Society of Cutaneous Oncology · Proposed Fall 2026 Special Session
Inside the Review Room
From patient-level data to FDA regulatory decision-making — a 90-minute moderated conversation about how oncology evidence is reviewed, reconstructed, debated, and ultimately translated into regulatory action.
The idea
Recent regulatory discussions in cutaneous oncology have highlighted an important gap between how clinicians encounter clinical-trial evidence and how FDA reviewers evaluate the same evidence. A publication may present a clean summary estimate; a regulatory review may involve reconstruction of patient-level efficacy and safety data, adjudication of discrepancies, sensitivity analyses, multidisciplinary debate, and a different evidentiary question altogether.
The proposed session would not be a retrospective debate about a particular application. The RP1 experience is better used as the prompt for a broader educational question:
What actually happens inside FDA between receipt of an oncology application and a regulatory decision — and when does an advisory committee become part of that process?
The goal would be to make the review process visible to practicing clinicians and investigators, with particular attention to the parts that are usually hidden from view: patient-level data review, independent analyses, multidisciplinary deliberation, and the role of advisory committees.
Format
Conversational, moderator-led panel
Length
90 minutes
Proposed guests
Katherine + Sandeep, with David moderating
Proposed format
The session would be structured as a conversational, moderator-led panel rather than a series of formal presentations. One or two simple orienting slides could be useful — for example, a schematic of the review pathway and a schematic of what an application contains — but the discussion would otherwise move through moderator questions, dialogue between the panelists, and audience discussion.
Moderator questions would be shared in advance so the discussion can stay focused on process and general principles, while still leaving room for give-and-take between the panelists and the audience. The intent would not be to ask the speakers to discuss confidential deliberations or re-litigate a specific product decision.
90-minute agenda
Act I · approximately 50 minutes Inside the FDA Review Room
0–5 min Opening and
framing
Why this matters to clinicians: the difference between reading a paper
and reviewing an application.
5–15 min What actually arrives at
FDA?
What an oncology application contains, what is available at the patient
level, and how regulatory evidence differs from the final
manuscript.
15–28 min Who reviews the
application?
How the work moves from individual discipline review through
multidisciplinary discussion, supervisory review, office-level
deliberation, and Center-level decision-making.
28–45 min Reconstructing the
evidence
The centerpiece of the session: how efficacy and safety are
independently interrogated, how reviewers work with patient-level data,
and what happens when FDA and sponsor analyses do not align.
45–55 min From analysis to
regulatory action
How unresolved questions become information requests, additional
analyses, labeling discussions, benefit–risk judgments, approval, or a
complete response.
Act II · approximately 25 minutes When the Review Goes Public: Advisory Committees
55–65 min Why convene an advisory
committee?
Who initiates the process, what kinds of uncertainty lead to outside
consultation, and how FDA decides what it wants the committee to
address.
65–77 min Inside the
meeting
How questions are built, discussion versus voting questions, what the
review team listens for, and how the public meeting fits into an already
mature internal review.
77–80 min The morning
after
How the committee’s advice is interpreted and incorporated into the
ongoing regulatory process.
80–90 min Audience discussion +
closing question
How should clinicians interpret an FDA review, a product label, and a
peer-reviewed manuscript as different windows onto the same trial?
Moderator question bank
Act I — Inside the FDA Review Room
1 · What actually comes through the door?
- When a sponsor submits a BLA or NDA, what does the clinical review team actually receive?
- At the patient level, what information is available beyond the sponsor’s tables and figures?
- How standardized are the submitted datasets, and how much work is required before a reviewer can interrogate them?
- What can a reviewer inspect when an individual patient looks unusual — dates, narratives, adverse events, concomitant therapy, imaging assessments, protocol deviations, or other source-level information?
- How different is the evidence package available to FDA from the evidence package that clinicians eventually encounter in a publication?
2 · Who actually reviews the data?
- Who is responsible for the first-pass clinical review of efficacy and safety?
- What is the division of labor between clinical reviewers, statisticians, pharmacologists, safety reviewers, regulatory project managers, and other disciplines?
- Who is actually writing code or reproducing analyses, and who is looking at individual patients?
- How often do clinical and statistical reviewers arrive at different interpretations of the same signal?
- How are disagreements handled within the multidisciplinary review team?
- At what points does the review move from an individual reviewer to team-level, office-level, and Center-level discussion?
3 · What does “FDA independently reviewed the data” actually mean?
- If a sponsor reports an objective response rate, progression-free survival estimate, or safety event rate, how independently does FDA reproduce that number?
- Does FDA effectively construct its own analytic version of the trial from the submitted patient-level data?
- Which analyses are routinely reproduced, and which are generally accepted from the sponsor unless a concern emerges?
- How are evaluability, missing assessments, censoring, progression dates, response confirmation, protocol deviations, and analysis populations handled?
- What happens when the sponsor’s estimate and FDA’s estimate differ?
- How large or consequential does a discrepancy need to be before it changes the regulatory conversation?
- Is the key issue usually the numerical difference itself, or what the discrepancy reveals about uncertainty in the underlying evidence?
4 · Following one patient down the rabbit hole
- What makes a reviewer stop at an individual patient and investigate more deeply?
- How much manual patient-level review occurs in a modern application?
- Can a single patient materially change the interpretation of an efficacy endpoint or a safety signal?
- How are ambiguous cases adjudicated across clinical and statistical reviewers?
- When FDA asks the sponsor for clarification, what does that exchange look like in practice?
- How often does a patient-level discrepancy lead to a broader re-analysis of the dataset?
5 · From analysis to a regulatory judgment
- At what point does the review move from “What does the dataset show?” to “Is the benefit–risk profile sufficient for approval?”
- How is uncertainty carried into that benefit–risk judgment?
- How are different disciplinary reviews synthesized into a single regulatory position?
- What kinds of unresolved issues can be addressed through labeling, postmarketing commitments, or additional analyses — and what kinds of issues are fundamental enough to prevent approval?
- How does a complete response letter fit into the broader review process?
- When an application returns after a complete response, what is effectively reopened and what prior work carries forward?
Act II — When the Review Goes Public: Advisory Committees
6 · Why does an application go to an advisory committee?
- Who first raises the possibility of an advisory committee, and who ultimately decides to convene one?
- At what stage of the review is that decision usually made?
- What kinds of problems are best suited to outside expert advice: scientific uncertainty, disagreement about benefit–risk, novel technology, precedent-setting questions, or something else?
- Does referral to an advisory committee imply that the internal review team is divided, or can it occur even when the team has a fairly developed view?
- What does FDA hope to learn from an advisory committee that it cannot obtain from its internal review alone?
7 · How are the committee questions created?
- Who writes the questions that ultimately appear in the briefing documents and at the meeting?
- How does FDA decide whether it wants a discussion question, a formal vote, or both?
- What makes a good voting question from the regulatory team’s perspective?
- How much should observers infer from the wording of FDA’s questions?
- By the time the advisory committee meets, how mature is the internal review team’s thinking?
8 · What does the review team hear that the audience may miss?
- Is the numerical vote the most important output, or is the discussion often more informative?
- How does the team interpret a strongly favorable vote with important reservations? A divided vote? A unanimous vote with substantial uncertainty?
- What kinds of comments from committee members can materially change the team’s thinking?
- How should clinicians understand the fact that the committee is advisory rather than determinative?
9 · What happens the next morning?
- Does the review team formally reconvene after the advisory committee?
- How is the committee’s advice incorporated into the remaining review?
- Can the meeting change the requested analyses, labeling discussions, postmarketing requirements, or ultimate regulatory recommendation?
- When FDA ultimately reaches a conclusion that differs from the committee vote, what is the best way for clinicians to understand that apparent divergence?
Closing discussion
A final synthesis could return to the question that motivated the session:
Why might FDA and a peer-reviewed publication describe the same clinical trial differently without either necessarily being “wrong”?
Possible threads:
- different data cuts or analysis populations;
- independent reconstruction of efficacy or safety outcomes;
- different handling of missing data, censoring, evaluability, or protocol deviations;
- sensitivity analyses that matter more in a regulatory review than in a manuscript;
- regulatory benefit–risk questions that are not identical to the scientific question posed in a publication;
- the difference between summarizing a trial and deciding whether the evidence supports a particular regulatory action.
The take-home message would be that the FDA review, product label, and peer-reviewed manuscript are related but distinct views of the evidence — each generated for a different purpose.
Candidate fall dates
The fall scheduling survey identified the following 90-minute windows:
| Option | Date | Time |
|---|---|---|
| September | Friday, September 25, 2026 | 3:30–5:00 PM ET |
| October | Wednesday, October 28, 2026 | 3:30–5:00 PM ET |
| November | Friday, November 20, 2026 | 3:30–5:00 PM ET |
These were selected through the SoCO fall scheduling survey to maximize broad participation across the three fall meetings.
Working title options
Preferred: Inside the Review Room: From Patient-Level Data to FDA Regulatory Decision-Making
Alternative: Inside the FDA Review: How Oncology Evidence Becomes a Regulatory Decision