Diffusion-weighted imaging of the liver: usefulness of ADC values in the differential diagnosis of focal lesions and effect of ROI methods on ADC measurements

Abstract

Object: By measuring the apparent diffusion coefficient (ADC) of liver parenchyma and focal hepatic lesions (FHL) we proposed to investigate the utility of ADC in the differential diagnosis of hepatic disease and to determine the influence of region of interest (ROI) characteristics in those measurements.

Materials and methods: Ninety-three patients with at least one supracentimetric FHL, or parenchymal abnormality, were retrospectively evaluated. Diagnosis was based on histopathologic data or, alternatively, on a combination of consensus between imaging methods and 24 months of follow-up. Ninety lesions were evaluated with respiratory-triggered diffusion-weighted imaging (b values: 50 and 700 s/mm(2)): 14 hepatocellular carcinomas, 18 metastases, 10 focal nodular hyperplasias, four adenomas, 30 hemangiomas and 14 cysts. ADC of hepatic parenchyma was measured by placing ROIs in four different segments, and in FHLs by using three circular 1 cm(2) ROIs and one ROI encompassing the full lesion. Data was statistically analyzed (p < 0.05 considered significant), and a receiver operating characteristic curve was assessed to evaluate the accuracy for the diagnosis of malignancy.

Results: Our measurements showed that parenchyma ADC was significantly higher in segment II and that ADCs of malignant lesions were significantly lower than those of benign lesions (p < 0.001). There was significant overlap between benign solid lesions and malignant lesions and the area under the curve for malignancy was 0.939 (sensitivity 89.7 %, specificity 90.6 %), using a cutoff of 1.43 × 10(-3) mm(2)/s. No significant difference was found between ROIs of different characteristics.

Conclusion: ADC measurements can help to characterize FHLs and differentiate normal from pathological parenchyma. Any ROI above 1 cm(2) can provide accurate ADC measurements in homogenous lesions.