Clinical Outcomes and Safety of Spatially Fractionated Radiotherapy for Malignant Tumors: A Systematic Review and Evidence Map

Spatially fractionated radiotherapy (SFRT) concentrates high doses in selected tumor regions while delivering lower doses between them. This systematic review evaluated clinical outcomes, treatment delivery, and harms across SFRT techniques, with particular attention to comparative evidence. PubMed, Europe PMC, and ClinicalTrials.gov were last searched on September 19, 2026; CINAHL, CENTRAL, Scopus, and WHO ICTRP were last searched on August 7, 2026. The original searches and selection were conducted by the corresponding author, with substantive artificial intelligence (AI) assistance and human checks across review stages. The September extension introduced further AI-assisted source verification and analysis. Eligible reports described patients receiving intentional SFRT for malignant tumors. Results were grouped by clinical question and synthesized without meta-analysis. Appraisal used the Cochrane risk-of-bias tool for randomized trials (RoB 2), Risk of Bias in Non-randomized Studies-of Interventions (ROBINS-I), and design-matched Joanna Briggs Institute (JBI) tools; selected comparative questions underwent Grading of Recommendations Assessment, Development and Evaluation (GRADE) evaluation. The final corpus comprised 151 reports representing 137 investigations. Evidence was predominantly noncomparative, with heterogeneous treatments, outcome definitions, and follow-up. A prematurely terminated randomized trial documented one response among nine participants assigned to GRID plus standard radiotherapy, with one unknown outcome, and one among three assigned to standard radiotherapy. A nonrandomized study in bulky unresectable non-small-cell lung cancer reported responses in 19/20 PATHY participants versus 4/20 receiving conventional palliative radiotherapy, but prognosis-based allocation prevented causal interpretation. Certainty was very low for four defined comparative questions. Several cohorts reported early tumor shrinkage or symptom improvement, while prospective evidence also showed nonuniform pain changes, progression, and relapse after initial response. Severe complications included obstruction, late genitourinary injury, and fatal carotid hemorrhage. Differences in endpoints, treatment combinations, analysis units, missingness, and possible cohort overlap precluded a defensible pooled effect or unique participant total. Restrictions by source access, design, era, and overlap preserved descriptive evidence of activity and delivery but did not establish comparative superiority, durable benefit, or general safety. SFRT warrants indication-specific evaluation with defined patient-important outcomes, reproducible dosimetry, and complete toxicity follow-up. The protocol and subsequent robustness amendment were retrospective and unregistered. No specific funding was received.