10 Cancer Drugs Died Targeting the PI3K Pathway. Drug 11 Just Got FDA Approval.

Count the dead: buparlisib, pictilisib, taselisib, BEZ235, apitolisib. Five pan-PI3K or dual PI3K/mTOR inhibitors that entered late-stage solid tumor trials and came back in body bags, killed by liver toxicity, dose-limiting side effects, or flat-out inferiority to existing options. Novartis tried it, Genentech tried it, Roche tried it, and all three walked away empty-handed.

Then came the hematologic chapter: four PI3K inhibitors won FDA approval for blood cancers between 2014 and 2021, and all four subsequently had their indications withdrawn after showing the same lethal pattern. Idelalisib, duvelisib, umbralisib, copanlisib. Each drug shrank tumors and delayed progression, then patients started dying faster than those who never took the medication at all. Richard Pazdur, director of the FDA Oncology Center of Excellence, called the findings across six randomized trials “unprecedented in oncology”: drugs that improved progression-free survival while worsening overall survival. No other drug class in oncology has ever produced that dissociation.

On July 14, Celcuity Inc. of Minneapolis announced that the FDA had approved REVTORPYK, its brand name for gedatolisib, a drug originally discovered at Pfizer as PKI-587. It is the first compound ever to simultaneously inhibit all four class I PI3K isoforms (α, β, δ, γ) and both mTOR complexes (mTORC1 and mTORC2) and survive the regulatory gauntlet. Its approved indication covers HR-positive, HER2-negative metastatic breast cancer in patients whose tumors do not carry a PIK3CA mutation, the so-called wild-type majority representing approximately 60% of this cancer subtype.

That last detail matters more than the mechanism, and here is why.

The 60% Nobody Could Treat

Breast cancer is not one disease, and within the enormous HR-positive, HER2-negative subtype that accounts for about 70% of all breast cancer cases and makes this the largest single category in the field, approximately 40% of patients carry a mutation in the PIK3CA gene while the other 60% do not.

For the 40% with the mutation, two drugs already exist: Novartis won FDA approval for alpelisib (Piqray) in 2019, and Roche followed with inavolisib (Itovebi) in 2024, both isoform-selective PI3Kα inhibitors designed specifically for tumors carrying PIK3CA mutations.

For the 60% without the mutation, there was nothing: no PI3K-targeting therapy, no mTOR-targeting combination that had cleared a Phase 3 trial for this population, and no targeted option at all when these patients progressed past first-line CDK4/6 inhibitor therapy, leaving fulvestrant alone as the best available alternative, an estrogen receptor antagonist that works by blocking hormone signaling rather than by attacking the PI3K pathway directly. VIKTORIA-1 measured exactly how well that option performs in this setting: median progression-free survival of 2.0 months, with an overall response rate of 1%, meaning one patient in a hundred showed measurable tumor shrinkage of 30% or more while the median patient progressed within roughly eight weeks.

VIKTORIA-1: The Numbers

Celcuity’s Phase 3 VIKTORIA-1 trial tested gedatolisib in two configurations against fulvestrant alone, both in PIK3CA wild-type patients who had progressed on prior CDK4/6 inhibitor therapy. Results were stark.

Measure	Triplet	Doublet	Fulvestrant Alone
Regimen	Gedatolisib + palbociclib + fulvestrant	Gedatolisib + fulvestrant	Fulvestrant
Median PFS	9.3 months	7.4 months	2.0 months
Hazard ratio	0.24 (p<0.0001)	0.33 (p<0.0001)	—
Risk reduction	76%	67%	—
ORR	32%	28%	1%
Median DOR	17.5 months	12.0 months	N/A

A hazard ratio of 0.24 means patients on the triplet regimen had a 76% lower risk of disease progression or death compared to fulvestrant alone, while the response rate jumped from 1% to 32%, with those responses lasting a median of 17.5 months. In a disease setting where the control arm progressed in eight weeks, responses lasting nearly a year and a half represent a different category of clinical benefit.

Side effects are real and worth attention. Stomatitis (mouth sores) occurred in 72% of triplet patients, with 22% at Grade 3 severity requiring prophylactic steroid-containing mouthwash for management, while hyperglycemia hit 46% of the triplet arm and rash affected 30%. None of these are trivial burdens, and they will shape real-world adherence, but they are also manageable toxicities in a population whose alternative was watching their cancer progress in two months.

Counting the Patient-Years

How large is the population that REVTORPYK just opened up? Start with the American Cancer Society’s 2024 projection of roughly 310,000 new breast cancer cases per year in the United States, of which approximately 70% are HR-positive, HER2-negative, yielding about 217,000 new cases of this subtype annually. An estimated 150,000 women are living with metastatic breast cancer at any given time, and approximately 73% have the HR-positive, HER2-negative subtype based on SEER data, producing a prevalent population of roughly 110,000 women with HR-positive, HER2-negative metastatic disease. Apply the 60% PIK3CA wild-type frequency: that puts the target population at roughly 66,000 women.

Not all 66,000 are eligible for REVTORPYK immediately, because the approval covers second-line-and-beyond treatment, meaning patients must have already progressed on CDK4/6 inhibitor therapy. A reasonable estimate of the annual cohort entering this treatment line falls between 25,000 and 30,000 patients.

Now apply the VIKTORIA-1 progression-free survival gains: at 25,000 patients per year, each gaining a median of 7.3 months of progression-free survival (the triplet’s 9.3 months minus the control’s 2.0), that is 182,500 additional patient-months without disease progression per annual cohort, or roughly 15,200 patient-years. Because the median is not the mean, individual gains vary, but this estimate is likely conservative since mean progression-free survival typically exceeds the median for right-skewed survival distributions. Before Monday, these patients had a median of 2.0 months before their cancer advanced, so every one of those 15,200 patient-years represents time that simply did not exist in the treatment landscape.

Why Gedatolisib Survived the Graveyard

If every prior pan-PI3K inhibitor failed, what makes gedatolisib different? Two factors stand out, and both involve how the drug is administered rather than what it targets.

First, the dosing schedule mattered enormously: most failed PI3K inhibitors were daily oral pills, flooding the body with continuous PI3K suppression that drove cumulative toxicity, but gedatolisib is administered intravenously once per week in the triplet regimen, giving normal cells time to recover between exposures and appearing to widen the therapeutic window between efficacy and toxicity.

Second, the dual PI3K/mTOR mechanism matters because prior pan-PI3K inhibitors blocked PI3K alone, and cancer cells responded by rerouting signaling through the mTOR pathway, a well-documented resistance mechanism that gedatolisib circumvents by hitting both PI3K and mTOR simultaneously, closing the escape route. BEZ235, Novartis’s earlier dual PI3K/mTOR inhibitor, attempted the same strategy but was an oral compound with continuous dosing, and dose-limiting toxicities terminated its development in Phase 1, while gedatolisib’s intermittent IV administration appears to solve the tolerability problem that killed BEZ235.

The Novartis Irony

While Celcuity was running VIKTORIA-1, Novartis was trying to replace its own PI3K inhibitor, because Piqray (alpelisib), approved in 2019 for PIK3CA-mutant patients, has been declining with $382 million in 2025 revenue, down 15% year-over-year for the second consecutive year of contraction. In January 2026, Novartis paid $2 billion upfront, with up to $3 billion in milestones, to acquire SNV4818 from Sun Pharma as a next-generation PI3Kα replacement for Piqray.

Novartis spent $2 billion to cover the 40% of patients who carry PIK3CA mutations, and Celcuity just captured the other 60%, the larger population, with a drug the market barely saw coming from a company whose pre-approval market capitalization was roughly $4 billion. Structural irony: the pharma giant that invested most heavily in the PI3K pathway over the past decade does not own the largest addressable patient population within it.

The Ghost That Haunts Every PI3K Drug

REVTORPYK’s PFS data are striking, perhaps the strongest PFS signal in the PI3K class to date, but the question that should keep every oncologist cautious is not about progression-free survival but about overall survival, and this is not academic caution.

Four hematologic PI3K inhibitors lost their FDA approvals not because they failed to shrink tumors but because, when mature survival data arrived, patients on PI3K inhibitors were dying faster than those on comparator arms despite effective tumor control and measurable delays in progression. Six randomized trials showed this same pattern, and no other drug class in oncology history has produced such a consistent dissociation between PFS benefit and OS harm.

VIKTORIA-1’s overall survival data are not yet mature, with Celcuity expecting OS results in 2027 or 2028. Until that readout arrives, the most important question remains unanswered: does REVTORPYK break the PI3K curse, or has it merely delayed the moment when the curse becomes visible? Open-label trial design (patients and physicians knew who received which treatment) adds another layer of uncertainty, as open-label PFS assessments are susceptible to bias in timing of imaging and progression calls.

There are reasons for cautious optimism, starting with the fact that the hazard ratio of 0.24 is far stronger than anything the withdrawn hematologic inhibitors showed for PFS, and the intermittent dosing schedule produces a different toxicity profile than the daily oral drugs that worsened survival. Different mechanism, different dosing, different disease setting. But “this time is different” is a phrase that has preceded most of oncology’s most expensive disappointments.

Limitations

Overall survival data from VIKTORIA-1 are immature, and given the PI3K class history of PFS-OS dissociation, this is not a routine caveat but the central open question surrounding the drug’s long-term value. Patient population estimates combine incidence data from ACS projections, metastatic breast cancer prevalence estimates, SEER subtype distributions, and PIK3CA wild-type frequency from Anderson et al. (2020) rather than a single integrated epidemiologic dataset, so the 25,000 annual eligible patient figure and the 66,000 prevalent population figure are approximations, not registry counts. Celcuity has not disclosed REVTORPYK pricing, making cost-effectiveness analysis impossible at this stage. VIKTORIA-1 was an open-label trial, a design that meta-analyses have shown can inflate PFS measurements by 20–40% in some settings. A stomatitis rate of 72% in the triplet arm may reduce real-world adherence below trial conditions, where patients receive intensive prophylaxis and monitoring that community oncology practices may not replicate.

What You Can Do

For patients with HR-positive, HER2-negative metastatic breast cancer who have progressed on CDK4/6 inhibitor therapy, the immediate action is confirming PIK3CA mutation status with your oncologist, since REVTORPYK’s approval specifically covers wild-type patients who lack that mutation, with commercial launch expected in late Q3 2026. PIK3CA mutation carriers should continue discussing alpelisib and inavolisib as their established targeted options.

Investors watching the PI3K space should track two catalysts with particular attention. First, Celcuity plans to file a supplemental NDA in Q3 2026 to extend REVTORPYK’s approval to PIK3CA-mutant patients, based on the mutant cohort of VIKTORIA-1, which if approved would make gedatolisib the first PI3K inhibitor indicated across both mutant and wild-type populations, covering 100% of HR-positive, HER2-negative metastatic disease rather than 40%. Second, VIKTORIA-2 is an ongoing Phase 3 trial testing gedatolisib in first-line treatment, which would move it earlier in the treatment sequence and substantially expand the eligible population.

Oncologists and researchers should follow the VIKTORIA-1 overall survival data as they mature over the next 12 to 24 months, because if overall survival confirms the PFS signal, REVTORPYK will have done something no PI3K drug has ever done: proved that inhibiting this pathway extends life and does not just delay progression. If OS disappoints, the PI3K graveyard will have claimed its most promising member yet. Either way, the answer matters.

The Bottom Line

For twenty years, the PI3K pathway has been the most important target in cancer that nobody could hit without causing more harm than benefit, a graveyard of drugs where ten compounds died trying and four more won approvals only to lose them after patients started dying faster on treatment than off it. FDA’s own oncology director called that pattern unprecedented. Then a company with a $4 billion market cap, working with a drug that Pfizer originally discovered and abandoned, ran a trial showing a 76% reduction in disease progression for 66,000 patients who had no targeted therapy at all. Whether gedatolisib truly breaks the PI3K curse depends on survival data that will not arrive for another year or two. But for patients whose previous best option offered two months before their cancer advanced, 9.3 months of progression-free survival on the triplet regimen is not a marginal improvement but the difference between having a treatment and having none.