KEY POINTS
- This single-institution retrospective cohort included 232 patients with advanced differentiated thyroid cancer treated between 2004 and 2020. 133 received radical neck IMRT/VMAT, while 99 comparable high-risk patients did not receive EBRT; indications included T4 disease, residual or unresectable disease and repeated neck recurrence poorly controlled with radioactive iodine.
- The RT cohort actually carried several adverse baseline features, including older age and a higher proportion of men. Median follow-up from diagnosis was 84 months.
- Locoregional relapse-free survival was significantly better with EBRT: HR 0.41 (95% CI 0.22–0.74; p=0.0028). Median time to first locoregional recurrence or death was 6.6 years with RT versus 3.2 years without RT.
- The association persisted after adjustment for potential confounders, with multivariable analysis showing approximately a 61% lower hazard of locoregional relapse or death (HR 0.39) in the RT cohort. No significant overall survival advantage was observed.
- Results were consistent among pT3–T4 patients, where EBRT remained associated with better locoregional relapse-free survival (HR 0.41, 95% CI 0.21–0.79; p=0.0062).
- Conformal planning did not appear to create a major geographic-miss problem. Among 18 evaluable locoregional failures after completed RT, only one was clearly marginal; most failures represented progression of previously treated gross disease rather than recurrence just outside the field.
- Late toxicity remained clinically relevant: 12/133 patients (9.0%) required esophageal dilatation, 5/133 (3.8%) developed worsening dysphonia, and one patient with a previously repaired postoperative fistula developed fatal tracheoesophageal refistulization after RT. Acute toxicity could not be reliably graded retrospectively.
CLINICAL TAKEAWAY
These data strengthen the rationale for modern neck EBRT in carefully selected patients with high-risk, residual, unresectable or locally invasive differentiated thyroid cancer, where locoregional failure can be clinically devastating. The magnitude of association is notable, but treatment selection was nonrandomized and the cohorts were imbalanced, so the study cannot establish a causal benefit or define exactly which resected patients should receive EBRT.