Is DWI/ADC a Useful Tool in the Characterization of Focal Hepatic Lesions Suspected of Malignancy?

PloS one

PubMedID: 25025151

Testa ML, Chojniak R, Sene LS, Damascena AS, Guimarães MD, Szklaruk J, Marchiori E. Is DWI/ADC a Useful Tool in the Characterization of Focal Hepatic Lesions Suspected of Malignancy?. PLoS ONE. 2014;9(7):e101944.
Apparent diffusion coefficient (ADC) values calculated through magnetic resonance imaging have been proposed as a useful tool to distinguish benign from malignant liver lesions. Most studies however included simple cysts in their analysis. Liver cysts are easy to diagnose, have very high ADC values and their inclusion facilitates differentiation in the ADC values between benign and malignant liver lesions groups. We prospectively evaluated the ability of ADC values to differentiate metastatic liver lesions from all benign or only solid benign liver lesions.

Sixty-seven adult cancer patients with 188 liver lesions were evaluated. Lesions were categorized as metastatic or benign throughout imaging and clinical evaluation. One hundred and five (105) metastatic lesions and 83 benign lesions including hemangiomas (37), cysts (42), adenomas (2) and focal nodular hyperplasias (2) were evaluated. ADC values were calculated for each lesion utilizing two b values (0 and 600 sec/mm2).

The average ADC value for cysts was 2.4×10-3 mm2/sec (CI: 2.1-2.6), for solid benign lesions was 1.4×10-3 mm2/sec (CI: 1.1-1.7) and for metastases was 1.0×10-3 mm2/sec (CI: 0.8-1.3). There was a difference between the ADC values of metastases and benign solid lesions (p<0.0001). With the ADC value of 1.5×10-3 mm2/sec as a cut off it is possible to distinguish metastatic from benign liver lesions, including cysts, with an accuracy of 78%. But to distinguish metastatic from benign solid liver lesions the best ADC cut off value was 1.2×10-3 mm2/sec and the accuracy drops to 71%.

ADC values proved to be helpful in the distinction between metastasis and benign solid hepatic lesions. But the exclusion of cysts in the analysis point out to a lower cut off value and lower accuracy than previously reported.