Abstract PD1-06: Comparison of mutational landscapes of primary breast cancer and first metastatic relapse: Results from the ESOPE study

Background

Genomic profile of breast cancer metastases (M) may differ from that of the primary tumor (PT). In a multicenter prospective study (ESOPE, NCT 01956552) including 130 patients with biopsies of the first metastatic deposit, we have shown that luminal breast cancers are the most prone to phenotypical subtype changes (Comte et al, ASCO 2016#550). We report here the first results of a comparative PT/M targeted next generation sequencing (NGS) mutational analysis.

Methods

Of 130 patients, 117 paired PT/M samples obtained before any treatment were available for analysis. Targeted Sequencing was done using Illumina Hiseq2500 technology with a custom made 95 breast cancer associated genes panel. Sequence data were aligned to the human reference genome (hg19) using Bowtie2 algorithm. Median depth was 607X and 87% of targets achieved 100X depth. SNVs and indels were called using GATK UnifiedGenotyper. We retained COSMIC confirmed non synonymous, exonic/splice variants and observed at a frequency lower than 0,1% in population. Further confirmation of detected variants was performed with comparison to public databases (cbioportal, tumorportal), and potential pathogenicity was evaluated with 4 different public algorithms. We present here the results obtained from the first 35 matched PT/M samples (liver mets 68%), focusing analysis on 40 genes including PIK3CA (20 genes), ER (6 genes) and MAPK (11 genes) pathways, RUNX1, CDH1 and TP53 genes.

Results

Patients characteristics are representative of patients with first line metastatic breast cancer (Comte et al, ASCO 2016#550). Among the 40 genes analyzed in the 70 samples, we detected 134 somatic mutations (70 in PT and 64 in M) including 15 indels and 119 SNV. Among these 134 mutations there were 74 different mutations (66SNV and 8 indels) classified pathogenic for 26 and of unknown pathogenicity for 48 of them. We detected at least 1 mutation in 31 PT and in 28 M. Median numbers of mutations were 1 in PT (range 1-9) and 1 in M (range1-22) samples (p=0.295, Wilcoxon rank sum test). Top ten mutated genes in PT included PIK3CA, TP53, NCOR1, NF1, GATA3, CDH1, ERBB3, PTEN, HRAS, INPP4B. In M samples, the 10 top genes were PIK3CA, TP53, ERBB3, AKT3, CDH1, ERBB4, GATA3, INPP4B, MET, MTOR. Only 3 ESR1 mutations were detected, including 1 PT/M pair and 1 M. Beyond highly shared PIK3CA and TP53 mutations, overall crude PT/M discordance rate was 31%. Analysis by histological subtypes showed PT and M specific mutational profiles, suggesting a role in ERB gene family (notably ERBB3) and MAPK driven pathways in early metastatic progression. Specific metastatic site analysis suggested enrichment in MAPK pathway mutations in liver metastases when compared to other sites. Variant allelic fractions were globally not significantly different between PT and M samples.

Conclusion

In this prospective multicenter series of systematic biopsies of first metastases, we report a targeted mutational analysis of matched PT and M samples not modified by previous therapy exposure. Early analyses suggest specific genotypical changes according to tumor subtype and/or metastatic site. Extended and updated results will be reported at the meeting.

Citation Format: Cottu PH, Boulai A, Callens C, Baulande S, Legoix-Ne P, Bernard V, Vincent-Salomon A, Benhamo V, Brain EGC, Chemlali W, Campone M, Bachelot TD, Giacchetti S, Bonneterre J, Bidard F-C, Servois V, Comte A, Belin L, Sigal B, Bièche I. Comparison of mutational landscapes of primary breast cancer and first metastatic relapse: Results from the ESOPE study [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr PD1-06.