Reproducibility and non-redundancy of radiomic features extracted from arterial phase CT scans in hepatocellular carcinoma patients: impact of tumor segmentation variability.

BACKGROUND: The reproducibility and non-redundancy of radiomic features are challenges in accelerating the clinical translation of radiomics. In this study, we focused on the robustness and non-redundancy of radiomic features extracted from computed tomography (CT) scans in hepatocellular carcinoma (HCC) patients with respect to different tumor segmentation methods. METHODS: Arterial enhanced CT images were retrospectively randomly obtained from 106 patients. As a training data set, 26 HCC patients were used to calculate the features' reproducibility and redundancy. Another data set (55 HCC patients and 25 healthy volunteers) was used for classification. The GrowCut and GraphCut semiautomatic segmentation methods were implemented in 3D Slicer software by two independent observers, and manual delineation was performed by five abdominal radiation oncologists to acquire the gross tumor volume (GTV). Seventy-one radiomic features were extracted from GTVs using Imaging Biomarker Explorer (IBEX) software, including 17 tumor intensity statistical features, 16 shape features and 38 textural features. For each radiomic feature, intraclass correlation coefficient (ICC) and hierarchical clustering were used to quantify its reproducibility and redundancy. Features with ICC values greater than 0.75 were considered reproducible. To generate the number of non-redundancy feature subgroups, the R2 statistic method was used. Then, a classification model was built using a support vector machine (SVM) algorithm with 10-fold cross validation, and area under ROC curve (AUC) was used to evaluate the utility of non-redundant feature extraction by hierarchical clustering. RESULTS: The percentages of excellent reproducible features in the manual delineation group, GraphCut and GrowCut segmentation group were 69% [49], 73% [52] and 79% [56], respectively. Sixty-five percent [46] of the features showed strong robustness for all segmentation methods. The optimal number of cluster subgroup were 9, 13 and 11 for manual delineation, GraphCut and GrowCut segmentation, respectively. The optimal cluster subgroup number was 6 for all groups when the collectively high reproducibility features were selected for clustering. The receiver operating characteristic (ROC) analysis of radiomics classification model with and without feature reduction for healthy liver and HCC had an AUC value of 0.857 and 0.721 respectively. CONCLUSIONS: Our study demonstrates that variations exist in the reproducibility of quantitative imaging features extracted from tumor regions segmented using different methods. The reproducibility and non-redundancy of the radiomic features rely greatly on the tumor segmentation in HCC CT images. We recommend that the most reliable and uniform radiomic features should be selected in the clinical use of radiomics. Classification experiments with feature reduction showed that radiomic features were effective in identifying healthy liver and HCC.

Heterogeneity analysis of triphasic CT scan perfusion parameters in differential diagnosis of hepatocellular carcinoma and hemangioma.

This study is to investigate quantitative measures and heterogeneity of perfusion parameters in the differential diagnosis of hepatocellular carcinoma (HCC) and hemangioma.In total, 32 HCC and 44 hemangioma (types 1, 2, and 3) cases were included in this retrospective study. Hepatic artery coefficient (HAC), portal vein coefficient (PVC), and arterial enhancement fraction (AEF) were calculated. Tumor heterogeneity was analyzed. Perfusion parameters and corresponding percentiles were compared between the HCC and hemangioma (especially atypical hemangioma) cases, as well as between the substantial lesion part and surrounding normal tissue.The mean value, and the 10th, 50th, 75th, and 90th percentiles of PVC were significantly lower in the HCC cases than the types 1 and 2 hemangioma cases (P < .01). Moreover, the 90th percentile PVC in the HCC cases was also significantly lower than the type 3 hemangioma case (P < .01), while the mean value, and all the percentiles of AEF in the HCC cases were higher than the types 2 and 3 hemangioma cases (P < .01). The 10th percentile HAC in the HCC cases was higher than the type 2 hemangioma cases (P < .05). The mean value, and the 10th and 50th percentile HAC in the HCC cases were higher than the type 3 hemangioma case (P < .05). However, there was no statistically significant difference in HAC between the HCC and type 1 hemangioma cases (P > .05).Quantitative measurement of perfusion parameters and heterogeneity analysis show significance differences in the early detection and differential diagnosis of HCC and hemangioma cases, which might contribute to increasing the diagnostic accuracy.

Correlation between CT features of adrenocortical and adrenal medullary tumors and expression of miR-96 in serum.

Correlation between CT features of adrenocortical and adrenal medullary tumors and the expression of miR-96 in serum were investigated. A total of 230 patients with adrenocortical tumors and 194 patients with adrenal medullary tumors were selected in Dongying People's Hospital from August 2013 to August 2017. The two groups of patients underwent CT examination, and the signs and symptoms were recorded. The expression of miR-96 in the serum of the two groups was detected by RT-PCR, and the correlation between the expression of serum miR-96 and CT features was analyzed. In patients with adrenocortical tumor, serum miR-96 expression levels were significantly higher in patients with tumor diameter ≥5 cm than those with tumor diameter <5 cm (p<0.001). In patients with adrenal medullary tumor, serum miR-96 expression levels were significantly higher in patients with tumor diameter ≥3 cm than those with tumor diameter <3 cm (p<0.001). In patients with adrenocortical or adrenal medullary tumor, serum miR-96 expression levels were significantly higher in patients with peripheral infiltration than those without peripheral infiltration (p<0.001), and serum miR-96 expression levels were also significantly higher in patients with distant metastasis than those without distant metastasis (p<0.001). Serum levels of miR-96 in patients with benign adrenocortical and adrenal medullary tumors were significantly lower than those with malignant tumors in the same group (p<0.001). miR-96 may have oncogenic functions in patients with adrenocortical or adrenal medullary tumors. Increased expression level of miR-96 may promote proliferation, invasion and metastasis of tumors, and serum levels of miR-96 provide references for the diagnosis of adrenocortical and adrenal medullary tumors.