Automatic assessment of bowel preparation by an artificial intelligence model and its clinical applicability.

BACKGROUND AND AIM: Reliable bowel preparation assessment is important in colonoscopy. However, current scoring systems are limited by laborious and time-consuming tasks and interobserver variability. We aimed to develop an artificial intelligence (AI) model to assess bowel cleanliness and evaluate its clinical applicability. METHODS: A still image-driven AI model to assess the Boston Bowel Preparation Scale (BBPS) was developed and validated using 2361 colonoscopy images. For evaluating real-world applicability, the model was validated using 113 10-s colonoscopy video clips and 30 full colonoscopy videos to identify "adequate (BBPS 2-3)" or "inadequate (BBPS 0-1)" preparation. The model was tested with an external dataset of 29 colonoscopy videos. The clinical applicability of the model was evaluated using 225 consecutive colonoscopies. Inter-rater variability was analyzed between the AI model and endoscopists. RESULTS: The AI model achieved an accuracy of 94.0% and an area under the receiver operating characteristic curve of 0.939 with the still images. Model testing with an external dataset showed an accuracy of 95.3%, an area under the receiver operating characteristic curve of 0.976, and a sensitivity of 100% for the detection of inadequate preparations. The clinical applicability study showed an overall agreement rate of 85.3% between endoscopists and the AI model, with Fleiss' kappa of 0.686. The agreement rate was lower for the right colon compared with the transverse and left colon, with Fleiss' kappa of 0.563, 0.575, and 0.789, respectively. CONCLUSIONS: The AI model demonstrated accurate bowel preparation assessment and substantial agreement with endoscopists. Further refinement of the AI model is warranted for effective monitoring of qualified colonoscopy in large-scale screening programs.

Comparative Cost Analysis Between Endoscopic Resection and Surgery for Submucosal Colorectal Cancer.

BACKGROUND: There are few studies analyzing the cost of endoscopic resection and surgical resection in the treatment of submucosal colorectal cancer. OBJECTIVE: The objective was to perform a detailed cost analysis of endoscopic resection and surgical resection for submucosal colorectal cancer. DESIGN: This was a retrospective observational study. SETTING: This study was conducted at a tertiary academic center. PATIENTS: Medical records of 484 patients with submucosal colorectal cancer who underwent endoscopic resection or surgical resection between July 2003 and July 2015 were reviewed. MAIN OUTCOME MEASUREMENTS: The total costs during index admission and follow-up as well as clinical outcomes between the 2 groups were compared in the whole cohort and propensity score-matched cohort. RESULTS: In the propensity score-matched analysis ( n = 155 in each group), the endoscopic resection and surgical resection groups did not show significant differences in the rates of procedure-related adverse events (6.5% vs 3.9%; p = 0.304) and recurrence (0.6% vs 1.3%; p > 0.99). Readmission was more common in the endoscopic resection group (40.6% vs 11.0%; p < 0.001) because 64 (41.3%) patients underwent additional surgery for endoscopic noncurative resection. The endoscopic resection group had a lower cost during the index admission (1335.6 vs 6698.4 USD; p < 0.001), whereas the surgical resection group had a lower cost during follow-up (2488.7 vs 5035.7 USD; p < 0.001). The total cumulative cost was lower in the endoscopic resection group (6371.3 vs 9187.1 USD; p < 0.001). The same trend was observed in the whole cohort without propensity score matching. LIMITATIONS: A limitation of this study was the retrospective nature of analysis. CONCLUSIONS: The total cumulative cost for treatment and follow-up for submucosal colorectal cancer was lower in the endoscopic resection group, which had comparable oncologic outcomes as the surgical resection group. Endoscopic resection can be considered a cost-effective option for initial treatment for submucosal colorectal cancer. See Video Abstract at http://links.lww.com/DCR/B881 . ANLISIS COMPARATIVO DE COSTOS ENTRE LA RESECCIN ENDOSCPICA Y LA CIRUGA PARA EL CNCER COLORRECTAL SUBMUCOSO: ANTECEDENTES: Existen pocos estudios que analizan el costo de la resección endoscópica y la resección quirúrgica en el tratamiento del cáncer colorrectal submucoso.OBJETIVO: El objetivo fue realizar un análisis detallado de costos tanto de la resección endoscópica y la resección quirúrgica para el cáncer colorrectal submucoso.DISEÑO: Este fue un estudio observacional retrospectivo.AJUSTE: Este estudio se realizó en un centro académico terciario.PACIENTES: Se revisaron las historias clínicas de 484 pacientes con cáncer colorrectal submucoso que fueron sometidos a resección endoscópica o resección quirúrgica entre julio de 2003 y julio de 2015.PRINCIPALES MEDICIONES DE RESULTADOS: Los costos totales durante la admisión índice y el seguimiento, así como los resultados clínicos entre los dos grupos, fueron comparados en toda la cohorte y la cohorte emparejada por puntuación de propensión.RESULTADOS: En el análisis emparejado por puntuación de propensión ( n = 155 en cada grupo), los grupos de resección endoscópica y resección quirúrgica no mostraron diferencias significativas en las tasas de eventos adversos relacionados con el procedimiento (6,5% vs 3,9%, p = 0,304) y recurrencia (0,6% vs 1,3%, p > 0,99). La readmisión fue más común en el grupo de resección endoscópica (40,6% vs 11,0%, p < 0,001) porque 64 (41,3%) pacientes fueron sometidos a una cirugía adicional para lograr la resección en aquellos casos en que la resección endoscópica no fue curativa. El grupo de resección endoscópica tuvo un costo menor durante el ingreso índice (1335.6 vs 6698.4 USD, p < 0.001), mientras que el grupo de resección quirúrgica tuvo un costo menor durante el seguimiento (2488.7 vs 5035.7 USD, p < 0.001). El costo total acumulado fue menor en el grupo de resección endoscópica (6371,3 vs 9187,1 USD, p < 0,001). La misma tendencia se observó en toda la cohorte sin emparejamiento por puntuación de propensión.LIMITACIONES: La naturaleza retrospectiva del análisis.CONCLUSIONES: El costo total acumulado para el tratamiento y seguimiento del cáncer colorrectal submucoso fue menor en el grupo de resección endoscópica, que tuvo resultados oncológicos comparables a los del grupo de resección quirúrgica. La resección endoscópica puede considerarse una opción rentable para el tratamiento inicial del cáncer colorrectal submucoso. Consulte Video Resumen en http://links.lww.com/DCR/B881 . (Traducción-Dr Osvaldo Gauto ).

Cost of Endoscopic Submucosal Dissection Versus Endoscopic Piecemeal Mucosal Resection in the Colorectum.

BACKGROUND: Few studies have compared the costs of colorectal endoscopic submucosal dissection (ESD) and endoscopic piecemeal mucosal resection (EPMR). AIMS: Here, we aimed to investigate the cost-effectiveness of these approaches by analyzing clinical outcomes and costs. METHODS: Data from patients undergoing colorectal ESD and EPMR were retrospectively reviewed. Clinical outcomes (procedure time, complete resection, and recurrence) were compared, and total direct costs (procedural and follow-up) were assessed. RESULTS: Data from 429 ESD and 115 EPMR patients were included in the analysis. The complete resection rate was significantly higher (83.9% vs. 32.2%, p < 0.001), recurrence rate was lower (0.5% vs. 7.1%, p < 0.001), procedure time was longer (55.4 ± 47.0 vs. 25.6 ± 32.7 min, p < 0.001), and total direct procedural costs at the initial resection were higher (1480.0 ± 728.0 vs. 729.8 ± 299.7 USD, p < 0.001) in the ESD group than in the EPMR group. The total number of surveillance endoscopies was higher in the EPMR group (1.7 ± 1.5 vs. 1.3 ± 1.1, p = 0.003). The cumulative total costs of ESD and EPMR were comparable at 3 and 2 years' follow-up in the adenoma and mucosal/superficial submucosal cancer subgroups, respectively. CONCLUSIONS: Colorectal ESD was associated with higher complete resection and lower recurrence rates. EPMR showed shorter procedure times and similar cumulative total direct costs. ESD or EPMR should be chosen based on both clinical outcomes and cost-effectiveness.

Assessment of factors associated with smoking cessation at diagnosis or during follow-up of Crohn's disease.

BACKGROUND AND AIM: Smoking cessation is known to improve the course of Crohn's disease (CD). However, the factors associated with smoking cessation after CD diagnosis have not been well established. METHODS: Clinical characteristics and change in smoking status were evaluated in 445 current smokers at the time of CD diagnosis. Patients were classified into three subgroups based on their final smoking status and time of smoking cessation: non-quitters, quitters at diagnosis, and quitters during follow-up. RESULTS: The overall smoking cessation rate was 55.7% (248 of 445 patients). The diagnosis of CD was the main reason for quitting (41.5%, 103 of 248 patients). Smoking cessation at the time of CD diagnosis was associated with intestinal resection within 3 months from CD diagnosis (odds ratio [OR] 2.355, 95% confidence interval [CI] 1.348-4.116, P = 0.003), light smoking (OR 2.041, 95% CI 1.157-3.602, P = 0.014), and initiation of smoking before 18 years of age (OR 0.570, 95% CI 0.327-0.994, P = 0.047). Light smoking (OR 1.762, 95% CI 1.019-3.144, P = 0.043) and initiation of smoking before 18 years (OR 0.588, 95% CI 0.381-0.908, P = 0.017) were also associated with overall smoking cessation. CONCLUSION: Quitters after CD diagnosis, including quitters at diagnosis and quitters during follow-up, had features distinct from those of non-quitters. Given the motivation at CD diagnosis, a detailed history of smoking habits should be taken and all current smokers should be encouraged to quit smoking at the time of CD diagnosis.

Intermediate-length colonoscope needs more training duration than long-length colonoscope.

OBJECTIVE: Few data are available on the influence of a colonoscope length for trainees, which could affect both the duration of training and colonoscopy quality. We conducted this study to validate which scope needs more duration for training to reach technical competence and to shows better quality indicators during diagnostic colonoscopy. MATERIALS AND METHODS: We conducted a prospective randomly assigned study from April 2010 to February 2011 at Asan Medical Center. Among the 1329 patients enrolled, 1200 colonoscopies were analyzed. We compared cecal intubation rate, adenoma detection rate, cecal intubation time (<20 min), and withdrawal time between the trainees using the intermediate-length colonoscope and those using long-length colonoscope. RESULTS: Trainees who used the long-length colonoscope showed a higher overall cecal intubation rate (88.2% vs. 81.0%, p = 0.001) and adenoma detection rate (49.7% vs. 34.2%, p < 0.001) than those using the intermediate-length colonoscope. The successful cecal intubation rate improved significantly and reached the requisite standard of competence (>90%) after 90 procedures in the long-length colonoscope group. However, the trainees using the intermediate-length colonoscope reached the requisite standard of competence after 150 procedures. Logistic regression analysis revealed that prolonged cecal intubation was associated with the use of the intermediate-length colonoscope, poor colon preparation, a small number of esophagastroduodenoscopy or sigmoidoscopy procedures conducted, and pain during procedures. CONCLUSIONS: During the same training period, use of the long-length colonoscope in trainee was better in terms of reaching competency and quality indicators, and was less painful for the patients during colonoscopic procedures.

Education can improve inter-observer agreement in double balloon endoscopy assessment.

BACKGROUND/AIMS: Double balloon endoscopy (DBE) diagnoses are difficult because of the frequent occurrence of tiny lesions or longitudinally arranged erythema, which may result from true pathologies or contact trauma during DBE insertion. This study was performed to explore interobserver agreement for DBE assessment and to determine whether this can be improved by educating endoscopists. METHODOLOGY: Three experienced endoscopists independently evaluated DBE pictures of an initial training set, consisting of 124 lesions in 78 patients with suspected small bowel bleeding. Each endoscopist made DBE diagnoses independently and assessed whether each lesion was or was not a source of bleeding. Tiny lesions were assessed twice, assuming they had been detected during DBE insertion and/or DBE withdrawal. After two sessions of consensus review by the endoscopists, they evaluated a second, validation set consisting of an additional 43 lesions in 30 patients. RESULTS: Inter-observer agreement with respect to DBE assessment was moderate in the initial training set, but improved to good in the second validation set. The proportion of diagnoses changed to DBE contact lesions when they were assumed to be detected during DBE withdrawal decreased in the second validation set. CONCLUSIONS: Although inter-observer agreement for DBE assessment is not excellent, it can be improved by education of endoscopists.