Combined Analysis of a Phase III Randomized Trial and Phase II Prospective Trial with Blind Control Matching of Patients Receiving Whole-Lung, Low-Dose Radiation for COVID-19: Full Results and Immunologic Correlates of the RESCUE 1-19 Trial

Abstract

Whole-lung, low-dose radiation (LD-RT) for COVID-19 requires randomization and biologic correlates to determine causality and mechanism. A phase III trial randomized COVID-19 patients to physician's choice of drug therapy with or without LD-RT. Primary endpoint was intubation-free survival (IFS). The trial was designed with 80% power (two-sided log rank, alpha 0.05) to detect a hazard ratio of 0.3 after 12 intubation events. Estimating a 25% event rate, the planned sample size was 84 patients plus 25 to account for declining intubation rates and screen failures. Due to hospitalization declines and other barriers, the trial closed prematurely. Available randomized data were analyzed by intention-to-treat and combined with phase II results and immunologic correlates, using one-sided significance and an alpha of 0.1 to inform future trial design. From Jun 2020-Jun 2022, 14 patients were randomized on a phase III trial. From Apr 2020-Dec 2020, 42 patients were enrolled on a phase II trial and blindly matched to 40 controls from contemporaneous trials. 96 total patients and 193 blood samples were available for analysis. Mean hospital duration with LD-RT was 12.9 vs 15.4 days in controls (p = 0.12). Oxygen flow rate >15 L/min (26% vs 38%, p = .27), high-flow oxygen >30 L/min (24% vs 38%, p = 0.18), non-invasive positive-pressure >60 L/min (9% vs 27%, p = 0.03), and mechanical ventilation (9% vs 24%, p = 0.05) reduced with LD-RT. Mean supplemented oxygen volume was 171,759 vs 547,626 liters in controls, with daily means of 10 vs 23 L/min (p = 0.03). Radiographs worsened in 43% vs 71% of controls (p = 0.03). Arterial blood gas mean P/F ratios improved 22% after LD-RT vs declined 8% in controls (p = 0.12). Mean days febrile were 1.8 vs 2.9 in controls (p = 0.10). Rate of myocardial injury was 47% vs 40% in controls (p = 0.77). Flow cytometry revealed 4-fold and 30-fold larger expansions, respectively, in CD8- and CD4-positive CD3+PD1+Ki67-high proliferating cytotoxic T-cells (300% vs 75% expansion, p = 0.07) and helper T-cells (200% expansion vs 6% contraction, p = 0.03) at day 7. In the randomized cohort, mean oxygen volume fell 75% with LD-RT to 78,336 vs 316,786 liters in controls (p = 0.13), mean flow rates were 5.1 vs 18.4 L/min (p = 0.13), radiographs worsened in 50% vs 100% (p = .17), P/F ratios improved 31% vs declined 68% in controls (p = 0.03), hospital duration was 8.9 vs 11.5 days (p = 0.22), and zero LD-RT patients vs one control intubated. Combined analysis of a phase II/III randomized trial suggests that LD-RT prevents ventilation, reduces supplemental oxygen need, improves clinical course, and enhances immune response. LD-RT may have both immediate direct effects and delayed enhanced immunity in COVID-19. Larger multi-institutional trials are justified.