KEY POINTS
- This retrospective Shanghai Proton and Heavy Ion Center study included 76 patients with recurrent nasopharyngeal carcinoma previously irradiated with photons and salvaged with intensity-modulated carbon-ion reirradiation between 2015 and 2019. All received a standardized 63 Gy(RBE) in 21 fractions.
- At median follow-up of 38.3 months, 26 patients (34.2%) developed local recurrence. Two-year overall survival was 71.7%, local control 73.5%, regional control 92.1% and distant control 92.6%.
- RBE-weighted target-dose metrics did not significantly distinguish controlled from recurrent tumors. In contrast, multiple dose-averaged LET metrics were higher in tumors remaining locally controlled, suggesting that apparently equivalent physical/RBE-weighted plans may deliver biologically different high-LET distributions.
- In patients with sufficiently long follow-up, the proportion of GTV receiving at least 50 keV/μm was almost twice as high with local control: median 57.5% versus 31.75% (p=0.013). Similar significant differences extended across approximately the 49–63 keV/μm range.
- GTV LETd99—the near-minimum LET received by the tumor—was the most robust parameter. After multivariable adjustment, each increase in LETd99 was associated with lower recurrence risk (HR 0.91, 95% CI 0.84–0.99; p=0.023) and remained significant in both conventional Cox and competing-risk analyses.
- Restricted-spline analysis identified 39.4 keV/μm as a potential threshold. Patients with GTV LETd99 ≥39.4 keV/μm had 94.4% two-year local control versus 65.3% below the threshold (p=0.0031); predictive performance peaked at 24 months with AUC 0.77.
- Tumor volume was a major determinant of intratumoral LET distribution, showing correlations of approximately r=−0.16 to −0.67 across LET metrics. Wider beam angular spans preferentially increased coverage with moderately high LET around 50–70 keV/μm, suggesting one potential route toward prospective LET-guided planning. These observations remain retrospective and post hoc.
CLINICAL TAKEAWAY
This study strengthens the argument that carbon-ion planning may need to optimize where the LET lands, rather than relying exclusively on RBE-weighted dose. A near-minimum tumor LET around 39–40 keV/μm is an intriguing candidate planning target, but it should not be treated as a validated constraint until prospectively tested.
SOURCE
International Journal of Radiation Oncology, Biology, Physics