Wu Zhou1, Lijuan Zhang1, Kaixin Wang1, Shuting Chen2, Guangyi Wang2, Zaiyi Liu2, Changhong Liang2. 1. Key Laboratory for Health Informatics, Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China. 2. Department of Radiology, Guangdong General Hospital, Guangdong Academy of Medical Sciences, Guangdong Province, China.
Abstract
PURPOSE: To investigate the performance of texture analysis in characterizing malignancy of hepatocellular carcinomas (HCCs) based on contrast-enhanced magnetic resonance imaging (MRI). MATERIALS AND METHODS: Gd-DTPA contrast-enhanced MRI data of 46 consecutive subjects with resected HCC were retrieved. The mean intensity and gray-level run-length nonuniformity (GLN) were quantified as the discriminative features based on the arterial phase images and compared between groups with different histological grading using independent Student's t-test or Welch-Satterthwaite approximate t-test for data following a normal distribution and Mann-Whitney U-test for data violating the normal distribution. The performance of texture features in differentiating the biological aggressiveness of HCC was assessed using receiver operating characteristic (ROC) analysis. P < 0.05 was set as the significance level. RESULTS: Low-grade HCCs had increased mean intensity and decreased GLN in four directions, as compared with high-grade HCCs (P < 0.0005). A cutoff value of 739.37 for the average intensity resulted in an optimal sensitivity of 76% and specificity of 100% for histological grading discrimination. Cutoff values of 34.18, 66.59, 36.82, and 80.31 for the GLN in four directions (0°, 45°, 90°, 135°) resulted in an optimal sensitivity of 92%, 84%, 68%, 76% and specificity 66.70%, 71.40%, 85.70%, 76.20%, respectively. The areas under the ROC curve for the average intensity and GLN in four directions were 0.918, 0.846, 0.836, 0.827, and 0.838, respectively. CONCLUSION: Texture features indexed by mean and GLN based on the arterial phase images reflect biologic aggressiveness, and may have potential applications in predicting the histological grading of HCC preoperatively. Evidence level: 4 J. MAGN. RESON. IMAGING 2017;45:1476-1484.
PURPOSE: To investigate the performance of texture analysis in characterizing malignancy of hepatocellular carcinomas (HCCs) based on contrast-enhanced magnetic resonance imaging (MRI). MATERIALS AND METHODS:Gd-DTPA contrast-enhanced MRI data of 46 consecutive subjects with resected HCC were retrieved. The mean intensity and gray-level run-length nonuniformity (GLN) were quantified as the discriminative features based on the arterial phase images and compared between groups with different histological grading using independent Student's t-test or Welch-Satterthwaite approximate t-test for data following a normal distribution and Mann-Whitney U-test for data violating the normal distribution. The performance of texture features in differentiating the biological aggressiveness of HCC was assessed using receiver operating characteristic (ROC) analysis. P < 0.05 was set as the significance level. RESULTS: Low-grade HCCs had increased mean intensity and decreased GLN in four directions, as compared with high-grade HCCs (P < 0.0005). A cutoff value of 739.37 for the average intensity resulted in an optimal sensitivity of 76% and specificity of 100% for histological grading discrimination. Cutoff values of 34.18, 66.59, 36.82, and 80.31 for the GLN in four directions (0°, 45°, 90°, 135°) resulted in an optimal sensitivity of 92%, 84%, 68%, 76% and specificity 66.70%, 71.40%, 85.70%, 76.20%, respectively. The areas under the ROC curve for the average intensity and GLN in four directions were 0.918, 0.846, 0.836, 0.827, and 0.838, respectively. CONCLUSION: Texture features indexed by mean and GLN based on the arterial phase images reflect biologic aggressiveness, and may have potential applications in predicting the histological grading of HCC preoperatively. Evidence level: 4 J. MAGN. RESON. IMAGING 2017;45:1476-1484.
Authors: Katja Pinker; Fuki Shitano; Evis Sala; Richard K Do; Robert J Young; Andreas G Wibmer; Hedvig Hricak; Elizabeth J Sutton; Elizabeth A Morris Journal: J Magn Reson Imaging Date: 2017-11-02 Impact factor: 4.813
Authors: Samantha M Ruff; Luke D Rothermel; Laurence P Diggs; Michael M Wach; Reed I Ayabe; Sean P Martin; David Boulware; Daniel Anaya; Jeremy L Davis; John E Mullinax; Jonathan M Hernandez Journal: HPB (Oxford) Date: 2019-11-13 Impact factor: 3.647