Suyan Bi, Changyou Zhong, Ying Long, Zhongqing Huang, Zhitao Dai
Abstract Background Spatially fractionated radiation therapy (SFRT)—mainly including GRID, LATTICE, minibeam, and microbeam—widens the therapeutic window for bulky or radioresistant tumors. Yet clinical translation faces two persistent hurdles: technical heterogeneity across modalities and a lack of evidence-based, tumor-specific dose prescription guidelines. Methods This scoping review maps the SFRT evidence landscape from 2005 to 2025. Following PRISMA-ScR guidelines, we synthesized 60 studies (preclinical, clinical, technical) and identified valley dose (D valley ) as a consistently reported strong candidate predictor of the therapeutic window. Results Based on this pattern, we outline an evidence-informed organizational framework. Synthesizing the extracted data, we delineate a tripartite structure that emerged from the literature: (1) a tumor-and modality-specific dosimetric decision frame work (Level 1) capturing histology-dependent D valley benchmarks (e.g., ≥ 2 Gy for cervical LATTICE, ~ 10 Gy for brain proton minibeam) with PVDR guidance; (2) a Mechanism- Biomarker-Outcome (MBO) chain (Level 2) linking spatial dose gradients to quantifiable cascades—vascular normalization (DCE-MRI Ktrans), immune preservation (CD8⁺ density), and bystander signaling (HMGB1, TNF-α, ROS); (3) an exploratory surrogate marker model (Level 3) identifying early short-term metrics (6-month metabolic response ≥ 50% SUVmax reduction, cytokine dynamics) as candidate correlates of long-term local control. Conclusion By mapping the shift from peak-centric ablation to a valley-dose-driven approach reflected in the literature, this review provides a structured, evidence-informed scaffold for future clinical trial design and technical harmonization. The framework presented is intended as a hypothesis-generating tool to inform the standardized implementation of SFRT in precision radiation oncology.