Surgical resection is the primary treatment for hepatocellular carcinoma (HCC). However, studies indicate that nearly 70% of patients experience HCC recurrence within five years following hepatectomy. The earlier the recurrence, the worse the prognosis. Current studies on postoperative recurrence primarily rely on posto perative pathology and patient clinical data, which are lagging. Hence, develop ing a new pre-operative prediction model for postoperative recurrence is crucial for guiding individualized treatment of HCC patients and enhancing their pro gnosis. To identify key variables in pre-operative clinical and imaging data using ma chine learning algorithms to construct multiple risk prediction models for early postoperative recurrence of HCC. The demographic and clinical data of 371 HCC patients were collected for this retrospective study. These data were randomly divided into training and test sets at a ratio of 8:2. The training set was analyzed, and key feature variables with predictive value for early HCC recurrence were selected to construct six different machine learning prediction models. Each model was evaluated, and the best-performing model was selected for interpreting the importance of each variable. Finally, an online calculator based on the model was generated for daily clinical practice. Following machine learning analysis, eight key feature variables (age, intratumoral arteries, alpha-fetoprotein, pre-operative blood glucose, number of tumors, glucose-to-lymphocyte ratio, liver cirrhosis, and pre-operative platelets) were selected to construct six different prediction models. The XGBoost model outperformed other models, with the area under the receiver operating characteristic curve in the training, validation, and test datasets being 0.993 (95% confidence interval: 0.982-1.000), 0.734 (0.601-0.867), and 0.706 (0.585-0.827), respectively. Calibration curve and decision curve analysis indicated that the XGBoost model also had good predictive performance and clinical application value. The XGBoost model exhibits superior performance and is a reliable tool for predicting early postoperative HCC recurrence. This model may guide surgical strategies and postoperative individualized medicine.

Purpose: To construct a 2.5-dimensional (2.5D) CT radiomics-based deep learning (DL) model to predict early postoperative recurrence of hepatocellular carcinoma (HCC). Patients and Methods: We retrospectively analyzed the data of patients who underwent HCC resection at 2 centers. The 232 patients from center 1 were randomly divided into the training (162 patients) and internal validation cohorts (70 patients); 91 patients from center 2 formed the external validation cohort. We developed a 2.5D DL model based on a central 2D image with the maximum tumor cross-section and adjacent slices. Multiple views (transverse, sagittal, and coronal) and phases (arterial, plain, and portal) were incorporated. Multi-instance learning techniques were applied to the extracted data; the resulting comprehensive feature set was modeled using Logistic Regression, RandomForest, ExtraTrees, XGBoost, and LightGBM, with 5-fold cross validation and hyperparameter optimization with Grid-search. Receiver operating characteristic curves, calibration curves, DeLong test, and decision curve analysis were used to evaluate model performance. Results: The 2.5D DL model performed well in the training (AUC: 0.920), internal validation (AUC: 0.825), and external validation cohorts (AUC: 0.795). The 3D DL model performed well in the training cohort and poorly in the internal and external validation cohorts (AUCs: 0.751, 0.666, and 0.567, respectively), indicating overfitting. The combined model (2.5D DL+clinical) performed well in all cohorts (AUCs: 0.921, 0.835, 0.804). The Hosmer-Lemeshow test, DeLong test, and decision curve analysis confirmed the superiority of the combined model over the other signatures. Conclusion: The combined model integrating 2.5D DL and clinical features accurately predicts early postoperative HCC recurrence.

Aim: To develop a habitat imaging method for preoperative prediction of early postoperative recurrence of hepatocellular carcinoma.Methods: A retrospective cohort study was conducted to collect data on 344 patients who underwent liver resection for HCC. The internal subregion of the tumor was objectively delineated and the clinical features were also analyzed to construct clinical models. Radiomics feature extraction was performed on tumor subregions of arterial and portal venous phase images. Machine learning classification models were constructed as a fusion model combining the three different models, and the models were assessed.Results: A comprehensive retrospective analysis was conducted on a cohort of 344 patients who underwent hepatic cancer resection at one of the two centers. it was found that the combined SVM model yielded superior results after comparing various metrics, such as the AUC, accuracy, sensitivity, specificity, and DCA.Conclusions: Habitat analysis of sequential CT images can delineate distinct subregions within a tumor, offering valuable insights for early prediction of postoperative HCC recurrence.

Objective: To explore the surgical effect of three-dimensional (3D) image reconstruction technology in pancreatoduodenectomy. Methods: The clinical records of 47 cases who underwent pancreatoduodenectomy between January 2018 and December 2019 at the department of hepatobiliary surgery of the General Hospital of Ningxia Medical University were retrospectively examined, including 23 males and 24 females, with an average age of 55.00 ± 10.06 years. All patients underwent enhanced computed tomography (CT), and the 3D images were reconstructed by uploading the CT imaging data. The pre-operation evaluation and treatment strategy were planned according to CT imaging and 3D data, respectively. The change of treatment strategy based on 3D evaluation, actual surgical procedure, tumor volume measured by 3D model, actual tumor volume, variants of hepatic artery, operation time, intraoperative blood loss, post-operation hospital stay and post-operation complications was recorded. Results: The treatment strategies were changed after 3D visualization in 10 (21.3%) out of 47 patients because of blood vessel and organ invasion by tumor. The surgical procedure was changed in three cases, and the surgical procedure was optimized and improved in seven cases. All surgical plans based on 3D visualization technology were matched with the actual surgical procedures. Tumor volume measured by 3D model was 19.69 ± 23.47 mL, post-operation actual tumor volume was 17.07 ± 20.29 mL, with no significant difference between them (t = 0.54, p = 0.59). Pearson’s correlation analysis showed statistical significance (r = 0.766, p = 0.00). The average operation time was 4.85 ± 1.75 h, median blood loss volume was 447.05 (50–5000) mL, and post-operation hospital stay was 26.13 ± 11.13 days. Six cases had pancreatic fistula, two cases had biliary leakage, and four cases had delayed gastric emptying. Ascites and pleural effusion was observed in three cases. Conclusions: 3D visualization technology can offer a precise and individualized surgical plan before operation, which might improve the safety of pancreatoduodenectomy, and has application value in preoperative planning.

Background: Recurrence remains common after resection of hepatocellular carcinoma (HCC) even when tumor diameter is limited. We evaluated routinely available clinicopathological predictors of recurrence within 2 years after resection of HCC ≤5 cm and developed an exploratory nomogram with bootstrap internal validation. Methods: This retrospective cohort included patients who underwent hepatic resection from 2014 to 2022. Two-year recurrence was analyzed using logistic regression, and recurrence-free survival (RFS) was analyzed using Cox regression. The revised full model contained seven explicitly specified predictors: Ki-67, preoperative TACE/RFA, microvascular invasion (MVI), satellite nodules, tumor diameter, prothrombin time (PT), and preoperative ascites. Missing model covariates were addressed in a 50-dataset multiple-imputation sensitivity analysis. Internal validation used 1,000 bootstrap resamples with the complete seven-predictor model refitted in each resample. Results: Among 215 patients, 93 (43.3%) developed recurrence within 24 months. The complete-case model included 190 patients and 83 events. In adjusted logistic analysis, Ki-67, preoperative TACE/RFA, MVI, satellite nodules, tumor diameter, and PT were associated with recurrence; preoperative ascites was retained but was not statistically significant. In adjusted Cox analysis, Ki-67, MVI, satellite nodules, and PT were associated with shorter 2-year RFS. The nomogram had an apparent AUC of 0.775 and an optimism-corrected AUC of 0.744; the optimism-corrected Brier score and calibration slope were 0.206 and 0.808, respectively. Multiple-imputation estimates were directionally consistent with the complete-case results. Conclusions: In patients with resected HCC ≤5 cm, proliferative activity, microscopic dissemination, tumor burden, and hepatic synthetic function contributed to early-recurrence risk. The internally validated nomogram showed moderate discrimination but remains exploratory and requires independent external validation before clinical use.