KEY POINTS
- This retrospective cohort included 269 patients with hepatocellular carcinoma treated with proton beam therapy between 2009 and 2015: 105 aged ≥75 years and 164 aged <75 years. Median follow-up among survivors was 37 months.
- PBT fractionation was anatomy-adapted: common schedules included 66 Gy(RBE)/10 fractions for peripheral tumors, 72.6 Gy(RBE)/22 fractions for hilar lesions, and 74 Gy(RBE)/37 fractions for tumors near the gastrointestinal tract. All patients completed their planned treatment.
- Three-year overall survival was 65.9% in patients ≥75 versus 56.9% in younger patients (p=0.20). Three-year local control was 86.7% vs 90.5% (p=0.24) and PFS 30.3% vs 29.9% (p=0.81).
- After adjustment for liver function, tumor size, portal vein tumor thrombosis, performance status and other covariates, age ≥75 remained unrelated to mortality (HR 1.25, 95% CI 0.89–1.74; p=0.198) or PFS (HR 1.15, 95% CI 0.86–1.54; p=0.331).
- Comorbidity burden was considerably more informative. Compared with CCI 0, CCI ≥3 was associated with more than threefold higher mortality (HR 3.31, 95% CI 1.54–7.15; p=0.002). Its association with PFS was weaker and narrowly missed significance (HR 1.87; p=0.057).
- Formal interaction testing did not show that the impact of comorbidity differed significantly between younger and older patients: age × CCI interaction was p=0.094 for OS and p=0.903 for PFS. Treating age continuously produced the same overall message.
- Treatment was well tolerated across age groups. No acute grade ≥3 toxicity occurred. Grade 1–2 dermatitis was more common in older patients (42.9% vs 26.8%), while only two late grade ≥3 events occurred, both in younger patients: one gastrointestinal bleed and one dermatitis.
CLINICAL TAKEAWAY
Being 75 or older was not, by itself, associated with worse local control, progression, survival, or severe toxicity after proton therapy for HCC in this selected cohort. For treatment selection and counseling, comorbidity burden may therefore be more informative than chronological age alone, although these retrospective data do not prove that age can be ignored in broader unselected populations.