Hepatocellular carcinoma with macrovascular invasion: Navigating heterogeneous evidence

Macrovascular invasion remains one of the most challenging features in HCC, defining an advanced disease stage associated with poor prognosis and complex therapeutic decision-making. Although systemic therapy is widely regarded as the standard of care, particularly in Western clinical practice, alternative approaches, including surgery and locoregional therapies, continue to be proposed for carefully selected patients, especially in Eastern guidelines, with encouraging results.1–4 This divergence reflects not only differences in health care systems and technical expertise but also persistent gaps and inconsistencies in the available clinical evidence. In this context, Fortuny et al5 report a comprehensive systematic review and survival meta-analysis evaluating patients with HCC and macrovascular invasion treated with upfront locoregional therapies. Patients with macrovascular invasion do not represent a homogeneous clinical entity. The extent and location of vascular invasion (segmental vs. main portal vein), tumor burden, liver function, presence of portal hypertension, performance status, and access to subsequent lines of therapy all have a major influence on outcomes. In observational studies, these factors often drive treatment allocation, leading to substantial selection bias that is difficult to fully address, even with advanced statistical adjustments. Accordingly, among the 73 included studies encompassing more than 10,300 patients with macrovascular invasion and treated with surgery or locoregional therapies (alone or in combination or with a sequential approach), most treatment modalities were associated with pronounced heterogeneity, with I2 values exceeding 80%.5 Notably, patients treated with systemic therapies were not included in this meta-analysis, further limiting the comparability of outcomes across the evolving therapeutic landscape. By excluding studies lacking minimal baseline information on liver function and performance status, the authors attempted to reduce some of these biases. Nevertheless, their analysis shows that substantial residual heterogeneity persists across most treatment strategies, particularly surgical approaches, transarterial chemoembolisation (TACE)-based combinations, and sequential therapies. These findings underscore the need for extreme caution when interpreting pooled survival estimates in this setting and highlight the inherent limitations of meta-analyses based on heterogeneous observational data. One important contribution of this work is the identification of 1-year overall survival (OS) as the most reliable and comparable endpoint across studies. Beyond the first year, data completeness declines sharply, with more than 70% of survival data missing at 5 years, substantially limiting the interpretability of longer-term outcomes. This observation reflects the clinical reality of advanced HCC with macrovascular invasion, where liver function deterioration, treatment discontinuation, and loss to follow-up are frequent. In this context, 1-year OS represents a pragmatic endpoint. However, recent prospective real-world data in patients with macrovascular invasion treated with atezolizumab-bevacizumab report a median OS of ~16 months,6 suggesting that longer time points may be informative when comparing surgery and locoregional approaches with modern immunotherapy-based regimens. Thus, while an emphasis on short-term survival is justified by the limitations of the existing literature, future studies should aim to incorporate longer-term outcomes, aligning endpoints with evolving therapeutic standards in this high-risk population. Among the evaluated modalities, transarterial radioembolization (TARE) emerges as the only treatment associated with low heterogeneity (I2 = 0%) and a pooled 1-year OS of 34%.5 This consistency is a notable finding and likely reflects more homogeneous patient selection and greater standardization of technical procedures compared with other locoregional strategies. Nevertheless, reproducibility should not be equated with clinical superiority. Randomized trials comparing TARE with sorafenib have failed to demonstrate a survival advantage,7,8 and no head-to-head studies have evaluated TARE against contemporary immunotherapy-based regimens. In addition, the pooled 1-year OS reported in this meta-analysis appears lower than that observed in recent real-world cohorts treated with atezolizumab-bevacizumab, particularly among patients with preserved liver function.9 The authors appropriately caution against indirect comparisons and overinterpretation of these findings. Taken together, the available data do not support replacing systemic therapy with TARE but rather highlight the need for dedicated prospective trials directly comparing TARE with immunotherapy-based approaches. Ultimately, these findings suggest that the future may lie not in choosing between locoregional and systemic therapies, but in their rational integration to optimize tumor control while preserving liver function, with the goal of moving a subset of patients closer to curative strategies. One of the most intriguing aspects of the analysis is the apparent survival advantage observed in sequential treatment strategies and surgical series. At first glance, these approaches appear to outperform other modalities, with some pooled estimates suggesting strikingly high 1-year OS rates. However, these findings are almost certainly driven by immortal-time bias. Only patients who survive long enough, preserve adequate liver function, and maintain good performance status can proceed through multiple lines of therapy or be considered for surgical intervention. Consequently, apparent survival benefits reported in retrospective series should not be interpreted as evidence of treatment efficacy, but rather as reflections of patient selection and differential access to subsequent therapies. Across studies and treatment modalities, liver function consistently emerges as a dominant prognostic determinant. Recent observational data in patients treated with immunotherapy further reinforce this point, demonstrating marked survival differences according to ALBI grade.6 These observations underscore the need for therapeutic strategies that prioritize preservation of the hepatic reserve alongside tumor control, as any approach that compromises liver function risks negating potential oncologic benefit. In this regard, composite endpoints such as time to treatment failure, including liver decompensation features, may more accurately reflect clinically meaningful outcomes in future trials.10 Finally, this meta-analysis highlights the urgent need for stratified, head-to-head randomized studies comparing surgery or locoregional approaches with contemporary immunotherapy-based regimens, using standardized prognostic criteria and clinically relevant endpoints. Where randomized trials are not feasible, individual patient data meta-analyses may help address residual confounding that cannot be resolved at the aggregate level. Importantly, this study does not provide definitive answers. Rather, its strength lies in rigorously delineating the limits of the existing evidence and clarifying where conclusions can, and cannot, be drawn. In a field often shaped by institutional practice and regional preferences, such methodological clarity is essential.