In order to solve the pathological grading of hepatocellular carcinomas (HCC) which depends on biopsy or surgical pathology invasively, a quantitative analysis method based on radiomics signature was proposed for pathological grading of HCC in non-contrast magnetic resonance imaging (MRI) images. The MRI images were integrated to predict clinical outcomes using 328 radiomics features, quantifying tumour image intensity, shape and text, which are extracted from lesion by manual segmentation. Least absolute shrinkage and selection operator (LASSO) were used to select the most-predictive radiomics features for the pathological grading. A radiomics signature, a clinical model, and a combined model were built. The association between the radiomics signature and HCC grading was explored. This quantitative analysis method was validated in 170 consecutive patients (training dataset: n = 125; validation dataset, n = 45), and cross-validation with receiver operating characteristic (ROC) analysis was performed and the area under the ROC curve (AUC) was employed as the prediction metric. Through the proposed method, AUC was 0.909 in training dataset and 0.800 in validation dataset, respectively. Overall, the prediction performances by radiomics features showed statistically significant correlations with pathological grading. The results showed that radiomics signature was developed to be a significant predictor for HCC pathological grading, which may serve as a noninvasive complementary tool for clinical doctors in determining the prognosis and therapeutic strategy for HCC.
Objective To explore the clinical value of artificial intelligence (AI) quantitative parameters in distinguishing pathological grades of stageⅠ invasive adenocarcinoma (IAC). Methods Clinical data of patients with clinical stageⅠ IAC admitted to Yantaishan Hospital Affiliated to Binzhou Medical University from October 2018 to May 2023 were retrospectively analyzed. Based on the 2021 WHO pathological grading criteria for lung adenocarcinoma, IAC was divided into gradeⅠ, grade Ⅱ, and grade Ⅲ. The differences in parameters among the groups were compared, and logistic regression analysis was used to evaluate the predictive efficacy of AI quantitative parameters for grade Ⅲ IAC patients. Parameters were screened using least absolute shrinkage and selection operator (LASSO) regression analysis. Three machine learning models were constructed based on these parameters to predict grade Ⅲ IAC and were internally validated to assess their efficacy. Nomograms were used for visualization. ResultsA total of 261 IAC patients were included, including 101 males and 160 females, with an average age of 27-88 (61.96±9.17) years. Six patients had dual primary lesions, and different lesions from the same patient were analyzed as independent samples. There were 48 patients of gradeⅠ IAC, 89 patients of grade Ⅱ IAC, and 130 patients of grade Ⅲ IAC. (1) Comparison among the three groups: the differences in parameters such as consolidation/tumor ratio (CTR), long diameter, short diameter, malignancy probability, CT average value, CT maximum value, CT minimum value, CT median value, CT standard deviation, kurtosis, skewness, and entropy were statistically significant (P<0.05). (2) Comparison between two groups: gradeⅠ and grade Ⅱ were combined and compared with grade Ⅲ, and univariate analysis showed that the differences in all variables except age were statistically significant (P<0.05). Multivariate analysis suggested that CTR and CT standard deviation were independent risk factors for identifying grade Ⅲ IAC, and the two were negatively correlated. (3) Pathological comparisons: no lymph node metastasis was found in gradeⅠpatients, two gradeⅡ patients were of lymph node metastasis with micro-papillary components, and 19 grade Ⅲ patients were of lymph node metastasis. Grade Ⅲ IAC exhibited advanced TNM staging, more pathological high-risk factors, higher lymph node metastasis rate, and higher proportion of advanced structure. (4) Correlation analysis: CTR was positively correlated with the proportion of advanced structures in all patients. This correlation was also observed in grade Ⅲ but not in gradeⅠand grade ⅡIAC. (5) CTR and CT median value were selected by using LASSO regression, and logistic regression, random forest, and XGBoost models were constructed and validated. Among them, the XGBoost model demonstrated the best predictive performance. Conclusion Cautious consideration should be given to grade Ⅲ IAC when CTR is higher than 39.48% and CT standard deviation is less than 122.75 HU. The XGBoost model based on combined CTR and CT median value has good predictive efficacy for grade Ⅲ IAC, aiding clinicians in making personalized clinical decisions.