KEY POINTS
- This prospective, single-centre observational study collected 640 skin samples from 86 patients receiving adjuvant three-dimensional conformal radiotherapy after breast-conserving surgery.
- Whole-breast treatment delivered 45 Gy in 18 fractions. Fifty patients received an additional boost of 10–12.5 Gy in four or five fractions, producing total doses of 55–57.5 Gy.
- Maximum acute radiation dermatitis was grade 0–1 in 25 patients (29%) and grade 2–3 in 61 (71%), including 49 grade 2 and 12 grade 3 cases. No grade 4 dermatitis occurred.
- At mid-treatment, the Clostridiaceae family was found in 30% of samples from patients who developed moderate-to-severe dermatitis and in none from the mild group (p = 0.0020). Anaerococcus US436 was similarly restricted to the moderate-to-severe group at 21.7%.
- Across all treated samples, Clostridiaceae, Clostridium, Anaerococcus and Anaerococcus US436 remained enriched in the moderate-to-severe group. Granulicatella elegans was also enriched in this group, whereas Bacillus thuringiensis was associated with mild dermatitis.
- Presence of at least two among Clostridium, Anaerococcus US436 and Granulicatella elegans was associated with moderate-to-severe dermatitis (odds ratio 11.88; 95% confidence interval, 2.57–54.86). Adding absence of Bacillus thuringiensis increased the odds ratio to 24.8, but with a very wide confidence interval of 3.15–195.05.
- The treated and contralateral breasts had different microbiome profiles even before radiotherapy, and both changed during treatment. Results used a false-discovery-rate threshold of 0.25, were not adjusted for clinical confounders and were not externally validated.
CLINICAL TAKEAWAY
The skin microbiome may eventually help identify patients at increased risk of clinically significant radiodermatitis or provide a target for preventive intervention. These signatures are not ready for clinical testing: the analysis was exploratory, the dermatitis rate was unusually high, effect estimates were imprecise and causality cannot be inferred from bacterial association.