CHICAGO—Patients with newly diagnosed hormone-sensitive metastatic prostate cancer (mPC) who had prostate radiotherapy before their androgen deprivation therapy (ADT) lived longer than those treated with standard ADT alone in a retrospective analysis—the largest single-center experience to date of primary tumor-directed radiotherapy in mPC—reported at the 2018 ASCO Annual Meeting (Abstract 5017).
“In this single-institution retrospective cohort—after adjusting for the baseline demographics—there was an association between receipt of prostate radiotherapy and improved overall survival,” said lead author Scott Carlyle Morgan, MD MSC, a radiation oncologist at the Ottawa Hospital Cancer Centre, University of Ottawa, Canada.
He said the findings were “consistent with the hypothesis” that radiotherapy directed to the primary tumor may have altered the natural history of hormone-sensitive mPC. “But this is by no means definitive. And we await results of the randomized studies.”
The researchers reviewed data from their own institution in the light of recent registry analyses that had suggested external-beam radiotherapy directed at the prostate might have improved overall survival (OS).
The Canadian analysis was from a total of 304 men with newly diagnosed mPC who had been referred between 2005 and 2015 for initial treatment with ADT. Patients were included in the analysis if they had been given prostate radiotherapy at biologically effective doses of at least the equivalent of a course of 40 Gy in 15 fractions, or if they had no prostate radiotherapy. By comparing these two groups—with allowance for confounding variables—it was possible to assess the association of prostate radiotherapy and OS.
Morgan said the two groups of patients (those who received and those who did not receive radiotherapy) were “substantially different” since the decision to include prostate radiotherapy was taken on clinical grounds. “Principally, patients who were offered prostate radiotherapy had bulky primary tumors. And the aim of radiotherapy was to prevent or delay local disease progression at the time of castration resistance.”
It was possible, statistically, to allow for the differences between the two groups and assess the impact of radiation as an independent factor on outcome, he said.
About a third of the patients (105) had prostate radiotherapy. The median age at diagnosis for the whole cohort was 75 years and median follow-up was 72.2 months. On univariate analysis prostate RT was associated significantly with improved OS with a hazard ratio (HR) of 0.62.
In patients treated with radiotherapy, the 2-year survival was 74.7 percent, and 5-year OS was 41.8 percent. Among those who had systemic treatment alone, the figures were lower: 56.9 percent survived 2 years and 27.6 percent 5 years. Median OS in patients receiving radiotherapy was 48.3 months compared with 29.2 months for those not treated with radiotherapy.
In a multivariate model—taking account of age at diagnosis, year of diagnosis, presenting prostate specific antigen (PSA), tumor stage, nodal status, and metastasis subdivision—radiotherapy was still associated with improved survival (HR=0.64) but with borderline statistical significance.
“There's no doubt that there was a certain amount of selection of patients in this cohort for prostate radiotherapy,” said Morgan. And he noted that patients having radiotherapy had, on average, more favorable characteristics.
“We attempted so far as possible to adjust for all the factors that we could adjust for—like age at presentation, presenting PSA, biopsy Gleason score, T stage, N stage, and M1 category.” But he said they were not able to adjust for some other variables, such as performance status and the volume of metastatic disease.
“We also did analyses to look at patients who had survived at least 1 year and at least 2 years,” said Morgan. This was to remove selection bias by excluding patients who were less likely to have received radiation therapy because of their disease status.
“The hazard ratio for overall survival was attenuated in patients who survived more than 1 year and more than 2 years,” he said. “The hazard ratio was 0.78 in at least 1-year survivors and 0.73 in at least 2-year survivors. And it was no longer significant. So, it does suggest that there was a degree of selection bias operating.”
Morgan said the 19-month difference in median OS between the two groups was clinically significant, but caution was needed to distinguish genuine treatment effect. So, he regarded the findings as “hypothesis generating” and said they were awaiting results from the large-scale randomized trials under way, in particular, from a section of the soon-to-be released Stampede randomized trial that was looking at “exactly this question.”
Peter M. Goodwin is a contributing writer.