Questions for a targeted-therapy proposal
You will be able to: Turn a proposal into an assessable claim with evidence, dependencies and an explicit decision.
A useful proposal states what it hopes to change, what supports that expectation and what would count as failure. A platform name or target shortlist is too little information to assess benefit, safety or readiness.
Before starting: Mechanism and reading cancer evidence.
Ask in this order
| Question | What a useful answer contains |
|---|---|
| What clinical problem is being addressed? | Disease setting, treatment intent, available alternatives and measurable goal |
| What is the target? | Exact molecule or peptide–HLA (human leukocyte antigen) combination, specimen and method |
| Is it accessible and sufficiently selective? | Protein localization and normal-tissue evidence, with sampling limits |
| What is the proposed format? | Binder, payload or cell design and a product-specific mechanism |
| What evidence predicts benefit? | Relevant human study, or a clear description of preclinical evidence and uncertainty |
| What is the safety plan? | Expected risks, monitoring, management capability and stopping rules |
| What must happen before use? | Manufacture, quality release, sponsor, site and access pathway |
| What is being decided now? | Data collection, research, manufacturing or administration; each has different consequences |
Product-specific questions
For an ADC (antibody-drug conjugate), ask how the target and linker deliver the payload, which resistance mechanisms matter and what indication or protocol supports use. Internalization is often important but is not a universal binary disqualifier independent of product design.
For an engager or cellular therapy, ask about CRS (cytokine release syndrome) and neurotoxicity, target expression on healthy tissue, trafficking and persistence. Stopping a protein infusion does not immediately reverse its biological effects. A living cell product also needs its own persistence and management plan.
For a radioligand, ask which imaging agent informs its distribution, how imaging relates to therapy, and what dosimetry and organ limits apply. Tracer uptake alone does not predict either clinical benefit or a CAR (chimeric antigen receptor)-T safety threshold.
For TCR (T-cell receptor)-based therapy, ask which peptide and HLA allele are recognized, how specificity was established and whether the target complex is present. A list of expanded clonotypes does not answer these questions.
Assess the proposed decision
An exploratory experiment, a committed manufacturing batch and treatment administration are distinct decisions. Ask what each costs in time, tissue and opportunity, what is reusable and what requires new consent or review. Avoid describing speculative therapy development as guaranteed insurance against recurrence.
Access check: An investigational new drug (IND) application governs investigation or expanded access; a biologics license application (BLA) concerns marketing approval. An IND is not a BLA. Manufacturer participation and a capable clinical site remain necessary.
Try it
A proposal says the target is highly expressed and “the drug can be prescribed.” What evidence is missing?
Explain it back: Accessible protein, normal-tissue risk, actual product availability, the indication or access pathway, and evidence for the patient's disease setting. RNA abundance does not establish these.
Explain it back
A good question reveals a missing premise before it becomes a treatment assumption.
Takeaway
A good question reveals a missing premise before it becomes a treatment assumption.
Current application
Treatment status belongs in the care plan, unresolved choices in questions, and provider-specific capabilities in companies. A mechanism lesson does not establish eligibility or a treatment plan.
Sources
Checked October 8, 2026.