Radiologic-pathologic correlation of hepatocellular carcinoma treated with chemoembolization.

To correlate posttreatment radiologic and pathologic findings in patients who underwent transarterial chemoembolization before transplantation or resection. Thirty-five patients with postchemoembolization follow-up imaging underwent liver transplantation/resection. Pre- and posttreatment contrast-enhanced magnetic resonance imaging were used to evaluate radiologic findings. Imaging characteristics using World Health Organization (WHO) and European Association for the Study of the Liver (EASL) criteria after treatment were evaluated. Treated lesions were examined by pathology (gold standard) for the assessment of necrosis. Radiologic findings on magnetic resonance imaging were correlated to pathologic findings to assess the predictability by imaging of actual necrosis. Kappa (κ) statistics were used to determine intermethod agreement between WHO and EASL criteria. Fourteen (40%) of 35 lesions had biopsy-proven hepatocellular carcinoma. Thirteen (37%) of 35 target lesions showed complete pathologic necrosis. Complete pathologic necrosis was seen in 35% of lesions with pretreatment size <3 cm. Complete pathologic necrosis was seen in 1 (100%) of 1, 6 (67%) of 9, 6 (33%) of 18, and 0 (0%) of 7 of the lesions that exhibited complete response (CR), partial response (PR), stable disease (SD), or progressive disease (PD) by WHO criteria, respectively. Complete pathologic necrosis was seen in 9 (82%) of 11, 4 (36%) of 11, 0 (0%) of 8, and 0 (0%) of 5 of the lesions that showed CR, PR, SD, or PD by EASL criteria, respectively. EASL CR and WHO response were shown to have ≥85% specificity for predicting complete pathologic necrosis. The κ coefficient for agreement between WHO and EASL was 0.29. EASL and WHO criteria had minimal intermethod agreement. EASL CR and WHO response were able to predict pathologic necrosis.