Background/Aims: Colorectal polyps are precursor lesions of colorectal cancer, and their histopathological types are critical for deter mining malignant potential. Predicting polyp histopathological types may support early and appropriate clinical management. Machine learning (ML) algorithms based on accessible demographic, clinical, and lifestyle data can contribute to individualized screening strategies. Materials and Methods: This retrospective cross-sectional study included 491 individuals who underwent colonoscopy for the first time between 2022 and 2025 at University of Health Sciences, Antalya Training and Research Hospital. Demographic and clinical data were recorded, and dietary habits were assessed using the Food Frequency Questionnaire. Patients were classified into 3 groups according to histopathology: adenomatous polyp, hyperplastic polyp, and no polyp. Four ML algorithms—decision tree, random forest, support vector machines (SVMs), and extreme gradient boosting—were applied. Model performance was evaluated using accuracy, sensitivity, specific ity, kappa statistic, and McNemar’s test. Variable contributions were further analyzed with SHapley Additive exPlanations. Results: Accuracy ranged from 70.9% to 76.4%, with the highest performance from SVM (76.4%) and random forest (75.7%). Extreme gradient boosting showed lower overall accuracy (70.9%) but was the only model that identified hyperplastic polyps. The no polyp groupwas consistently predicted with high accuracy (sensitivity 85.6%-95.9%). Precision for adenomatous polyps was highest with SVM (71.4%). SHapley Additive exPlanations analysis highlighted frequent bulgur consumption (>2 times/week), red meat intake, age, and body mass index as major predictors. Conclusion: Machine learning algorithms can predict colorectal polyp histopathological types using routine demographic, clinical, and dietary data, enabling more personalized and effective screening beyond age-based protocols. Cite this article as: Koker G, Gorgulugil GZ, Coskuner MA, Durmus ME. Machine learning–based prediction of histopathological classification in colorectal polyps. Turk J Gastroenterol. 2025;36(10):700-707

Background/Aims: To determine the role of serum procalcitonin levels and ascites/subcutaneous echogenicity ratio (ASER) in predicting ascites infection in hospitalized cirrhotic patients. Materials and Methods: A total of 50 patients hospitalized because of cirrhosis-related ascites were included in this study. In these patients, 44% of ascites were infected (peritonitis), whereas 56% of ascites were sterile. These two groups were compared in terms of procalcitonin levels and ASER for predicting ascites infection. Receiver operating characteristic (ROC) curves were used to evaluate the diagnostic performance of ASER, and the predicting outcome of ASER was compared with procalcitonin levels. Results: The ASER values of the patients with the diagnosis of infected ascites were significantly higher than in those with the diagnosis of sterile ascites (p<0.001). ROC analysis was performed to determine the diagnostic ASER value for infected ascites. An ASER greater than 0.0019 determined peritonitis with 95.5% sensitivity and 100% specificity. A procalcitonin level greater than 0.05 determined peritonitis with 86.4% sensitivity and 75% specificity. Using ROC analysis, an ASER greater than 0.0019 [area under curve (AUC): 0.974, 95% confidence interval (CI) (0.884–0.999, p0.5 mg/dL [AUC: 0.860, 95% CI (0.884–0.999, p<0.001) (p<0.045)]. Conclusion: According to our findings, the determination of ASER and serum procalcitonin levels seems to provide satisfactory diagnostic accuracy in differentiating ascites infections in hospitalized cirrhotic patients. ASER values significantly differentiate ascites infections better than procalcitonin levels.

Background/Objectives: CD71⁺ erythroid cells (CECs) are immature red blood cells (proerythroblasts, erythroblasts, and reticulocytes). CECs play an important role in the development of sepsis and cancer by causing immunosuppression. We examined the CEC levels in the peripheral blood of beta thalassemia (βThal) patients and investigated the relationship between CECs and the clinical status of the patients, especially splenectomy.Methods: Sixty‐eight patients with βThal (46 splenectomized and 22 nonsplenectomized) and 15 healthy controls were included in this study. The hemogram parameters, ferritin, and CECs (flow cytometry method) were measured.Results: It was observed that the CEC level in the patient group was significantly higher than the control group (p < 0.05). CEC levels were found to be significantly higher in patients with splenectomy than in patients with nonsplenectomy (p < 0.05). CEC levels were higher in patients with nontransfusion‐dependent βT (NTD‐βThal) than in patients with transfusion‐dependent βT (TD‐βThal) (p < 0.05). CEC levels were found to be significantly higher in patients with splenectomy than in patients with nonsplenectomy in both TD‐βThal and NTD‐βThal groups (p < 0.05). There was a moderate‐negative correlation was detected between CECs and Hb levels (r = −0.467; p < 0.05).Conclusions: High CEC levels in βThal patients develop as a result of ineffective erythropoiesis. We think that keeping CEC levels under control is important for prognosis, especially in patients with splenectomy.

Introduction: Obesity has become a global health issue, with its prevalence steadily increasing. It is closely linked to several metabolic disorders, cardiovascular diseases, non-alcoholic fatty liver disease, and chronic low-grade inflammation and can progress to more severe liver conditions. This study evaluates the relationship between obesity and inflammatory markers in individuals with different obesity levels. Methods: A cross-sectional study was conducted among 50 patients categorized into three obesity classes based on body mass index (BMI). Blood samples were taken to evaluate inflammatory and metabolic markers, including white blood cell (WBC) count, neutrophil-to-lymphocyte ratio (NLR), and C-reactive protein (CRP). Results: There were no statistically significant differences in inflammatory markers such as WBC, NLR, or CRP; a trend toward higher CRP levels was observed in Class 3 obesity. Conclusion: In our study, no statistically significant association was observed between inflammatory markers and the degree of obesity. Although the sample size was relatively small, it is essential to acknowledge that obesity is a multifaceted condition, and genetic variations may play a role in these results.

Objective: Alcohol intoxication frequently leads to emergency department admissions with a wide range of clinical severity. The prognostic role of the neutrophil-to-lymphocyte ratio in this setting remains unclear. This study investigated the association between the neutrophil-to-lymphocyte ratio and disease severity, as well as its predictive value for adverse clinical outcomes in patients with alcohol intoxication.Methods: This multicenter, retrospective, cross-sectional study included patients hospitalized for alcohol intoxication. Demographic, clinical, and laboratory data were obtained from electronic health records. Group comparisons, correlation analyses, multivariable logistic regression, and ROC analyses were performed to assess the association between the neutrophil to lymphocyte ratio and clinical outcomes.Results: A total of 64 patients were included. Hemodialysis was required in 42.2% of patients, and in-hospital mortality was 20.3%. NLR was significantly higher in patients requiring hemodialysis (p = 0.014) and showed a positive correlation with length of hospital stay (rho = 0.29, p = 0.019). Ocular and neurological findings were strongly associated with hemodialysis requirement (both p < 0.001). In ROC analysis, NLR demonstrated moderate discriminative ability for predicting hemodialysis requirement (AUC = 0.681). Multivariable logistic regression identified ocular findings as the strongest independent predictor, while the independent effect of NLR was limited.Conclusion: NLR is associated with markers of clinical severity and inflammatory burden in alcohol intoxication and correlates with hemodialysis requirement and length of hospital stay. However, its standalone predictive performance is limited. NLR may be used as a supportive biomarker alongside clinical assessment rather than as a sole decision-making tool.

Background: Osteoporosis has a rising global incidence and social burden. Serum uric acid’s dual roles in oxidative stress and inflammation may influence bone health, but findings are inconsistent and require further research. This study aimed to evaluate the relationship between SUA levels and osteoporosis in a multicenter cohort obtained from different regions of Türkiye. Methods: This multi-center retrospective study included 3280 individuals, postmenopausal women and men aged 45 and older, from 16 centers in Türkiye. Individuals were excluded if they recently consumed alcohol, had severe renal dysfunction, certain hormonal or mineral disorders, specific medications, or certain menopausal statuses. Bone mineral density (BMD) at the hip and lumbar spine was measured using dual-energy X-ray absorptiometry (DXA), and participants were classified as normal or having osteopenia or osteoporosis based on T-score thresholds. Results: Overall, 34.8% were male, and 65.2% were female. For the lumbar spine, 36.8% had osteopenia, and 13.5% had osteoporosis; similarly, for the total hip, 40.8% had osteopenia, and 7.9% had osteoporosis. ROC analysis identified a threshold of 3.9 mg/dL serum uric acid (SUA) (AUC 0.374; p < 0.001), which was positively associated with both lumbar and total hip BMD. Osteoporosis rates were higher in patients with SUA < 3.9 mg/dL compared to those with SUA ≥ 3.9 mg/dL at the lumbar spine (29.1% vs. 14.2%, p < 0.001) and total hip sites (23.6% vs. 15.9%, p = 0.003). After adjustment for potential confounders, SUA was a significant independent predictor of osteoporosis in the lumbar spine (OR 0.70; p < 0.001) and the hip (OR 0.80; p < 0.001). Conclusions: Serum uric acid levels are inversely linked to bone mineral density and osteoporosis risk, indicating a potential role in bone health. However, due to study limitations, causal relationships remain unproven, and further research is needed.