Reply to: “The “six-and-twelve score” for TACE treatment: Does it really help us?”

Development of a prognostic score for recommended TACE candidates with hepatocellular carcinoma: A multicentre observational studyJournal of HepatologyVol. 70Issue 5PreviewAccording to the guidelines of the American Association for the Study of Liver Disease (AASLD) and the European Association for the Study of Liver (EASL), transarterial chemoembolization (TACE) is currently the only recommended treatment option for patients with intermediate stage hepatocellular carcinoma (HCC),1–3 with well-preserved liver function and performance status.4–6 These patients, as well as those at early stage but considered unresectable due to tumour size, location, patient age, and suggestions of stage migration are considered “recommended” or “ideal” TACE candidates, i.e. Full-Text PDF The “six-and-twelve score” for TACE treatment: Does it really help us?Journal of HepatologyVol. 71Issue 5PreviewWe have read with interest the manuscript by Wang et al.,1 wherein they propose a new scoring system named “six-and-twelve score” for patients with unresectable hepatocellular carcinoma (HCC) undergoing transarterial chemoembolization (TACE). This prognostic index is simply calculated by the sum of tumor size and number. It determines 3 groups with different prognosis (G1, the sum ≤6; G2, the sum >6 but ≤12; G3, the sum >12). A survival benefit of more than 20 months can be expected for patients with HCC and an index ≤12, and more than 40 months for those with an index ≤6. Full-Text PDF We sincerely appreciate Bourlière et al. for the external validation of our prognostic model in a French centre,[1]Bourliere M. Penaranda G. Adhoute X. Bronowicki J. “The six-and-twelve score” for TACE treatment: does it really help us?.J Hepatol. 2019; 71: 1051-1052Abstract Full Text Full Text PDF Scopus (7) Google Scholar with essential results derived from a different setting. Before interpretation of these results, we would like to clarify that our model is a continuous model in the first place, with linear predictor = largest tumour diameter + tumour numbers.[2]Wang Q. Xia D. Bai W. Wang E. Sun J. Huang M. et al.Development of a prognostic score for recommended TACE candidates with hepatocellular carcinoma: a multicentre observational study.J Hepatol. 2019; 70: 893-903Abstract Full Text Full Text PDF PubMed Scopus (99) Google Scholar Only with this presentation can we achieve individualized outcome prediction, whereas “6” and “12” are cut-off values for quick stratified prediction. The use of cut-offs provides convenience yet sacrifices accuracy, as they underestimate heterogeneity within the strata (e.g., tumour size + number being 2 and 6) but exaggerate the difference around the cut-off points (e.g., 6 and 8).[3]Steyerberg E.W. Clinical prediction models: a practical approach to development, validation, and updating. Springer Science & Business Media, 2008Google Scholar In the findings of Bourlière et al., it is somehow intriguing to see that G2 stratum, with twice the upper limit of tumour burden compared to G1 stratum, had similar median overall survival as the latter stratum (29 vs. 23 months).[1]Bourliere M. Penaranda G. Adhoute X. Bronowicki J. “The six-and-twelve score” for TACE treatment: does it really help us?.J Hepatol. 2019; 71: 1051-1052Abstract Full Text Full Text PDF Scopus (7) Google Scholar A likely explanation is that a majority of patients in these 2 strata concentrated around the cut-off point, especially when 66% patients are multi-nodular, the sum of tumour size and number in G1 strata may easily exceed 3 and approach 6. This loss of information may account for the reduced performance and is the reason why we provided a nomogram for individualized prediction prior to stratified prediction with cut-offs. Therefore, for external validation, the continuous presentation of the model may be more accurate. A more in-depth look into this question leads us to the matter of population. As we stated in our study, population is a determinant of the performance of the model. A well-performing model or staging system may become less effective in a totally different population. This problem becomes more intense in hepatocellular carcinoma (HCC), a disease with notorious heterogeneity among different populations, which is why the generalisation of most staging systems other than Barcelona Clinic Liver Cancer (BCLC) remains controversial,4Liu P.H. Hsu C.Y. Hsia C.Y. Lee Y.H. Su C.W. Huang Y.H. et al.Prognosis of hepatocellular carcinoma: assessment of eleven staging systems.J Hepatol. 2016; 64: 601-608Abstract Full Text Full Text PDF PubMed Scopus (175) Google Scholar, 5Parikh N.D. Scaglione S. Li Y. Powell C. Yerokun O.A. Devlin P. et al.A comparison of staging systems for hepatocellular carcinoma in a multicenter US cohort.Clin Gastroenterol Hepatol. 2018; 16: 781-782Abstract Full Text Full Text PDF PubMed Scopus (11) Google Scholar and for BCLC staging, further refinements are still required.[6]European Association for the Study of the Liver EASL clinical practice guidelines: management of hepatocellular carcinoma.J Hepatol. 2018; 69: 182-236Abstract Full Text Full Text PDF PubMed Scopus (4093) Google Scholar Similarly, the tumour characteristics of this external cohort receiving transarterial chemoembolisation (TACE) were mainly multi-nodular (66%) with smaller diameters (32 [25–44] mm), unlike our cohort where single-nodular tumours predominated (57%) and tumour diameters were larger on average (61 [38–98] mm). This difference may have caused the reduced performance of our model in the external validation set. However, even with the information loss caused by the cut-off values and the small number of patients in G3 strata (only 2 patients), our model, utilising tumour burden parameters only and without liver function or performance status, exhibited similar performance as the BCLC staging system. This might indicate that tumour burden is still an essential predictor in this cohort representative of European patients. However, it must be noted that our “six-and-twelve” score is not developed to replace the BCLC staging system, but rather to refine it by further stratifying optimal TACE candidates with tumour burden heterogeneity that is unclassified by the BCLC system (Fig. 1). This stratification can provide referential criteria to define the target population for better clinical decision-making in future trial design. For example, patients with tumour burden >6 but ≤12 may be considered as a target population in trials involving TACE and other additional therapies such as loco-regional therapy, whereas those with tumour burden >12 may be candidates for trials comparing TACE alone and combined with systemic therapy. Moreover, despite the result that liver function parameters are not predictors in our cohort, we also agree with Bourlière et al. that when populations differ, heterogeneity in liver function, aetiology, alpha-fetoprotein, and performance status, etc. will increase. These parameters might thus become more influential to survival after TACE and their addition to tumour burden criteria may improve predictive performance. Just like the situation with the HAP (hepatoma arterial embolization prognostic) score and its modified versions,7Kadalayil L. Benini R. Pallan L. O'Beirne J. Marelli L. Yu D. et al.A simple prognostic scoring system for patients receiving transarterial embolisation for hepatocellular cancer.Annal Oncol. 2013; 24: 2565-2570Abstract Full Text Full Text PDF PubMed Scopus (219) Google Scholar, 8Pinato D.J. Arizumi T. Allara E. Jang J.W. Smirne C. Kim Y.W. et al.Validation of the hepatoma arterial embolization prognostic score in European and Asian populations and proposed modification.Clin Gastroenterol Hepatol. 2015; 13 (1204-1208.e1202)Abstract Full Text Full Text PDF PubMed Scopus (39) Google Scholar, 9Park Y. Kim S.U. Kim B.K. Park J.Y. Kim D.Y. Ahn S.H. et al.Addition of tumor multiplicity improves the prognostic performance of the hepatoma arterial-embolization prognostic score.Liver Int. 2016; 36: 100-107Crossref PubMed Scopus (53) Google Scholar, 10Cappelli A. Cucchetti A. Cabibbo G. Mosconi C. Maida M. Attardo S. et al.Refining prognosis after trans-arterial chemo-embolization for hepatocellular carcinoma.Liver Int. 2016; 36: 729-736Crossref PubMed Scopus (43) Google Scholar our model may not exhibit satisfying performance in its current form when greater multi-dimensional heterogeneity is introduced by a totally different HCC cohort. It can be expected that this deviation may be partly corrected by further external validation with a large sample size in different settings, and corresponding modification of the model to improve its performance, which is our next step. It will be hard to find a panacea to all the difficulties associated with developing prognostic models for a disease as heterogeneous as HCC, but we hope to get closer to it with improvements through trial and error. In conclusion, we thank Bourlière et al. for their external validation. The generalisability of our model might be improved with further modifications derived from validations in different settings. The authors received no financial support to produce this manuscript. The authors declare no conflicts of interest that pertain to this work. Please refer to the accompanying ICMJE disclosure forms for further details. Drafting and revision of the manuscript: QW, DX, WB, and GH. The following are the Supplementary data to this article: Download .pdf (.26 MB) Help with pdf files Supplementary Data