A natural language processing algorithm accurately classifies steatotic liver disease pathology to estimate the risk of cirrhosis.

BACKGROUND: Histopathology remains the gold standard for diagnosing and staging metabolic dysfunction-associated steatotic liver disease (MASLD). The feasibility of studying MASLD progression in electronic medical records based on histological features is limited by the free-text nature of pathology reports. Here we introduce a natural language processing (NLP) algorithm to automatically score MASLD histology features. METHODS: From the Mass General Brigham health care system electronic medical record, we identified all patients (1987-2021) with steatosis on index liver biopsy after excluding excess alcohol use and other etiologies of liver disease. An NLP algorithm was constructed in Python to detect steatosis, lobular inflammation, ballooning, and fibrosis stage from pathology free-text and manually validated in >1200 pathology reports. Patients were followed from the index biopsy to incident decompensated liver disease accounting for covariates. RESULTS: The NLP algorithm demonstrated positive and negative predictive values from 93.5% to 100% for all histologic concepts. Among 3134 patients with biopsy-confirmed MASLD followed for 20,604 person-years, rates of the composite endpoint increased monotonically with worsening index fibrosis stage (p for linear trend <0.005). Compared to simple steatosis (incidence rate, 15.06/1000 person-years), the multivariable-adjusted HRs for cirrhosis were 1.04 (0.72-1.5) for metabolic dysfunction-associated steatohepatitis (MASH)/F0, 1.19 (0.92-1.54) for MASH/F1, 1.89 (1.41-2.52) for MASH/F2, and 4.21 (3.26-5.43) for MASH/F3. CONCLUSIONS: The NLP algorithm accurately scores histological features of MASLD from pathology free-text. This algorithm enabled the construction of a large and high-quality MASLD cohort across a multihospital health care system and disclosed an accelerating risk for cirrhosis based on the index MASLD fibrosis stage.

Impact of a Citywide Benchmarking Intervention on Colonoscopy Quality Performance.

INTRODUCTION: There is marked variability in colonoscopy quality, limiting its effectiveness in colorectal cancer prevention. Multiple indicators have been established as markers for colonoscopy quality; however, there are conflicting data on the effects of quality reporting programs on endoscopist performance. In this study, we investigate the impact of a multicenter quarterly report card initiative on colonoscopy quality metric performance. METHODS: Data were collected from 194 endoscopists at 10 participating sites throughout New York City using a Qualified Clinical Data Registry from January 2013 to December 2014. Adenoma detection rate (ADR), cecal intubation rate, withdrawal time, bowel preparation quality and appropriate interval recommendations were tracked. Report cards were distributed to each site on a quarterly basis and technical assistance was provided as needed. Performance trends were analyzed using the Cochran-Armitage trend and analysis of variance tests. RESULTS: 37,258 screening colonoscopies were performed during the study period. There was a positive performance trend for ADR over time from the first quarter of 2013 to the last quarter of 2014 (15.6-25.7%; p < 0.001). There were also increases in cecal intubation rates (78.2-92.6%; p < 0.001), bowel preparation adequacy rates (77.5-92.8%; p < 0.001), and adherence to appropriate screening intervals (28.0-55.0%; p < 0.001). There was no clinically significant change in mean withdrawal time. CONCLUSIONS: The implementation of a quarterly report card initiative resulted in statistically significant improvements in adenoma detection, cecal intubation, bowel preparation adequacy rates, and appropriate recommended screening intervals.

Temporal Trends and Risk Factors for Postcolonoscopy Colorectal Cancer.

BACKGROUND: Colonoscopy is effective for colorectal cancer (CRC) prevention, yet patients may develop CRC despite adhering to screening/surveillance intervals. There are limited data on predictive factors associated with these postcolonoscopy CRCs (PCCRCs). We aimed to measure PCCRC rates and identify risk factors for PCCRC. METHODS: We performed a case-control study, comparing patients with PCCRCs to spontaneous CRCs diagnosed during a 12.5-year period at an academic medical center. PCCRCs were defined as CRCs diagnosed in between guideline-recommended screening/surveillance intervals. RESULTS: During the 12.5-year period, of 1266 CRCs diagnosed, 122 (10%) were PCCRCs. 70% of PCCRCs were diagnosed within 5 years of a prior colonoscopy. There was an increasing trend for PCCRC rates in recent years [odds ratio (OR), 2.78; 95% confidence interval (CI), 1.51-5.09], with PCCRCs comprising 13.6% of cancers diagnosed in 2016 as compared with 5.7% of cancers diagnosed in 2005. Older age (OR per year, 1.02; 95% CI, 1.01-1.04), proximal colonic location (OR, 1.99; 95% CI, 1.20-3.33) and early stage (OR, 2.57; 95% CI, 1.34-4.95) were associated with PCCRCs. In total, 41% of PCCRCs were diagnosed by a different physician from the physician who did the prior colonoscopy, and 42% of physicians did not diagnose any of their PCCRC cases. CONCLUSIONS: PCCRC rates are rising in recent years, likely reflecting the widespread adoption of colonoscopy as a primary screening tool, and are more common in older patients and those with proximal, early-stage tumors. The finding that a large proportion of PCCRCs are diagnosed by a different physician raises the concern that physicians are unaware of their own patients' PCCRCs.