Patterns of genomic change in residual disease after neoadjuvant chemotherapy for estrogen receptor-positive and HER2-negative breast cancer

Abstract

Background: Treatment of patients with residual disease after neoadjuvant chemotherapy for breast cancer is an unmet clinical need. We hypothesised that tumour subclones showing expansion in residual disease after chemotherapy would contain mutations conferring drug resistance.

Methods: We studied oestrogen receptor and/or progesterone receptor-positive, HER2-negative tumours from 42 patients in the EORTC 10994/BIG 00-01 trial who failed to achieve a pathological complete response. Genes commonly mutated in breast cancer were sequenced in pre and post-treatment samples.

Results: Oncogenic driver mutations were commonest in PIK3CA (38% of tumours), GATA3 (29%), CDH1 (17%), TP53 (17%) and CBFB (12%); and amplification was commonest for CCND1 (26% of tumours) and FGFR1 (26%). The variant allele fraction frequently changed after treatment, indicating that subclones had expanded and contracted, but there were changes in both directions for all of the commonly mutated genes.

Conclusions: We found no evidence that expansion of clones containing recurrent oncogenic driver mutations is responsible for resistance to neoadjuvant chemotherapy. The persistence of classic oncogenic mutations in pathways for which targeted therapies are now available highlights their importance as drug targets in patients who have failed chemotherapy but provides no support for a direct role of driver oncogenes in resistance to chemotherapy. CLINICALTRIALS.GOV: EORTC 10994/BIG 1-00 Trial registration number NCT00017095.

Conflict of interest statement

The authors declare no competing interests.

© 2021. The Author(s), under exclusive licence to Springer Nature Limited.

Figures

Fig. 1. Commonest driver mutations before and after treatment.

The mutation count is non-redundant (i.e., genes are only counted once if multiple mutations were identified in a tumour, or the same mutation was present in multiple samples from a tumour).

Fig. 2. Variants gained and lost after neoadjuvant chemotherapy.

a Number of coding variants gained and lost in each tumour. b, c Examples of tumours showing different patterns of clonal change after treatment (b, PD30309, FEC arm; c, PD26285, T-ET arm). The size of the plotting symbol reflects the Fisher P value (mutations supported by fewer reads have a smaller symbol).

Fig. 3. Frequency and VAF of the most frequently mutated oncogenes and tumour suppressor genes before and after neoadjuvant chemotherapy.

a Percentage of tumours with specific genes mutated before and after treatment. b–f VAF for the most commonly mutated drivers. There are substantial changes in both directions after treatment. The size of the plotting symbol reflects the Fisher P value (mutations supported by fewer reads have a smaller symbol).

Fig. 4. Copy number variants in pre- and post-neoadjuvant chemotherapy samples.

a Heatmap showing regions gained and lost coloured red and blue, respectively. The y axis is ordered by patient ID with pre- followed by post-treatment samples. b Log2 ratio before and after treatment for the most commonly amplified genes (CCND1 and FGFR1). The dashed grey line corresponds to a copy number of 4. The dotted grey line corresponds to likely gains (3 sd above a copy number of 2). If two pretreatment samples were tested, the second sample is indicated by an X and joined to the first sample by a horizontal line. Note that the FGFR1 sample with log ratio 0.6 pre/1.6 post was not scored as amplified because it was not a focal gain. FEC: 5-fluorouracil + epirubicin + cyclophosphamide x6; T-ET: docetaxel x3 then docetaxel + epirubicin x3.

Fig. 5. Pathway analysis.

Driver oncogenes mutated or amplified after treatment are grouped according to the biological pathway. CCND1, FGFR1, AURKA, MDM2 and NCOA3 are within amplified regions; all of the other genes shown contained driver mutations. Note that for amplified regions the true driver may be a nearby gene rather than the one shown.