Differences in liver and mortality outcomes of non-alcoholic fatty liver disease by race and ethnicity: A longitudinal real-world study.

BACKGROUND/AIMS: Understanding of nonalcoholic fatty liver disease (NAFLD) continues to expand, but the relationship between race and ethnicity and NAFLD outside the use of cross-sectional data is lacking. Using longitudinal data, we investigated the role of race and ethnicity in adverse outcomes in NAFLD patients. METHODS: Patients with NAFLD confirmed by imaging via manual chart review from any clinics at Stanford University Medical Center (1995-2021) were included. Primary study outcomes were incidence of liver events and mortality (overall and non-liver related). RESULTS: The study included 9,340 NAFLD patients: White (44.1%), Black (2.29%), Hispanic (27.9%), and Asian (25.7%) patients. For liver events, the cumulative 5-year incidence was highest among White (19.1%) patients, lowest among Black (7.9%) patients, and similar among Asian and Hispanic patients (~15%). The 5-year and 10-year cumulative overall mortality was highest for Black patients (9.2% and 15.0%, respectively, vs. 2.5-3.5% and 4.3-7.3% in other groups) as well as for non-liver mortality. On multivariable regression analysis, compared to White patients, only Asian group was associated with lower liver-related outcomes (aHR: 0.83, P=0.027), while Black patients were at more than two times higher risk of both non-liver related (aHR: 2.35, P=0.010) and overall mortality (aHR: 2.13, P=0.022) as well as Hispanic patients (overall mortality: aHR: 1.44, P=0.022). CONCLUSION: Compared to White patients, Black patients with NAFLD were at the highest risk for overall and non-liver-related mortality, followed by Hispanic patients with Asian patients at the lowest risk for all adverse outcomes. Culturally sensitive and appropriate programs may be needed for more successful interventions.

Differential Mortality Outcomes in Real-world Patients with Lean, Nonobese, and Obese Nonalcoholic Fatty Liver Disease.

Background and Aims: Nonalcoholic fatty liver disease (NAFLD) is commonly associated with obesity but can develop in normal-weight people (lean NAFLD). We compared outcomes in lean, overweight, and obese NAFLD. Methods: This retrospective chart review included patients at Stanford University Medical Center with NAFLD confirmed by imaging between March 1995 and December 2021. Lean, overweight, and obese patients had body mass index of <25.0, >25.0 and <29.9, and ≥30.0 kg/m2 for non-Asian and >23.0 and ≥27.5 for overweight and obese Asian patients. Results: A total of 9061 lean (10.2%), overweight (31.7%), and obese (58.1%) patients were included. Lean patients were 5 years older than obese patients (53±17.4 vs. 48.7±15.1 years), more were female (59.6% vs. 55.2%), white (49.1% vs. 46.5%), had NASH (29.2% vs. 22.5%), cirrhosis (25.3% vs.19.2%), or nonliver cancer (25.3% vs. 18.3%). Fewer had diabetes (21.7% vs. 35.8%) or metabolic comorbidities (all p<0.0001). Lean NAFLD patients had liver-related mortality similar to other groups but higher overall (p=0.01) and nonliver-related (p=0.02) mortality. After multivariable model adjustment for covariates, differences between lean and obese NAFLD in liver-related, nonliver-related, and overall mortality (adjusted hazard ratios of 1.34, 1.00, and 1.32; p=0.66, 0.99, and 0.20, respectively) were not significant. Conclusions: Lean NAFLD had fewer metabolic comorbidities but similar adverse or worse outcomes, suggesting that it is not benign. Healthcare providers should provide the same level of care and intervention as for overweight and obese NAFLD.

Artificial Intelligence System for Automatic Quantitative Analysis and Radiology Reporting of Leg Length Radiographs.

Leg length discrepancies are common orthopedic problems with the potential for poor functional outcomes. These are frequently assessed using bilateral leg length radiographs. The objective was to determine whether an artificial intelligence (AI)-based image analysis system can accurately interpret long leg length radiographic images. We built an end-to-end system to analyze leg length radiographs and generate reports like radiologists, which involves measurement of lengths (femur, tibia, entire leg) and angles (mechanical axis and pelvic tilt), describes presence and location of orthopedic hardware, and reports laterality discrepancies. After IRB approval, a dataset of 1,726 extremities (863 images) from consecutive examinations at a tertiary referral center was retrospectively acquired and partitioned into train/validation and test sets. The training set was annotated and used to train a fasterRCNN-ResNet101 object detection convolutional neural network. A second-stage classifier using a EfficientNet-D0 model was trained to recognize the presence or absence of hardware within extracted joint image patches. The system was deployed in a custom web application that generated a preliminary radiology report. Performance of the system was evaluated using a holdout 220 image test set, annotated by 3 musculoskeletal fellowship trained radiologists. At the object detection level, the system demonstrated a recall of 0.98 and precision of 0.96 in detecting anatomic landmarks. Correlation coefficients between radiologist and AI-generated measurements for femur, tibia, and whole-leg lengths were > 0.99, with mean error of < 1%. Correlation coefficients for mechanical axis angle and pelvic tilt were 0.98 and 0.86, respectively, with mean absolute error of < 1°. AI hardware detection demonstrated an accuracy of 99.8%. Automatic quantitative and qualitative analysis of leg length radiographs using deep learning is feasible and holds potential in improving radiologist workflow.

Meta-analysis: global prevalence, trend and forecasting of non-alcoholic fatty liver disease in children and adolescents, 2000-2021.

BACKGROUND: NAFLD is increasing in children. AIMS: To determine the recent trend and forecast the future global prevalence of paediatric NAFLD METHODS: We searched PubMed, Embase, Web of Science and Cochrane library databases from inception to 1 May 2021 for studies of children and adolescents (≤21 years) with NAFLD. Obesity was defined with weight at ≥95th percentile and overweight as 85th to <95th percentile as per the Center for Disease Control BMI-for-age percentile cut-offs. RESULTS: From 3350 titles and abstracts, we included 74 studies (276,091 participants) from 20 countries/regions. We included 14 studies in the general NAFLD prevalence analysis, yielding an overall prevalence of 7.40% (95% CI: 4.17-12.81) regardless of the diagnostic method, and 8.77% (95% CI: 3.86-18.72) by ultrasound. Among continents with more than one study, the prevalence of NAFLD was 8.53% (95% CI: 5.71-12.55) for North America, 7.01% (95% CI: 3.51-13.53) for Asia, and 1.65% (95% CI: 0.97-2.80) for Europe. NAFLD prevalence regardless of the diagnostic method was 52.49% (95% CI: 46.23-58.68, 9159 participants) and 39.17% (95% CI: 30.65-48.42, 5371 participants) among obese and overweight/obese participants, respectively. For the general population, trend analysis from 2000 to 2017 indicates an increasing global prevalence of paediatric NAFLD from 4.62% to 9.02% at a yearly increase of 0.26%, whereas forecast analysis predicts a prevalence of 30.7% by 2040. CONCLUSION: The prevalence of paediatric NAFLD varies by region and is 52.49% overall among the obese population and 7.40% in the general population. It is predicted to reach 30.7% by 2040.

Automating Scoliosis Measurements in Radiographic Studies with Machine Learning: Comparing Artificial Intelligence and Clinical Reports.

Scoliosis is a condition of abnormal lateral spinal curvature affecting an estimated 2 to 3% of the US population, or seven million people. The Cobb angle is the standard measurement of spinal curvature in scoliosis but is known to have high interobserver and intraobserver variability. Thus, the objective of this study was to build and validate a system for automatic quantitative evaluation of the Cobb angle and to compare AI generated and human reports in the clinical setting. After IRB was obtained, we retrospectively collected 2150 frontal view scoliosis radiographs at a tertiary referral center (January 1, 2019, to January 1, 2021, ≥ 16 years old, no hardware). The dataset was partitioned into 1505 train (70%), 215 validation (10%), and 430 test images (20%). All thoracic and lumbar vertebral bodies were segmented with bounding boxes, generating approximately 36,550 object annotations that were used to train a Faster R-CNN Resnet-101 object detection model. A controller algorithm was written to localize vertebral centroid coordinates and derive the Cobb properties (angle and endplate) of dominant and secondary curves. AI-derived Cobb angle measurements were compared to the clinical report measurements, and the Spearman rank-order demonstrated significant correlation (0.89, p < 0.001). Mean difference between AI and clinical report angle measurements was 7.34° (95% CI: 5.90-8.78°), which is similar to published literature (up to 10°). We demonstrate the feasibility of an AI system to automate measurement of level-by-level spinal angulation with performance comparable to radiologists.