Minimization of Geometric Uncertainties and Setup Errors Improves Tumor Control for Intracranial Stereotactic...

To assess the lesion-level effect of 1-2-mm geometric uncertainty on tumor control probability (TCP) in brain-metastasis stereotactic radiosurgery (SRS), assuming a uniform 2-mm planning target volume margin.

We retrospectively analyzed 75 brain metastases in 50 patients who underwent linear accelerator-based SRS. Geometric uncertainty was simulated by applying positional shifts to the treatment plan, and the resulting reduction in TCP was quantified. The resulting dosimetric and radiobiological characteristics were compared between tumors ≤20 mm and those >20 mm.

Tumor size modulated the impact of geometric uncertainty. Small lesions (≤20 mm) exhibited steeper dose gradients (gradient index, 3.6 vs 2.6; P = .001) and therefore larger reductions in gross tumor volume D100% under a 2-mm shift (median, 4.5 Gy vs 3.3 Gy; P < .001). However, because their initial TCP was high (median 96.6%), the TCP decrement remained modest (median 3.4%). In contrast, large lesions (>20 mm) began with a lower TCP (median 82.1%) and lay on the steep portion of the dose-response curve, so similar or smaller dose losses produced a substantially greater TCP decline (median 10.9% at 2 mm; P < .001). The same pattern was observed with 1-mm shifts (ΔTCP 3.5% vs 1.0%).

Patients with larger tumors (>20 mm) may be most vulnerable to a significant loss of tumor control from geometric uncertainties in SRS. This analysis shows promise for personalizing treatment delivery, supporting a shift from applying fixed planning target volume margins toward strategies such as fractionated radiation therapy tailored to individual tumor size and radiobiological characteristics. This hypothesis-generating study requires external validation, with fractionated stereotactic radiation therapy being a subject for future investigation.