The Impact of Co-Alterations on Outcomes after Local Therapy for Patients with KRAS-Mutant Lung Adenocarcinoma Brain Metastases

PURPOSE/OBJECTIVE(S)

Brain metastases are common in NSCLC with up to 25% of patients having brain metastases (BMs) at the time of diagnosis and 30% developing BMs during their disease course. KRAS is an oncogenic driver in approximately 25% of lung adenocarcinomas. Genomic alterations co-occurring with KRAS are associated with distinct biological landscapes which may influence prognosis. Herein, we sought to identify correlations between genomic profiles, intracranial progression free survival (iPFS), and overall survival (OS).

MATERIALS/METHODS

We retrospectively reviewed 156 patients with KRAS-mutant lung adenocarcinoma BM who underwent SRS for their BMs at MSKCC from 2010-2022. Each patient had at least one tumor sample profiled with MSK-IMPACT, a custom FDA-cleared next-generation sequencing. Mutations, copy number alterations, and fusions were filtered for driver alterations using OncoKB. Survival outcomes were calculated from date of MRI indicating metastatic brain disease.

RESULTS

Of the 156 patients, 80 patients presented with BMs at diagnosis whereas 76 developed BMs during their disease course, with a median 2 lines of therapy prior to BM diagnosis. The most common KRAS mutation was G12C (n = 64; 41%), G12V (n = 26, 17%), G12D (n = 17; 11%), and G12A (n = 11; 7%). The most frequently co-altered genes were TP53 (n = 71, 46%), STK11 (n = 51, 33%), CDKN2A (n = 27, 17%), KEAP1 (n = 17, 11%), and SMARCA4 (n = 10, 6%). The presence of a KEAP1 co-occurring alteration was associated with inferior iPFS (HR 1.95, 95% CI 1.05 - 3.59, p = 0.035) and the presence of SMARCA4 was also associated with inferior iPFS (HR 2.28, 95% CI 1.05 - 4.95, p = 0.038). The presence of an STK11 mutation was associated with worse OS (HR 1.57, 95% 1.01 - 2.43, p = 0.045). In a multi-variate clinico-genomic model, KEAP1 and STK11 co-occurring alterations remained significantly associated with iPFS. Patients with KEAP1-altered tumors had an increased incidence of intracranial regional progression. The 24-month cumulative incidence of regional progression amongst KEAP1-altered tumors was 57% (95% CI, 29%-77%) compared with 37% (95% CI, 29%-46%) among KEAP1-wildtype tumors (P = 0.041). Patients with CDKN2A-altered tumors had an increased incidence of leptomeningeal disease (LMD) as a form of intracranial progression. The 24-month cumulative incidence of LMD amongst CDKN2A-altered tumors was 11% (95% CI, 2.7%-27%) compared with 4.1% (95% CI, 1.5%-8.8%) among CDKN2A-wildtype tumors (P = 0.023).

CONCLUSION

In our cohort of molecularly profiled KRAS-mutant lung adenocarcinoma BM patients treated with SRS, we found that co-occurring KEAP1 and STK11 were significantly associated with worse iPFS. We also observed that CDKN2A co-altered tumors had an increased incidence of LMD. These findings have implications for future efforts to personalize brain metastasis management based on comprehensive genomic profiling.

Salvage Radiotherapy as a Bridge for Relapsed Secondary CNS Lymphoma

PURPOSE/OBJECTIVE(S)

Secondary CNS lymphoma (SCNSL) is a challenging clinical scenario observed in 2-5% of non-Hodgkin lymphoma patients, for which a standard of care has not been defined. We studied the indications for, and outcomes of SCNSL patients referred for radiotherapy (RT).

MATERIALS/METHODS

We identified patients with aggressive B cell lymphoma who received brain RT for SCNSL between 1999-2023 at a tertiary cancer center. Patients were grouped and analyzed by RT indication. Overall survival (OS) was determined from RT start using the Kaplan-Meier method. OS analysis comparing patients who did and did not receive therapy after RT was landmarked at 60 days from start of RT to minimize immortal time bias. "SCNSL-directed therapy" is defined as systemic therapy for the treatment of SCNSL, as opposed to CNS prophylaxis.

RESULTS

We identified 99 SCNSL patients treated with RT. To account for the heterogeneity of RT referrals, we focused on the most common indication: salvage of radiographic progression after SCNSL-directed systemic therapy (n = 58). Among this group, median age was 62 (interquartile range [IQR]: 48-69) and 86% had diffuse large B cell histology. At initial lymphoma diagnosis, 10% of patients had CNS involvement, 90% received Rituximab-based therapy, and 25% received prior CNS prophylaxis. For SCNSL directed therapy, 90% received methotrexate (MTX)-based regimen. Median time from initial SCNSL diagnosis to RT was 4.4 months (IQR 1.7-7.0), with a median of 2.0 lines of therapy prior to RT (IQR 1.0-3.0). 86% of patients were symptomatic at RT with median KPS of 70 (IQR: 60-80). RT targets included whole brain (86%) and partial brain (14%). 1 patient had craniospinal RT. Median RT dose was 30 Gy (IQR: 24-30) over 10 fractions. Median OS for the entire salvage cohort was 3.5 months (m). Landmark analysis 2m post RT showed that median OS differed when patients were stratified by receipt of further therapy: CAR-T (9.4m, n = 4), hematopoietic cell transplant (8.5m, n = 6), other systemic therapy (4.4m, n = 17), no systemic therapy (0.6m, n = 10) (p = 0.0004). 29% of patients who received further therapy after RT achieved long term survival.

CONCLUSION

In our cohort, most SCNSL patients are referred for salvage RT, with a median OS of 3.5m. 86% of patients had neurologic symptoms after having failed a median of 2 lines of SCNSL-directed therapy; the clinical urgency of this scenario implies that without RT, patients may not have been suitable candidates for further therapy. However, among patients for whom RT was successfully used to bridge to additional therapy, 29% could achieve long-term survival. This study supports further investigation of RT as a combined modality strategy for relapsed/refractory SCNSL, including with emerging cellular therapies.