ESMO 2017-GENETICS OF TNBC

Patients experiencing disease progression after being treated with a neoadjuvant EGFR targeting agent plus chemotherapy for triple negative breast cancer (TNBC) showed distinctive genomic alterations rather than a comprehensive genomic pattern, according to findings from a hypothesis-generating study presented during ESMO 2017, the Annual Congress of the European Society for Medical Oncology in Madrid, Spain.

Nina Radosevic-Robin, from French National Institute for Health and Medical Research (INSERM) group U1240 (University Clermont Auvergne, Centre Jean Perrin, Department of Pathology) in Clermont-Ferrand, France and colleagues from Memorial Sloan Kettering Cancer Centre in New York, NY, USA investigated the possible genomic determinants of disease recurrence in two neoadjuvant trials (NCT00933517 and NCT00600249) testing the activity of the anti-EGFR antibodies panitumumab and cetuximab combined with chemotherapy in locally advanced TNBC.

The investigators analysed pre-treatment tumour tissue samples and post-treatment residual tumours by next-generation sequencing (NGS), using a targeted exome panel (MSK-IMPACT) containing 410 cancer-related genes. The analysis included data of 15 patients who achieved pathological complete response (pCR) and 33 patients who did not.

In the analysed cohort, 9 out of 33 patients without pCR experienced recurrence within 2 years after surgery.

Five of those patients experienced early fatal metastatic recurrences (within less than one year post–surgery). Five other patients (1 pCR, 4 non-pCR) demonstrated later non-fatal recurrence, at more than one but less than 2 years post-surgery.

Twenty-three patients having residual tumours did not experience a recurrence up to 5 years post-surgery.

Among the 15 patients achieving pCR, one patient had a later non-fatal recurrence consisting of a solitary brain metastasis.

Analysis of patients having disease recurrence yielded patient specific genomic alterations

Most frequent genomic alterations by group of recurrence.

© Nina Radosevic-Robin.

The association between disease recurrence and NGS findings was analysed in a case-by-case fashion, which yielded patient-specific genomic alterations, without obvious genomic pattern predictive of recurrence.

Interestingly, SOX9 amplification was found in tumours of 3 out of 5 patients with early fatal metastatic recurrence and nowhere else in this series.

Post-neoadjuvant residual tumours of each patient experiencing an early fatal metastatic recurrence had an individual set of genomic anomalies. For example, patient one had no mutations, but SOX9, AKT1 and TGFBR2 amplified, together with TERT loss. Patient two had SOX9, RAD21 and NOTCH2 amplified, as well as mutations in PTEN, PIK3CA, ERBB3 and ARID1B. Patient three had SOX9 and MYC amplified, PTEN lost and KDM6A mutated. Patient four had mutated RAF1, FGFR2, MLL2, and GLI1, with numerous copy number alterations of other genes. Patient five had SETD2 mutation, HIST2H3D, HIST2H3C, MCL1 and EZH2 amplification, as well as loss of MAP3K1.

The patient who achieved pCR but later developed a single brain metastasis had BRCA1, MLL2, CDK12, and PPM1D mutation, in the pre-therapy sample.

Various genomic aberrations were detected in other non-pCR patients, for example 3 different activating mutations in PIK3CA in one single residual tumour, and the post-treatment appearance of HRAS G12S mutation or PARP1 amplification.

Conclusions

This study draws attention at SOX9 amplification as a potential mechanism of resistance to therapy and metastatic progression of TNBC. Although small, the study confirmed important inter-tumour heterogeneity of TNBC and the need to interpret NGS findings in the context of other tumour features, such as histology, gene and protein expression, immune infiltrates.

Aleix Prat of the Medical Oncology Department, Hospital Clínic of Barcelona, University of Barcelona who discussed the study results said although strong preclinical data exists, further studies looking at the mechanistic role of SOX9 in patients with breast cancer are needed. The prognostic value of SOX9 amplification in TNBC beyond tumour load, proliferation and immune infiltration must be established.

From a predictive point of view, SOX9 itself might be difficult to target. Instead, targeting SOX9 upstream/downstream signalling pathways should be explored (e.g. PML with arsenic trioxide).