Search

Background and Aim: A hemostatic gel, PuraStat (3‐D Matrix, Tokyo, Japan), is used for various gastrointestinal hemostasis. In this study, we analyzed the efficacy of PuraStat for perioperative bleeding (POB) and prevention of delayed bleeding (DB) to colorectal cold snare polypectomy (CSP) with continuous anticoagulant. Methods: This was a single‐center, retrospective study. Subjects were lesions of 2–9 mm under continuous anticoagulant from 2021 to 2023 and treated with PuraStat for POB. The definition of POB was bleeding which did not stop spontaneously by 1.0–1.5 min after resection and needed hemostasis. Successful hemostasis was defined as cessation of bleeding within 1.0–1.5 min after spraying PuraStat and the rate of it and risk factors of POB were analyzed. For comparison, cases receiving previous CSP without PuraStat were extracted from all cases with CSP (2018‐2021), and POB and DB rate (DBR) were analyzed after propensity score matching. Results: One hundred twenty‐two lesions (91: direct oral anticoagulant (DOAC), 31: warfarin) with anticoagulant were analyzed and the rate of successful hemostasis with PuraStat was 92.6% (DOAC/warfarin: 93.4%/80.6%, P = 0.01). The rate of DB was 0.0%. Multivariate analysis showed that significant risk factors about unsuccessful hemostasis for POB with PuraStat were lesion size 8–9 mm (P < 0.01), warfarin (P = 0.01), and combination of antiplatelet (P = 0.01). Regarding the comparison about CSP with/without PuraStat, the clipping rate and DBR were 8.5%/94.9% (P < 0.01) and 0%/1.7% (P = 1.0). Conclusion: The effects of PuraStat for POB and DB in colorectal CSP with continuous anticoagulant were acceptable. [ABSTRACT FROM AUTHOR]

Objectives: In light‐emitting diode (LED) and laser colonoscopy, linked color imaging (LCI) superiority to white‐light imaging (WLI) for polyp detection is shown separately. We analyzed the noninferiority of LCI between LED and laser colonoscopy and that of WLI (LECOL study). Methods: We prospectively collected nonpolypoid lesions with WLI and LCI using LED and laser colonoscopy from January 2021 to August 2021. All images were evaluated randomly by 12 endoscopists (six nonexperts and six experts in three institutions) using the polyp visibility score: 4, excellent; 3, good; 2, fair; and 1, poor. The comparison score (LED better/similar/laser better) for redness and brightness was evaluated for WLI and LCI pictures of each lesion. Results: Finally, 63 nonpolypoid lesions were evaluated, and the mean polyp size was 24.5 ± 13.4 mm. Histopathology revealed 13 serrated lesions and 50 adenomatous/cancerous lesions. The mean polyp visibility scores of LCI pictures were significantly higher than those of WLI in the LED (3.35 ± 0.85 vs. 3.08 ± 0.91, P < 0.001) and the laser (3.40 ± 1.71 vs. 3.05 ± 0.97, P < 0.001) group, and the noninferiority of LCI pictures between LED and laser was significant (P < 0.001). The comparison scores revealed that the evaluation of redness and brightness (LED better/similar/laser better) were 26.8%/40.1%/33.1% and 43.5%/43.5%/13.0% for LCI pictures (P < 0.001) and 20.6%/44.3%/35.1% and 60.3%/31.7%/8.0% for WLI pictures (P < 0.001), respectively. Conclusion: The noninferiority of polyp visibility with WLI and LCI in LED and laser colonoscopy was shown. WLI and LCI of LED tended to be brighter and less reddish than those of laser. [ABSTRACT FROM AUTHOR]

Background and Aim Methods Results Conclusion There are no previous studies in which computer‐aided diagnosis (CAD) diagnosed colorectal cancer (CRC) subtypes correctly. In this study, we developed an original CAD for the diagnosis of CRC subtypes.Pretraining for the CAD based on ResNet was performed using ImageNet and five open histopathological pretraining image datasets (HiPreD) containing 3 million images. In addition, sparse attention was introduced to improve the CAD compared to other attention networks. One thousand and seventy‐two histopathological images from 29 early CRC cases at Kyoto Prefectural University of Medicine from 2019 to 2022 were collected (857 images for training and validation, 215 images for test). All images were annotated by a qualified histopathologist for segmentation of normal mucosa, adenoma, pure well‐differentiated adenocarcinoma (PWDA), and moderately/poorly differentiated adenocarcinoma (MPDA). Diagnostic ability including dice sufficient coefficient (DSC) and diagnostic accuracy were evaluated.Our original CAD, named Colon‐seg, with the pretraining of both HiPreD and ImageNET showed a better DSC (88.4%) compared to CAD without both pretraining (76.8%). Regarding the attentional mechanism, Colon‐seg with sparse attention showed a better DSC (88.4%) compared to other attentional mechanisms (dual: 79.7%, ECA: 80.7%, shuffle: 84.7%, SK: 86.9%). In addition, the DSC of Colon‐seg (88.4%) was better than other types of CADs (TransUNet: 84.7%, MultiResUnet: 86.1%, Unet++: 86.7%). The diagnostic accuracy of Colon‐seg for each histopathological type was 94.3% for adenoma, 91.8% for PWDA, and 92.8% for MPDA.A deep learning‐based CAD for CRC subtype differentiation was developed with pretraining and fine‐tuning of abundant histopathological images. [ABSTRACT FROM AUTHOR]