onc brain

The longer read

This paper fills a defined gap. The mobile-spine consensus contours, Cox et al. for intact disease and Redmond et al. for the postoperative spine, left the sacrum out, and no sacrum-specific pattern-of-failure analysis exists. What it supplies is a vocabulary, the eight-segment sacral classification system, plus CTV rules derived from expert contours. It does not supply outcome data. Read it as a standardization instrument: its value lies less in whether each rule is optimal than in making sacral SBRT volumes reportable in a shared language, so that future failure analyses can be pooled across centres.

The agreement numbers carry a predictable shape. GTV agreement (kappa 0.73) exceeded CTV agreement (kappa 0.59), as expected when one volume is defined by imaging and the other encodes judgment about microscopic spread. The informative part is where agreement broke down. Cases 2 and 4, with kappas of 0.41 and 0.48, were ala lesions encroaching on the vertebral body and posterior elements. That is exactly where the compartmental approach pushes the CTV toward nerve roots and bowel, and it is where three experts preferred a 0 to 5 mm isotropic margin, citing nerve root and bowel toxicity and sacral insufficiency fracture. So the consensus is firmest for central vertebral body lesions, where the decision is easy, and weakest at the lateral and posterior boundary, where it matters.

The acceptance vote should also be weighed carefully. Consensus in 9 of 10 cases sounds decisive, but the panel voted on a contour derived from its own submissions. That shows the panel is internally coherent. It does not show that the contours are correct. Highly significant kappa p-values only establish that agreement beat chance, a low bar for nine experienced academic spine SBRT clinicians. The panel was also narrow: all academic, with dedicated spine SBRT programmes, and the 10 cases were selected for representativeness rather than accrued consecutively. The core anatomic argument, that the fused sacrum lacks the disc barrier thought to limit spread in the mobile spine, is plausible, but the authors call the supporting literature sparse. The S1 ala exception and the refusal to use an involved ala's ossification line as a boundary are reasoning from anatomy, not from failure data.

Scope is the biggest constraint on transfer. The rules cover bone-confined disease only, excluding epidural extension, paraspinal disease, re-irradiation and the postoperative sacrum. Yet the authors' own review of a sacral SBRT series describes bulkier disease at presentation, with progression in the vertebral body together with paraspinal tissue, followed by isolated epidural progression, as the dominant failure patterns. That series reported cumulative LC of 86.5% at 1 year and 78.7% at 2 years. The scenarios this guideline excludes may therefore account for much of the sacral failure burden, and the margins for them remain individual judgment.

The rules would be wrong if sacral failures turn out to be mainly epidural or paraspinal rather than marrow-contiguous. In that case, compartmental bony expansion would add plexopathy and fracture risk without improving local control. The classification system is what makes that question testable. The practical read is to adopt the nomenclature for reporting, use the CTV rules for bone-confined disease, and treat the ala and posterior-element boundary as unresolved.