Genomically personalized radiation dose de-escalation: A phase II basket radiation study in patients with pathogenic mutations in <i>ATM</i> .

3015 Background: ATM (ataxia-telangiectasia mutated) is a key DNA damage response gene, and pathogenic variants are known to cause a rare hereditary syndrome notable for extreme radiosensitivity. Somatic ATM inactivation is associated with markedly improved tumor control following RT. We hypothesized that ATM -mutant tumors could achieve effective local control with substantially lower RT doses, reducing normal tissue toxicity. This concept of genomically personalized radiation dose de-escalation was tested in a phase II trial. Methods: We conducted a single-arm phase II trial of low-dose RT in patients with metastatic solid tumors harboring pathogenic ATM mutations. Eligible patients had radiographic progression suitable for palliation and all patients received 4Gy x 2 fractions- 60% reduced compared to conventional 4Gyx5 palliation- delivered to a target lesion. The primary endpoint was 6-month local failure rate (radiographic progression or re-irradiation) and benchmarked against historical local failure control rates of 82% for 4Gyx5 standard palliation, with pre-specified analysis conducted at this initial dose level. Toxicity was graded by CTCAE criteria. Results: Twenty-one patients (across 9 histologies; 3 germline ATM , 18 somatic) were enrolled. At 6 months, 12 patients were evaluable; 10 of 12 (83.3%) achieved local control with the 4Gyx2 regimen. Notably, this included a 4.4cm angiosarcoma with a complete response by 6 months- an exceptional outcome with such a low dose. Two patients (16.7%) had local failure by 6 months. Both of these tumors retained ATM protein expression by immunohistochemistry and harbored concurrent pathogenic TP53 mutations. In contrast, molecular analysis confirmed biallelic ATM loss in four patients, including the top three responders. No grade ≥2 toxicities or other significant RT-related side effects were observed. Conclusions: ATM -mutant tumors were effectively controlled with markedly reduced RT dose (4Gyx2), achieving local control in 83% of lesions at 6 months, comparable to historical response rates with standard palliation (4Gyx5), but with minimal toxicity. Additionally, FACETS analysis showed a stronger correlation with treatment response than germline vs. somatic mutation status, suggesting that FACETS analysis could be more informative in determining eligibility criteria for future clinical trials. These preliminary findings establish ATM as a novel biomarker for personalized RT dose de-escalation and compel further investigation into ATM as a therapeutic target. Clinical trial information: NCT05010031 .