onc brain

The longer read

The value of HEAT is not that it shows five-fraction prostate SBRT works. That case was made by HYPO-RT-PC and PACE-B, and five-fraction treatment is already routine in many departments. What HEAT does is remove two objections that survived those trials, and both of them are objections a radiation oncologist actually hits in clinic.

The first is the comparator. HYPO-RT-PC's control arm used limited IMRT, and PACE-B and NRG-GU005 carried heterogeneous control arms, which is the presenters' own framing of the gap. If your alternative to five fractions is 70.2 Gy in 26 fractions delivered with IMRT to every patient, as it is here, the relevant question is whether SBRT matches a modern moderately hypofractionated schedule rather than whether it matches something closer to historical practice. HEAT is designed around that comparison specifically, which makes its answer more directly transferable than a trial whose control arm no one still uses.

The second is ADT. The prior randomized trials in this space did not permit androgen deprivation, which left a real population unrepresented: the intermediate-risk man for whom short-course ADT is on the table. HEAT permitted up to six months in both arms, and 28% of the cohort received it. That is not a large enough subgroup to answer whether SBRT and ADT interact, but it does mean the trial is not silent on the population where the question comes up.

What should temper confidence is the arithmetic of the interim read. With 142 patients analyzed against a planned 420 evaluable, and failure rates of 7% and 7.4%, the trial is excluding a 12% absolute difference using very few events. A non-inferiority margin that wide is defensible for a surrogate endpoint where the absolute event rate is low and the two arms are both expected to perform well, but it means the result rules out a fairly coarse degree of inferiority rather than establishing equivalence. The formal p-value of 0.007 reflects the design working as specified, not a tight estimate of the true difference. The prespecified final analysis is what should move practice, and this interim should be read as consistent with expectation rather than as settling anything.

The toxicity read carries a caveat that is easy to miss and worth carrying into patient conversation. Acute G2+ GI toxicity was lower with the five-fraction arm, which is the opposite of the intuition that a larger fraction size should cost more acutely, and is plausible given the shorter overall treatment time and smaller integrated dose to rectum. But the trial scored use of supportive medication such as laxatives or psyllium as grade 2, which raises event counts in both arms and makes the comparison sensitive to how liberally supportive medication was offered. Late GI and both acute and late GU were comparable, which is the more consequential finding: the concern with 7.25 Gy fractions has always been late effects, and at a median follow-up of 59.7 months there is no signal of a late penalty.