Machine learning models to predict submucosal invasion in early gastric cancer based on endoscopy features and standardized color metrics.

Conventional endoscopy is widely used in the diagnosis of early gastric cancers (EGCs), but the graphical features were loosely defined and dependent on endoscopists' experience. We aim to establish a more accurate predictive model for infiltration depth of early gastric cancer including a standardized colorimetric system, which demonstrates promising clinical implication. A retrospective study of 718 EGC cases was performed. Clinical and pathological characteristics were included, and Commission Internationale de l'Eclariage (CIE) standard colorimetric system was used to evaluate the chromaticity of lesions. The predicting models were established in the derivation set using multivariate backward stepwise logistic regression, decision tree model, and random forest model. Logistic regression shows location, macroscopic type, length, marked margin elevation, WLI color difference and histological type are factors significantly independently associated with infiltration depth. In the decision tree model, margin elevation, lesion located in the lower 1/3 part, WLI a*color value, b*color value, and abnormal thickness in enhanced CT were selected, which achieved an AUROC of 0.810. A random forest model was established presenting the importance of each feature with an accuracy of 0.80, and an AUROC of 0.844. Quantified color metrics can improve the diagnostic precision in the invasion depth of EGC. We have developed a nomogram model using logistic regression and machine learning algorithms were also explored, which turned out to be helpful in decision-making progress.

A novel optical coherence tomography­based calcium scoring system can predict the stent expansion of moderate and severe calcified lesions.

Coronary calcified lesions can exert serious effects on stent expansion. A calcium scoring system, based on optical coherence tomography (OCT), has been previously developed to identify relatively mild calcified lesions that would benefit from plaque modification procedures. Therefore, the present study aimed to establish a novel OCT-based scoring system to predict the stent expansion of moderate and severe calcified lesions. A total of 33 patients who underwent percutaneous coronary intervention (PCI; 34 calcified lesions were observed using coronary angiography) were retrospectively included in the present study. Coronary angiography and OCT images were subsequently reviewed and analyzed. Furthermore, a calcium scoring system was developed based on the results of multivariate analysis before the optimal threshold for the prediction of stent underexpansion in patients with moderate and severe calcified lesions was determined. The mean age of the patients was 67±10 years. The present analysis demonstrated that the final post-PCI median stent expansion was 70.74%, where stent underexpansion (defined as stent expansion <80%) was observed in 23 lesions. The mean maximum calcium arc, length and thickness, which were assessed using OCT, were found to be 230˚, 25.10 mm and 1.18 mm, respectively. A multivariate logistic regression model demonstrated that age and the maximum calcium arc were independent predictors of stent underexpansion. A novel calcium scoring system was thereafter established using the following formula: (0.16 x age) + (0.03 x maximum calcium arc) according to the ß-coefficients in the multivariate analysis, with the optimal cut-off value for the prediction of stent underexpansion being 16.87. Receiver operating characteristic curve analysis demonstrated that this novel scoring system yielded a larger area under the curve value compared with that from a previous study's scoring system. Therefore, in conclusion, since the calcium scoring system of the present study based on age and the maximum calcium arc obtained from OCT was specifically developed in the subjects with moderate and severe calcified lesions, it may be more accurate in predicting the risk of stent underexpansion in these patients.

The Association between Non-Alcoholic Fatty Liver Disease (NAFLD) and Advanced Fibrosis with Serological Vitamin B12 Markers: Results from the NHANES 1999-2004.

BACKGROUND: There is evidence that vitamin B12 and associated metabolite levels are changed in non-alcoholic fatty liver disease (NAFLD) and non-alcoholic steatohepatitis (NASH); however, their association has been in dispute. METHODS: We included 8397 individuals without previous liver condition or excess alcohol intake from the National Health and Nutrition Examination Survey (NHANES) 1999-2004. NAFLD was diagnosed with Fatty Liver Index (FLI) ≥ 60 or USFLI ≥ 30, and participants with advanced fibrosis risks were identified with elevated non-alcoholic fatty liver disease fibrosis score (NFS), fibrosis 4 index (FIB-4), or aspartate aminotransferase (AST)/platelet ratio index (APRI). Step-wide logistic regression adjusting for confounders was used to detect the association between NAFLD or advanced fibrosis with serum vitamin B12, folate, red blood cell folate (RBC folate), homocysteine (HCY), and methylmalonic acid (MMA). RESULTS: The weighted prevalence of NAFLD was 44.2%. Compared with non-NAFLD participants, patients with NAFLD showed significantly increased RBC folate level and RBC counts, decreased serum vitamin B12 and folate, and similar HCY and MMA levels. NAFLD with advanced fibrosis risk had higher MMA and HCY, reduced serum vitamin B12, and similar serum folate and RBC folate levels than NAFLD with low fibrosis risk. Only RBC folate was independently associated with an increased risk of NAFLD (OR (95% CI): 2.24 (1.58, 3.18)). In all participants, MMA (OR: 1.41 (1.10, 1.80)) and HCY (OR: 2.76 (1.49, 5.11)) were independently associated with increased risk for advanced fibrosis. In participants with NAFLD, this independent association still existed (OR: 1.39 (1.04, 1.85) for MMA and 1.95 (1.09, 3.46) for HCY). In all participants, the area under the receiver operating characteristic curve (ROC AUC) on fibrosis was 0.6829 (0.6828, 0.6831) for MMA and 0.7319 (0.7318, 0.7320) for HCY; in participants with NAFLD, the corresponding ROC AUC was 0.6819 (0.6817, 0.6821) for MMA and 0.6926 (0.6925, 0.6928) for HCY. CONCLUSION: Among vitamin B12-associated biomarkers, RBC folate was independently associated with elevated NAFLD risk, whereas MMA and HCY were associated with increased risk for advanced fibrosis in the total population and NAFLD participants. Our study highlighted the clinical diagnostic value of vitamin B12 metabolites and the possibility that vitamin B12 metabolism could be a therapeutic target for NASH. Further studies using recent perspective data with biopsy proven NASH could be conducted to validate our results.