Oncology · Medical Literature
Overall Survival of Patients With PSA-Recurrent Prostate Cancer: Long-Term Analysis of a Randomized Phase 2 Trial of pTVG-HP DNA Vaccine Versus Placebo.
Abstract
We previously reported results from a randomized phase 2 trial evaluating a DNA vaccine (pTVG-HP) encoding prostatic acid phosphatase (PAP) versus placebo (GM-CSF vaccine adjuvant alone) in patients with PSA-recurrent, nonmetastatic, noncastrate prostate cancer (NCT01341652). The primary endpoint of 2-year metastasis-free survival (MFS) was not different between the treatment arms. This trial was reopened for long-term evaluation. The median follow-up for long-term overall survival was 4.1 years (range 0.1-13.8+ years). Median overall survival of 48 patients treated with pTVG-HP was 13.4 years versus 8.6 years for 49 patients treated with placebo (hazard ratio [HR] = 0.42, 95% confidence interval [CI], 0.20-0.90). Long-term follow-up information on survival and subsequent treatments was available for 59 patients treated at one site. Of these, 56 were treated with subsequent androgen deprivation therapy (ADT). When analyzing overall survival for this cohort, the adjusted HR was 0.60 (95% CI, 0.36-1.00) when comparing pTVG-HP versus placebo, after adjusting for baseline characteristics and for ADT and subsequent treatment exposure. The median time to beginning ADT was 2.2 years for patients randomized to pTVG-HP versus 1.4 years for patients randomized to placebo. The median time to the next therapy for castration-resistant disease after beginning ADT was 4.7 years for patients randomized to pTVG-HP versus 2.1 years for patients randomized to placebo. These results suggest that pTVG-HP may have had a single-agent benefit that could not be appreciated using an MFS endpoint. This is consistent with results from other trials of anticancer vaccines, which suggest that these may have more impact on longer-term endpoints such as survival.
Clinical relevance
In patients with PSA-recurrent, nonmetastatic, noncastrate prostate cancer who were originally enrolled in a randomized phase 2 trial of pTVG-HP DNA vaccine versus placebo, does long-term follow-up reveal a difference in overall survival between treatment arms? __SNAPSHOT__ {"studyDesign":"Long-term follow-up analysis of a randomized, placebo-controlled phase 2 trial (NCT01341652); overall survival was a secondary/exploratory endpoint evaluated after trial reopening for extended follow-up","setting":"Single academic institution (University of Wisconsin, for the 59-patient cohort with complete long-term data); broader multi-site enrollment for the full 97-patient randomized cohort; United States","sampleSize":"n=97 randomized (pTVG-HP: n=48; placebo/GM-CSF: n=49); long-term survival and treatment data available for n=59 (single-site cohort)","population":"Men with PSA-recurrent, nonmetastatic (M0), noncastrate (castration-sensitive) prostate cancer; PSA doubling time <12 months per original trial entry criteria","intervention":"pTVG-HP — a plasmid DNA vaccine encoding prostatic acid phosphatase (PAP), co-administered intradermally with GM-CSF adjuvant at 14-day intervals for six doses, then quarterly for up to 2 years","comparator":"Placebo: GM-CSF (200 µg) vaccine adjuvant alone, administered on the same schedule","primaryOutcome":"The original trial primary endpoint was 2-year metastasis-free survival (MFS), which was not different between arms. This long-term analysis reports overall survival (OS) as the key endpoint; p-value for OS not reported in the abstract.","followUp":"Median long-term follow-up for OS: 4.1 years (range: 0.1–13.8+ years)","fundingSource":"Not reported in the abstract; prior phase 2 publication (PMID 31644357) acknowledged NIH/NCI support — applicability to this follow-up analysis is unconfirmed"}
Key findings
In the full randomized cohort (n=97), median overall survival was 13.4 years in the pTVG-HP arm (n=48) versus 8.6 years in the placebo arm (n=49), yielding an unadjusted hazard ratio (HR) of 0.42 (95% CI: 0.20–0.90), favoring the vaccine. The absolute difference in median OS was 4.8 years (calculated). A p-value for this primary OS comparison is not reported in the abstract. In the single-site cohort with complete long-term data (n=59), of whom 56 received subsequent androgen deprivation therapy (ADT), the adjusted HR for OS (pTVG-HP vs. placebo), after adjustment for baseline characteristics and ADT/subsequent treatment exposure, was 0.60 (95% CI: 0.36–1.00). The median time to initiation of ADT was 2.2 years in the vaccine arm; the corresponding figure for the placebo arm is not reported in the provided abstract text (the abstract appears truncated). NNT cannot be calculated from available data. Adverse event data from this long-term follow-up are not reported in the abstract. The original phase 2 trial found no statistically significant difference in the primary endpoint of 2-year metastasis-free survival (41.8% vaccine vs. 42.3% placebo; P=0.97).
Practice implications
This long-term reopened follow-up of a randomized phase 2 trial provides hypothesis-generating evidence that pTVG-HP DNA vaccination may be associated with improved overall survival in men with PSA-recurrent, nonmetastatic, castration-sensitive prostate cancer. The unadjusted HR of 0.42 (95% CI: 0.20–0.90) is statistically notable but must be interpreted with significant caution: (1) OS was not the prespecified primary endpoint of the original trial; (2) the original primary endpoint (2-year MFS) was negative; (3) the sample size is small (n=97 full cohort, n=59 with complete long-term data), substantially limiting statistical power and precision; (4) this is a phase 2 trial, not powered or designed to detect OS differences; (5) the confidence interval for the adjusted HR (0.60; 95% CI: 0.36–1.00) includes 1.00, indicating that after adjusting for baseline factors and subsequent treatments (including ADT), the survival advantage does not reach conventional statistical significance; (6) subsequent treatment heterogeneity — including variable timing and type of ADT — may have confounded OS comparisons despite adjustment in the single-site subcohort; and (7) only the abstract was available for this review, limiting full methodological assessment. These findings do NOT establish that pTVG-HP improves OS and should not change current clinical practice. Patients enrolled had PSA-recurrent, nonmetastatic, castration-sensitive prostate cancer with PSA doubling time <12 months; results are not applicable to metastatic or castration-resistant disease. The potential signal of delayed ADT initiation (median 2.2 years in the vaccine arm) is of biological interest but remains unconfirmed. __LIMITATIONS__ Overall survival was not the prespecified primary endpoint; this is a post-hoc long-term analysis of a trial that was negative on its original primary endpoint (2-year MFS).,Very small sample size (n=97 total; n=59 with complete long-term data) with wide confidence intervals, inadequate to reliably detect or quantify OS differences.,Phase 2 design — not powered for OS; risk of type I error is substantially elevated for secondary/exploratory analyses.,Adjusted HR confidence interval (0.36–1.00) crosses 1.00, meaning the adjusted OS benefit is not statistically significant at the 0.05 threshold.,No p-value for the OS comparison is reported in the abstract, limiting statistical interpretation.,Confounding by subsequent treatments (ADT timing, type, and duration) is a major concern; adjustment was only partially feasible and confined to the single-site subcohort (n=59).,Potential for selection bias: the 59-patient single-site cohort may not be representative of all 97 randomized patients.,Abstract appears truncated — complete data for the placebo arm's median time to ADT and additional efficacy/safety metrics are unavailable for this review.,Long follow-up duration (up to 13.8 years) introduces risk of informative censoring and loss of treatment allocation balance over time.,Only abstract available for review — full methodological details, covariate adjustment strategy, Kaplan-Meier curves, and safety reporting cannot be assessed.,Generalizability is limited: modern PSA-recurrent prostate cancer management now includes PSMA-PET staging, salvage radiotherapy intensification, and novel hormonal agents not available at trial initiation.,Funding source not confirmed for this follow-up analysis, precluding assessment of potential industry influence. __BOTTOM_LINE_CATEGORY__ promising_but_preliminary __BOTTOM_LINE__ This long-term follow-up of a randomized phase 2 trial reports a numerically large but statistically uncertain OS benefit with pTVG-HP DNA vaccine (median OS 13.4 vs. 8.6 years; unadjusted HR 0.42, 95% CI 0.20–0.90) in PSA-recurrent, nonmetastatic, castration-sensitive prostate cancer, but the original primary endpoint was negative and the adjusted OS benefit did not reach significance. These results are hypothesis-generating only and do not constitute sufficient evidence to adopt this vaccine into clinical practice. Confirmatory randomized trials with OS as a prespecified primary endpoint and adequate statistical power are required before any practice change can be considered.