Adavosertib

Adavosertib (AZD1775/MK-1775) is an oral WEE1 kinase inhibitor in Phase 2 development for platinum-resistant ovarian, fallopian tube, and primary peritoneal cancers. The drug has a complicated commercial history: originated at Merck as MK-1775, licensed to AstraZeneca as AZD1775, then largely shelved by AstraZeneca around 2021 after mixed efficacy and tolerability data [1]. The active trials, including NCT02101775 and several NCI-MATCH subprotocols, are now run by the National Cancer Institute rather than industry [2][3]. A 2021 Lancet readout showed PFS benefit when combined with gemcitabine (4.6 vs 3.0 months, HR ~0.55) in platinum-resistant ovarian, but no biopharma is currently positioned to file this molecule [1]. The mechanism itself is well-validated and lives on - Zentalis's azenosertib (ZN-c3), which selects patients by Cyclin E1 (CCNE1) positivity rather than running all-comers, carries the commercial torch for WEE1 inhibition into registrational trials [4][11]. Adavosertib at this point is a mechanism-validation story, not a commercial bet.

Phase 2, not novel from a chemistry standpoint - the molecule has been in human trials for over a decade. Originally a Merck oncology asset out-licensed to AstraZeneca; AstraZeneca effectively wound down development around 2021 after the Lancet Phase 2 readout failed to convert into a registrational program, shifting WEE1 focus elsewhere [1]. No FDA designations have been granted for the ovarian indication. Active trials are predominantly academic and NCI-sponsored. NCI runs NCT02101775 (gemcitabine ± adavosertib in platinum-resistant ovarian/peritoneal/fallopian) plus several MATCH subprotocols screening genomically selected populations: BRCA mutants in NCT04439227 (primary completion now December 2026) and SETD2 mutants in NCT03284385 (Phase 2 results published 2024) [2][3][5]. Two recent Phase 2 NSCLC readouts (docetaxel and carboplatin/pemetrexed combinations) were also published in 2025 [6]. No regulatory submission for any indication is on the public timeline. The commercial torch for WEE1 has effectively passed to Zentalis's azenosertib, which has cleaner kinase selectivity, prospectively enriches for CCNE1+ tumors, and is now running both the DENALI Phase 2 (accelerated-approval-supporting) and the ASPENOVA Phase 3 (NCT07546500) [4][11]. Adavosertib's value at this point is mechanistic: proof-of-concept evidence for the target, not a marketable asset.

WEE1 is a brake on cell division. When a cell's DNA is damaged, WEE1 phosphorylates CDK1, the master enzyme that drives entry into mitosis. That phosphorylation pauses the cell at the G2/M checkpoint, giving it time to repair its DNA before splitting [7]. Block WEE1 with adavosertib and that pause disappears. Damaged cells barrel into mitosis with broken chromosomes and die from 'mitotic catastrophe' - the cellular equivalent of a car crash mid-turn, where the cell tries to split with broken chromosomes and the mitotic machinery falls apart before division completes. The case for WEE1 is strongest in tumors that have lost the upstream G1 checkpoint - p53-mutant cancers. Healthy cells can stop at G1 to repair; p53-mutant tumors rely almost entirely on the G2 checkpoint. Knock out that last brake and they're particularly vulnerable. High-grade serous ovarian cancer is roughly 96% p53-mutant, which is why it became the lead indication [8]. The biology is well-validated by human genetics, mouse models, and a decade of clinical signal across multiple tumor types [1][6][8]. The hard part has never been 'does blocking WEE1 kill p53-mutant tumors?' It does. The hard part is doing it without also killing healthy proliferating tissue: bone marrow, gut lining, hair follicles. That tradeoff defines the narrow therapeutic window the entire class is fighting.

NCT02101775 is a randomized Phase 2 of gemcitabine plus adavosertib versus gemcitabine plus placebo in platinum-resistant ovarian, primary peritoneal, and fallopian tube cancers [2]. Patient-level randomization, double-blind, placebo-controlled. Primary endpoint is progression-free survival, a reasonable proxy in this population where overall survival readouts get noisy from post-progression therapies. NCI-sponsored, academic-investigator-led. Population: women whose disease progressed within six months of platinum therapy, a textbook unmet-need group with poor expected survival on standard care. The gemcitabine combination is mechanistically deliberate, not opportunistic: gemcitabine is a nucleoside analog that stalls replication forks and induces replication stress, which forces cells to depend even more on the G2/M checkpoint for survival. Knocking out WEE1 on top of replication-stressed tumors should produce more lethal mitotic entry than either agent alone - a synthetic-lethal logic, not simple additivity. The trial has already read out a positive signal. The published Phase 2 (Lheureux et al., Lancet 2021) randomized 99 patients (94 eligible, 61 to adavosertib + gemcitabine and 33 to gemcitabine + placebo) and showed median PFS of 4.6 versus 3.0 months favoring adavosertib + gemcitabine, HR ~0.55 [1]. That is a real signal but not a registrational one: small N and tolerability was rough, with significant rates of Grade 3+ neutropenia, thrombocytopenia, and nausea. No biomarker selection in the main trial, despite preclinical rationale that p53-mutant and CCNE1-amplified tumors should be the responders. The NCI MATCH subprotocols (Z1I for BRCA, SETD2-mutant cohort) are smaller exploratory baskets without comparator arms, so they generate signal-detection data rather than confirmatory evidence [3][5].

Our model gives this drug a 4% chance of eventually being approved. That is below the historical baseline of about 13% for drugs at this stage, meaning the model sees reasons to be more cautious than average. It reached that figure by weighing ten facts about the trial and sponsor, and the biggest drags were a weak or thin sponsor approval record, limited earlier-phase results, a randomized design, and the use of a comparator or control arm. The remaining facts were close to average and did not move the estimate much in either direction.

Safety has been the consistent problem. The on-target toxicity is mechanism-based - WEE1 inhibition hits proliferating normal cells, so myelosuppression, GI toxicity, and fatigue show up across trials at clinically meaningful rates [1][6]. Combining with DNA-damaging chemotherapy compounds those effects. The narrow therapeutic window is the central efficacy risk: doses high enough for tumor activity often require reductions or schedule changes that erode the signal. Patient selection is the other open question, and arguably the bigger one. The NCT02101775 trial does not enrich for p53 mutation status, CCNE1 amplification, or any other biomarker, despite strong preclinical rationale that these subsets are the true responders. If responders are a subset, an all-comers design dilutes the apparent benefit and makes the molecule look weaker than the mechanism really is. Zentalis built its entire azenosertib program around CCNE1+ selection (~50% of platinum-resistant ovarian) [11], which is probably a larger differentiator than the cleaner kinase profile and may explain why their program is advancing where AstraZeneca's stalled. A dosing-optimization paper in 2026 still finds the field is searching for the right schedule in uterine serous carcinoma a decade after first-in-human [9]. Execution risk is low for trial completion (NCI runs these competently) but high for commercialization - no industrial sponsor is positioned to file. Commercial risk dominates. Even with a positive readout, who runs the Phase 3 and submits the NDA? Adavosertib is a returned asset without a clear owner. Zentalis is competing for the same indication with azenosertib, a molecule they actually want to develop [4]. Payer reception for a marginal-benefit cytotoxic combination in a small population would be cool at best.

Watch the mechanism, not the molecule. WEE1 as a target is real: the genetics, the preclinical data, and the Lheureux 2021 readout all line up [1][10]. But adavosertib itself is a stranded asset. AstraZeneca walked. Merck has not signaled re-engagement. NCI trials will continue to read out, but no industrial sponsor is teed up to convert a positive readout into a filing. The actionable signal here is Zentalis's azenosertib (ZN-c3), which has the same mechanism with better selectivity, prospectively selects CCNE1+ tumors, and has a sponsor committed to taking it through registration [4][11]. The two Zentalis readouts to distinguish: DENALI is the Phase 2 CCNE1+ platinum-resistant ovarian trial with topline expected year-end 2026 and the potential to support accelerated approval; ASPENOVA (NCT07546500) is the confirmatory Phase 3 that enrolled its first patient in May 2026 and is years from a readout, likely 2028 at earliest [11]. The 2026 DENALI readout is the near-term WEE1 inflection - not the Phase 3, and not adavosertib. The remaining adavosertib readouts from NCI are worth tracking for what they say about the class. NCT04439227 (BRCA-selected, primary completion now December 2026) and the already-published NCT03284385 (SETD2-mutant, results in Cancer Research Communications 2024) are the cleanest mechanistic questions [3][5]. As a commercial entity for this compound: pass. As a science readout for the target: worth tracking, because every adavosertib data point recalibrates expectations for azenosertib.