Hypofractionation in Unresectable Stage II-III NSCLC: A Systematic Technical Review and Practical Recommendations

Background: Concurrent chemoradiotherapy (cCRT) followed by 1 year of consolidation durvalumab remains the standard of care for unresectable stage III non-small cell lung cancer (NSCLC), yet long-term survival is still limited and around 20% of patients present locoregional relapse. Hypofractionation may improve the therapeutic ratio by increasing the biologically effective dose (BED), shortening overall treatment time (OTT) and potentially mitigating radiation-induced lymphopenia. Objective: To identify which patients with unresectable stage II-III NSCLC may benefit most from radical-intent hypofractionated radiotherapy (HypoRT), with or without systemic therapy, and to define the technical aspects, contouring margins, and dosimetric requirements necessary for the safe use of hypofractionated schedule. Methods: A systematic search of PubMed, EMBASE, CENTRAL and Web of Science (January 2010-January 2024, English language) was performed following PRISMA principles. Eligible studies included unresectable NSCLC treated with a BED10 > 40 Gy, allowing induction systemic or targeted therapy; reviews were included for context. Records were screened by six independent reviewers. Studies were classified as non-photon vs. photon, and photon studies were further divided into non-concurrent and concurrent strategies. Clinical outcomes (dose, fractionation, biologically effective Dose for alpha/beta:10 (BED10), equivalent dose of 2 Gy/fraction (EQD2), overall treatment time (OTT), follow-up (FU), median progression-free survival (mPFS), median Overall Survival (mOS), local control (LC), grade ≥ 3 toxicity (G3) and technical variables (immobilization, simulation, GTV imaging, CTV/PTV margins, delivery, image-guided radiotherapy (IGRT), organ at risk (OAR) constraints) were extracted. Results: Of 340 records (170 duplicates), 26 studies including stage II-III disease were analyzed (1155 patients): 1 prospective cohort, 5 phase I, 12 phase II, 1 phase III and 2 phase I/II dose-escalation trials, 2 carbon-ion and 3 proton series. Of these, 13 were photon-based studies with contemporary simulation techniques. Two clinically distinct scenarios emerged. Group A (non-concurrent platinum-based strategies; 166 patients with modern 4DCT simulation) showed wide variability, ranging from a simultaneous-integrated-boost regimen to 75 Gy with 0% grade ≥ 3 toxicity and mPFS 40 months to a chemotherapy-driven schedule with 70% grade 4 chemo-related toxicity. Only one phase III trial was identified, reporting a 2-year local control of 85.8%. Group B (concurrent strategies; nine of the 14 platinum-based studies retrieved, namely those using modern simulation techniques) achieved a median BED of 89.7 Gy to the highest dose-volume, median of OS 20 months and toxicity GIII ≥ ranging 10-20%, with a median study-level grade 5 toxicity of 5.5% (range 0-17.8%; 34 fatal events among 431 evaluable patients, 7.9%), rising to 20.8% (11/53) in the most escalated randomized arm and 22.2% (2/9) in the highest dose-escalation cohort; adaptive SABR boost to 70 Gy/15 fr delivered high local control but clinically meaningful fatal toxicity. Fatal toxicity was confined to schedules exceeding BED10 100 Gy delivered to the whole tumor volume with concurrent platinum, whereas an equivalent BED restricted to a metabolically defined subvolume without concurrent chemotherapy produced no grade ≥3 events. Conclusions: Radical-intent hypofractionation is feasible across both unfit and fit unresectable stage II-III NSCLC populations, but the randomized evidence remains scarce and heterogeneous. Benefit appears most consistent when high BED is delivered with strict normal-tissue sparing in carefully selected patients using modern 4DCT-based simulation, IGRT and intensity-modulated delivery. The pattern of fatal toxicity with aggressive concurrent systemic treatments and dose escalation schedules argues for caution and for prospective trials with harmonized minimum technical standards.