Advancing Automatic Gastritis Diagnosis: An Interpretable Multilabel Deep Learning Framework for the Simultaneous Assessment of Multiple Indicators.

The evaluation of morphologic features, such as inflammation, gastric atrophy, and intestinal metaplasia, is crucial for diagnosing gastritis. However, artificial intelligence analysis for nontumor diseases like gastritis is limited. Previous deep learning models have omitted important morphologic indicators and cannot simultaneously diagnose gastritis indicators or provide interpretable labels. To address this, an attention-based multi-instance multilabel learning network (AMMNet) was developed to simultaneously achieve the multilabel diagnosis of activity, atrophy, and intestinal metaplasia with only slide-level weak labels. To evaluate AMMNet's real-world performance, a diagnostic test was designed to observe improvements in junior pathologists' diagnostic accuracy and efficiency with and without AMMNet assistance. In this study of 1096 patients from seven independent medical centers, AMMNet performed well in assessing activity [area under the curve (AUC), 0.93], atrophy (AUC, 0.97), and intestinal metaplasia (AUC, 0.93). The false-negative rates of these indicators were only 0.04, 0.08, and 0.18, respectively, and junior pathologists had lower false-negative rates with model assistance (0.15 versus 0.10). Furthermore, AMMNet reduced the time required per whole slide image from 5.46 to 2.85 minutes, enhancing diagnostic efficiency. In block-level clustering analysis, AMMNet effectively visualized task-related patches within whole slide images, improving interpretability. These findings highlight AMMNet's effectiveness in accurately evaluating gastritis morphologic indicators on multicenter data sets. Using multi-instance multilabel learning strategies to support routine diagnostic pathology deserves further evaluation.

Digital Image Analysis of Ki67 Heterogeneity Improves the Diagnosis and Prognosis of Gastroenteropancreatic Neuroendocrine Neoplasms.

Ki67 is a reliable grading and prognostic biomarker of gastroenteropancreatic neuroendocrine neoplasms (GEP-NENs). The intratumor heterogeneity of Ki67, correlated with tumor progression, is a valuable factor that requires image analysis. The application of digital image analysis (DIA) enables new approaches for the assessment of Ki67 heterogeneity distribution. We investigated the diagnostic utility of Ki67 heterogeneity parameters in the classification and grading of GEP-NENs and explored their clinical values with regard to their prognostic relevance. The DIA algorithm was performed on whole-slide images of 102 resection samples with Ki67 staining. Good agreement was observed between the manual and DIA methods in the hotspot evaluation (R2 = 0.94, P < .01). Using the grid-based region of interest approach, score-based heat maps provided a distinctive overview of the intratumoral distribution of Ki67 between neuroendocrine carcinomas and neuroendocrine tumors. The computation of heterogeneity parameters related to DIA-determined Ki67 showed that the coefficient of variation and Morisita-Horn index were directly related to the classification and grading of GEP-NENs and provided insights into distinguishing high-grade neuroendocrine neoplasms (grade 3 neuroendocrine tumor vs neuroendocrine carcinoma, P < .01). Our study showed that a high Morisita-Horn index correlated with poor disease-free survival (multivariate analysis: hazard ratio, 56.69), which was found to be the only independent predictor of disease-free survival in patients with GEP-NEN. These spatial biomarkers have an impact on the classification and grading of tumors and highlight the prognostic associations of tumor heterogeneity. Digitization of Ki67 variations provides a direct and objective measurement of tumor heterogeneity and better predicts the biological behavior of GEP-NENs.

Sex-specific prediction value of common carotid artery diameter for stroke risk in a hypertensive population: a cross-sectional study.

BACKGROUND: The importance of sex as a risk factor for stroke has been established. This study aimed to assess sex-related disparities in carotid artery diameter and stroke in a hypertensive population. METHODS: The cross-sectional survey was conducted in rural areas of northeast China. A multistage cluster sampling method was employed to select a representative population. The study comprised 3,245 individuals with hypertension. The common carotid artery (CCA) interadventitial diameter was measured by ultrasound. A linear model of restricted cubic spline function was used to characterize the concentration-response (C-R) relationship between CCA diameter and stroke. RESULTS: The overall prevalence of stroke was 8.9% among hypertensive individuals, with a higher rate in men than in women (10.8% vs. 7.6%). When the women's CCA diameters were divided into quartiles, the top quartile (>8.10 mm) had a 2.49 (95% CI: 1.36-4.56) times greater risk of stroke compared to the bottom quartile (≤6.80 mm) after adjustment was made for other variables. The C-R relationship further confirmed a positive association between CCA diameter and stroke prevalence in women. Moreover, a category-free net reclassification index (0.325; 95% CI: 0.173-0.476; P<0.001) and an integrated discrimination index (0.008; 95% CI: 0.004-0.012, P<0.001) showed improvement in predicting the probability of stroke from CCA diameter. However, no significant relationship between CCA diameter and prevalence of stroke was found in men. CONCLUSIONS: The risk of stroke increased proportionally with the enlargement of the CCA diameter in women, supporting the sex-specific value of CCA diameter in optimizing the risk stratification of stroke.

Prevalence and determinants of metabolic syndrome based on three definitions in rural northeast China.

OBJECTIVE: To gain more comprehensive understanding of metabolic syndrome (Mets) among in general Chinese population. DESIGN: Cross-sectional study. Mets was defined by three widely accepted definitions including modified Adults Treatment Panel (ATP) III criteria, International Diabetes Federation (IDF) criteria and harmonized definition. Risk factors were evaluated by using multivariate logistic regression. SETTING: Nineteen rural villages in northeast China. PARTICIPANTS: The survey was conducted in September 2017 and May 2018 on 10 926 individuals. RESULTS: According to modified ATP III criteria, IDF criteria and harmonised definition, the overall prevalence of Mets was 41·3 % (95 % CI 40·3, 42·2), 34·2 % (95 % CI 33·2, 35·1) and 44·1 % (95 % CI 43·1, 45·1), respectively. Females had a higher prevalence, and elevated blood pressure was the most frequent. Age, female sex, non-peasant worker, higher BMI and lower-annual income were independent risk factors of Mets in all three definitions (all ps < 0·05). Based on modified ATP III criteria and harmonised definition, heavy drinking was positively correlated with Mets. In contrast, former drinking was inversely associated with Mets. CONCLUSIONS: Mets is highly prevalent in rural areas of northeast China. Its independent risk factors include higher age, female sex, non-peasantry worker, higher BMI and lower-annual income. Modified ATP III criteria and harmonised definition may be superior definitions of Mets.

Prevalence of Obesity and Associated Risk Factors and Cardiometabolic Comorbidities in Rural Northeast China.

OBJECTIVE: To investigate the epidemiological features of obesity in rural Northeast China. DESIGN: This was a 2017-2018 cross-sectional study of 10,891 participants aged ≥40 years that was designed to investigate the prevalence of obesity in rural areas of Liaoning Province. Demographic data, biochemical parameters, and physical examinations were completed by well-trained personnel. Logistic regression analyses were then carried out to investigate independent risk factors and associated cardiometabolic comorbidities of obesity. RESULTS: The proportions of general obesity only, central obesity only, and combined obesity were 1.0%, 31.3%, and 17.4%, respectively. Overall, 49.8% of our subjects were obese. Female gender, being married, being separated/divorced/widowed, or eating more meat were significantly associated with obesity. Smoking, higher family income, or regular physical exercise were negatively associated with obesity. General obesity only was significantly correlated with hypertension, diabetes, and high triglycerides (OR = 2.79, OR = 2.79, and OR = 3.37, resp.). General obesity only was irrelevant to high total cholesterol, low high-density lipoprotein cholesterol, and high low-density lipoprotein cholesterol, although central obesity only, or combined obesity, was relevant to these factors. Prehypertension and prediabetes showed a positive association with different types of obesity. CONCLUSIONS: We identified a high prevalence of general and central obesity in rural Northeast China, with similar independent risk factors. Participants with combined obesity had the highest risk of cardiometabolic comorbidities, indicating that the combined use of both waist circumference and body mass index is useful in practice.