Proton therapy does not show lower brain necrosis risk than IMRT in nasopharyngeal cancer

In 256 nasopharyngeal cancer patients, brain necrosis occurred in 6.3% after proton therapy versus 5.1% after photon IMRT, without a significant difference.

KEY POINTS

  • This retrospective cohort study from Memorial Sloan Kettering Cancer Center included 256 patients with nasopharyngeal carcinoma treated with definitive RT between June 2013 and March 2025. 79 patients (30.9%) received proton therapy and 177 (69.1%) photon IMRT. Median follow-up was substantially shorter with protons (43 months) than photons (87 months), an important consideration when comparing delayed brain necrosis.
  • Patients received curative-intent treatment with a median prescribed dose of approximately 70 Gy, usually combined with systemic therapy. Proton treatment included earlier uniform-scanning techniques and subsequent pencil beam scanning/IMPT. Although T-stage distributions were not statistically different, T3–T4 disease was numerically more frequent in the proton cohort, while induction chemotherapy was administered to 38% of proton versus 14% of photon patients (p<0.0001), indicating potentially meaningful treatment-selection differences.
  • Crude brain necrosis incidence was 6.3% with protons (5/79) versus 5.1% with photons (9/177; p=0.7674). When death without brain necrosis was treated as a competing event, the two-year cumulative incidence was 6.2% versus 1.8%, and the five-year incidence was 9.7% versus 5.2%, respectively (Gray's p=0.2342). The numerical excess with protons was not statistically significant and cannot be interpreted as either proven increased risk or equivalence.
  • Clinically significant grade ≥2 brain necrosis occurred in three proton-treated and five photon-treated patients. Two-year cumulative incidence was 4.7% with protons versus 1.2% with photons, while five-year incidence was 4.7% versus 2.5% (p=0.3132). One photon-treated patient experienced grade 3 brain necrosis; no proton-treated patient had a grade 3 event, and no grade 4–5 brain necrosis occurred in either cohort.
  • Advanced T stage was the strongest reported risk factor. In exploratory univariable cause-specific Cox regression, T4 disease was associated with substantially higher brain necrosis risk (HR 8.48, 95% CI 1.83–39.30; p=0.0063) compared with T1 disease. Proton versus photon modality was not significantly associated with risk (HR 1.85, 95% CI 0.60–5.68; p=0.2832). Only 14 total brain necrosis events occurred, so reliable multivariable adjustment was not performed.
  • All 14 brain necrosis events involved the temporal lobe, and dosimetry among affected patients was similar between modalities. Median maximum temporal lobe dose was 72.3 Gy with protons versus 73.9 Gy with photons, while D0.5cc was 70.5 versus 71.4 Gy, without statistically significant differences. Median time to diagnosis was 19 months after proton RT versus 32 months after photon RT (p=0.1601). Estimated five-year OS was 90.0% versus 80.7%, respectively, also without a significant difference (p=0.1573).
  • The study cannot establish comparative equivalence or causal differences between modalities. The proton cohort was smaller, followed for approximately half as long, and included numerically more advanced primary tumors and substantially more induction chemotherapy. The exploratory hazard model was unadjusted, and temporal lobe dose data were most detailed among patients who developed necrosis rather than across the full cohort. No patient-level LET or variable-RBE analysis established a mechanistic explanation for the observed events.

CLINICAL TAKEAWAY

Proton therapy was not associated with a statistically demonstrated reduction in radiation-induced brain necrosis compared with photon IMRT in this retrospective nasopharyngeal cancer cohort, despite its potential normal-tissue dosimetric advantages. Advanced tumor stage and temporal lobe exposure remain important considerations, while prospective comparisons with balanced follow-up and detailed brain dosimetry are required to establish whether proton therapy meaningfully changes late brain toxicity risk.

SOURCE

European Journal of Cancer

Browse more research Suggest a correction