Molecular markers of metastatic disease in KRAS-mutant lung adenocarcinoma

Abstract

Background: Prior studies characterized the association of molecular alterations with treatment-specific outcomes in KRAS-mutant (KRAS^MUT) lung adenocarcinoma (LUAD). Less is known about the prognostic role of molecular alterations and their associations with metastatic disease.

Patients and methods: We analyzed clinicogenomic data from 1817 patients with KRAS^MUT LUAD sequenced at the Dana-Farber Cancer Institute (DFCI) and Memorial Sloan Kettering Cancer Center (MSKCC). Patients with metastatic (M1) and nonmetastatic (M0) disease were compared. Transcriptomic data from The Cancer Genome Atlas (TCGA) were investigated to characterize the biology of differential associations with clinical outcomes. Organ-specific metastasis was associated with overall survival (OS).

Results: KEAP1 (DFCI: OR = 2.3, q = 0.04; MSKCC: OR = 2.2, q = 0.00027) and SMARCA4 mutations (DFCI: OR = 2.5, q = 0.06; MSKCC: OR = 2.6, q = 0.0021) were enriched in M1 versus M0 tumors. On integrative modeling, NRF2 activation was the genomic feature most associated with OS. KEAP1 mutations were enriched in M1 versus M0 tumors independent of STK11 status (KEAP1^MUT/STK11^WT: DFCI OR = 3.0, P = 0.0064; MSKCC OR = 2.0, P = 0.041; KEAP1^MUT/STK11^MUT: DFCI OR = 2.3, P = 0.0063; MSKCC OR = 2.5, P = 3.6 × 10^-05); STK11 mutations without KEAP1 loss were not associated with stage (KEAP1^WT/STK11^MUT: DFCI OR = 0.97, P = 1.0; MSKCC OR = 1.2, P = 0.33) or outcome. KEAP1/KRAS-mutated tumors with and without STK11 mutations exhibited high functional STK11 loss. The negative effects of KEAP1 were compounded in the presence of bone (HR = 2.3, P = 4.4 × 10^-14) and negated in the presence of lymph node metastasis (HR = 1.0, P = 0.91).

Conclusions: Mutations in KEAP1 and SMARCA4, but not STK11, were associated with metastatic disease and poor OS. Functional STK11 loss, however, may contribute to poor outcomes in KEAP1^MUT tumors. Integrating molecular data with clinical and metastatic-site annotations can more accurately risk stratify patients.

Keywords: KEAP1; KRAS; STK11; lung adenocarcinoma; metastatic.

Figure 1

Molecular characterization of metastatic KRAS^MUT LUAD. (A) Percent of M0 vs M1-Primary vs M1-Metastasis tumors in DFCI and MSKCC cohorts. (B) TMB and (C) FGA in M0 vs M1-Primary vs M1-Metastasis tumors in DFCI and MSKCC cohorts. (D) Volcano plots comparing the somatic alterations enriched in M1 vs M0 tumors.

Figure 2.

Comparison of KRAS^MUT LUAD by KEAP1 vs STK11 mutational status. (A) Forest plots showing the rates of KEAP1^MUT/STK11^MUT (KEAP1/STK11), KEAP1^MUT/STK11^WT (KEAP1) and KEAP1^WT/STK11^MUT (STK11) tumors amongst M1 vs M0 tumors in the DFCI and MSKCC chorts. (B) Survival curves comparing overall survival between KEAP1/STK11, KEAP1, STK11 and KEAP1^WT/STK11^WT (WT) groups in the DFCI and MSKCC cohorts. Hazard ratios (HR) are computed with reference to WT group. (C) Boxplot comparing STK11 Loss signature score between KEAP1/STK11, KEAP1, STK11 and WT groups in the TCGA cohort. (D) Heatmap and barplot comparing ssGSEA signatures scores between KEAP1 or KEAP1/STK11 vs STK11 groups in the TCGA cohort. Signatures with q < 0.05 are shown. (E) Boxplot comparing the signatures with the highest and lowest difference in median signature score in figure D by KEAP1 and STK11 mutational status.

Figure 3.

Association of OS with presence of metastasis in specific organ sites in MSKCC cohort. (A) Forest plot showing hazard ratios from multivariable cox proportional hazards model by metastasis to specific organ sites. (B) Bubble plot comparing the rates of somatic alterations between M1 vs M0 tumors by presence of metastasis in specific organ sites. Survival curves by presence of metastasis in distant lymph nodes vs (C) bone and (D) liver. Survival curves by KEAP1 mutational status and presence of metastasis in (E) bone and (F) lymph nodes.

Figure 4.

Predictive modeling of overall survival. (A) Graphical illustration of random forest survival (RSF) model. (B) Bar graph showing top 10 most important features by mean difference in concordance-index (c-index) between c-index of the training (MSKCC) and validation (DFCI) cohorts when a given feature is permuted. (C) Cumulative-dynamic area under receiving operating curve (AUC) for training and validation cohort. (D) Predicted patient-level survival curves in the training and validation cohort by KEAP1 and STK11 mutational status.

Figure 5.

Molecular characterization of metastatic KRAS^MUT vs KRAS^WT LUAD. (A) TMB and (B) FGA in KRAS^MUT vs KRAS^WT LUAD in M0, M1-Primary, and M1-Metastasis tumors sequenced at MSKCC. (C) Volcano plot comparing the somatic alterations enriched in M1 KRAS^MUT vs KRAS^WT tumors sequenced at MSKCC. For visualization, alterations with Log10 q-value > 10, were set to have -Log10 q-value = 10. (D) Comparison of ssGSEA signature scores between KRAS^MUT and KRAS^WT tumors by stage in the TCGA cohort. All ssGSEA signatures were compared between stage III/IV KRAS-mutant vs stage III/IV KRAS-wildtype tumors, and only those with q < 0.05 are shown. (E) Survival curves by KRAS, KEAP1 and STK11 mutational status in M1 tumors sequenced at MSKCC. (F) STK11 Loss score by KRAS mutational status amongst KEAP1/STK11, KEAP1, STK11, and WT tumors in the TCGA cohort.