Dr. L. Michael Glodé
L. Michael Glodé, MD, FACP, FASCO, of the University of Colorado, Denver, presented results of the SWOG S9921 study, which he described as originating in quite a different era than exists today. At the time of its design in 1999, certain preoperative characteristics including a PSA of 20 ng/mL or above, among others, defined a high-risk group of patients with a 50% chance of biochemical relapse at 5 years. Mitoxantrone was among the only chemotherapy options at the time, also differentiating this trial from the current landscape.
The study included patients with localized disease considered “operable for cure.” If patients had a bone scan (required for PSA of 20 ng/mL and above), it had to be negative. All patients underwent a radical prostatectomy within 120 days prior to registration in the trial. Of 983 patients who were randomly selected to a study arm, 22 were ineligible, leaving 481 patients who received ADT alone (bicalutamide and goserelin) and 480 who received ADT along with MP. The median age was 60 years, and most patients were white. The trial accrued more slowly than anticipated, and accrual was stopped early due to detection of three cases of leukemia in the chemotherapy group. Ultimately six patients developed leukemia, often within 2 to 3 years after mitoxantrone exposure.
The primary endpoints were overall and disease-free survival. In both cases, the addition of MP to ADT did not change the outcome. For disease-free survival (DFS), the 10-year estimate for both groups was 72%, and the hazard ratio was 1.01 (95% CI [0.80, 1.27]; p = 0.94). In the ADT arm, there were 85 recurrences (57% of total DFS events) and 64 deaths without recurrence (43%); with the addition of MP, there were 91 recurrences (62%) and 56 deaths without recurrence (38%).
Overall survival was also similar between the groups, with a 10-year rate of 87% with ADT and 86% with the addition of MP (HR 1.06, 95% CI [0.79, 1.43]; p = 0.70). The cause of death with ADT alone was prostate cancer in 18%, other cancer in 18%, other causes in 28%, unknown cause but not prostate cancer in 22%, and unknown cause in 14%. The biggest difference in the patients receiving MP was in those who died due to other cancers, accounting for 36% of the deaths in that group; six patients in the MP group developed leukemia, compared with none in the ADT-alone group.
The difference in deaths due to other cancers was statistically significant (p = 0.011). Other common secondary cancers that developed included lung, pancreas, and gastrointestinal malignancies. The remainder of deaths in the MP group were due to prostate cancer in 22%, other causes in 16%, unknown but not prostate cancer in 14%, and unknown causes in 12%.
Other adverse events of grade 3 or higher were also more frequent with the addition of MP. There were a total of 47 grade 4 adverse events with MP, compared with only five without it; for grade 3 events, these totals were 210 and 136, respectively (among adverse events with at least one grade 4 reported in the study). Leukopenia was common with MP, with 39 grade 4 events, compared to none without it.
Dr. Glodé said that though survival was greater than anticipated in both study arms, the MP combination’s increased risk for leukemia and other cancers is notable. “There was no evidence that this form of chemotherapy improved prostate cancer survival or freedom from progression,” he said.
Susan F. Slovin, MD, PhD, of Memorial Sloan Kettering Cancer Center, the discussant for the abstract, said that this study adds to a list of adjuvant chemotherapy trials that had trouble accruing and, thus, remain underpowered for firm conclusions. “They keep getting cut right in the middle, because if nothing is seen early on, they get shut down,” she said.
To improve that situation, Dr. Slovin said a new grading system beyond the Gleason score would go a long way toward patient selection and stratification. Genomic profiling is becoming more relevant and potentially useful, as are the use of novel directed imaging tracers such as fluciclovine. “Are these trials feasible?” she asked. “The answer, I think, is yes. Are there enrollment challenges? Maybe. But using other methods of interrogation will be much more helpful.”
– David Levitan