Prognostic and predictive value of loss of nuclear RAD51 immunoreactivity in resected non-small cell lung cancer patients

Objectives: In response to DNA damage, recombination proteins are relocalized into sub-nucleaf complexes that are microscopically detected as RAD51-containing nuclear foci. We aimed for assessing the prognostic and predictive value of loss of nuclear RAD51 immunoreactivity ('RAD51 loss') in 2 independent stage I to III non-small cell lung cancer (NSCLC) patient cohorts undergoing surgical resection and eventual perioperative chemo-/radiotherapy (CT/RT). Materials and methods: The discovery set included 69 evaluable patients (19 adenocarcinomas, ADC, 50 squamous cell carcinomas, SCC) from Palacky University Hospital, 45/69 (65.2%) with additional platinum-based CT. The replication set entailed 845 evaluable patients (446 ADC, 399 SCC) from University Hospital Zurich, 308/845 (36.5%) with platinum based CT or RT. RAD51 loss was defined as <= 20% of tumor cell nuclei having any nuclear RAD51 expression. We assessed the prognostic value of RAD51 loss in all patients and its predictive value in patients receiving CT/RT. Results: RAD51 loss was observed in 40/69 (58.0%) and 439/845 (51.9%) evaluable tumors in the discovery and replication set, respectively (p = 0.34). It was more frequent in ADC compared to SCC (57.2% vs 47.4%, p = 0.003). RAD51 loss was significantly associated with worse OS in both the discovery (adjusted HR = 2.39, p = 0.039) and replication set (adjusted HR = 1.31, p = 0.008). The unfavourable prognostic effect of RAD51 loss seen in the overall population was not observed in patients receiving perioperative CT (adjusted HR = 1.07, p = 0.73) or perioperative RT (adjusted HR= 1.05, p = 0.82). Conclusion: RAD51 loss has an unfavourable prognostic impact in NSCLC patients undergoing curative surgical resection, but it may have a favourable predictive value in the subgroup of patients receiving perioperative platinum-based CT or RT, most likely as a consequence of deficient DNA repair. (C) 2017 Elsevier B.V. All rights reserved.