KEY POINTS
- This retrospective single-institution study from the University of Maryland analyzed 37 patients with bulky or high-grade nonmetastatic soft-tissue sarcoma treated between 2009 and 2023. Patients received proton-based spatially fractionated radiotherapy (SFRT; n=16) or photon SFRT (n=21), followed by conventional fractionated RT and definitive surgery. Median tumor size was 11.8 cm, 75.7% had extremity tumors, and median follow-up was 40 months.
- Both treatment groups received SFRT as an initial high-dose treatment before conventional RT. The institutional approach delivered approximately 12–15 Gy in a single SFRT fraction, followed by conventional RT, most commonly 50.4 Gy. Surgical resection followed approximately 5–7 weeks after conventional RT. Treatment selection was not randomized, and 13 patients received neoadjuvant systemic therapy while nine underwent hyperthermia, introducing important potential confounding.
- Major postoperative wound complications were almost identical between modalities, occurring in 6/16 proton-treated patients (37.5%) and 8/21 photon-treated patients (38.1%). The absolute risk difference was −0.6 percentage points (95% CI −32.2 to 32.0), with RR 0.98 (95% CI 0.42–2.21; p=1.000). The very wide confidence interval prevents conclusions that the two techniques have equivalent wound safety.
- Other treatment-related toxicity rates were not significantly different. Acute grade 2–3 toxicity occurred in 37.5% with protons versus 52.4% with photons (p=0.508), while late toxicity was documented in 18.8% versus 42.9% (p=0.166). Only one patient developed grade 3 acute radiation dermatitis, and no grade 4–5 toxicity was reported. The study was not sufficiently powered to exclude clinically important differences.
- Major pathological response (MPR), defined as at least 80% tumor necrosis, occurred in 12/35 evaluable patients (34.3%; 95% CI 20.8–50.8%). MPR was not significantly associated with SFRT modality (p=0.489). In exploratory survival analysis, patients achieving MPR had longer median disease-free survival (80 versus 27 months; log-rank p=0.014), without a significant overall survival difference. MPR should not be confused with pathological complete response.
- Proton GRID produced greater spatial dose heterogeneity than photon GRID. Among 34 patients with reviewable plans, median D90/D10 was 0.038 with protons versus 0.093 with photons (p=0.04). The median proportion of GTV receiving at least 95% of the SFRT prescription was 1.94% versus 0.24% (p<0.001), consistent with the deliberately spatially heterogeneous approach. However, greater dose modulation was not associated with superior pathological response or clinical outcomes in this cohort.
- Oncological outcomes remained heterogeneous and were not clearly related to treatment modality. Local recurrence occurred in 7/37 patients (18.9%), while 17 (45.9%) experienced local or distant progression. Negative surgical margins were achieved in 32 patients (86.5%), and positive margins were strongly associated with local recurrence (p=0.002). Interpretation is limited by the small cohort, varying systemic treatments, predominantly GRID rather than lattice techniques, and nonrandom treatment allocation.
CLINICAL TAKEAWAY
Preoperative proton-based SFRT followed by conventional radiotherapy and surgery did not demonstrate higher wound complication rates than photon-based SFRT in this small retrospective sarcoma cohort, despite greater spatial dose heterogeneity. The findings provide preliminary clinical reassurance about feasibility but cannot establish comparative safety or superiority, and larger prospective studies are required.