HDR skin brachytherapy achieves 97% two-year local control with favorable early cosmesis

In 174 head and neck skin lesions, HDR brachytherapy achieved 97.3% two-year local control, with slightly worse early cosmesis after eight fractions.

KEY POINTS

  • This retrospective single-institution study from UCLA evaluated 136 patients with 174 head and neck nonmelanoma skin cancer lesions treated with definitive high-dose-rate brachytherapy between 2013 and 2024. Median age was 77 years, and 98.3% of lesions were T1, with a median diameter of 0.7 cm. Basal cell carcinoma accounted for 74.1%, cutaneous squamous cell carcinoma for 19.5%, and the nose was the most frequent anatomical site (51.7%).
  • Two definitive HDR brachytherapy regimens were used: 40 Gy in eight fractions, delivered two to three times weekly, and 48 Gy in 16 fractions, generally delivered five times weekly. Treatment employed iridium-192 Leipzig applicators (46.6%) or custom surface molds (53.4%), selected according to lesion geometry and depth. The target included visible disease with a 5–10 mm margin, and dose was prescribed to a depth of 3 mm. The eight-fraction regimen was used in 43.7% of lesions.
  • Local tumor control was high, with Kaplan–Meier estimates of 97.3% at two years (95% CI 94.3–100%) and 90.4% at five years (95% CI 83.1–98.4%). Nine local failures were recorded, including two cases with equivocal pathology conservatively classified as recurrence. No regional or distant failures were documented. However, median follow-up was only 21.1 months, and 30 lesions remained at risk at five years, limiting the precision of long-term estimates.
  • Acute radiation dermatitis was predominantly mild. Among 167 lesions with evaluable toxicity data, grade 1 dermatitis occurred in 77.2%, grade 2 in 12.6%, and no grade ≥3 acute dermatitis was reported. Grade 2 dermatitis was more frequent with custom molds than Leipzig applicators (20.2% vs 3.8%; p=0.01), although custom molds were preferentially used for larger and more complex lesions. Acute toxicity did not differ significantly between fractionation schedules (p=0.99).
  • Cosmetic outcomes were assessed independently by six blinded reviewers from dermatology, facial plastic surgery, and radiation oncology using paired photographs from 124 lesions in 95 patients. The validated Global Aesthetic Improvement Scale (GAIS) was compared with a newly developed five-point Overall Cosmesis Score (OCS), where lower scores indicated better results. Both scales produced a central lesion-level score of approximately 2.14, consistent with generally favorable early cosmetic outcomes, and showed strong agreement (R²=0.99; p<0.001).
  • The new OCS demonstrated higher interobserver reliability than GAIS, with weighted Fleiss' κ of 0.66 versus 0.52 and Krippendorff's α of 0.72 versus 0.57. The eight-fraction regimen was associated with slightly worse cosmetic ratings than 16 fractions (median OCS 2.45 vs 2.00; p=0.01), with an adjusted mean difference of 0.62 (p=0.03). However, shorter treatment was preferentially selected for older patients, larger lesions, and less cosmetically sensitive locations, so the association cannot establish a causal fractionation effect.
  • Timing strongly influenced cosmetic assessment. Post-treatment photographs were obtained a median of 49 days after completion, with 70.2% acquired within 90 days. Lesions photographed within 1–21 days had substantially worse scores (mean OCS 3.64) than those assessed later, consistent with acute treatment effects. These were cross-sectional comparisons rather than serial assessments of the same lesions. Long-term cosmesis, late fibrosis, pigment changes, telangiectasia, and patient-reported satisfaction were not systematically evaluated, and the OCS requires external validation.

CLINICAL TAKEAWAY

HDR surface brachytherapy achieved high local control and generally favorable early cosmetic outcomes in selected patients with small head and neck nonmelanoma skin cancers who were unsuitable for or declined surgery. Eight-fraction treatment was associated with modestly worse early cosmetic scores than 16 fractions, but nonrandom treatment selection and limited late cosmetic follow-up prevent conclusions about the optimal fractionation schedule.

SOURCE

Clinical and Translational Radiation Oncology

Browse more research Suggest a correction