KEY POINTS
- Investigators retrospectively analyzed 101 patients treated with single-fraction, monoisocentric LINAC SRS for 1–22 brain metastases between 2022 and 2023. Median age was approximately 66 years, NSCLC was the most common primary (41%), prescriptions ranged from 15–21 Gy, and a 1-mm PTV margin was routinely used.
- The proposed memory circuit extended beyond the hippocampi to include the bilateral hippocampi, bilateral amygdalae, fornix and corpus callosum. None of these structures had been explicitly spared during the original clinical optimization.
- Dose increased substantially as intracranial disease burden increased. Patients with ≥10 metastases received memory-circuit Dmean 2.9 versus 0.41 Gy, Dmedian 2.9 versus 0.34 Gy, and Dmax 9.0 versus 1.9 Gy compared with patients with one lesion; all comparisons were p<0.0001.
- The 13 patients with ≥10 lesions were then replanned using Sparing Memory with Advanced Radiosurgical Targeting, or SMART. Memory-circuit Dmean fell from 2.9 to 1.8 Gy (p=0.0001), Dmedian from 2.9 to 1.8 Gy (p=0.0002), Dmax from 9.0 to 4.7 Gy (p=0.0007) and D100 from 1.5 to 0.98 Gy (p=0.001).
- The reduction did not come at the expense of measured target coverage. Median PTV D95 remained 19 versus 19 Gy (p=0.77), D98 18 versus 18 Gy (p=0.76) and Dmean 21 versus 21 Gy (p=0.74), with >95% prescription coverage retained.
- SMART also reduced some low-dose exposure: median memory-circuit V5Gy declined from 0.80 to 0.0048 cc (p=0.037) and V8Gy from 0.18 to 0.11 cc (p=0.0078). Maximum doses to conventional organs at risk such as the optic apparatus and brainstem did not significantly worsen.
- The critical unanswered question is clinical benefit. This was a 13-patient retrospective replanning exercise, patients were not prospectively treated according to a validated memory-circuit constraint, and no neurocognitive testing demonstrated that reducing dose to the amygdalae, fornix or corpus callosum improves memory after SRS.
CLINICAL TAKEAWAY
Treating many metastases with a single-isocenter SRS plan increases incidental dose to neural structures involved in memory, and that dose appears modifiable without sacrificing target coverage. SMART is therefore a technically plausible planning strategy, but whether broader memory-circuit sparing improves cognition beyond hippocampal avoidance remains unproven.