OBJECTIVETo evaluate the value of MR diffusion-weighted imaging (DWI) histogram analysis for predicting tumor progression in patients with hepatocellular carcinoma (HCC) treated with radiofrequency ablation (RFA).

METHODSIn a retrospective study, both 35 stable patients and 38 progressive patients with biopsy-proven HCC underwent breath-hold DWI before RFA treatment. The pre-treatment apparent diffusion coefficient (ADC) values were averaged from the lowest to 10th, 30th, 50th, and 100th percentile using DWI histogram analysis software respectively, and were called ADC10, ADC30, ADC50 and ADC100. The ratios of ADC10, ADC30, ADC50 and ADC100 to mean ADC of non-lesional area were calculated, called RADC10, RADC30, RADC50 and RADC100, respectively.

RESULTSBefore RFA treatment, the ADC30, ADC50, ADC100, RADC30, RADC50 and RADC100 values of the tumors in the progression group were significantly higher than those of the stable group (P < 0.05 for all), respectively. The progression-free survival (PFS) of all patients was (16.0 ± 10.6) months. Univariate Cox regression analysis indicated that RADC10, RADC30, RADC50 values of the tumor were significantly associated with PFS (P < 0.05 for both). In multivariate analysis, only the RADC50 value of tumor was a significant predictor for tumor progression (P < 0.05). When the cut-off value of RADC50 (0.72) was used, the PFS of below the cut-off value group [(30.2 ± 3.0) months] was significantly higher than that of those above the cut-off value group [(20.0 ± 3.1) months] (P = 0.01).

CONCLUSIONSPre-RFA DWI histogram analysis may serve as a biomarker for predicting tumor progression in patients with HCC treated with RFA.