Volume 36, No. 02, Month MARCH, Year 2021, Pages 149 - 156

The performance of the diffusion weighted magnetic resonance imaging using apparent diffusion coefficient measurement in detecting malignant liver lesions

Julaluck Promsorn, Praewpan Na-udom, Panita Limpawattana, Nittaya Chamadol, Kulyada Somsap, Muckesh Harisinghani

Abstract Download PDF

Background and Objective: ADC (Apparent diffusion coefficient) values have been shown to be helpful for liver lesion characterization. There are; however, discrepancies in the ADC values and controversies regarding the optimal cutoff ADC values to differentiates malignant from benign liver lesions. The purpose of this study was to measure ADC values of malignant liver lesions and to identify the optimal cutoff ADC value to differentiate malignant from benign liver lesions. Material and Methods: A retrospective study of 180 MRI (Magnetic resonance imaging) of liver during June 1, 2012 to December 31, 2014. ADC value was measured and compared between benign and malignant liver lesions. The optimal ADC value to differentiated between malignant and benign liver lesions was calculated. Results: Seventy-nine malignant liver lesions included 52 CCAs, 20 HCCs, 7 liver metastases had median ADC value 1.06x10-3 mm2/sec. 101 benign liver lesions included 44 hemangiomas, 11 FNHs, 7 hepatic adenomas and 39 cysts had median ADC value 1.93x10-3 mm2/sec. The differences between the median ADC values of malignant liver lesions (1.06x10-3 mm2/sec) and benign liver lesions (1.93x 10-3 mm2/sec) was statistically significant (p<0.05). The ADC value of <1.49x10-3 mm2/sec was the optimal cut-off values to indicate malignant liver mass with the sensitivity of 84.8%, specificity of 81.2%. Conclusion: ADC value is useful for differentiating malignant from benign liver lesions with1.49x10-3 mm2/s as optimal cutoff ADC value


Magnetic resonance imaging (MRI); Apparent diffusion coefficient (ADC) value; Cholangiocarcinoma (CCA); hepatocellular carcinoma


Published by : Faculty of Medicine, Khon Kaen University
Contributions welcome at :