<b><i>Background and Aims:</i></b> For patients with intermediate-stage hepatocellular carcinoma (HCC), the definition of high tumor burden remains controversial. This study aimed to compare the prognostic value of different criteria of tumor burden in patients with intermediate-stage HCC undergoing transarterial chemoembolization (TACE). <b><i>Methods:</i></b> From 2007 to 2019, 632 treatment-naive patients with intermediate-stage HCC undergoing TACE were retrospectively enrolled. We compared different criteria of tumor burden in discriminating radiologic response and survival, including up-to-7, up-to-11, 5–7, 7 lesions criteria, and newly proposed 7–11 criteria. <b><i>Results:</i></b> The proportions of patients classified as high tumor burden were varied by different criteria. Among the 5 criteria, 7–11 criteria have the best performance to discriminate complete response (CR) and overall survival (OS) after TACE. In patients with low, intermediate, and high tumor burden classified by 7–11 criteria, the CR rate was 21, 12, and 2.5%, respectively (<i>p</i> &#x3c; 0.001), and the median OS was 33.1, 22.3, and 11.9 months, respectively (<i>p</i> &#x3c; 0.001). By multivariate analysis, 7–11 criteria were significantly associated with CR (intermediate vs. high burden, odds ratio = 4.617, <i>p</i> = 0.002; low vs. high burden, odds ratio = 8.675, <i>p</i> &#x3c; 0.001) and OS (intermediate vs. high burden, hazard ratio = 0.650, <i>p</i> &#x3c; 0.001; low vs. high burden, hazard ratio = 0.520, <i>p</i> &#x3c; 0.001). Seven to 11 criteria also had the lowest corrected Akaike information criteria, highest homogeneity value, and highest area under the receiver operating characteristic curve in predicting 1-, 2-, and 3-year mortality among all criteria. <b><i>Conclusion:</i></b> Conventional definitions of tumor burden were not optimal for patients with intermediate HCC. The new 7–11 criteria had the best discriminative power in predicting radiologic response and survival in patients with intermediate-stage HCC undergoing TACE.

Abstract. Use of discrimination nets for many-to-one pattern matching has been shown to dramatically improve the performance of the KnuthBendix completion procedure used in rewriting. Many important applications of rewriting require associative-commutative (AC) function symbols and it is therefore quite natural to expect performance gains by using similar techniques for AC-completion. In this paper we propose such a technique, called AC-discrimination net, that is a natural generalization of the standard discrimination net in the sense that if no AC-symbols are present in the pattern, it specializes to the standard discrimination net. Moreover we show how AC-discrlmination nets can be augmented so as to further improve the performance of AC-matching on problems that are typically seen in practice.

No abstract

No abstract