KEY POINTS
- This retrospective two-center study evaluated 108 patients with retinoblastoma treated with radiotherapy between 1979 and 2025, including 42 receiving photon RT and 66 receiving proton RT. Median age at diagnosis was 0.7 years, median age at irradiation was 1.1 years, and 80 patients (74%) had bilateral retinoblastoma, used as a clinical surrogate for hereditary disease when genetic testing was unavailable.
- Follow-up was substantially longer after photon therapy: median 28.7 years versus 18.2 years after proton therapy, with 20.8 years for the overall cohort. Secondary malignant neoplasms (SMNs) were classified as occurring within the radiation field if they overlapped tissue receiving at least 50% of the prescription dose. The analysis treated death as a competing risk when estimating cumulative SMN incidence.
- The primary clinically important finding was a significant difference in 15-year cumulative incidence of in-field SMNs: 11.1% after photon RT versus 0% after proton RT (p=0.005). Across the complete follow-up period, seven photon-treated patients developed in-field secondary cancers, whereas no in-field SMNs were observed among the 66 proton-treated patients. The association was also observed in analyses restricted to patients with bilateral disease.
- When all secondary malignancies were considered, including those outside the irradiated field, the 15-year cumulative incidence was 11.1% after photon therapy versus 2.0% after proton therapy. This difference did not reach conventional statistical significance (p=0.055). Overall, 7/42 photon-treated patients (16.7%) and 2/66 proton-treated patients (3.0%) developed an SMN during available follow-up.
- All seven in-field SMNs after photon therapy occurred outside the clinical target volume, but within the defined radiation field. Three additional photon-associated in-field SMNs emerged more than 20 years after RT, emphasizing the importance of prolonged surveillance. The two SMNs observed after proton therapy were both out-of-field, in patients with a family history of retinoblastoma, suggesting a possible contribution from hereditary cancer predisposition rather than high-dose local irradiation.
- Overall survival remained high in both groups. Estimated 15-year OS was 94.4% after photon therapy and 97.9% after proton therapy (p=0.224), with no statistically significant survival difference. The study therefore provides evidence of an association with lower in-field second cancer incidence, not proof of improved overall survival.
- Important baseline imbalances limit causal interpretation. Proton-treated patients were more likely to have received chemotherapy before RT (62% vs 12%), to have been older than one year at irradiation (68% vs 36%), and to have experienced disease progression before RT (26% vs 2%). Germline mutation data were unavailable for most patients, modern treatment selection differed from historical practice, and the small number of second cancers prevented reliable multivariable adjustment.
CLINICAL TAKEAWAY
Among children with retinoblastoma who require radiotherapy, this long-term comparison supports considering proton therapy when feasible because of its association with substantially fewer in-field secondary malignancies. However, the retrospective treatment-era differences, unequal follow-up and incomplete genetic characterization prevent a definitive causal comparison, while continued lifelong surveillance remains important regardless of radiation modality.