• Roman Kloeckner, Michael B Pitton, Christoph Dueber, Irene Schmidtmann, Peter R Galle, Sandra Koch, Marcus A Wörns, and Arndt Weinmann.
• Department of Diagnostic and Interventional Radiology, University Medical Center of the Johannes Gutenberg University Mainz, Langenbeckstrasse 1, 55131 Mainz, Germany. Electronic address: roman.kloeckner@unimedizin-mainz.de.
• J Vasc Interv Radiol. 2017 Jan 1; 28 (1): 94-102.

PurposeTo perform an external validation of the Assessment for Retreatment with Transarterial Chemoembolization (ART) and α-fetoprotein (AFP), Barcelona Clinic Liver Cancer (BCLC), Child-Pugh, and response (ABCR) scores and to compare them in terms of prognostic power.Materials And MethodsFrom 2000 to 2015, 871 patients with hepatocellular carcinoma underwent transarterial chemoembolization at a tertiary referral hospital, and 176 met all inclusion and exclusion criteria for both scores and were analyzed. Nineteen percent (n = 34) had BCLC stage A disease and 81% had stage B disease. Thirty-nine patients (22%) presented with elevated AFP levels. Overall survival was calculated. Scores were validated and compared with a Harrell C-index, integrated Brier score (IBS), and prediction error curves.ResultsBefore the second chemoembolization procedure, 22 patients (12%) showed an increase of 1 point in Child-Pugh score and 51 patients (22%) had an increase of ≥ 2 points. Thirty-one patients (23%) showed a > 25% increase in aspartate aminotransferase level, and 114 (65%) showed a response to treatment. Consequently, 127 patients (72%) had a low ART score and 49 (28%) had a high ART score. One hundred fifty-eight patients (90%) had a low ABCR score, whereas 18 (10%) had a high ABCR score. Low and high ART score groups had median survival durations of 20.8 and 15.3 mo, respectively. Harrell C-indexes were 0.572 and 0.608, and IBSs were 0.135 and 0.128, for ART and ABCR, respectively. For both scores, an increase in Child-Pugh score ≥ 2 points and a radiologic response were significantly associated with survival.ConclusionsBoth scores were of limited predictive value, and neither was sufficient to support clear-cut clinical decisions. Further effort is necessary to determine criteria for making valid clinical predictions.Copyright © 2017 SIR. Published by Elsevier Inc. All rights reserved.

41 keywords...

hide…