Mutational signatures of synchronous and metachronous brain metastases from lung adenocarcinoma

Brain metastases (BM) are observed in 30% to 50% of patients at initial diagnosis with more patients developing metastases during treatment [1]. Synchronous brain metastases (SBM) and metachronous brain metastases (MBM) are two distinct clinical conditions. Understanding the intrinsic factors of discrepancy and developing therapeutic strategies is an urgent need. The immune-suppressive microenvironment of SBM and MBM has recently been investigated closely [2]. However, the mutational profiles of SBM and MBM are still uncharacterized. Herein, we outlined the mutational characteristics of SBM and MBM, and explored the molecular pathways to search for potential treatment strategies.

The clinical characteristic of patients diagnosed with BM was summarized in Additional file 6: Table S1. Fifty-three patients who had SBM at the first diagnosis were defined as the synchronous group. Forty-three MBM patients were in stage I to IV stage at the first diagnosis and then developed BM over the course of the disease. Long-term follow-up overall survival (OS) of two cohorts showed that patients in the synchronous group had an OS of 31 (95% CI 21.19–40.81) months and were shorter compared to patients in the metachronous group with an OS of 71 (95% CI 34.14–107.86) months (p = 0.0005) (Additional file 1: Figure S1). We further calculated the brain overall survival (BOS), defined as the time from the diagnosis of brain metastases to the end of follow-up. The median BOS was 31 (95% CI 21.19–40.81) months in the SBM group and 39 (95% CI 21.80–56.20) months in the MBM group, with no statistically significant difference between the two groups (P = 0.32) (Additional file 2: Figure S2). In MBM group, the median interval time of brain parenchymal metastases (BPM) and Leptomeningeal metastases (LM) from diagnosis was 21.3 months (IQR, 14.8 – 37.5) and 38.4 months (IQR, 19.6 – 56.0).

We have identified mutations in the plasma and CSF samples from ninety-six patients enrolled in our study. Figure 1 displayed the mutation profiles of the most frequently detected mutations from plasma (Fig. 1A) and CSF (Fig. 1B) samples. In our research, we have also analyzed the plasma samples of 11 patients with liver metastases (Additional file 3: Figure S3) and 29 patients with bone metastases (Additional file 4: Figure S4). The top 3 gene mutation were EGFR, TP53, and MYCN in 29 patients, EGFR TP53, GRIN2A were most frequently mutated in 11 patients with liver metastasis. Compared to blood plasma, CSF samples revealed a more comprehensive genomic characterization of BM. The detection rate of EGFR in CSF was 58% and in plasma was 32%. TP53 mutation was detected in 48% and 13% of CSF and plasma, respectively. According to the subtypes of EGFR, alterations were more detected in CSF while they were less detected or not available in plasma samples (Fig. 1C). Within the individual patient, we noted that some mutations (15.5%) can be identified in both sample types which is called the shared mutation (Fig. 1D). Besides, only a few private mutations (7.6%) can be detected in plasma. In a very large proportion of BM patients, CSF can provide more unique genetic mutations (76.9%) than plasma (7.6%). Hence, CSF identification of BM mutation spectrum is more advantageous.

Subsequently, we removed copy number alterations and extracted the information on single nucleotide variations to perform the further analysis. We focused on the molecular features of altered somatic mutations between synchronous and metachronous metastases. The most common mutations were EGFR and TP53. SBM patients harbor more mutation in CTNNB1 while APC and ALK were more mutated genes in the MBM group (Fig. 2A). Variant allele frequency (VAF) in the synchronous metastatic tumor showed high dissimilarity to those in the metachronous metastatic tumor. And, we noted that EGFR and TP53 have a high allelic frequency in the metachronous group (Fig. 2B). Next, we characterize the molecular profile of variants in EGFR and TP53, In the SBM cohort, the most common EGFR mutation type is exon 19 deletion, while in the MBM cohort, L858R point mutation is the most frequent EGFR mutation type (Fig. 2C). Therefore, for synchronous brain metastasis patients with the 19 deletion mutation, we recommend the first-line use of osimertinib. For non-synchronous brain metastasis patients with L858R mutation, the combination of bevacizumab and erlotinib could be considered as a first option. Four missense TP53 mutations were found in SBM patients, including D74, S116, D153 and T377. R273 was a unique TP53 mutation site in MBM patients (Fig. 2D). Then, we performed gene enrichment analysis to identify enriched signaling pathways. Mutated genes of both cohorts were enriched in RTK-RAS and TP53 pathways. NFR2 accounted for a large fraction of affected pathways in the MBM group (Fig. 2E and F). In addition, we have interrogated the pharmacogenomic interactions in both groups to explore potential druggable medicine. The clinically actionable genes in the MBM group were ALK, APC, ARID1A, AXIN2, BRAF, while APC, ATRX, CBL, CDK6, CTNNB1 were the clinically actionable genes for the SBM group. The distinct drug-gene associations also revealed the heterogeneous nature of the two cohorts (Additional file 5: Figure S5).

Altogether, our study strengthened the rationale for the use of CSF to detect clinical drive alterations of brain parenchymal metastases. The comparison of genomic characterization revealed the similarities and disparities between synchronous and metachronous brain metastases. These potentially clinically actionable targets present in the brain metastases are critical for the development of more effective treatment strategies to improve outcomes for these two distinct populations.