For decades, cancer treatment and drug reimbursement have largely been organized around the anatomical origin of a tumor. Lung cancer, melanoma, breast cancer, ovarian cancer, and hundreds of rarer malignancies have followed separate regulatory pathways, even as molecular oncology and immunotherapy have increasingly challenged the idea that organ of origin should always determine therapeutic access. Tumor-agnostic approvals have already begun to reshape this framework, but they have generally remained linked to defined molecular biomarkers and specific regulatory indications.
Australia has now taken a considerably broader step. On March 1, 2026, the Australian Pharmaceutical Benefits Scheme (PBS) introduced what Vivek Subbiah and Razelle Kurzrock describe in a recent Journal of Clinical Oncologycommentary as the first national pan-tumor, multi-indication reimbursement framework for nivolumab plus ipilimumab. Under this approach, dual immune checkpoint blockade can be reimbursed for patients with advanced or metastatic cancer when the treating oncologist determines that treatment is appropriate. Pembrolizumab followed through a similarly broad recommendation from the Pharmaceutical Benefits Advisory Committee (PBAC).
The policy raises an important question for immuno-oncology: whether access to checkpoint inhibition should always remain tied to tumor histology when the biology determining immune response can extend across conventional cancer classifications. Australia is now approaching this question at the level of a national reimbursement system, with potential implications for rare cancers, biomarker selection, real-world evidence generation, and the future economics of checkpoint inhibition.
From Tumor-Agnostic Approval to Tumor-Agnostic Access
Tumor-agnostic treatment itself is not a new concept. Precision oncology has already established therapies whose indications are determined by molecular characteristics rather than anatomical site, and immunotherapy has contributed substantially to this evolution. The Australian framework, however, moves beyond the conventional model of biomarker-defined tumor-agnostic treatment.
According to the commentary, the PBS listing places greater responsibility for treatment selection with the treating oncologist rather than requiring every rare tumor type to independently pass through the traditional indication-by-indication reimbursement pathway. The policy developed following a PBAC recommendation for nivolumab plus ipilimumab and a risk-sharing arrangement between Bristol Myers Squibb Australia and the Australian Department of Health, Disability and Ageing. Pembrolizumab subsequently received a broad multi-indication recommendation after a revised submission, and approximately 5,000 Australian patients annually are expected to benefit from the expanded framework.
This approach may be particularly consequential for patients with rare and ultra-rare malignancies. Patients with common cancers can benefit from large randomized trials, established regulatory pathways, multiple guideline-supported therapies, and dedicated reimbursement decisions. For very uncommon cancers, comparable evidence may never be generated, even when biological rationale and prospective clinical observations suggest that checkpoint inhibition can produce meaningful activity.
Rare Cancers Expose a Structural Problem in Evidence Generation
Subbiah and Kurzrock argue that indication-by-indication reimbursement creates a particular disadvantage for rare cancers. Rare and ultra-rare malignancies collectively represent a substantial disease burden, yet individual populations can be too small to support the development pathway routinely used for common cancers. Dedicated randomized trials may be impractical, recruitment may take years, and commercial incentives for obtaining a separate indication can be limited.
The NCI/SWOG S1609 DART study illustrates both the potential of checkpoint blockade in these diseases and the difficulty of generating conventional evidence. DART was a prospective multicenter basket study evaluating nivolumab plus ipilimumab across rare and ultra-rare cancers. The program enrolled 798 patients across 53 rare or ultra-rare cohorts at 1,083 US sites, providing an unusually broad prospective evaluation of dual checkpoint blockade in malignancies that are often underrepresented in conventional drug development.
The results also illustrate why conventional summary endpoints may not fully characterize immunotherapy benefit in very small populations. Across several rare malignancies, objective response rates were modest, often approximately 10%–25%, while a subset of responding patients experienced prolonged disease control. In refractory metastatic metaplastic breast cancer, for example, the ORR was 18%, with responses in 3 of 17 patients, and median PFS was only two months; however, all responders remained disease-free or alive at two to three years. In refractory clear-cell ovarian cancer, the ORR was 16% and median PFS was 3.7 months, but all three responders and one additional patient with stable disease had ongoing benefit extending beyond three to five years. In ovarian granulosa cell tumors, responses were observed in approximately one-quarter of patients and lasted three to five years.
These patterns are particularly relevant to checkpoint inhibition, where population-level median outcomes can coexist with durable benefit in a relatively small subgroup. In diseases with several established therapeutic options, a low overall response rate may limit the clinical relevance of a treatment. The interpretation can be different in an ultra-rare, aggressive malignancy for which effective systemic options are extremely limited.
DART (SWOG S1609) Phase II Trial: Nivolumab Plus Ipilimumab in Gynecologic Clear Cell Carcinomas
When the Trial Population May Never Be Large Enough
The challenge becomes even clearer in some of the smallest DART cohorts. Despite the study being available across more than 1,000 US sites, only six patients with small cell carcinoma of the ovary, hypercalcemic type were enrolled. Two of those six reportedly achieved complete remissions that remained ongoing beyond four years. Only four patients with chemotherapy-refractory gestational trophoblastic disease were accrued, three of whom experienced 100% tumor reduction that remained ongoing beyond three years.
These cohorts cannot provide the certainty expected from large randomized trials, but they illustrate the practical difficulty of applying the same evidentiary requirements across cancers with profoundly different incidence. Even large cooperative networks may struggle to complete enrollment when only a very small number of eligible patients exist. The Australian framework attempts to address this limitation by allowing available clinical evidence, individual patient characteristics, and oncologist judgment to contribute more directly to reimbursement decisions for advanced disease.
Clinical Judgment Becomes Part of the Access Framework
Under the model described by Subbiah and Kurzrock, broader reimbursement does not imply that nivolumab plus ipilimumab or pembrolizumab is expected to be effective across all cancers. Instead, eligibility determination is placed with the treating oncologist, who can consider tumor biology, treatment history, available alternatives, performance status, potential for benefit, and the patient’s risk of immune-related toxicity. The Medical Oncology Group of Australia supported this framework, according to the authors.
This changes where uncertainty is managed. Under an indication-specific reimbursement system, absence of a dedicated regulatory or reimbursement pathway can prevent access even when limited prospective evidence suggests activity. The Australian approach permits greater clinical discretion in situations where conventional evidence generation may be unrealistic, while placing greater responsibility on clinicians to distinguish a biologically and clinically reasonable use of checkpoint blockade from treatment with little expectation of benefit.
Such discretion will inevitably produce variation in treatment decisions, making systematic outcome collection particularly important. The long-term value of the model will depend not only on the number of patients gaining access to treatment but also on whether the resulting experience can identify which tumor types and patient populations derive meaningful benefit.
The Biomarker Question Remains Unresolved
One of the most controversial aspects of broad pan-tumor immunotherapy reimbursement is the absence of mandatory predictive biomarker selection within the framework described by the authors. Without a universal biomarker requirement, patients with tumors that are relatively insensitive to checkpoint blockade could receive treatment without benefit while remaining exposed to immune-related adverse events and the healthcare costs associated with therapy.
Biomarker-guided access might appear to provide a more precise alternative, but current immunotherapy biomarkers do not offer a universal solution. MSI-high status, high tumor mutational burden, and PD-L1 expression can enrich for response in specific clinical settings, yet their predictive value varies substantially across tumor types. The authors also point to cancers such as Kaposi sarcoma, where checkpoint inhibitor activity can occur despite the absence of these conventional predictive characteristics.
The complexity reflects the biology of checkpoint inhibition. Antitumor immunity is influenced by tumor antigenicity, antigen presentation, T-cell states, immune exclusion, myeloid populations, stromal organization, previous therapies, and multiple other components of the tumor microenvironment. Until predictive models can integrate these variables reliably across histologies, restricting pan-tumor access to a small number of established biomarkers could exclude patients capable of deriving durable benefit.
At the same time, biomarker uncertainty cannot justify indiscriminate treatment. The Australian model will therefore depend on careful clinical selection and on accumulating evidence that can progressively identify cancers in which broad checkpoint inhibitor access is more or less likely to provide meaningful benefit.
Broader Access Requires Equally Broad Safety Surveillance
Expanded access also means that more patients will be exposed to immune-related toxicity, including some who will not experience antitumor benefit. This is particularly relevant for nivolumab plus ipilimumab, where dual PD-1 and CTLA-4 inhibition can produce clinically significant immune-mediated adverse events.
The framework described in the commentary includes a mandated three-year review, creating an opportunity to evaluate both efficacy and toxicity after implementation. The authors propose prospective tracking by tumor type of immune-related adverse events and their management, objective response rates, durable disease control, duration of response, PFS, OS, and treatment discontinuation with the reasons for stopping therapy.
Linking PBS prescribing information with cancer registry and mortality data could provide population-level follow-up that would be difficult to obtain through individual studies of rare cancers. Such surveillance would be particularly valuable if it captures not only whether immune-related toxicity occurs but also its severity, management, and relationship to treatment discontinuation.
Can Real-World Evidence Refine Pan-Tumor Access?
The Australian model also creates an opportunity to use reimbursement as part of an ongoing evidence-generation process. Traditional oncology drug development generally moves from clinical trials to regulatory approval and then reimbursement, but that sequence becomes difficult when patient populations are too small to support conventional trials.
Systematically collected real-world outcomes cannot remove the limitations inherent to non-randomized evidence, including selection bias, differences in previous treatment, heterogeneity across rare tumor types, and residual confounding. However, these limitations need to be considered within the practical constraints of studying ultra-rare cancers, where sufficiently powered randomized trials may be difficult or impossible to conduct. In this setting, longitudinal real-world outcomes can complement evidence generated by small prospective studies and basket trials while helping identify tumor types in which checkpoint blockade produces reproducible and durable activity. The value of the Australian framework will therefore depend heavily on the quality of this post-reimbursement evidence and on whether emerging efficacy and safety data are used to refine treatment access over time.
The authors propose linking prescribing records with cancer registries and mortality data, potentially allowing national evaluation of treatment outcomes. If implemented rigorously, this could provide information on cancers that benefit consistently, populations with little apparent activity, patterns of exceptional response, and safety signals that emerge as treatment reaches broader patient groups.
Moving Beyond Organ of Origin Without Abandoning Precision
The United States has already established regulatory precedent for histology-independent cancer treatment, although these approvals have generally been linked to defined molecular characteristics. The Australian approach described by Subbiah and Kurzrock extends the concept by allowing checkpoint inhibitor reimbursement across advanced and metastatic cancers without requiring every histology to obtain a separate indication or every patient to meet a single biomarker criterion.
This does not make tumor histology irrelevant. Different cancers have profoundly different immune environments and probabilities of responding to checkpoint inhibition, and existing disease-specific evidence should continue to guide treatment whenever it is available. Rather, the framework addresses circumstances in which a rare cancer lacks a dedicated reimbursement pathway despite prospective or biological evidence suggesting potential sensitivity to immunotherapy.
As real-world experience accumulates, the broad initial framework could potentially become more precise. Tumor type, molecular features, immune biomarkers, treatment history, and clinical characteristics could eventually be integrated to identify populations with the greatest probability of durable benefit while limiting unnecessary exposure in tumors with consistently poor outcomes.
The Biosimilar Era Could Change the Economics of Access
The timing of the Australian decision may also be important. Subbiah and Kurzrock argue that the development of checkpoint inhibitor biosimilars could substantially change the economic context of broad immunotherapy access over the coming years. As lower-cost versions of established PD-1 and related therapies become available, some financial barriers that currently shape reimbursement decisions may diminish.
Reduced drug costs, however, will not by themselves create equitable access. A less expensive checkpoint inhibitor can remain unavailable when reimbursement is restricted to specific tumor indications, while broader availability without appropriate patient selection could increase unnecessary treatment and toxicity. The authors therefore propose that future pan-tumor frameworks could become agent-agnostic and accommodate multiple approved checkpoint inhibitors and biosimilars while preserving competition and affordability.
For lower- and middle-income countries, this intersection between biosimilar availability and reimbursement architecture may become especially relevant. Lower prices could expand the practical possibility of checkpoint inhibitor treatment, but health systems would still need mechanisms determining which patients can receive these therapies and how outcomes are monitored.
Read Full Article Here
