Maya Chen: I have been sitting with this one since I read it, and I genuinely couldn't decide if it was a good news story or a complicated one.
Dr. Nathan Hayes: Mm. Which part specifically?
Maya Chen: You know what caught me about this? The same drug failed in one type of prostate cancer and just got FDA approval for a different type, and it's the exact same molecule. What does that actually mean for how we think about treating cancer at all?
Dr. Nathan Hayes: Right, and that's — that question is exactly the right one, because what we know is that the drug, capivasertib, targets a kinase called AKT, and AKT is only doing the thing you need to block it from doing if the upstream regulator, PTEN, is already gone.
Maya Chen: So it's not that the drug changed — it's that the patient's genetic situation is what determines whether the drug has anything to actually hit.
Dr. Nathan Hayes: Exactly — and importantly, PTEN deficiency in metastatic prostate cancer isn't rare, it's present in somewhere around forty to fifty percent of cases, so this approval carves out a genuinely large subset of patients who previously had no targeted option at this stage of the disease.
Dr. Nathan Hayes: AstraZeneca ran two trials for the same drug in prostate cancer. Same molecule — capivasertib, an AKT inhibitor. CAPItello-280 got discontinued in April 2025 after the interim analysis. Futility. Then CAPItello-281, different population, the FDA approves it June 12, 2026. Same drug. Opposite outcomes.
Maya Chen: That's — wait, the same drug that just got approved also failed a trial the year before?
Dr. Nathan Hayes: Correct. And the difference isn't dose or formulation. The difference is who got it. CAPItello-280 gave capivasertib to a broader mCRPC population alongside docetaxel — patients weren't screened for any particular genetic profile. CAPItello-281 enrolled only patients with PTEN deficiency. Specifically, tumors where 90% or more of viable malignant cells showed no PTEN staining on immunohistochemistry. That cut-off matters enormously.
Maya Chen: So the drug isn't just broadly useful — it only works if the tumor has lost this specific gene.
Dr. Nathan Hayes: That's the mechanism. PTEN is a tumor suppressor — it normally puts the brakes on the PI3K/AKT signaling pathway. When PTEN is lost, AKT becomes constitutively active. It's a parallel growth signal, completely independent of androgen signaling. You can block the androgen receptor completely with abiraterone and ADT, but if AKT is running in the background, the tumor still has fuel. Capivasertib blocks all three AKT isoforms — AKT1, 2, and 3. You're closing the escape route.
Maya Chen: Okay, but — and I want to stay here for a second — about 25% of men with this disease have PTEN deficiency. So for three out of four, this approval changes nothing.
Dr. Nathan Hayes: That's accurate. The majority of mAPMN/S prostate cancer patients — what we used to call hormone-sensitive prostate cancer — do not have PTEN loss. For them, this drug is molecularly irrelevant. Which is actually why CAPItello-280 failed. You gave an AKT inhibitor to a population where most tumors don't have hyperactive AKT. Of course it cratered.
Maya Chen: Hmm. So it's not that the drug is weak — it's that it was aimed at the wrong people.
Dr. Nathan Hayes: Exactly. And CAPItello-281 was designed from the start with prospective biomarker selection. 1,012 patients, all PTEN-deficient confirmed by central testing. The primary data were presented at ESMO 2025 in Berlin. 19% reduction in risk of radiographic disease progression or death versus placebo plus the same abiraterone and ADT backbone.
Maya Chen: Here's what I keep circling — 19% sounds significant, but I'm not sure what it feels like for someone sitting across from their oncologist. Like, what does 19% actually buy you?
Dr. Nathan Hayes: It's a fair push, and I'll be precise. 19% reduction in the hazard ratio for radiographic progression or death — that's a relative risk reduction, and it is statistically significant. But the primary endpoint of CAPItello-281 is rPFS — how long before the cancer visibly progresses on imaging. Overall survival is a key secondary endpoint. And that OS data is not yet mature. So we cannot say this regimen extends how long patients live. Not yet.
Maya Chen: Wait — so the FDA approved it, but we don't have survival data as the primary result?
Dr. Nathan Hayes: That's how oncology approvals often work — rPFS is a validated surrogate endpoint. The FDA accepted it. But anyone framing this as a proven survival benefit right now is outrunning the evidence. The benefit is real. It's measured. It's also bounded.
Maya Chen: And on top of that — Daniel George at Duke presented tolerability data at the 2026 ASCO GU Cancers Symposium and diarrhea came up as a notable adverse event. For someone newly diagnosed, possibly feeling fine, adding a drug that causes significant diarrhea — that's a real conversation.
Dr. Nathan Hayes: It is. And it's worth naming directly: the approved regimen is capivasertib plus abiraterone plus prednisone plus ADT. Four agents. In a newly diagnosed population. The benefit-risk conversation is not trivial.
Maya Chen: Though — I want to hold onto the other side of this. For that one-in-four man who learns at diagnosis that his tumor has lost PTEN, that there's now a treatment matched to that specific biology — that's not nothing. That's the first time that's ever been true in prostate cancer.
Dr. Nathan Hayes: It is the first and only targeted therapy for PTEN-deficient mAPMN/S prostate cancer. I'm not minimizing that. Capivasertib was already approved for PTEN-altered metastatic breast cancer — this is its second tumor type. The precision oncology infrastructure transferred. But precision oncology is also splintering — HRR-mutated patients are getting PARP inhibitors, PSMA-positive patients are getting radioligand therapy. Each group has its own biomarker test, its own eligibility criteria, its own approval timeline. And the 75% without a matched target are still in conventional care.
Maya Chen: And that's sort of the uncomfortable part, right — because the patients who didn't respond in the unselected trial, they weren't wrong to try. The trial was wrong about who to ask.
Dr. Nathan Hayes: Correct. And if PTEN deficiency has gradations — partial loss, monoallelic versus biallelic, loss by mutation versus by methylation — then the next trial will have to draw a new line, and some patients who were inside today's approval will end up outside the next one. That's not a failure of the science. That's what refinement actually looks like in real time.
Maya Chen: So who decides where the line is — the tumor, or the test we built to read it?
