Why this matters
Moderate hypofractionation is well established for photon radiotherapy in localized prostate cancer, but prospective randomized evidence specifically supporting shorter proton therapy schedules has been more limited.
The randomized proton cohort embedded within COMPPARE directly addressed that evidence gap.
The practical question was straightforward: can a 4-week proton course replace nearly 8 weeks of conventionally fractionated proton therapy without worsening long-term bowel function or compromising disease control?
At 3 years, the answer for gastrointestinal outcomes is yes.
Study design
COMPPARE is a multi-institutional prospective study comparing outcomes after proton and photon radiotherapy for prostate cancer.
Within the proton cohort, 600 patients were randomized between:
- 60 Gy(RBE) in 20 fractions
- 78 Gy(RBE) in 39 fractions
Twenty-four patients were screen failures after randomization and another 23 were withdrawn before the 3-year assessment, leaving 553 patients for the current analysis:
- 277 treated with 20 fractions
- 276 treated with 39 fractions
Median follow-up was 4.1 years.
The population was predominantly favorable in terms of baseline disease characteristics:
- 77% had PSA below 10 ng/mL
- 94.5% had Gleason score 6-7
- 39.1% received androgen deprivation therapy
- 90.5% received rectal spacers
The co-primary endpoints were patient-reported moderate or major problems with bowel urgency and bowel frequency at 3 years using EPIC.
The prespecified noninferiority margins were:
- 6 percentage points for bowel urgency
- 4.5 percentage points for bowel frequency
Secondary endpoints included grade 2 or higher gastrointestinal toxicity and freedom from biochemical disease progression.
Key results
Moderately hypofractionated proton therapy met the prespecified criteria for noninferiority for gastrointestinal function at 3 years.
Median EPIC bowel summary scores were:
- 96.4 with 20 fractions
- 94.6 with 39 fractions
Moderate or major bowel urgency was reported by:
- 4.1% with 20 fractions
- 7.9% with 39 fractions
Moderate or major bowel frequency problems occurred in:
- 2.5% with 20 fractions
- 4.6% with 39 fractions
The confidence intervals remained within the prespecified noninferiority margins for both primary endpoints.
Clinician-reported gastrointestinal toxicity was also similar.
Grade 2 or higher GI toxicity occurred in:
- 4.4% with 20 fractions
- 4.8% with 39 fractions
After adjustment for rectal spacer use:
- HR 1.06
- 95% CI 0.49-2.33
Biochemical control was high in both groups.
Freedom from biochemical disease progression was:
- 97.8% with 20 fractions
- 99.25% with 39 fractions
The corresponding HR was 1.76, with a wide 95% CI of 0.52-5.93.
Interpretation
The clinical message is simple: prostate proton therapy does not appear to require 39 fractions to preserve bowel outcomes.
A 20-fraction schedule reduced treatment duration by almost half while meeting the prespecified noninferiority criteria for patient-reported bowel urgency and frequency at 3 years.
The secondary outcomes point in the same direction. Grade 2 or higher GI toxicity was almost identical between groups, and biochemical control remained very high with both schedules.
This is particularly relevant for proton therapy, where prolonged treatment courses can magnify travel, access, and cost burdens for patients referred to specialized centers.
There is also an important nuance in the disease-control result.
The trial did not formally establish noninferiority for biochemical control in this analysis. Freedom from biochemical progression was numerically 97.8% vs 99.25%, and the confidence interval around the hazard ratio was wide because relatively few biochemical events had occurred.
The appropriate conclusion is therefore that no significant difference in biochemical control has emerged so far, not that oncologic equivalence has been definitively established.
The bowel results are more robust because gastrointestinal function was the prespecified primary endpoint.
Limitations
The current analysis focuses on 3-year gastrointestinal outcomes.
Twenty-four randomized patients were screen failures and another 23 were withdrawn before the 3-year assessment, leaving 553 patients in the analysis.
Rectal spacers were used in more than 90% of patients, which may limit generalizability to proton programs where spacers are used less frequently.
The number of biochemical failures was low, resulting in a wide confidence interval around the disease-control estimate.
The current follow-up is also insufficient to address very late endpoints such as second malignancy, which the investigators specifically identify as requiring longer observation.
Finally, this trial compares two proton fractionation schedules. It does not address whether proton therapy itself provides superior outcomes to photon radiotherapy.