onc brain

The longer read

The value of this series is less the headline DFS than how it is distributed. A 5-year DFS of 93.36% across a mixed cohort of BCC and SCC, primary and recurrent, definitive and adjuvant, sits in the range most single-institution brachytherapy reports describe, and the authors place it alongside the Leipzig experience reported by Kohler and the mold technique reported by Guix. That is reassuring but not discriminating; nearly every retrospective NMSC RT series reports control in the 90s, because lesion selection does much of the work. The more useful signal is that recurrent disease separated clearly from primary (5-year DFS 85.9% vs 95.6%, p=0.003), while applicator type and definitive versus adjuvant intent did not separate meaningfully. For a radiation oncologist, that points to disease history, rather than device, as the variable worth weighing when offering surface brachytherapy as an alternative to re-excision or Mohs.

The applicator comparison is where a reader should be most cautious. Molds were used for larger lesions (mean 19.38 vs 13.07 mm) and for raised or curved surfaces, so any difference in DFS or toxicity between molds and Leipzig cones is confounded by size and anatomy. The authors attempted to address this by restricting to lesions within the Leipzig range, but the toxicity association with molds stayed borderline (p=0.067), and lesion size itself carried a significant odds ratio for acute toxicity. The 5-year DFS of 88% for molds versus 94.9% for Leipzig is better read as a size and site gradient than as evidence that one device controls disease less well. Compounding this, the paper's own accounting is not internally consistent: six recurrences are described as five Leipzig plus three mold, and the acute toxicity figure in the abstract does not match the results section. These are small errors, but in a series this size they are the kind that make subgroup p-values fragile.

The acute toxicity profile deserves more attention than the conclusion gives it. Roughly two in five lesions developed RTOG grade 3 acute reactions at 3 Gy per fraction three times weekly, and definitive treatment carried higher odds than adjuvant. Late outcomes were favourable, with grade 0-1 late toxicity in 95.2% and good or excellent cosmesis in 82%, which is the trade a surface brachytherapy patient is accepting: a brisk acute reaction that largely resolves. Follow-up of 33 months, though, is short for both late cosmesis and late BCC recurrence, so the late picture is incomplete.

What would make this wrong? If the high control reflects selection of small, favourable primaries, the result does not transfer to the recurrent or larger lesions where an RT alternative to surgery is most often requested, and the data in hand already hint that this is the case. The series supports HDR plesiotherapy as a reasonable definitive option for small primary NMSC in older pts, and gives a concrete reason to counsel recurrent-lesion pts on a lower control expectation. It does not establish a preferred fractionation, does not compare against excision or Mohs, and does not show that molds are inferior once size is accounted for.