Study Shows High Concordance of Recurrent Somatic Alterations in Primary and Matched Metastatic NSCLC

Get Permission

[O]ur results … [suggest] that genomic profiles of primary tumor reflect the genomic spectrum of the patient’s metastatic disease and can identify the key somatic alterations present in matched NSCLC metastases.

—Stéphane Vignot, MD, and colleagues

In a study reported in Journal of Clinical Oncology, Stéphane Vignot, MD, of Institut National de la Santé et de la Recherche Medicale (INSERM) Unit 981, Paris, and Gustave Roussy Institute, Villejuif, France, and colleagues used next-generation sequencing to identify somatic alterations in archived primary tumors and matched metastatic tumor samples in patients with non–small cell lung cancer (NSCLC).1 They found a high concordance of recurrent alterations, suggesting that primary tumors may contain “much of the relevant genomic information required to guide treatment on recurrence.”

Study Details

The study included 30 primary and matched metastatic tumor samples from 15 NSCLC patients who had undergone resection. Samples had to have tumor cellularity of greater than 50% (estimated median 70%). Histology was adenocarcinoma in 8 patients, squamous cell in 3, large-cell in 2, and basaloid in 2. No patient received molecular targeted therapy prior to biopsy of metastatic lesions.

Patients ranged in age from 41 to 82 years; 13 were male, 14 had a history of smoking, 10 had not received chemotherapy for early-stage disease, 2 had synchronous metastasis, and metastasis was locoregional in 7.

Genomic libraries were captured for 3,280 exons in 182 cancer-related genes and 37 introns from 14 genes often rearranged in cancer and were sequenced to high coverage (paired-end sequencing of 49 × 49 cycles). Tests for detecting alterations were validated to detect base substitutions at a ≥ 10% mutant allele frequency with ≥ 99% sensitivity and indels at a ≥ 0% mutant allele frequency with ≥ 95% sensitivity and a false discovery rate of < 1%. Recurrent alterations were defined as those occurring in ≥ 5% of NSCLC samples in the Catalogue of Somatic Mutations in Cancer (COSMIC) database or reported as amplified or deleted in ≥ 5% of samples in the literature; all other alterations were classified as “likely passenger alterations.”

Overall Concordance

A total of 311 somatic alterations were identified among the 189 evaluated genes, including 161 in primary tumors and 150 in metastases. Tumors in one patient (who had large-cell carcinoma, a 30 pack-year history of smoking, and had received neoadjuvant chemotherapy) yielded no targeted alterations.

Of the 311 alterations, 63 (20%) were known recurrent alterations (32 in primary tumors and 31 in metastases). Analysis of 170 unique mutations and 21 large structural alterations (irrespective of recurrent or passenger status) found in at least one of the paired tumor samples showed an overall rate of concordance between primary and metastatic tumors of 64% (64.5% for mutations and 59% for large structural changes).

Overall, there were 26 different short alteration events (substitutions and indels) in nine genes and five large structural alterations (amplifications and deletions) in four genes. TP53 mutations were the most frequent recurrent alteration, found in 12 patients. Tumors had two to four recurrent alterations in 10 patients. Other mutations identified were those in GNAS (1), KRAS (3), NOTCH1 (1), PIK3CA (3), RB1 (1), SMARCA4 (2), and STK11 (3). One patient had an EGFR mutation in both the primary and metastatic tumors in association with TP53 mutation. No ALK rearrangements were identified. Large structural alterations consisted of CDKN2A deletion and amplification of FGFR1, MYC, and SOX2. Two patients had no recurrent alterations.

Concordance of Recurrent Alterations

Analysis of 33 recurrent alterations and 159 likely passenger alterations showed concordance rates between primary and metastatic samples of 94% for recurrent alterations vs 63% for passenger alterations. Among mutations, concordance was 93% for recurrent mutations and 61% for passenger mutations. Among large structural alterations, concordance was 100% for recurrent alterations and 40% for passenger alterations.

Limitations of the study include the relatively small size of the cohort and its heterogeneity in histologic subtypes and metastatic sites, with these factors limiting comparisons according to disease or treatment characteristics. As noted by the investigators, the study focused on patients with recurrence after surgery and not on patients who presented with metastatic disease or had biopsies taken from multiple sites simultaneously or sequentially after targeted treatment; such patients might have inherent genomic differences or differences that emerge under the selective pressure of targeted therapies.

Overall, the findings indicate that the status of major alterations in oncogenesis in the primary tumor is “a robust surrogate” for status at first recurrence in the absence of molecular targeted therapy.

The investigators concluded, “[O]ur results … reveal a high level of concordance for recurrent somatic alterations, suggesting that genomic profiles of primary tumor reflect the genomic spectrum of the patient’s metastatic disease and can identify the key somatic alterations present in matched NSCLC metastases. In these patients, biopsy of metastatic lesions can be hazardous, and samples are often limited in size. Our data indicate that archived primary material could be a suitable specimen for clinical decision on development of metastatic disease.” ■

Disclosure: For full disclosures of the study authors, visit


1. Vignot S, Framptom GM, Soria J-C, et al: Next-generation sequencing reveals high concordance of recurrent somatic alterations between primary tumor and metastases from patients with non-small-cell lung cancer. J Clin Oncol. April 29, 2013 (early release online).

Related Articles

SIDEBAR: Mitigating the Anxiety over Tumor Heterogeneity

This collaborative study with Foundation Medicine (Cambridge, Massachusetts), using very sensitive deep sequencing, partially mitigates some of the anxiety generated by the identification of tumor heterogeneity. While our data in lung cancer confirm that such heterogeneity exists, they also...