The addition of capivasertib (Truqap) to abiraterone acetate (Zytiga) and androgen deprivation therapy (ADT) continued to show an overall survival (OS) advantage vs placebo plus abiraterone and ADT with longer follow-up in patients with PTEN-deficient de novo metastatic hormone-sensitive prostate cancer (mHSPC), according to updated data from the phase 3 CAPItello-281 trial (NCT04493853) presented at the
In the updated analysis, which was first presented at the 2026 ESMO Congress, the median OS was 57.4 months (95% CI, 47.0-not calculable [NC]) with capivasertib plus abiraterone (n = 507) vs 46.5 months (95% CI, 40.7-53.2) with placebo plus abiraterone (n = 505), a difference of 10.9 months (HR, 0.82; 95% CI, 0.68-0.99; P = .044).2 The descriptive analysis did not meet the threshold for statistical significance. At the March 9, 2026, data cutoff, which provided approximately 17 months of additional follow-up, OS maturity was 42%, with 195 deaths in the capivasertib arm and 226 deaths in the placebo arm.
The updated median radiographic progression-free survival (rPFS) was 35.5 months (95% CI, 29.7-40.5) with capivasertib vs 26.0 months (95% CI, 22.4-31.3) with placebo (HR, 0.83; 95% CI, 0.70-0.99; nominal P = .033). In an exploratory analysis using increasingly stringent PTEN-deficiency cutoffs, the OS HR was 0.76 (95% CI, 0.61-0.93) among patients with at least 95% PTEN loss (capivasertib, n = 404; placebo, n = 410), 0.75 (95% CI, 0.56-1.02) among those with at least 99% loss (capivasertib, n = 205; placebo, n = 196), and 0.74 (95% CI, 0.53-1.02) among those with 100% loss (capivasertib, n = 169; placebo, n = 162).
Reviewing the updated safety data, Daniel George, MD, noted that the on-target adverse effects (AEs) of capivasertib tended to emerge during the first months of treatment and did not accumulate with longer exposure.
"Most of that toxicity is happening early, and the patients are getting it under control and managing it for the most part," said George, the Eleanor Easley Distinguished Professor of Medicine, Surgery, and Urology at Duke University School of Medicine and co-chair of the Duke Cancer Institute Center for Prostate and Urologic Cancers in Durham, North Carolina.
The double-blind trial randomized 1012 patients with de novo mHSPC whose tumors were PTEN deficient, defined as at least 90% of viable malignant cells with no specific cytoplasmic PTEN staining by immunohistochemistry. Among approximately 6200 patients who submitted tumor tissue, 25% were PTEN deficient.2
Patients were randomly assigned 1:1 to receive capivasertib at 400 mg twice daily on a 4-days-on, 3-days-off schedule or placebo, each combined with abiraterone at 1000 mg plus prednisone or prednisolone at 5 mg once daily and ADT. Randomization was stratified by disease volume and visceral metastases and by geography. The primary end point was investigator-assessed rPFS; secondary end points included OS, symptomatic skeletal event-free survival, time to castration resistance, and time to prostate-specific antigen (PSA) progression.
At the primary data cutoff of October 7, 2024, with a median follow-up of approximately 18 months, the median rPFS was 33.2 months (95% CI, 25.8-44.2) with capivasertib vs 25.7 months (95% CI, 22.0-29.9) with placebo (HR, 0.81; 95% CI, 0.66-0.98; P = .034).3 OS was immature at 26% maturity (HR, 0.90; 95% CI, 0.71-1.15; P = .401).
The median time to castration resistance, a composite of radiographic progression, PSA progression, and symptomatic skeletal events, was 29.5 months (95% CI, 22.1-36.8) with capivasertib vs 22.0 months (95% CI, 18.2-24.9) with placebo (HR, 0.77; 95% CI, 0.63-0.94; nominal P = .009). Time to PSA progression yielded an HR of 0.73 (95% CI, 0.52-1.01; nominal P = .060), with 60 and 82 PSA progression events in the respective arms. George noted that many castration-resistance events in this population occurred without a preceding PSA rise, underscoring the need for imaging and symptom monitoring in addition to PSA testing.1
In patient-reported outcome analyses using the Functional Assessment of Cancer Therapy–Prostate (FACT-P), the median time to a clinically meaningful decline in physical well-being was 7.2 months (95% CI, 4.5-10.0) with capivasertib vs 14.6 months (95% CI, 8.1-NC) with placebo (HR, 1.43; 95% CI, 1.15-1.78). Time to clinically meaningful decline in functional well-being (HR, 1.06; 95% CI, 0.86-1.32) and FACT-P total score (HR, 1.10; 95% CI, 0.89-1.37) was similar between arms.
At the updated data cutoff, the median symptomatic skeletal event-free survival was 43.8 months (95% CI, 39.1-53.1) with capivasertib vs 35.9 months (95% CI, 32.1-42.4) with placebo (HR, 0.83; 95% CI, 0.69-0.99; nominal P = .044).2
Among patients who received capivasertib (n = 503), any-grade diarrhea occurred in 52.7%, hyperglycemia in 40.4%, and rash in 36.4%; the corresponding grade 3 or higher rates were 6.4%, 11.3%, and 12.7%. Diabetic ketoacidosis occurred in 6 patients (1.2%) in the capivasertib arm and 1 patient (0.2%) in the placebo arm, with no new cases in the capivasertib arm since the primary analysis. In the primary safety analysis, grade 3 or higher AEs occurred in 67.0% of patients in the capivasertib arm vs 40.4% of patients in the placebo arm (n = 503 each), and AEs led to discontinuation of capivasertib or placebo in 18.3% vs 4.8%, dose interruption in 62.8% vs 26.8%, and dose reduction in 29.0% vs 3.6%, respectively.3
George described a grade-based approach to managing these AEs, including holding capivasertib for grade 2 or higher diarrhea or rash until recovery to grade 1 or lower, using antihistamines and oral corticosteroids for rash, and starting metformin for early hyperglycemia, with dose reduction for recurrent toxicity.1
In June 2026, the FDA approved capivasertib plus abiraterone and prednisone for adult patients with PTEN-deficient metastatic androgen pathway modulation–naive or –sensitive prostate cancer, previously referred to as mHSPC, as detected by an FDA-authorized test, based on the rPFS data from CAPItello-281.4 The final OS analysis is planned after 522 deaths, corresponding to approximately 52% maturity.1
