Utility of apparent diffusion coefficient (ADC) values in differentiating benign and malignant skull lesions with histopathological (HPE) correlation

Abstract

AimAssess role of ADC in differentiating benign and malignant skull lesions and to evaluate the added value of ADC over conventional MRI in facilitating the differentiation. Materials and methods53 patients (24 males, 29 females; age 3–75 years) were subjected to both conventional and Diffusion weighted (DWI) MR imaging. DWI was performed using a single-shot SE EPI sequence with b-values of 0& 1000 s/mm2 on 1.5 T MR scanner. Margins of the lesion, number, soft-tissue component, local extension, periosteal reaction and enhancement pattern were the parameters used for differentiating benign & malignant lesions by conventional MRI. ADC values (mean of 3 ROIs over solid component) were calculated. Conventional MRI characteristics and ADC value of lesions were evaluated & compared using statistical analysis. These findings were compared and correlated with histopathology of the skull lesions. Results24 malignant and 29 benign lesions were identified on HPE (Histopathological examination) in 53 patients. ADC cut-off value of 0.96 × 10−3 mm2/s obtained from ROC curve was found to have 75.47% accuracy, 87.5% sensitivity, 65.52% specificity, 67.74% PPV and 86.36% NPV for differentiating malignant from benign lesions. Statistically significant differences (p < 0.05) were seen in the mean ADC values of malignant (0.64 ± 0.42 × 10−3 mm2/s) and benign lesions (1.14 ± 0.56 × 10−3 mm2/s). The sensitivity, specificity, PPV and NPV in differentiating benign & malignant skull lesions were found to be 58.33%, 62.07%, 56% and 64.29% respectively, with diagnostic accuracy of 60.38% on using conventional MRI alone and 75%, 72.41%, 69.23% and 77.78% respectively, with diagnostic accuracy of 73.58% on using conventional MRI with ADC. ConclusionADC is a promising non-invasive parameter that facilitates differentiation between benign and malignant skull lesions. It is a robust biomarker to narrow differentials when conventional imaging features are indeterminate.