Registered forecast
68% registered probability event by 2027-03-03 Unresolved
FDA approves zanzalintinib with atezolizumab for previously treated metastatic colorectal cancer on or before the disclosed PDUFA goal date of 2027-03-03
FF-063 Registered - Registered probability
- 68%
- Decision by
- 2027-03-03
- Reading window
- 2026-10-22 → 2027-03-03
- Decision stage
- 5 Guideline
- 6 Regulatory
- 7 Coverage
- Baselines it has to beat
-
- Authority base rate 80%
- Calendar cycle 80%
- Analyst proxy 72%
nextconsensus.com/forecast/ff-063/ registry snapshot 2026-11-07 Share image
NextConsensus registered this probability on 2026-11-04. It is one of 35 registered FDA goal-date forecasts.
The proposition
What was registered
The FDA takes an approval action on Exelixis’s New Drug Application for zanzalintinib, in combination with atezolizumab, for the treatment of patients with previously treated metastatic colorectal cancer, with an approval action date on or before 2027-03-03, the PDUFA target action date as extended by FDA on 2026-09-10 and disclosed in Exelixis’s Form 8-K filed 2026-09-11.
- Who decides
- US Food and Drug Administration (FDA), Center for Drug Evaluation and Research (CDER) (US)
- Decision stageThe kind of body that has to act: a guideline panel, a regulator, or a payer. The ladder, rung by rung
- Rung 6: Regulatory action — Label/indication change, approval action, REMS modification, or withdrawal by FDA, EMA, PMDA, Health Canada, etc.
- Decision typeThe type of decision this forecast is counted with, such as FDA actions by their goal dates
- FDA action by goal date
- Scope
- Previously treated metastatic colorectal cancer
The estimate
The number and the baselines it has to beat
The estimate that the named body takes the action by the deadline, from what was public at the cutoff.
The alternatives it was registered against, each with the reasoning that produced it. They were set at registration: how often this kind of decision happens on time, what the calendar alone suggests, and a reading of the same public record without this method. Each one's reasoning shows whether it is a count of past decisions or an estimate.
ESTIMATE FROM THE SHAPE OF THE CLASS, NOT AN ENUMERATED COUNT. Original oncology NDAs backed by a randomized phase 3 trial that met an overall survival endpoint are approved at their goal date in a large majority of cases; a goal-date extension after a major amendment is common and is not itself a refusal. The calendar’s own met-goal-date share for disclosed PDUFA dates does not exist yet (docs/foundations/short-horizon-reference-classes.md §4.1).
A clock-only model: the goal date is inside the window, FDA meets most goal dates, and most first-cycle actions on original NDAs are approvals. With no evidence input it cannot tell a strong application from a weak one, or an extended clock from an original one.
An analyst reading the same filings on 2026-10-21 would weigh a statistically significant overall survival benefit against an active comparator, published in The Lancet, and a large established sponsor, against a missed co-primary endpoint in patients without liver metastases and an FDA information request that turned into a major amendment and a three-month extension. Estimated by reconstructing that read, not by surveying analysts.
Regulatory action (FDA): offered on judgment, with no calibration figure until 15 in the class have resolved.
The method and every record are public, so you can check the work.
The clock
Dates that constrain the record
- Evidence as of
- 2026-10-21
Nothing after this date informed the estimate. - Registered
- 2026-11-04
Frozen and hashed. - Reading window
- 2026-10-22 → 2027-03-03
When the outcome can be read against the rule. - Event by
- 2027-03-03
The final qualifying action date; the frozen resolution rule defines what counts. Add to calendar
62b4a6317c9c2650b1bd1f3078ddfdcb788c70a4d0864d9c7db563dd7287286f · operator record jc-1793799419224-8f26ef1f Around this record
- This kind of deadline
- Check it, follow it
Confidential Asset & Milestone Evaluation
Evaluating an asset or milestone that the FDA could move?
Name the asset, the deciding authority, and the date your team commits to: a P&T review, an HTA submission, a label negotiation, a deal closing, or a milestone dispute. Your stake need not be direct sponsorship — an asset you evaluate, finance, advise on, or rely upon is enough. Every commercial evaluation applies the same rigorous pre-registration and primary-source verification shown on this page, delivered privately under mutual NDA.
Resolution
The rule that settles the forecast
The application number was not public at the cutoff, and openFDA held no record for the ingredient (SRC-004, SRC-005). The adjudicator identifies the qualifying approval by its indication plus its ingredient, never by a name the adjudicator infers: an approval action whose Drugs@FDA approval letter or label names metastatic colorectal cancer in previously treated patients in its indication, with zanzalintinib used in combination with atezolizumab, for a product whose active ingredient is zanzalintinib (XL092) or a salt of it, under any applicant (Exelixis, Inc. or an Exelixis subsidiary) and any proprietary name. Resolves YES if the openFDA Drugs@FDA records returned by https://api.fda.gov/drug/drugsfda.json?search=products.active_ingredients.name:%22ZANZALINTINIB%22&limit=100 and by https://api.fda.gov/drug/drugsfda.json?search=openfda.generic_name:%22zanzalintinib%22&limit=100, read together (openFDA does not always fill the generic-name field for a new application), contain such an approval: an application whose ORIG submission has submission_status AP and a submission_status_date on or before 2027-03-03. An indication narrower than the applicant sought (for example, limited to patients without active liver metastases, or by line of therapy or biomarker status) still resolves YES. An approval of zanzalintinib for another tumor type, or as monotherapy only, does not count and resolves partial_nonqualifying. Resolves NO if no qualifying approval action is dated on or before 2027-03-03: a complete response letter, a goal-date extension, withdrawal or no action. The goal date is frozen as disclosed at the cutoff: 2027-03-03, which already includes the three-month extension FDA notified on 2026-09-10 after deeming updated data a major amendment (SRC-002); the earlier date of 2026-12-03 (SRC-001) is superseded and is not the goal date, though an approval dated on or before it also resolves YES. A further extension does not move the date, and an approval dated after 2027-03-03 resolves NO as occurred_after_deadline. 2027-03-03 is a Wednesday. Read no earlier than 2027-03-17. If openFDA has not refreshed by then, or both queries return nothing, the Drugs@FDA web record (https://www.accessdata.fda.gov/scripts/cder/daf/, searched by the name zanzalintinib) governs, under the same identification test. Resolves unresolvable only if none of these sources can be read after the read floor. Resolves VOID, with the reason recorded, only if FDA withdraws every public record of approval actions named above without an equivalent public successor, so that no approval date for this product can be read; while any of them publishes it, the rule stands.
Source the rule names: https://api.fda.gov/drug/drugsfda.json?search=openfda.generic_name:%22zanzalintinib%22&limit=100, https://api.fda.gov/drug/drugsfda.json?search=products.active_ingredients.name:%22ZANZALINTINIB%22&limit=100, https://www.accessdata.fda.gov/scripts/cder/daf/
Readers: Resolver A, clinical methodology; Resolver B, payer evidence. The same two reader roles are assigned to every forecast registered so far.
What this record will show once it resolves
Beside every result, the count for its class (FDA action by goal date): registered, resolved, unresolved, and expired. One forecast is one instance, not a rate.
- The finding
- As the rule above reads it: occurred on time, did not occur, or left unscored.
- The readers and their sources
- Two reader roles read the source named above: Resolver A, clinical methodology; Resolver B, payer evidence. Who adjudicated it is named here, with the sources they read when the adjudication lists them. Being named as a reader means applying the frozen rule to the public record; it is not support for the estimate, the method, or NextConsensus.
- The number against its baselines
- 68% beside three baselines, each scored by the same rule (squared error), so a reader can see how the record did against each of them.
Reasoning
What the number rests on
The facts it was built from are below, as they stood at the cutoff, with every source archived; each baseline above carries the reasoning that produced it. The written mechanism and the conditions that would make the proposition ill-posed were frozen with the record and are held in the ledger, which the hash on this page covers.
Frozen covariatesThe facts about the evidence and the authority's process that were frozen with the forecast at its cutoff
The state of the record at the cutoff
Recorded as they stood on 2026-10-21, including the 2 fields whose values were unknown then. Those are left unknown rather than filled in.
- review pathway
- Standard review (SRC-001). Center CDER is the drafter’s understanding (a small-molecule NDA), unverified.
- submission type
- Original NDA for zanzalintinib in combination with atezolizumab, submitted December 2025, accepted February 2026 (SRC-001); application number not public, and openFDA holds no record for the ingredient (SRC-004, SRC-005).
- evidence class
- One randomized, active-controlled phase 3 trial, STELLAR-303, versus regorafenib, in metastatic refractory non-MSI-high or non-dMMR colorectal cancer (SRC-001).
- pivotal n
- unknown — not stated in the archived sources
- endpoint type
- Clinical: overall survival, as dual primary endpoints in the intention-to-treat population (met) and in patients without liver metastases (not met) (SRC-001, SRC-003).
- approved competitors in class
- unknown — the archived sources name regorafenib as the trial comparator but do not list the approved later-line therapies
- months since last guideline revision
- not applicable — no guideline is the subject of this proposition
- panel cycle position
- not applicable — the act is an FDA approval action
- sponsor request status
- NDA accepted February 2026 with a goal date of 2026-12-03 (SRC-001); extended three months to 2027-03-03 on 2026-09-10 after FDA deemed updated safety and efficacy data, submitted in response to its information request, a major amendment (SRC-002).
- odac referral
- none announced in the archived sources
- sponsor state change
- Goal-date extension after a major amendment, announced in the 8-K filed 2026-09-11 (SRC-002); Exelixis’s 2026 revenue guidance excludes any zanzalintinib launch (SRC-003).
Sources
Everything the estimate was built from
Each was retrieved on or before the 2026-10-21 cutoff and archived at the digest shown. If a publisher edits a page after the fact, the digest stops matching and the substitution is visible.
-
SRC-001Exelixis Form 10-Q for the quarter ended 2026-07-03, filed 2026-08-05: STELLAR-303 (zanzalintinib plus atezolizumab versus regorafenib, metastatic refractory non-MSI-high colorectal cancer) met its dual primary endpoint of OS in the ITT population (HR 0.80, 95% CI 0.69-0.93, P=0.0045 at final analysis, published in The Lancet); NDA submitted December 2025, accepted February 2026 under standard review with a PDUFA date of December 3, 2026; in June 2026 the other dual primary endpoint, OS in patients without liver metastases, showed a non-significant trend (HR 0.83, 95% CI 0.66-1.05, P=0.1185; median 15.9 versus 12.7 months). -
SRC-002Exelixis Form 8-K, filed 2026-09-11 (latest Exelixis filing before the cutoff): on 2026-09-10 FDA extended the review of the zanzalintinib plus atezolizumab NDA in metastatic colorectal cancer by three months; Exelixis had submitted updated safety and efficacy data in response to an FDA information request, which FDA deemed a major amendment; updated PDUFA action date March 3, 2027. -
SRC-003Exelixis Form 8-K Exhibit 99.1, filed 2026-08-05: Q2 2026 results release. FDA is reviewing the NDA for zanzalintinib with atezolizumab in previously treated metastatic colorectal cancer; 2026 revenue guidance excludes any zanzalintinib launch; restates the June 2026 non-significant OS trend in the non-liver-metastases subgroup of STELLAR-303 and the ITT OS endpoint met. -
SRC-004openFDA drugsfda products.active_ingredients.name:"ZANZALINTINIB", retrieved 2026-10-21: 404 No matches found. No application for the ingredient. -
SRC-005openFDA drugsfda openfda.generic_name:"zanzalintinib", retrieved 2026-10-21: 404 No matches found. -
SRC-006PDS Biotechnology Form 8-K Exhibit 99.1, filed 2026-10-21: a third party's October 2026 investor presentation that tabulates published zanzalintinib plus atezolizumab results in metastatic colorectal cancer beside its own program. It reports nothing about Exelixis's NDA or FDA's review; archived because the mint-day EDGAR search names it, and it has no bearing on the proposition.