2fe4f8af0077df7864439043d5dfbefadd3a0891","internal_anchors":1},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}