The China Paradox in LEAP-012 Trial Authors: Masatoshi Kudo Published: 2026 Introduction The treatment landscape of intermediate-stage hepatocellular carcinoma (HCC) has undergone substantial transformation in recent years. For decades, transarterial chemoembolization (TACE) has been the standard treatment for patients with Barcelona Clinic Liver Cancer (BCLC) stage B disease. However, the emergence of effective systemic therapies, particularly immune checkpoint inhibitors (ICIs) combined with anti-angiogenic agents, has reshaped therapeutic strategies and renewed interest in combining systemic therapy with locoregional approaches [1]. Among the clinical trials evaluating this strategy [2–5], the LEAP-012 study [3] represents an important milestone. The trial demonstrated that the addition of lenvatinib [6] and pembrolizumab to TACE significantly improved progression-free survival (PFS) compared with TACE alone, highlighting the potential of combining systemic immunotherapy with locoregional therapy in intermediate-stage HCC. Despite these promising findings, subgroup analyses revealed a striking regional signal. The magnitude of treatment benefit appeared substantially greater in the Chinese population than in the global intention-to-treat (ITT) cohort. This observation raises an important question regarding the interpretation of currently ongoing HCC trials: does the stronger treatment effect observed in China reflect enhanced biological efficacy of the experimental regimen, or could it partly arise from differences in the performance of the TACE control arm? The China Paradox in LEAP-012 In the global ITT population, the addition of lenvatinib and pembrolizumab to TACE improved median PFS from 10.0 months with TACE alone to 14.6 months, corresponding to a hazard ratio (HR) of 0.66 (95% CI: 0.51–0.84) (Fig. 1). Time to progression (TTP) also improved from 10.3 months to 16.6 months (HR 0.59), indicating a consistent benefit of the combination regimen (Fig. 2) [3]. Computer-simulated Kaplan-Meier curves of progression-free survival (PFS) comparing lenvatinib plus pembrolizumab combined with TACE and dual placebo plus TACE in the global cohort and the Chinese subgroup. PFS was longer in the Chinese subgroup than in the global cohort in the combination arm, whereas the opposite pattern was observed in the dual placebo plus TACE arm, with shorter PFS in the Chinese subgroup. Computer-simulated Kaplan-Meier curves of time to progression (TTP) comparing lenvatinib plus pembrolizumab combined with TACE and dual placebo plus TACE in the global cohort and the Chinese subgroup. TTP was shorter in the Chinese subgroup than in the global cohort in the dual placebo plus TACE arm. However, the Chinese subgroup demonstrated an even more pronounced treatment effect [7]. Median PFS in the combination arm reached 16.6 months, whereas the TACE-alone arm achieved only 6.5 months, yielding a hazard ratio of 0.52 (Fig. 1). A similar pattern was observed for overall survival (OS), with an HR of 0.80 in the global population but 0.58 in the Chinese subgroup (Fig. 3). Computer-simulated Kaplan-Meier curves of overall survival (OS) comparing lenvatinib plus pembrolizumab combined with TACE and dual placebo plus TACE in the global cohort and the Chinese subgroup. The hazard ratio for OS notably more favored the Chinese subgroup compared with the global cohort. Additionally, the 1-year OS rate in the lenvatinib plus pembrolizumab plus TACE arm was higher in the Chinese cohort than in the global cohort. At first glance, these findings might suggest that the lenvatinib plus pembrolizumab plus TACE combination is particularly effective in the Chinese population. However, closer evaluation reveals an apparent paradox. Baseline characteristics in the Chinese subgroup were broadly comparable to those in the global cohort and, in several respects, even more favorable. Patients in China were younger and had a higher proportion of preserved liver function, with a larger percentage of patients classified as ALBI grade 1 (Table 1). Under such circumstances, outcomes in the TACE-alone arm might reasonably be expected to be at least comparable to those observed in the global population. Table 1.Baseline characteristics in LEAP-012: global (ITT) and ChinaLEN+PEM+TACETACE aloneglobal (n = 237)China (n = 105)global (n = 243)China (n = 103)Age, median, years65596659ECOG PS0, %91.195.287.789.3Etiology, % HBV64.697.159.392.2 HCV17.76.716.05.8 Non-viral24.51.031.34.9AFP >400 ng/mL15.611.916.528.2Child-Pugh score A586.182.989.389.3BCLC stage A33.841.028.023.3 B57.054.360.166.0 C8.94.811.910.7ALBI grade 172.278.171.680.6Tumor burden score <647.344.847.740.8 Baseline characteristics in LEAP-012: global (ITT) and China However, median PFS in the Chinese TACE-alone arm was 6.5 months, substantially shorter than the 10.0 months in the global cohort (Table 2). Importantly, this discrepancy persisted even when examining the 1-year PFS rate, which was 35% in the Chinese subgroup compared with 43% in the global population. Because early time-point survival estimates are relatively unaffected by censoring, this difference suggests that the poorer outcomes in the Chinese control arm cannot be explained solely by statistical artifacts. Instead, the findings imply that factors beyond baseline patient characteristics may have influenced treatment outcomes. Table 2.PFS and OS in LEAP-012: global (ITT) vs. ChinaLEN+PEM+TACETACE aloneglobal (n = 237)China (n = 105)global (n = 243)China (n = 103)PFS median, months14.616.610.06.5 HR, median (95% CI)0.66 (0.51–0.84)0.53 (0.37–0.75) p value0.0002 1-year PFS rate, %62704335 2-year PFS rate, %39232817OS 1-year OS rate, %89958381 HR (95% CI)0.80 (0.57–1.11)0.58 (0.34–0.97) p value0.0867 PFS and OS in LEAP-012: global (ITT) vs. China Paradoxically, patients in the Chinese subpopulation treated with LEN+PEM+TACE demonstrated better PFS and OS than the global ITT population (Table 2). This phenomenon may be referred to as the “China Paradox” in LEAP-012. TACE Procedure Quality as a Hidden Determinant of Treatment Effect One possible explanation for this paradox lies in the procedure-dependent nature of TACE. Unlike systemic therapies, whose efficacy is largely independent of operator technique, the effectiveness of TACE depends heavily on technical execution. Key procedural elements include precise angiographic mapping of tumor-feeding arteries, identification of extrahepatic collateral supply, successful superselective catheterization, and appropriate embolization endpoints [5]. When these technical steps are not consistently achieved, embolization may be incomplete, leading to insufficient tumor necrosis and potential damage to surrounding liver parenchyma. Such procedural variability may substantially influence clinical outcomes. Indeed, several studies have demonstrated that superselective or precision TACE is associated with improved tumor response and survival compared with non-selective or lobar embolization [5, 8–10]. These considerations may help explain the unexpectedly poor performance of the TACE-alone arm in the Chinese subgroup of LEAP-012. China represents one of the highest-volume environments for TACE worldwide. Although operator expertise is generally high, the logistical pressures associated with large procedural volumes result in heterogeneity in technical execution. Superselective catheterization requires meticulous angiographic mapping and careful microcatheter navigation into tumor-feeding arteries. When this technically demanding step is not consistently achieved and embolization occurs more proximally, tumor control may be compromised while hepatic toxicity increases [5, 8–10]. In a recent multicenter study by Jiang et al. [11], substantial variability in TACE practice across China was identified, particularly with respect to the implementation of precision TACE. This heterogeneity is especially pronounced in high-volume clinical settings, where, in some centers, only approximately 30 min can be allocated per case to accommodate patient throughput. Under such constraints, the high procedural volume can preclude meticulous angiographic mapping for the evaluation of extrahepatic collateral vessels and limit the consistent achievement of superselective catheterization. As a result, procedures are frequently confined to lobar-level approaches. These differences in procedural quality translate into clinically meaningful disparities in outcomes, including lower tumor necrosis rates, deterioration of liver function, and ultimately shorter PFS and OS. Importantly, the study demonstrated significantly superior outcomes in institutions that perform precision TACE compared with those that do not [11]. Because hazard ratios are inherently relative measures, poorer outcomes in the control arm inevitably magnify the apparent benefit of the experimental regimen. Therefore, variability in procedural quality may represent an underrecognized determinant of treatment effect in trials where locoregional therapies serve as comparators. Superior Efficacy of Immunotherapy in HBV-Related HCC However, the favorable hazard ratios observed in the Chinese subgroup cannot be explained solely by potential underperformance of the control arm. Another important factor may be the high prevalence of hepatitis B virus (HBV)–related HCC in the Chinese population (Table 1). Accumulating evidence suggests that patients with HBV-related HCC may derive greater benefit from immunotherapy compared with those with other etiologies. Several systematic reviews and meta-analyses of randomized controlled trials have demonstrated that ICI-based therapies produce superior outcomes in HBV-related HCC [12–15] (Table 3). These studies consistently report favorable hazard ratios for both PFS and OS in HBV-associated tumors compared with HCV-related or non-viral HCC. The biological basis for this observation remains unclear. A higher prevalence of chronic HBV infection in the Chinese population may influence the tumor immune microenvironment in ways that enhance responsiveness to LEN+PEN combination immunotherapy. Moreover, the relatively younger age of HBV patients and the lower prevalence of cirrhosis with preserved liver function may improve treatment tolerability to LEN+PEM+TACE and further contribute to more favorable responses. Table 3.Efficacy of immunotherapy in HCC by etiologyCategoryTreatment lineEtiology (n)Odds ratio (OR) (ORR)PFS HROS HRShao and Akhmetzhanov [14]A systematic review and meta-analysis (9 RCTs)1st lineHBV (n = 3,336)––HBV: HR 0.74HCV (n = 1,427)HCV: HR 0.77Non-viral (n = 2,277)Non-viral: HR 0.86 (p = 0.13 for difference)Wong et al. [15]Systematic review, network meta-analysis (29 RCTs)1st-line and 2nd lineHBV (n = 5,918)– ICI + VEGF/VEGFRi ICI + VEGF/VEGFRi (1L: 16 studies)HCV (n = 2,204)HBV: HR 0.55HBV: HR 0.65Non-viral (n = 2,285)HCV: HR 0.70HCV: HR 0.72(1st line)Non-viral: HR 0.80Non-viral: HR 0.98 (p = 0.01 for difference) ICI (mono or combination) ICI (mono or combination) HBV: HR 0.81HCV: HR 0.79Non-viral: HR 0.83 (p = 0.94 for difference)Guan et al. [13]Meta-analysis (12 RCTs)1st line–HBV: OR 3.93HBV: HR 0.54HBV: HR 0.73HCV: OR 3.08HCV: HR 0.70HCV: HR 0.78Non-viral: OR 3.02Non-viral: HR 0.82Non-viral: HR 0.87Du et al. [12]A systematic review and meta-analysis (8 RCTs)1st lineHBV (n = 2,660)HBV: OR 3.84HBV: HR 0.53HBV: HR 0.71HCV (n = 1,029)HCV: OR 3.05HCV: HR 0.65HCV: HR 0.80Non-viral (n = 1,579)Non-viral: OR 2.99 (p = 0.743 for difference)Non-viral: HR 0.81 (p = 0.001 for difference)Non-viral: HR 0.87 (p = 0.079 for difference) Efficacy of immunotherapy in HCC by etiology Given that HBV infection accounts for the majority of HCC cases in China, this etiologic difference may have contributed to the favorable outcomes observed with the lenvatinib plus pembrolizumab plus TACE combination in the Chinese subgroup. Implications for Trial Design in Intermediate-Stage HCC The findings from LEAP-012 also raise broader questions regarding the appropriate trial design and endpoints for studies evaluating TACE combined with systemic therapy since the global ITT population did not prolong OS differently from the Chinese subpopulation [16]. Most previous TACE combination trials, all of which ended with negative results, have used RECIST-based progression-free survival as the primary endpoint [3, 4, 17–21]. However, in patients undergoing TACE combined with systemic therapy, radiological progression according to RECIST does not necessarily represent true treatment failure. Because TACE can be repeated, the appearance of new intrahepatic lesions or regrowth of previously treated tumors may still be amenable to additional locoregional therapy. To address this limitation, the TACTICS trial [22, 23] introduced the concept of TACE-specific PFS, a composite endpoint that captures clinically meaningful progression events such as tumor growth as compared with baseline, deterioration of liver function, vascular invasion, extrahepatic spread, and TACE refractoriness [24]. This endpoint better reflects real-world clinical practice, where systemic therapy may be continued in combination with repeated on-demand TACE beyond RECIST PD as long as the disease remains controllable by locoregional therapy, preferably resulting in improving OS [1, 25]. Actually, TACE-specific PFS correlated well with OS in the TACTICS and TACTICS-L trial; thus, TACE-specific PFS may be a surrogate endpoint of OS. Taken together, these considerations suggest that TACE-specific PFS may represent a more appropriate endpoint for future trials combining TACE with immunotherapy in intermediate-stage HCC [26]. To date, TACE-specific PFS has been used as an endpoint in several trials, including TACTICS-L [27], TALENTACE [2], and CARES-005 trials [5]. Notably, the EMERALD-3 trial incorporates this treatment strategy, namely, sequential therapy with dual ICIs (Durvalumab plus Tremelimumab) plus lenvatinib, combined with repeated TACE beyond clinical progression. Therefore, the results of EMERALD-3 are highly anticipated. Conclusion The stronger treatment effect observed in the Chinese subgroup of LEAP-012 likely reflects the interaction of multiple factors rather than a single explanation. Procedural quality of TACE may have contributed to underperformance of the control arm, while the high prevalence of HBV-related HCC and favorable patient characteristics in China may have enhanced responsiveness to lenvatinib plus pembrolizumab plus TACE treatment. Ultimately, the “China paradox” highlighted by LEAP-012 underscores two important lessons. First, the technical quality of TACE procedures can profoundly influence clinical outcomes in trials involving locoregional therapies. Second, effective systemic therapy such as lenvatinib plus pembrolizumab has the potential to substantially enhance the results of locoregional treatment strategies. Finally, the incorporation of TACE-specific PFS into future trial designs may be ideal for the successful conduct of global studies evaluating TACE plus immunotherapy combination therapy. Statement of Ethics No statement is needed because this study was based exclusively on published data. Conflict of Interest Statement The author declares lectures at Chugai, Eisai, and AstraZeneca; grants from Otsuka and Chugai; advisory consulting for Chugai, Roche, Eisai, and AstraZeneca. Masatoshi Kudo is the Editor-in-Chief of Liver Cancer. Funding Sources There was no funding for this editorial. Author Contributions Masatoshi Kudo conceived, wrote, and approved the final manuscript.