KEY POINTS
- This PRISMA systematic review searched PubMed, Embase, and Scopus for clinical and translational evidence published from 2015 through 2025, with narrative rather than pooled synthesis because of marked methodological heterogeneity.
- Prostate cancer provided the strongest clinical evidence for a low α/β ratio. CHHiP produced an estimate of 1.6 Gy (95% CI 1.4–1.9), while data from PROFIT, HYPRO, and other randomized studies clustered broadly around 1.2–2.0 Gy.
- PACE-B further supports pronounced fraction sensitivity: 5-year biochemical/clinical failure-free survival was 94.6% with conventional/moderate fractionation and 95.8% with 36.25 Gy in five fractions, consistent with modeled α/β estimates around 1.4–1.7 Gy.
- ADT may alter clinically derived estimates: prostate cohorts treated with RT alone produced effective α/β values around 1.8–2.0 Gy, compared with approximately 1.2–1.5 Gy when hormonal suppression was incorporated.
- Breast cancer data from START, FAST, and FAST-Forward support a lower value for microscopic postoperative disease than the traditional 10-Gy tumor assumption, with estimates around 2.7–3.5 Gy.
- Bladder cancer fractionation data from BC2001 and BCON were compatible with a tumor α/β around 5–6 Gy, while renal cell carcinoma also appears considerably more fraction-sensitive than classical “high α/β tumor” assumptions suggest.
- By contrast, high α/β estimates derived from lung SBRT may partly reflect hypoxia, tumor kinetics, extreme dose per fraction, and limitations of extrapolating the conventional linear-quadratic model into ablative dose ranges.
- The authors argue that α/β should increasingly be considered a context-dependent phenotype, potentially influenced by disease state, systemic therapy, microenvironment, and intrinsic genomic radiosensitivity rather than treated as a universal constant.
CLINICAL TAKEAWAY
The linear-quadratic framework remains clinically useful, but the familiar “tumor = 10 Gy, late tissue = 3 Gy” shorthand is increasingly difficult to defend across all disease sites. The strongest lesson is not that every α/β value should now be replaced with a new fixed number, but that fractionation sensitivity depends on tumor biology, treatment context, and the data used to estimate it.