We read with interest the letter of Kudo et al. from one of the large Japanese series of patients undergoing transarterial chemoembolization (TACE), who tried to validate the Assessment for Retreatment (ART) score in their patient sample. When the authors applied the same selection criteria to their 513 patients treated with >2 TACE procedures, only 49 patients were eligible for ART score-based analysis for having received two TACE procedures within 90 days (all others had longer intervals between TACE sessions) and the ART score was not able to differentiate between two prognostic groups.
Looking at the two ART score groups in the Japanese patient sample more closely, it is quite obvious that the Japanese low ART score group was very similar to our own European cohort (median survival 22.4 versus 23.7/27.6 months), while the big difference was in the high ART score group between Kudo et al.'s and our cohorts (median survival 16.5 versus 6.6/8.1 months). In trying to understand this, one would have to take a closer look at the patient characteristics in both cohorts.
First of all, the ART score does yield two groups with different survival in the Japanese patients as well, but the difference is much less pronounced than in the European patients and did not reach statistical significance. This could well be due to the small patient number in this analysis, which was only performed in 44 of the 49 patients (see fig. 1 of Kudo et al.). Regarding the overall small sample size (n = 49), the unexplained exclusion of 10% of patients may have had a significant impact on survival analysis.
Besides that, and as pointed out by the authors, there are obvious differences in the patient and procedural characteristics between European and Japanese centers. In particular, Japanese patients usually present with smaller tumors due to nationwide screening programs and the TACE procedure is often carried out more meticulously than in European centers, allowing for less frequent TACE procedures and supposedly longer survival in patients undergoing TACE in Japan. This could potentially introduce a special selection bias, as indicated by the fact that less than 10% of Kudo et al.'s cohort was eligible for ART score calculation compared to almost 90% in the Austrian cohorts. But for a universally applicable score, these differences should be taken care of by diverting a different fraction of patients into the appropriate prognostic group. Surprisingly, this was not the case as much as we would expect: despite presumably smaller tumors and a more selective TACE procedure, still 27.3% of the Japanese patients retreated with TACE within 3 months presented with a high ART score before TACE 2, not so different from the 38% in the Austrian cohort.
Considering that overall survival in TACE patients seems to be longer in Japan compared to Europe, it is not so much the longer survival in the Japanese high ART score group but the surprisingly low survival in the Japanese low ART score group that is hard to explain. Since we do not have any data on the BCLC-staging, Child-Pugh scores, which type of Child-Pugh score was applied, the differences in etiology of liver disease, the reason for performing a second TACE within 3 months, causes of death, total number of TACE cycles applied, etc., in the Japanese patients, it is impossible to reach any firm conclusions from the data presented by Kudo et al. In particular, BCLC-stage A could be much more prevalent in Japan, where cadaveric liver transplantation is very rare and living donors will likely not be able to fill in the demand, leaving more BCLC-A patients for TACE compared to Europe.
With such a small patient number, it is conceivable that there was a significant imbalance between baseline variables in the Japanese cohort, e.g., BCLC-stage A in high ART score versus B in low ART score, which would have a major impact on survival. Likewise, a higher number of patients with a more marginal liver function or a significantly higher tumor burden in the low ART score group could explain the comparatively low survival, the atypical need for an early second TACE, and the less pronounced survival difference in the Japanese cohorts.
We would definitely be very interested in seeing a well-matched complete effort to validate the ART score in Asian patients as well as in conducting a well-matched comparative analysis between Asian and Western patients, which could help to even better establish the universal applicability of the ART score as a selection tool for repeat TACE.
Markus Peck-Radosavljevic, M.D.
Wolfgang Sieghart, M.D.
Florian Hucke, M.D.
for the Vienna HCC Study Group
Department of Internal Medicine III
Division of Gastroenterology and Hepatology
Medical University of Vienna