FDA finalised three cancer clinical trial eligibility guidances on the same day. All three were issued on 27 July 2026 by the Oncology Center of Excellence with CDER and CBER, and announced in the Federal Register on 28 July 2026. Each finalises a draft that had been sitting since 26 April 2024:
- Laboratory Values — 91 FR 47245, Docket FDA-2024-D-1402 (finalises the draft at 89 FR 32450).
- Washout Periods and Concomitant Medications — 91 FR 47247, Docket FDA-2024-D-1376 (finalises the draft at 89 FR 32440).
- Performance Status — 91 FR 47251, Docket FDA-2024-D-1377 (finalises the draft at 89 FR 32442).
All three are titled “Guidance for Industry, IRBs, and Clinical Investigators,” which is the detail that matters most. FDA has addressed these documents not only to sponsors but explicitly to institutional review boards and site investigators — the people who see a protocol after it is written and decide whether to approve it.
The changes from draft to final are small. FDA states that the Laboratory Values and Washout guidances contain “minor revisions for clarification,” and that the Performance Status guidance adds “updates on additional considerations for sponsors (i.e., impact of including patients with lower performance status on trial retention and sample size)” plus minor clarity revisions. The substance is what was proposed in 2024. What is new is that it is now final, and that all three landed together.
The common premise
Each guidance opens with the same reasoning. Eligibility criteria exist to select the intended population and reduce risk to participants — but, in FDA’s words, “some eligibility criteria may have become commonly accepted over time or used as a template across trials.” Unnecessarily restrictive criteria slow accrual, limit patient access, and produce results that do not represent the population that will actually use the drug.
These are non-binding recommendations issued under FDA’s good guidance practices regulation at 21 CFR 10.115. They establish no rights and bind neither FDA nor the public, and “should” in FDA guidance means suggested, not required. That said, a protocol that departs from a final FDA guidance without a stated rationale is a protocol that has to explain itself — to a review division, and now also to an IRB that has been handed FDA’s expectations in writing.
Performance status: PS2 should be in unless you say why not
This is the guidance with the sharpest single recommendation. FDA defines low performance status as ECOG PS2–4 or KPS ≤70, and then states:
“Patients with ECOG PS2 (or KPS 60-70) should be included unless there is a scientific and/or clinical rationale for exclusion justified by established safety considerations.”
Around that:
- PS eligibility criteria “should be based on the patient population in which the treatment is expected to be applied in clinical practice.”
- Criteria should be re-evaluated and modified throughout development. Later-phase trials (phase 2/3) “should generally mirror the intended use population” and include PS2 patients unless safety concerns emerged in earlier phases — and the rationale for any exclusion “should be justified and stated explicitly in the protocol.”
- Baseline performance status data should be collected in all clinical trials, to characterise the enrolled population, and PS can be used as a stratification factor where the baseline range is wide.
FDA is candid about why PS is a weak gatekeeper. Determination is “inherently subjective,” which affects inter-rater reliability and invites bias at category borderlines. It cites evidence that clinicians assign patients aged 65 and over higher numeric ECOG scores than younger patients despite no difference in objectively measured physical activity, that PS is less predictive of cancer outcomes in older adults, and that it may be less relevant for modern agents whose toxicity profile differs from cytotoxic chemotherapy. Current scales also cannot distinguish low PS caused by disease burden — which effective treatment may reverse — from low PS caused by anything else.
Where a sponsor is genuinely worried, FDA offers routes other than exclusion: pre-specified low-PS cohorts (ECOG ≥2) exempt from the primary analysis, “generally small in size and exploratory in nature,” enrolled incrementally with an early stopping rule based on safety data; and the option of discussing with the review division a primary efficacy analysis restricted to the conventionally eligible subset while enrolling a broader population. It cross-references the December 2019 adaptive design guidance for the former.
It also recommends supplementing ECOG and KPS with patient-reported physical and role function (per the October 2024 core patient-reported outcomes guidance), digital health technologies such as activity trackers for objective activity data, and a geriatric assessment tool for older adults — noting studies showing comprehensive geriatric assessment is more descriptive than PS for older adults and better than KPS at predicting chemotherapy toxicity. These additional assessments “should not limit trial eligibility.”
The retention point added at finalisation is a practical one: lower-functioning participants may depend more on family or caregivers to attend visits, so sponsors “should consider the potential effect of inclusion of patients with low PS on trial retention and necessary sample size,” and may use decentralised study elements to mitigate it.
Laboratory values: stop copying forward the last protocol
The headline recommendation is that laboratory values “should be used as exclusionary criteria only when clearly necessary to mitigate potential safety concerns.” Beneath it, four points with immediate protocol-writing consequences:
- Criteria must be customised to the drug. Requirements should follow from mechanism of action, PK/PD and anticipated toxicities. FDA’s example: if a drug is not hepatically metabolised and is not expected to cause hepatic toxicity in the intended population, hepatic entry criteria should be broad enough to avoid unnecessary exclusions — excluding only multiple-fold elevations of ALT or bilirubin above the upper limit of normal. And explicitly: “laboratory value-related restrictions from earlier clinical trials should not be carried forward automatically.”
- Only as restrictive as the risk requires. FDA works a concrete example. For drugs that may prolong QTc, low potassium, calcium or magnesium can raise arrhythmia risk — and the common protocol response is to require electrolytes “within normal limits.” That excludes patients whose levels are slightly above normal, who carry no increased QTc risk. The correct drafting is to require values above the lower limit of normal.
- Account for inter-laboratory variation. Defining eligibility by broad reference-based ranges rather than specific numerical values is offered as one way to handle it.
- Account for population variation in normal values. FDA points to the Duffy null phenotype and absolute neutrophil count, suggesting sponsors consider broadening ANC criteria accordingly, and notes that reference intervals by construction capture only 95 percent of results from a presumably healthy population — while abnormal values are far more common in people with cancer and frequently carry no clinical significance for the treatment being studied. Protocols can build in the ability to repeat a test within a defined period.
The guidance closes with a requirement to routinely reassess laboratory-based exclusions as a programme moves from early to late phase, loosening or removing anything no longer justified by a specific concern.
Washout periods: replace the calendar with a measurement
The shortest of the three, and the most directly usable. FDA’s core position is that time-based washouts are a proxy for something measurable, and the measurement is better:
- Clinical and laboratory parameters should be used in place of time-based washout periods to address safety concerns — FDA’s model phrasing being that a given laboratory value “must have returned to within normal limits or acceptable baseline prior to enrollment.”
- Where a time-based washout is used (FDA’s example: “at least 14 days must have elapsed since last treatment”), it should be scientifically justified, should take the PK/PD of the prior therapy into account, and the justification “should be clearly specified in the protocol.” A washout may legitimately be appropriate where prior therapy can produce delayed anti-tumour effects and the trial is estimating anti-tumour effect.
- Participants should have recovered from clinically significant adverse events of their most recent anti-cancer therapy where relevant to the investigational drug — for instance where toxicities overlap.
On concomitant medications, exclusion should occur “only when clinically relevant known or predicted drug-drug interactions and potential overlapping toxicities will impact the safety of trial subjects.” The alternative FDA prefers is dose or regimen modification — of either the investigational agent or the concomitant medication — justified and specified in the protocol, with subjects and caregivers adequately informed. It also notes that conducting drug-drug interaction evaluations early can widen enrolment in mid- and late-stage trials.
There is one piece of drafting FDA singles out as unacceptable. Exclusion criteria should carry a disease- and drug-specific scientific rationale “as opposed to vague statements,” and FDA gives the example to avoid verbatim: “Exclude patients taking concomitant medications expected to increase the risk for a clinically significant adverse event.” If that sentence, or something like it, is in your template, it is now specifically named in a final FDA guidance.
A footnote worth catching: for certain complex biological products such as cell or gene therapy products, other considerations may apply and should be discussed with the appropriate CBER office.
What this means for sites and IRBs
Sponsors write protocols, but three of the practical consequences land on institutions.
- IRBs now have a citable basis to ask for a rationale. FDA has said, in final guidance addressed to IRBs, that a PS2 exclusion needs a scientific or clinical rationale stated in the protocol, and that laboratory exclusions carried forward from an earlier trial are not automatically justified. An IRB asking “why is this criterion here?” is applying FDA’s own recommendation, not inventing a hurdle. That fits alongside the review responsibilities under 45 CFR 46 rather than replacing them.
- Feasibility assessments change. If later-phase oncology protocols begin including ECOG PS2 patients as a default, the screen-eligible population at your site widens, and so does the acuity of the enrolled population. Sample size and retention assumptions move with it — which is precisely the consideration FDA added at finalisation.
- Laboratory operations get a say. The inter-laboratory variation point and the repeat-testing allowance are questions your laboratory can answer better than the sponsor can. Reference ranges are local; a protocol written to broad reference-based ranges rather than fixed numeric cut-offs will screen differently at your site than one written to hard numbers.
None of this is a change in law, and none of it has a deadline. It is a change in what FDA has committed to paper as its current thinking, at the moment when a series that began in July 2020 with brain metastases, minimum age for paediatric inclusion, HIV/HBV/HCV infection, and organ dysfunction or prior malignancies — continued in July 2022 with available therapy in non-curative settings — adds its last three drafts as finals. The series is now substantially complete, and general good clinical practice expectations under ICH E6(R3) already push in the same direction on protocol justification.
Frequently asked questions
Are these binding?
No. All three are issued under 21 CFR 10.115 and state that they represent FDA’s current thinking, establish no rights, and bind neither FDA nor the public. An alternative approach is permitted if it satisfies applicable statutes and regulations.
Do they change anything for trials already open?
Nothing is required. The guidances recommend that eligibility criteria be reassessed as a programme moves from earlier to later phase, which is a natural point to revisit an open protocol — but there is no transition requirement and no deadline.
Does the performance status guidance apply to paediatric oncology trials?
The recommendations are stated to be specific to inclusion of adult patients, because paediatric performance status may use a different scale such as Lansky and paediatric inclusion raises its own considerations. FDA adds that many of the general considerations may nonetheless apply to paediatric oncology trials.
Was there a comment period?
Yes, on the 2024 drafts. These are notices of availability of final guidances, not proposals, so there is no open comment deadline attached to them. FDA states it considered comments received on the drafts.
What changed between draft and final?
Very little. Laboratory Values and Washout Periods received “minor revisions for clarification.” Performance Status received additional considerations on the effect of including lower-PS patients on trial retention and sample size, plus minor clarity revisions.
Primary sources: Food and Drug Administration, “Cancer Clinical Trial Eligibility Criteria: Laboratory Values; Guidance for Industry, Institutional Review Boards, and Clinical Investigators; Availability,” 91 FR 47245 (28 July 2026), Docket No. FDA-2024-D-1402; “Cancer Clinical Trial Eligibility Criteria: Washout Periods and Concomitant Medications,” 91 FR 47247 (28 July 2026), Docket No. FDA-2024-D-1376; “Cancer Clinical Trial Eligibility Criteria: Performance Status,” 91 FR 47251 (28 July 2026), Docket No. FDA-2024-D-1377; and the three final guidance documents themselves, each dated July 2026 and issued by the Oncology Center of Excellence with CDER and CBER. Drafts finalised: 89 FR 32450, 89 FR 32440 and 89 FR 32442, all 26 April 2024. Also cited within the guidances: FDA guidance for industry “Adaptive Design Clinical Trials for Drugs and Biologics” (December 2019); “Conducting Clinical Trials with Decentralized Elements” (September 2024); “Core Patient-Reported Outcomes in Cancer Clinical Trials” (October 2024).
