Introduction: Tumor sequencing efforts in breast cancer have historically focused on primary tissue samples. With the development of biomarker-based clinical trials and FDA-approved indications for anti-neoplastic agents in metastatic breast cancer (MBC), there is a deeper interest in elucidating genomic profiles of metastatic disease. We aimed to describe mutational profiles, tumor mutational burden (TMB), and microsatellite instability (MSI) status by subtype in a cohort of patients with MBC.

Methods: We conducted a retrospective analysis of genomic data from 198 patients with MBC who had next generation sequencing (NGS) of tumor tissue by the TEMPUS assay (TEMPUS, Chicago, IL). The majority of patients (n=115) were treated at Northwestern Medicine. Tumors were sequenced to a minimum of 500x depth with classification of genomic alterations as actionable mutations, pathogenic mutations without known treatment, germline alterations, variants of unknown significance, copy number alterations, and chromosomal rearrangements. TMB was defined as nonsynonymous mutations per megabase (mut/MB) sequenced. Statistical analysis of TMB by subtype was conducted using independent samples T-tests. Associations were tested through two-sided Fisher’s exact test.

Results: The cohort of 198 MBC patients had a median age of 51 years (range 23-88) and 52% were white, 9% black, 4% Asian, and 36% unknown. By histology, 52% had invasive ductal carcinoma, 5% had invasive lobular carcinoma, and the remainder of cases were unknown or other. The biopsy site of sequenced tissue was primary breast site in 35% and a metastatic site in the remaining. Sixty nine percent had hormone receptor positive (HR+), 21.3% had triple negative (TN), and 17.2% had HER2-positive (HER2+) MBC. Of the patients with MSI testing (n=183), 180 were MSI-stable, 3 were MSI-equivocal, and none were MSI-high. The median TMB in the whole cohort was 1.8 mut/MB (interquartile range 0.2-2.8), with a median of 2.9 (1.6-4.2) in TN compared to 1.7 (1.0-2.3) in HR+, and 1.4 (0.2-2.5) in HER2+ MBC. The mean TMB for the whole cohort and by TN, HR+, and HER2+ subtypes were 3.0, 4.8, 2.4 and 2.4, respectively. The difference in TMB was statistically significant for TN compared to HR+ (p=0.0006) and equivocal for TN compared to HER2+ (p=0.05). The most frequently seen aberrations in HR+ MBC were PIK3CA (34%), followed by TP53 (28%), ERBB2 (23%), GATA3 (23%), PTEN (13%). In HER2+ MBC alterations were seen in ERBB2 (70%), CDK12 (47%), TP53 (40%), PIK3CA (23%), ABCC3 (16%), PTEN (16%). In TN MBC TP53 (81%), KMT2C (19%), ZFHX3 (16%), PIK3CA (16%), and NCOR2 (14%) were most frequently altered. When comparing the three disease subtypes (HR+, HER2+, TN), statistically significant associations were found with TP53 mutations, most commonly observed in TN cases followed by HER2+ cases (p<0.00001), and ERBB2 amplifications, most commonly observed in HER2+ cases (p<0.00001). PIK3CA alterations were least frequently exhibited with TN compared to HR+ and/or HER2+ tumors (p=0.01). HR+ tumors had a greater proportion of GATA3 (p < 0.0001), and ESR1 (p=0.034) alterations compared to HR- tumors. Conversely, HER2+ tumors had a greater proportion of CKD12 (p< 0.0001) amplifications compared to HER- tumors. ​

Conclusions: MBC demonstrates a rich mutational profile with variations by subtype. Although TMB is lower than in several other solid malignancies, TMB greater than 10 mut/MB was observed in each subgroup, with TN MBC enriched for a higher TMB than HR+ MBC.

VIEW THE PUBLICATION