Clonal Evolutionary Analysis during HER2 Blockade in HER2-Positive Inflammatory Breast Cancer: A Phase II Open-Label Clinical Trial of Afatinib +/- Vinorelbine

Abstract

Background: Inflammatory breast cancer (IBC) is a rare, aggressive form of breast cancer associated with HER2 amplification, with high risk of metastasis and an estimated median survival of 2.9 y. We performed an open-label, single-arm phase II clinical trial (ClinicalTrials.gov NCT01325428) to investigate the efficacy and safety of afatinib, an irreversible ErbB family inhibitor, alone and in combination with vinorelbine in patients with HER2-positive IBC. This trial included prospectively planned exome analysis before and after afatinib monotherapy.

Methods and findings: HER2-positive IBC patients received afatinib 40 mg daily until progression, and thereafter afatinib 40 mg daily and intravenous vinorelbine 25 mg/m2 weekly. The primary endpoint was clinical benefit; secondary endpoints were objective response (OR), duration of OR, and progression-free survival (PFS). Of 26 patients treated with afatinib monotherapy, clinical benefit was achieved in 9 patients (35%), 0 of 7 trastuzumab-treated patients and 9 of 19 trastuzumab-naïve patients. Following disease progression, 10 patients received afatinib plus vinorelbine, and clinical benefit was achieved in 2 of 4 trastuzumab-treated and 0 of 6 trastuzumab-naïve patients. All patients had treatment-related adverse events (AEs). Whole-exome sequencing of tumour biopsies taken before treatment and following disease progression on afatinib monotherapy was performed to assess the mutational landscape of IBC and evolutionary trajectories during therapy. Compared to a cohort of The Cancer Genome Atlas (TCGA) patients with HER2-positive non-IBC, HER2-positive IBC patients had significantly higher mutational and neoantigenic burden, more frequent gain-of-function TP53 mutations and a recurrent 11q13.5 amplification overlapping PAK1. Planned exploratory analysis revealed that trastuzumab-naïve patients with tumours harbouring somatic activation of PI3K/Akt signalling had significantly shorter PFS compared to those without (p = 0.03). High genomic concordance between biopsies taken before and following afatinib resistance was observed with stable clonal structures in non-responding tumours, and evidence of branched evolution in 8 of 9 tumours analysed. Recruitment to the trial was terminated early following the LUX-Breast 1 trial, which showed that afatinib combined with vinorelbine had similar PFS and OR rates to trastuzumab plus vinorelbine but shorter overall survival (OS), and was less tolerable. The main limitations of this study are that the results should be interpreted with caution given the relatively small patient cohort and the potential for tumour sampling bias between pre- and post-treatment tumour biopsies.

Conclusions: Afatinib, with or without vinorelbine, showed activity in trastuzumab-naïve HER2-positive IBC patients in a planned subgroup analysis. HER2-positive IBC is characterized by frequent TP53 gain-of-function mutations and a high mutational burden. The high mutational load associated with HER2-positive IBC suggests a potential role for checkpoint inhibitor therapy in this disease.

Trial registration: ClinicalTrials.gov NCT01325428.

Conflict of interest statement

I have read the journal's policy and the authors of this manuscript have the following competing interests: CS declares advisory board or speaker fees on laboratory research over the last 3 years for Roche, Pfizer, Celgene, Boehringer Ingelheim, Novartis, Glaxo Smithkline and Eli Lilly. CS sits on the scientific advisory board and holds stock options for Epic Biosciences, APOGEN Biotech, Grail and is a founder of Achilles Therapeutics. RS KG NG and MUF are employees of Boehringer Ingelheim. SRJ has research funding from Pfizer and is on the advisory boards of Novartis, AstraZenaca and Genentech/Roche. All other authors have declared that no competing interests exist.

Figures

Fig 1. Study flow and patient disposition.

This figure describes the study and number of patients in each part of the clinical trial (blue outline), reasons for patients being excluded or discontinuing treatment (purple outline), and number of patients with genomic analysis performed (green outline). PD, progressive disease.

Fig 2. Somatic mutations in HER2-positive IBC.

(A) Top panel shows number of somatic mutations (SNVs and indels) identified across the 22 IBC patients. Data tracks below indicate if patient was: treated with trastuzumab prior to afatinib monotherapy (orange); oestrogen-receptor (ER) or progesterone-receptor (PgR) positive (yellow); derived confirmed clinical benefit from afatinib monotherapy (red); tumour underwent whole-genome doubling (WGD) (pink). Mutational signatures identified in IBC tumours were predominantly age-related (Signatures 1A and 1B) (blue), APOBEC-related (Signatures 2 and 13) (salmon), and others (grey). NEV, not evaluated; NA, no information available. (B) TP53, PIK3CA, AKT1, and ERBB2 mutations identified in samples are indicated if present (blue) or absent (grey). Gain-of-function mutations (TP53 p.R248, PIK3CA p.H1047R, AKT1 p.E17K, ERBB2 p.V777L) are indicated by a yellow dot. Clonal and subclonal mutations are indicated by dark blue and yellow outlines, respectively. Amplifications (≥2x ploidy), gains (≥1 copy number relative to ploidy), and losses (≤1 copy number relative to ploidy) in ERBB2 (HER2), PIK3CA, EGFR, and PTEN are indicated by red, pink, and dark blue, respectively. Somatic activation of PI3K/AKT/mTOR pathway (defined as PIK3CA activating mutation or gain, PTEN deletion, AKT1 mutation) indicated in orange. (C) Plots showing results of GISTIC analysis identifying recurrent focal gains (left panel in red) and losses (right panel in blue); y-axis is genomic position and x-axis is GISTIC q-value; green line represents significance threshold (q-value = 0.25). Gene names are indicated where significantly mutated cancer driver genes were previously associated with the GISTIC peak in a pan-cancer analysis of SCNAs [38]. (D) Box plot showing higher numbers of somatic nonsynonymous (NS) mutations identified in IBC patients compared to non-IBC patients. The band inside the box denotes median. (E) Bar plot showing an enrichment of TP53 mutations in IBC patients versus non-IBC patients. Yellow bar is proportion of gain-of-function TP53 p.R248 mutations. (F) Boxplot showing higher numbers of neoantigens predicted in IBC patients compared to non-IBC patients. Asterisk (*) denotes significant p-value <0.05.

Fig 3. Genomic analysis of tumour biopsies before treatment and following disease progression on afatinib monotherapy.

(A) Somatic mutations (SNVs and indels) identified in pre- and post-treatment biopsies. Green, mutations identified in pre-treatment only; yellow, mutations identified in both pre- and post-treatment; blue, mutations identified in post-treatment only. Data tracks below denote: if patient derived confirmed clinical benefit from afatinib monotherapy (red); amplifications (≥2x ploidy), gains (≥1 copy number relative to ploidy), and losses (≤1 copy number relative to ploidy) in ERBB2 (HER2), EGFR, PIK3CA, and PTEN are indicated by red, pink, and blue, respectively. NEV, not evaluated; NA, no information available. (B) Two main patterns of clonal evolution following afatinib monotherapy observed, either branched evolution or shifting clonal structure. Numbers refer to mutation clusters from PyClone results, also in S10 Fig. T1, pre-treatment biopsy; T2, post-treatment biopsy; CCF, cancer cell fraction.