onc brain

The longer read

The question F-SHARP answers well is whether a re-irradiated prostate tolerates focal dose escalation with HDR. The answer is clearly yes in this cohort: no acute grade 3 events, late grade 3 toxicity in 2 pts (3.2%), and QoL changes on EPIC-26 and IPSS that never crossed the minimal clinically important difference. For a radiation oncologist weighing salvage options in intraprostatic radiorecurrence, toxicity has historically been the reason to hesitate over re-treating the whole gland, and a focal target is the obvious lever to reduce it. This trial gives prospective, multi-institutional support for pulling that lever.

The efficacy read is less settled, and the design explains why. Disease control endpoints were secondary in a single-arm study powered around toxicity, so the 3-year rates (bPFS 62.7%, LRFS 73.0%, MFS 85.9%) describe this cohort rather than estimate an effect against anything. The central tension of focal salvage is that treating only the visible recurrence leaves untreated gland at risk, and an LRFS of 73.0% at 3 years is the figure a skeptic would point to. Without a whole-gland arm, it is not possible to say whether the local failures reflect out-of-field disease that whole-gland salvage would have covered or biology that no local therapy would have controlled. RTOG 0526 tested whole-gland salvage LDR in a similar post-EBRT setting, and the natural comparison between focal and whole-gland approaches remains unanswered by a randomized trial.

The fractionation finding has more practical weight than its framing suggests. Twenty-four pts received a single fraction and 38 received two, with no difference across endpoints (all P > .05). This was not a randomized comparison and is far too small to establish equivalence, but it shows that a single-implant schedule did not produce an obvious efficacy or toxicity penalty. For programs where a second implant is a real burden on pts and on the brachytherapy suite, that is a reason to consider the shorter schedule in a protocol or registry context.

The genomic classifier result is the part most likely to be overread. Higher scores tracked with worse bPFS, rPFS, LRFS and MFS among the 31 pts with results, and all 4 pts with a score above 0.85 recurred biochemically, locally and distantly. The direction is biologically coherent: a high-score recurrence after primary RT is a radioresistant, metastasis-prone phenotype, and a focal boost is unlikely to change its course. But the threshold is defined in four people within half the cohort, and the authors appropriately call GC promising and in need of further study. Read it as a hypothesis about who should be offered intensification (systemic therapy, wider salvage fields, or both) rather than a cutpoint ready for clinic.

What would make the reassuring read wrong? Selection is the main threat: pts referred for focal salvage HDR at experienced centers likely had imaging-defined, limited recurrences and good baseline function, which flatters both toxicity and control. Longer follow-up beyond the median 39.4 months also matters, since late GU toxicity after re-irradiation can emerge years out. The next step the data point toward is a trial that randomizes focal against whole-gland salvage, or stratifies by GC score to test intensification in the high-risk tail.