Cost-effectiveness of artificial intelligence for screening colonoscopy: a modelling study.

BACKGROUND: Artificial intelligence (AI) tools increase detection of precancerous polyps during colonoscopy and might contribute to long-term colorectal cancer prevention. The aim of the study was to investigate the incremental effect of the implementation of AI detection tools in screening colonoscopy on colorectal cancer incidence and mortality, and the cost-effectiveness of such tools. METHODS: We conducted Markov model microsimulation of using colonoscopy with and without AI for colorectal cancer screening for individuals at average risk (no personal or family history of colorectal cancer, adenomas, inflammatory bowel disease, or hereditary colorectal cancer syndrome). We ran the microsimulation in a hypothetical cohort of 100 000 individuals in the USA aged 50-100 years. The primary analysis investigated screening colonoscopy with versus without AI every 10 years starting at age 50 years and finishing at age 80 years, with follow-up until age 100 years, assuming 60% screening population uptake. In secondary analyses, we modelled once-in-life screening colonoscopy at age 65 years in adults aged 50-79 years at average risk for colorectal cancer. Post-polypectomy surveillance followed the simplified current guideline. Costs of AI tools and cost for downstream treatment of screening detected disease were estimated with 3% annual discount rates. The main outcome measures included the incremental effect of AI-assisted colonoscopy versus standard (no-AI) colonoscopy on colorectal cancer incidence and mortality, and cost-effectiveness of screening projected for the average risk screening US population. FINDINGS: In the primary analyses, compared with no screening, the relative reduction of colorectal cancer incidence with screening colonoscopy without AI tools was 44·2% and with screening colonoscopy with AI tools was 48·9% (4·8% incremental gain). Compared with no screening, the relative reduction in colorectal cancer mortality with screening colonoscopy with no AI was 48·7% and with screening colonoscopy with AI was 52·3% (3·6% incremental gain). AI detection tools decreased the discounted costs per screened individual from $3400 to $3343 (a saving of $57 per individual). Results were similar in the secondary analyses modelling once-in-life colonoscopy. At the US population level, the implementation of AI detection during screening colonoscopy resulted in yearly additional prevention of 7194 colorectal cancer cases and 2089 related deaths, and a yearly saving of US$290 million. INTERPRETATION: Our findings suggest that implementation of AI detection tools in screening colonoscopy is a cost-saving strategy to further prevent colorectal cancer incidence and mortality. FUNDING: European Commission and Japan Society of Promotion of Science.

The Colon Endoscopic Bubble Scale (CEBuS): a two-phase evaluation study.

BACKGROUND: To date, no scale has been validated to assess bubbles associated with bowel preparation. This study aimed to develop and assess the reliability of a novel scale - the Colon Endoscopic Bubble Scale (CEBuS). METHODS: This was a multicenter, prospective, observational study with two online evaluation phases of 45 randomly distributed still colonoscopy images (15 per scale grade). Observers assessed images twice, 2 weeks apart, using CEBuS (CEBuS-0 - no or minimal bubbles, covering < 5 % of the surface; CEBuS-1 - bubbles covering 5 %-50 %; CEBuS-2 - bubbles covering > 50 %) and reporting the clinical action (do nothing; wash with water; wash with simethicone). RESULTS: CEBuS provided high levels of agreement both in evaluation Phase 1 (4 experts) and Phase 2 (6 experts and 13 non-experts), with almost perfect intraobserver reliability: kappa 0.82 (95 % confidence interval 0.75-0.88) and 0.86 (0.85-0.88); interobserver agreement - intraclass correlation coefficient (ICC) 0.83 (0.73-0.89) and 0.90 (0.86-0.94). Previous endoscopic experience had no influence on agreement among experts vs. non-experts: kappa 0.86 (0.80-0.91) vs. 0.87 (0.84-0.89) and ICC 0.91 (0.87-0.94) vs. 0.90 (0.86-0.94), respectively. Interobserver agreement on clinical action was ICC 0.63 (0.43-0.78) in Phase 1 and 0.77 (0.68-0.84) in Phase 2. Absolute agreement on clinical action per scale grade was 85 % (82-88) for CEBuS-0, 21 % (16-26) for CEBuS-1, and 74 % (70-78) for CEBuS-2. CONCLUSION: CEBuS proved to be a reliable instrument to standardize the evaluation of colonic bubbles during colonoscopy. Assessment in daily practice is warranted.

Gastric Hyperplastic Polyps: A Benign Entity? Analysis of Recurrence and Neoplastic Transformation in a Cohort Study.

INTRODUCTION: Hyperplastic polyps represent 30-93% of all gastric epithelial polyps. They are generally detected as innocuous incidental findings; however, they have a risk of neoplastic transformation and recurrence. Frequency and risk factors for neoplastic transformation and recurrence are not well established and are fields of ongoing interest. This study aims to evaluate the frequency of and identify the risk factors for recurrence and neoplastic change of gastric hyperplastic polyps (GHP). METHODS: A single-centre retrospective cohort study including consecutive patients who underwent endoscopic resection of GHP from January 2009 to June 2020. Demographic, endoscopic, and histopathologic data was retrieved from the electronic medical records. RESULTS: A total of 195 patients were included (56% women; median age 67 [35-87] years). The median size of GHP was 10 (3-50) mm, 62% (n = 120) were sessile, 61% (n = 119) were located in the antrum, and 36% (n = 71) had synchronous lesions. Recurrence rate after endoscopic resection was 23% (n = 26). In multivariate analysis, antrum location was the only risk factor for recurrence (odds ratio [OR] 3.0; 95% confidence interval [CI] 1.1-8.1). Overall, 5.1% (n = 10) GHP showed neoplastic transformation, with low-grade dysplasia in 5, high-grade dysplasia in 4, and adenocarcinoma in 1. In multivariate analysis, a size >25 mm (OR 84; 95% CI 7.4-954) and the presence of intestinal metaplasia (OR 7.6; 95% CI 1.0-55) and dysplasia (OR 86; 95% CI 10-741) in adjacent mucosa were associated with an increased risk of neoplastic transformation. Recurrence was not associated with neoplastic transformation (OR 1.1; 95% CI 0.2-5.9). DISCUSSION: Our results confirmed the risk of recurrence and neoplastic transformation of GHP. Antrum location was a predictor of recurrence. The risk of neoplastic change was increased in large lesions and with intestinal metaplasia and dysplasia in adjacent mucosa. More frequent endoscopic surveillance may be required in these subgroups of GHP.

INTRODUÇÃO E OBJETIVOS: Os pólipos hiperplásicos constituem 30­93% das lesões gástricas epiteliais benignas. Apresentam-se frequentemente como achados endoscópicos incidentais inocuos, no entanto, apresentam risco de recorrência e transformação neoplásica. A frequência e fatores associados à recorrência e transformação neoplásica não estão bem estabelecidos e são áreas de interesse crescente. Este estudo pretende avaliar a frequência e identificar fatores associados à recorrência e transformado maligna dos pólipos gástricos hiperplásicos (PGH). MÉTODOS: Estudo de coorte retrospectivo unicèntrico incluindo consecutivamente doentes com PGH submetidos a ressedo endoscópica entre janeiro de 2009 e junho de 2020. Efetuada análise das características demográficas, endoscópicas e anatomopatológicas através dos registos clínicos eletrónicos. SUMARIOS DOS RESULTADOS: Incluídos 195 doentes género feminino: 56%, idade mediana: 67 (35­87) anos. Os pólipos apresentavam tamanho mediano de 10 (3­50) mm, 61.5% (n = 120) eram sésseis, 61% (n = 119) apresentavam localização antral e 36% (n = 71) tinham lesões síncronas. A frequência de recorrência foi 23% (n = 26). Na análise multivariada, apenas a localização no antro se associou significativamente a recorrência odds ratio (OR): 3.0; intervalo de confiança (IC) 95%: 1.1­8.1). Em 5.1% (n = 10) dos casos verificou-se transformação neoplásica, correspondendo a 5 casos de displasia de baixo grau, 4 casos de displasia de alto grau e 1 caso de carcinoma. Na análise multivariada, dimensão superior a 25 mm (OR: 84; IC95%: 7.4­954), presença de metaplasia (OR: 7.6; IC95%: 1.0­55) e displasia (OR: 86; IC95%: 10­741) na mucosa adjacente associaram-se a transformação neoplásica. A recorrência não se associou a transformação neoplásica (OR: 1.1; IC95%: 0.2­5.9). DISCUSSÃO/CONCLUSÃO: Estes resultados corroboraram os riscos de recorrência e transformado neoplásica associados aos PGH. A localização antral foi preditor de recorrência. O risco de transformado neoplásica foi superior em lesões maiores e coexistência de metaplasia e displasia na mucosa adjacente. Poderá justificar-se uma estratégia de vigilancia mais frequente nestes subgrupos.

Duodenal lymphangiectasias - The first sign of pancreatic adenocarcinoma.

A 41-year-old caucasian female, with past medical history of pituitary adenoma medicated with cabergoline, presented with worsening dyspepsia and unintentional weight loss of 5%. Physical exam and laboratory results were unremarkable for pathological findings. Esophagogastroduodenoscopy revealed an oedematous and exuberant lymphangiectasia appearance in the duodenum, with no ulceration or suspected infiltration component. However, duodenal biopsies revealed infiltration by poorly differentiated carcinoma. In the meantime, infection and inflammatory/autoimmune causes were ruled out. A CT scan was performed revealing a thickened and enlarged pancreas with ill-defined limits and several intra-abdominal adenopathies, raising suspicion of pancreatic lymphoproliferative disease. EUS with FNB was performed with biopsy of the pancreas and one of the larger adenopathy. EUS also revealed an enlarged, non-nodular pancreas and a thickened duodenal wall. Mild ascites was detected. Both EUS-biopsies were concordant on the diagnosis of carcinoma with gastric or pancreatic-biliary origin, highly aggressive (Ki67 > 80 %). Therefore, the diagnosis of pancreatic adenocarcinoma was assumed (cT4N1Mx). The patient is currently on palliative chemotherapy and remains paucisymptomatic.

A 3-day low-fibre diet does not improve colonoscopy preparation results compared to a 1-day diet: A randomized, single-blind, controlled trial.

Background: Although a 1-day low-fibre diet before colonoscopy is currently recommended, some endoscopists prescribe a 3-day diet. Objective: The objective of this study was to compare the influence of a 3-day versus a 1-day low-fibre diet on bowel preparation quality, patient tolerability and adherence. Methods: Outpatients scheduled for total colonoscopy were randomized in two groups, 3-day versus 1-day low-fibre diet, performing a 4-litre polyethylene glycol split-dose. The primary outcome was a reduction of inappropriate preparations in the 3-day low-fibre diet arm from 15% to 5% (bowel preparation was assessed by the Boston Bowel Preparation Scale). Secondary outcomes were adherence to, difficulty to perform, difficulty to obtain and willingness to repeat the diet. Intention-to-treat (ITT) and per-protocol (PP) analyses were conducted for the primary outcome. Results: A total of 412 patients were randomized (206 per group). Bowel preparation quality was similar between groups. On ITT analysis (n = 412), adequate bowel preparation was 91.7% (3-day diet) versus 94.7% (1-day diet), p = 0.24 and on PP analysis (n = 400) 93.5% versus 96.5%, respectively, p = 0.16. Difficulty to perform the diet was significantly higher on the 3-day diet, p = 0.04. No differences were found on difficulty to obtain the diet, willingness to repeat the diet, adverse events and intra-colonoscopy findings. Conclusion: A 3-day low-fibre diet does not bring benefit to the bowel preparation quality and is harder to perform than a 1-day diet.

Epidemiology of Clostridium difficile infection in Portugal: Experience at a tertiary care hospital.

BACKGROUND: Clostridium difficile is the main cause of healthcare-associated diarrhoea. Its incidence, severity and relapse rates increased over the past two decades. AIM: To study epidemiologic characteristics and treatment of Clostridium difficile infection (CDI) and compare with a previous cohort from the same hospital. METHOD: Retrospective analysis of clinical records of CDI diagnosed from 2010 to 2015 and comparison with data from 2004 to 2009. RESULTS: 259 cases were diagnosed, compared to 83 in 2004-2009. There was no difference in mean annual incidence (8.66 versus 7.11 per 1000 patients; p = .116), but a dramatic increase was observed in 2009/2010 (peak incidence: 21.63 cases per 1000 admissions). Females were more affected (61.4% versus 69.9%; p = .177). Median age was 80 and 83 (p = .097). We observed an increase in median number of antibiotics previously used (2 versus 3; p = .147) and in community-associated CDI (6% versus 19.7%; p = .003). There was a continued increase in the use of carbapenems and quinolones until 2010 and a high percentage of refractory cases in 2010. Female gender (p = .043), long-term care facility (LTCF) residency (p = .022) and a higher number of previous antibiotics (median of 3; p = .025) were independent predictors for refractory and recurrent CDI. CONCLUSIONS: CDI incidence achieved a peak in 2009/2010 coinciding with the introduction of alcohol-based hand products, increase in quinolone and carbapenem prescription and a possible outbreak of an epidemic strain. Female gender, LTCF residency and exposure to three or more antibiotics are risk factors for refractory and recurrent CDI. We emphasize the need to restrict use of large spectrum antibiotics.

Quality in Colonoscopy: Beyond the Adenoma Detection Rate Fever.

BACKGROUND: Colonoscopy quality is a hot topic in gastroenterological communities, with several actual guidelines focusing on this aspect. Although the adenoma detection rate (ADR) is the single most important indicator, several other metrics are described and need reporting. Electronic medical reports are essential for the audit of quality indicators; nevertheless, they have proved not to be faultless. AIM: The aim of this study was to analyse and audit quality indicators (apart from ADR) using only our internal electronic endoscopy records as a starting point for improvement. METHODS: An analysis of electronically recorded information of 8,851 total colonoscopies from a single tertiary centre from 2010 to 2015 was performed. RESULTS: The mean patient age was 63.4 ± 8.5 years; 45.5% of them were female, and in 14.6% sedation was used. Photographic documentation was done in 98.4% with 10.7 photographs on average, and 37.4% reports had <8 pictures per exam. Bowel preparation was rated as adequate in 67%, fair in 27% and inadequate in 4.9% of cases. The adjusted caecal intubation rate (CIR) was 92%, while negative predictors were inadequate preparation (OR 119, 95% CI 84-170), no sedation (OR 2.39, 95% CI 1.81-3.15), female gender (OR 1.61, 95% CI 1.38-1.88) and age ≥65 years (OR 1.56, 95% CI 1.34-1.82). In 28% of patients, a snare polypectomy was performed, correlating with adequate preparation (OR 5.75, 95% CI 3.90-8.48), male gender (OR 1.82, 95% CI 1.64-2.01) and age ≥65 years (OR 1.25, 95% CI 1.13-1.37; p < 0.01) as positive predictors. An annual evolution was observed with improvements in photographic documentation (10.7 vs. 12.9; p < 0.001), CIR (91 vs. 94%; p = 0.002) and "adequate" bowel preparation (p = 0.004). Conclusions: There is much more to report than the ADR to ensure quality in colonoscopy practice. Better registry systematization and integrated software should be goals to achieve in the short term.

INTRODUÇÃO: A qualidade em colonoscopia tem sido um tópico de importante discussão com várias orientações publicadas nesta área. Embora a taxa de deteção de adenomas (ADR) seja o indicador mais importante, vários outros indicadores estão descritos e precisam ser estudados/publicados. Os registos médicos eletrónicos são essenciais para a auditoria de indicadores de qualidade, mas provaram não ser infalíveis. OBJETIVO: Analisar e auditar os indicadores de qualidade, além do ADR, recorrendo apenas à base de dados de endoscopia do nosso centro, como ponto de partida para a melhoria. MÉTODOS: Análise do registo eletrónico de 8,851 colonoscopias totais realizadas em um único centro no período 2010­2015. RESULTADOS: A idade média dos pacientes foi de 63.4 ± 8.5 anos, 45.5% do sexo feminino. Em 14.6%, foi utilizada sedação. Documentação fotográfica em 98.4%, com média de 10.7 fotografias por exame e 37.4% com menos de 8 fotos por exame. A preparação intestinal foi avaliada como "adequada" em 67%, "razoável" em 27% e "inadequada" em 4.9%. A taxa de intubação cecal ajustada (CIR) foi de 92%, com preparação inadequada (OR 119, 95% CI 84­170), exame sem sedação (OR 2.39, 95% CI 1.81­3.15), sexo feminino (OR 1.61, 95% CI 1.38­1.88) e idade ≥65 (OR 1.56, 95% CI 1.34­1.82) como preditores negativos. Em 28%, foi realizada polipectomia, correlacionando-se com preparação "adequada" (OR 5.75, 95% CI 3.90­8.48), sexo masculino (OR 1.82, 95% CI 1.64­2.01) e idade ≥65 (OR 1.25, 95% CI 1.13­1.37; p < 0.01). Da evolução anual destacar melhorias na documentação fotográfica (10.7 vs. 12.9; p < 0.001), taxa de intubação cecal (91 vs. 94%; p = 0.002) e preparação intestinal "adequada" (p = 0.004). CONCLUSÕES: Há muito mais para reportar, além do ADR, de modo a garantir colonoscopias com qualidade. Uma maior sistematização no registo e softwares integrados devem ser objetivos a alcançar no curto prazo.