From algorithms to adaptive frameworks: from square to circle in hepatocellular carcinoma management

The management of hepatocellular carcinoma (HCC) is at a pivotal point. The traditional method, reliant on fixed staging algorithms, is beginning to weaken due to real-world complexities, new treatments, and the rising demand for personalised care (1-3).

However, a recent influential review in CA: A Cancer Journal for Clinicians inadvertently endorses a strict, institution-specific treatment plan that conflicts with its own commendable call for personalised, multidisciplinary management (4). Specifically, Moris et al. propose in their Fig. 3 (page 10) a prescriptive algorithm developed at their institution, which remains unvalidated—that is, it has not undergone external testing in multicentre or prospective studies. The review by Moris et al. (4) accurately states that “therapeutic decision making should be individualised… rather than dictated simply by stage”. However, by adopting such an unvalidated institutional algorithm, the review contradicts its own premise. Scientific journals should exercise caution in promoting prescriptive models without sufficient validation, especially as adaptive paradigms are increasingly endorsed by guidelines.

This inconsistency highlights a broader conceptual inertia in HCC management.

Figure 1 symbolically illustrates this paradox, contrasting the “square” rigidity of algorithmic logic with the “circle” of adaptive frameworks. Although clinicians are increasingly making therapeutic decisions in a multi-parametric, trajectory-sensitive environment, published frameworks often remain based on earlier models, despite growing evidence supporting adaptive approaches. In this context, the prognostic competition between neoplastic events (i.e., progression) and hepatic decompensation presents a fundamental clinical challenge that necessitates continuous assessment and dynamic therapeutic adaptation among the involved specialists (5).

Figure 1 Conceptual shift from algorithmic to adaptive models in HCC management. Visual metaphor showing the transition from rigid, square-shaped algorithmic logic based on binary stage-treatment connections to a flexible, circular model of Multiparametric and Converse Therapeutic Hierarchies (MTH and CTH), where the patient’s journey and intent direct dynamic decision-making. For example, a patient who was initially unresectable may downstage after 1st-line systemic therapy and intra-arterial treatments, prompting the MDT to reconsider resection (‘what now/what next’ logic of CTH). BSC, best supportive care; ECOG PS, Eastern Cooperative Oncology Group performance status; HCC, hepatocellular carcinoma; MDT, multidisciplinary team.

The review by Moris et al., despite its multidisciplinary aim, ultimately relies on a single-centre algorithm. Moreover, it does not acknowledge key modern concepts, failing to recognise Therapeutic Stage Migration (TSM) (3), Multiparametric and Converse Therapeutic Hierarchies (MTH and CTH) (2,6), all principles endorsed in recent European Association for the Study of the Liver (EASL) and European Society for Medical Oncology (ESMO) guidelines (1,7).

TSM refers to the principle that patients may appropriately receive treatments beyond those formally assigned to their stage when justified by individual conditions. It offers greater flexibility for stage-based systems, such as Barcelona Clinic Liver Cancer (BCLC), by avoiding the strict exclusion of patients who might still benefit from therapies at different stages (3). The MTH expands this reasoning further, decisively moving beyond a stage-based hierarchy (7). As recently emphasised by the updated EASL guidelines (1), staging is only one of several parameters to be considered; MTH incorporates patient fitness, tumour burden (not only stage), liver function, and technical feasibility within a structured MDT framework to optimise survival benefit. The CTH further develops this model by incorporating temporal reassessment, allowing for dynamic escalation or de-escalation of therapy based on treatment response and disease progression (6). Together, these frameworks enhance MDT decision-making by encouraging a personalised, multi-parametric assessment of each patient. While the evidence for TSM and MTH is strong, supported by international guidelines and robust clinical experience (6), CTH remains an evolving concept with limited prospective validation and requires confirmation through large, preferably multicentre studies. Notably, the 2025 update of the BCLC strategy further acknowledges this need for structured flexibility by explicitly embedding Therapeutic Stage Migration within a broader clinical decision-making framework based on the CUSE concept (Complexity, Uncertainty, Subjectivity, and Emotion), reinforcing the limits of purely deterministic algorithms (8).

These frameworks support the realisation of the highly personalised approach that the authors themselves advocate. However, when unvalidated institutional algorithms are showcased in prestigious outlets, they risk gaining unwarranted authority and influencing clinician behaviour regardless of their generalisability. This is particularly concerning in the current era, where HCC management increasingly relies on conversion therapy, immunotherapy-induced resectability, and the complex interplay between tumour biology and liver function. In this context, promoting deterministic trees—such as the institutional algorithm proposed by Moris et al.—may unintentionally hinder innovation, delay prompt surgical referral, or diminish the uptake of downstaging strategies (1).

A clear example of the limitations of rigid algorithms is seen in recent surgical evidence showing that patients with early multinodular HCC, who were traditionally excluded from resection, gain significant survival benefits from liver resection (LR) compared to radiofrequency ablation or transarterial chemoembolization (TACE) (9). In the multicentre study by Vitale et al. (9), which included 720 patients with early multinodular HCC, the 1-, 3-, and 5-year survival rates after matching-adjusted indirect comparison (MAIC) were: 89%, 71%, and 56% following resection; 94%, 65%, and 40% after ablation; and 91%, 49%, and 29% after TACE. Weighted multivariable analysis confirmed a significant survival benefit for surgery [hazard ratio (HR) percutaneous radiofrequency ablation (PRFA) vs. LR: 1.41, 95% confidence interval (CI): 1.07–1.86, p=0.01; HR TACE vs. LR: 1.86, 95% CI: 1.29–2.68, p=0.001]. Although retrospective and subject to potential residual selection bias (due to possible “hidden” variables not collected in the centres’ databases), and limited by MAIC assumptions (such as only using dichotomised variables to create weights), these findings underscore how algorithmic exclusion criteria could have denied patients a curative option. It highlights that algorithms can limit therapeutic reasoning, while a multiparametric therapeutic hierarchy provides a more detailed, patient-centred approach to care (2).

Recent updates in international guidelines further endorse the shift towards adaptive reasoning. The 2025 EASL Clinical Practice Guidelines challenge the rigid determinism of earlier models, favouring a therapeutic sequence based on intent (curative versus non-curative), feasibility, biological response, and resource availability (1). Similarly, ESMO advocates a matrix-based decision-making framework that considers liver function, tumour burden, molecular profile, and logistical factors (7). These documents no longer see treatment pathways as fixed or purely algorithmic, but instead emphasise treatment trajectories as a dynamic process.

This evolution is also reflected in the latest British gastroenterology guidelines, which explicitly advise against strict adherence to algorithmic models, and promote a holistic and multidisciplinary approach based on surgical feasibility and clinical complexity (10). The Italian national multisociety guidelines go even further, structurally integrating the MTH and CTH logic into their therapeutic flowcharts and prioritising MDT-based stratification over stage-based classification (11).

In this context, MTH and its development into CTH offer a practical solution to the tension between structure and individualisation. These models provide a structured yet adaptable approach to selecting and sequencing treatments. Instead of rigidly assigning therapies to fixed stages, they emphasise a curative intent whenever it is technically and biologically feasible, based on various parameters including real-time responses, portal pressure measurements, volumetric limits, and surgical feasibility. They function as practical guides for tumour boards managing borderline cases, where the decision is not about choosing “right” or “wrong”, but about deciding “what now” and “what next”. Significantly, their strength lies not only in theory but in practice: these models are already shaping MDT logic across European centres, embedding flexibility into daily decision-making.

We also need to consider the implications of continuing with algorithmic approaches in a global setting. In regions where HCC care is developing rapidly but remains fragmented, prescriptive algorithms might reinforce a therapeutic ceiling, normalising under-treatment and limiting patient access to multidisciplinary assessment. Conversely, MTH/CTH approaches provide clinicians with a tool to encourage treatment escalation, justifying exceptions not as deviations but as informed choices.

We are not advocating for abandoning structure, but for embracing structured flexibility. As shown in Figure 1, therapeutic reasoning in HCC is shifting from the strict “square” of exclusionary, stage-based algorithms to the more adaptable “circle” of MTH and CTH frameworks. This change is not just theoretical. It influences who receives treatment, how they are treated, and with what aim. The square signifies exclusion and categorisation, while the circle denotes inclusion and personalisation. The future of HCC treatment relies not on rigid algorithms but on logic models that reflect the biology of the disease and the realities of care. It involves frameworks that accept ambiguity, encourage multidisciplinary discussion, and prioritise the patient’s therapeutic journey over strict classification. This transition, both symbolically and practically, signifies a shift from the “square” of rigid, binary algorithmic thinking to the “circle” of flexible, adaptive multiparametric reasoning. As shown in Figure 1, this evolution highlights a move from exclusion to intent, from constraint to opportunity. We promote a flexible structure in line with current guidelines. The future depends on adaptive reasoning, dynamic sequencing, and purpose-driven decision-making—models that direct more patients, more frequently, towards a cure.

Acknowledgments

None.

Footnote

Provenance and Peer Review: This article was a standard submission to the journal. The article has undergone external peer review.

Peer Review File: Available at https://hbsn.amegroups.com/article/view/10.21037/hbsn-2025-575/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://hbsn.amegroups.com/article/view/10.21037/hbsn-2025-575/coif). A.V. reports lecture fees and support for attending meetings from AstraZeneca, and Roche. G.C. reports travel support from Roche, AztraZeneca and EISAI and participation on advisory boards for Roche and AztraZeneca. L.R. reports consulting fees from AbbVie, AstraZeneca, Basilea, Bayer, BMS, Eisai, Elevar Therapeutics, Exelixis, Genenta, Hengrui, Incyte, Ipsen, Jazz Pharmaceuticals, MSD, Nerviano Medical Sciences, Roche, Servier, Taiho Oncology, Zymeworks; lecture fees from AstraZeneca, Bayer, BMS, Eisai, Guerbet, Incyte, Ipsen, Roche, Servier; travel expenses from AstraZeneca and Servier; research grants (to Institution) from AbbVie, AstraZeneca, BeiGene, Exelixis, Fibrogen, Incyte, Ipsen, Jazz Pharmaceuticals, MSD, Nerviano Medical Sciences, Roche, Servier, Taiho Oncology, TransThera Sciences, Zymeworks; leadership roles as follows: ILCA Head of External Relations, member of the Executive Committee and the Governing Board, Chair of the EORTC GITCG - HPB/NET TF, Special Expert - International Trials Europe on the HB TF of the NCI GISC, member of the Management Committee of the COST Action CA22125 Precision-BTC-Network. The other authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. EASL Clinical Practice Guidelines on the management of hepatocellular carcinoma. J Hepatol 2025;82:315-74. [Crossref] [PubMed]
2. Vitale A, Cabibbo G, Iavarone M, et al. Personalised management of patients with hepatocellular carcinoma: a multiparametric therapeutic hierarchy concept. Lancet Oncol 2023;24:e312-22. [Crossref] [PubMed]
3. Reig M, Forner A, Rimola J, et al. BCLC strategy for prognosis prediction and treatment recommendation: The 2022 update. J Hepatol 2022;76:681-93. [Crossref] [PubMed]
4. Moris D, Martinino A, Schiltz S, et al. Advances in the treatment of hepatocellular carcinoma: An overview of the current and evolving therapeutic landscape for clinicians. CA Cancer J Clin 2025;75:498-527. [Crossref] [PubMed]
5. Cabibbo G, Celsa C, Battaglia S, et al. Early Hepatic Decompensation Identifies Patients with Hepatocellular Carcinoma Treated with Atezolizumab plus Bevacizumab or Sorafenib at Highest Risk of Death. Clin Cancer Res 2025;31:543-50. [Crossref] [PubMed]
6. Vitale A, Cabibbo G, Rimassa L, et al. The Concept of "Converse Therapeutic Hierarchy" for Patients with Hepatocellular Carcinoma. Liver Cancer. 2025;14:743-757. [Crossref] [PubMed]
7. Vogel A, Chan SL, Dawson LA, et al. Hepatocellular carcinoma: ESMO Clinical Practice Guideline for diagnosis, treatment and follow-up. Ann Oncol 2025;36:491-506. [Crossref] [PubMed]
8. Reig M, Sanduzzi-Zamparelli M, Forner A, et al. BCLC strategy for prognosis prediction and treatment recommendations: The 2025 update. J Hepatol 2025; Epub ahead of print. [Crossref]
9. Vitale A, Romano P, Cillo U, et al. Liver Resection vs Nonsurgical Treatments for Patients With Early Multinodular Hepatocellular Carcinoma. JAMA Surg 2024;159:881-9. [Crossref] [PubMed]
10. Suddle A, Reeves H, Hubner R, et al. British Society of Gastroenterology guidelines for the management of hepatocellular carcinoma in adults. Gut 2024;73:1235-68. [Crossref] [PubMed]
11. Cabibbo G, Daniele B, Borzio M, et al. Multidisciplinary Treatment of Hepatocellular Carcinoma in 2023: Italian practice Treatment Guidelines of the Italian Association for the Study of the Liver (AISF), Italian Association of Medical Oncology (AIOM), Italian Association of Hepato-Bilio-Pancreatic Surgery (AICEP), Italian Association of Hospital Gastroenterologists (AIGO), Italian Association of Radiology and Clinical Oncology (AIRO), Italian Society of Pathological Anatomy and Diagnostic Cytology (SIAPeC-IAP), Italian Society of Surgery (SIC), Italian Society of Gastroenterology (SIGE), Italian Society of Medical and Interventional Radiology (SIRM), Italian Organ Transplant Society (SITO), and Association of Patients with Hepatitis and Liver Disease (EpaC) - Part I - Surgical treatments. Dig Liver Dis 2024;56:223-234. [Crossref] [PubMed]