In Reply: Stereotactic Radiosurgery Dose Reduction for Melanoma Brain Metastases Patients on Immunotherapy or Targeted Therapy: A Single-Center Experience

To the Editor: We thank Shankar et al1 for their commentary on our recent study examining stereotactic radiosurgery (SRS) dose-reduction in patients with melanoma brain metastases treated concurrently with immunotherapy or targeted agents.2 Their insights raise important considerations, and we welcome the opportunity to clarify aspects of our methodology, interpretation, and clinical context. The question on temporal bias is valid, particularly in retrospective studies such as ours. We acknowledge that evolving systemic therapies, improved imaging and radiosurgical devices, and institutional expertise may have influenced outcomes in the reduced dose (RD) group. Nevertheless, the observation that local control improved despite lower prescription doses, especially in a tumor type that historically has been considered radioresistant, suggests more than an era effect alone. This raises the possibility of a synergistic interaction between radiosurgery and contemporary systemic therapies that merits further prospective evaluation. Regarding dosimetric parameters, such as V12, mediating adverse radiation events (AREs), it would be expected that the per matrix V12 in the RD group would be lower than the SD group. The median lesion V12 was 0.6 cm3 in both groups; however, the interquartile range was narrower in the RD group (0.5-0.9 cm3) than in the ASTRO cohort (0.5-1.2 cm3), and this difference reached statistical significance (P < .001). Skull-level V12 was comparable between groups, with no significant difference (P = .12). Although cumulative intracranial dose burden and total treated volume are additional important factors, this analysis supports the premise that prescription dose itself remains a meaningful determinant of toxicity. Naturally, the inclusion of V12 in the multivariable statistical models would have shown that V12 is a strong predictor; as prescription dose, treatment volume, and V12 are dependent variables, inclusion of all 3 in the same model would not be statistically sound. Moreover, studying V12 was not the focus of this study. The readers are encouraged to examine other studies better designed and focused on V12 in brain metastasis patients undergoing SRS and concurrent systemic therapy.3 Regarding the interpretation of ASTRO guideline doses,4 we agree that these recommendations are histology-agnostic and not tailored to melanoma specifically. Our findings are not intended to challenge the guideline framework, but rather to explore the feasibility of dose de-escalation in a biologically favorable context. The absence of formal stratification by lesion location in our analysis reflects a limitation of available data. In clinical practice, although dose decisions at our center incorporate anatomic considerations, especially for tumors near critical structures, dosimetric parameters have higher influence on the decision-making process. The complexity of defining "concurrent" systemic therapy is acknowledged. Although we adopted a four-week window based on previous radiosurgical studies, the timing and sequencing of immunotherapy relative to radiation could potentially influence both efficacy and toxicity profiles.3,5-8 Our retrospective Data set lacked sufficient granularity to address this interaction, but we agree that future studies should aim to capture and analyze these dynamics more precisely. Regarding other radiographic and clinical outcomes, such as intracranial progression-free survival, distant brain failure, neurocognitive outcomes, and steroid dependence, this study was not designed to evaluate these. Nevertheless, stereotactic radiosurgery remains a focal treatment and its therapeutic benefit in distant CNS progression is limited. The commentary on genomic heterogeneity within the BRAF-positive subgroup highlights an area of growing importance. We concur that future studies incorporating detailed genomic profiling will be essential to tailoring radiosurgical strategies based on tumor biology. Regarding technical treatment parameters, such as isodose line selection and conformity indices, SRS plans are tailored according to lesion size. Lesions >1.5 cm use an isodose line 50%, consistent with standard Gamma Knife practice. Higher isodose lines (>80%) are typically used for smaller lesions. Although these parameters were not included in our published analysis, we acknowledge their potential influence on dose distribution and outcomes.9 Although we did not report these variables for the 2 groups in the study, the used isodose lines did not differ significantly between the 2 groups (SD 85 (54, 95) vs RD 85 (60, 95), P > .9). The possibility that longer survival might increase the likelihood of detecting AREs is another important consideration.10 For instance, the risk of AREs in brain metastasis patients will differ compared with AVM patients.11 All analyses were performed with mortality considered a competing event, thus survival was taken into consideration. Irrespective of the statistical method, the lower incidence of AREs in the RD group, despite longer follow-up in many cases, suggests a true reduction in toxicity rather than a detection artifact. Finally, we agree that dose reduction is one of several approaches to optimizing the therapeutic ratio in radiosurgery, particularly as systemic therapies evolve. Hypofractionated SRS, especially for larger or eloquent lesions, remains a promising strategy.12 Nevertheless, the current findings support the idea that lower single-fraction doses may be sufficient for many patients. In summary, our study provides valuable data, with the limitations that are expected of retrospective studies, on the potential safety and efficacy of reduced-dose SRS in the setting of concurrent immunotherapy or targeted therapy. The current work also is hypothesis generating and will no doubt serve to stimulate additional studies by our group and others in the SRS field.