1. TOPAZ-1 versus KEYNOTE-966: treatment duration and what the survival curves suggest
The key structural difference is that TOPAZ-1 used a finite chemotherapy component, whereas KEYNOTE-966 permitted continued chemotherapy and immunotherapy for longer periods.
In TOPAZ-1, patients received gemcitabine/cisplatin with durvalumab for up to 8 cycles. After the chemotherapy phase, durvalumab was continued as maintenance every 4 weeks until disease progression or another discontinuation criterion. Thus, the protocol did not require indefinite gemcitabine.
In KEYNOTE-966, cisplatin was capped at 8 cycles, but gemcitabine could continue without a maximum duration, while pembrolizumab could be administered for up to 35 cycles (approximately 2 years), or until progression/toxicity.
The survival curves from both trials are important because the benefit from adding immunotherapy is not simply an early separation followed by convergence. In TOPAZ-1, the curves separated progressively over time, with approximately 35% of patients alive at 18 months versus 26% with chemotherapy alone in the early analysis. Median OS was 12.8 versus 11.5 months, HR 0.80.
KEYNOTE-966 similarly showed a statistically significant OS benefit, with median OS of 12.7 versus 10.9 months and HR 0.83.
Interpretation: both trials support the concept that adding checkpoint inhibition produces a modest improvement in median survival but potentially a more meaningful long-term survival tail in a subset of patients. However, the trials do not demonstrate that giving immunotherapy indefinitely is superior to a finite duration. Rather, the protocols establish different treatment frameworks that achieved broadly comparable median survival benefits.
2. Discussing durvalumab versus pembrolizumab with a self-funding Nigerian family
For this patient, I would make the conversation explicitly shared and financially informed, rather than presenting the more prolonged treatment schedule as automatically better.
I would explain to the family that both regimens are evidence-based first-line options for unresectable biliary tract cancer. TOPAZ-1 demonstrated an OS benefit with durvalumab plus gemcitabine/cisplatin, while KEYNOTE-966 demonstrated an OS benefit with pembrolizumab plus gemcitabine/cisplatin.
The major practical difference is treatment commitment and cumulative cost. With durvalumab, the chemotherapy component is limited to eight cycles and immunotherapy subsequently becomes maintenance. With pembrolizumab, the protocol permits treatment for up to 35 cycles, potentially creating a substantially longer period of drug expenditure.
For a family paying entirely out of pocket in Nigeria, I would therefore say something like:
"Our aim is not simply to choose the treatment that can theoretically be continued for the longest period. We need a treatment that gives you a reasonable chance of benefit while remaining financially sustainable. We should agree from the beginning on what the family can realistically afford and what we will do if the money becomes limited."
I would discuss:
- Expected benefit: both regimens improve survival compared with chemotherapy alone.
- Duration and cost: a longer permitted treatment duration does not necessarily mean greater survival benefit for every individual.
- Response: reassess radiologically after several treatment cycles and use the response to guide the discussion.
- Toxicity: immune-related adverse events can themselves generate additional costs through investigations, hospitalisation and treatment.
- Financial ceiling: establish what the family can realistically sustain before commencing treatment.
- Contingency planning: agree in advance what happens if funds become insufficient.
In this particular setting, financial toxicity should be treated as a legitimate component of clinical decision-making, alongside efficacy and physical toxicity. A treatment that a family can sustain is preferable to a theoretically attractive regimen that results in treatment interruption because resources are exhausted.
I would avoid telling the family that pembrolizumab must be continued indefinitely. In KEYNOTE-966, the maximum was 35 cycles, not indefinite treatment.
3. What if the family's funds are exhausted during maintenance?
This is where I would be particularly careful. I would not recommend empirically reducing the immunotherapy dose or extending dosing intervals simply to save money, because such modifications are not established equivalent strategies in this disease.
Instead, I would use a response-adapted and resource-conscious strategy.
If the patient has a sustained response or stable disease:
- Stop unnecessary chemotherapy once the protocol-defined chemotherapy course is completed. Continuing gemcitabine indefinitely is not necessary in TOPAZ-1, and cisplatin was capped at eight cycles in both trial designs.
- Continue maintenance immunotherapy when financially feasible, but discuss the family's financial ceiling openly.
- If treatment becomes unaffordable, do not substitute an unvalidated reduced dose or arbitrary extended interval. Instead, consider a planned treatment interruption with close clinical observation, recognizing that this is outside the evidence base of the pivotal trials.
- Use clinically appropriate rather than excessive surveillance. Assessment should focus on symptoms, physical examination, liver function and appropriately timed imaging rather than repeated expensive investigations without a clear clinical indication.
- If progression occurs, reassess the treatment strategy rather than continuing an ineffective and unaffordable maintenance regimen.
- Explore institutional assistance, compassionate-access programmes, pharmaceutical assistance where available, charitable support, and other legitimate financing mechanisms before abandoning an effective therapy.