The Radiation Oncology Feed For radiation oncologists · physicists · RTTs · dosimetrists
Worth Attention prostate Protons

Patient-informed design supported rapid and representative COMPPARE trial accrual

COMPPARE enrolled 2,524 patients from 51 institutions in 52 months, with Black participation reaching 16% of the cohort.

RadOncToday · ·1 min read

KEY POINTS

  • COMPPARE was designed after patient engagement showed that men prioritized survival, gastrointestinal quality of life, and disease control. Only 15% were expected to accept randomization between protons and photons, whereas 36% were expected to accept randomization between two proton fractionation schedules.
  • The study therefore used parallel pragmatic proton and IMRT cohorts rather than mandatory modality randomization. A randomized comparison of standard and moderately hypofractionated proton therapy was nested within the proton cohort.
  • COMPPARE enrolled 2,524 patients from 51 institutions over 52 months, despite major delays from site activation and four COVID-19 surges. Site activation required a median of 18 months, rather than the planned six months, and ranged from 5 to 33 months to first consent.
  • Among patients documented as eligible but not enrolled, only 1,308 of 5,382 (24%) declined participation. Many others selected a different treatment or received care outside a participating institution.
  • The nested proton fractionation trial enrolled 600 randomized patients. Among patients enrolled while both pragmatic and randomized options were available, the randomization rate was 35%, closely matching the patient-engagement prediction of 36%.
  • Institutional preference remained a major obstacle: nine of 30 proton centres declined to offer the randomized fractionation comparison despite participating in the broader prospective study.
  • Black patients accounted for 404 participants, or 16% of the entire cohort, exceeding the prespecified 10% goal; representation was 21% in the IMRT cohort and 13% in the proton cohort. Hispanic participation remained low at 5%.
  • There was no comparator trial designed without patient engagement. The analysis therefore supports feasibility and predictive value but cannot isolate engagement from pragmatic design, institutional enthusiasm, recruitment resources, or other factors.

CLINICAL TAKEAWAY

Patient input before protocol finalization can identify which comparisons people are genuinely willing to accept and can prevent an elegant but non-accruing trial. COMPPARE also shows that patient-centred design cannot eliminate slow contracting, institutional treatment bias, or unequal geographic access to proton facilities.

SOURCE

International Journal of Radiation Oncology, Biology, Physics