Array comparative genomic hybridization analysis discloses chromosome copy number alterations as indicators of patient outcome in lymph node-negative breast cancer

Ryoko Kikuchi-Koike,
Kazunori NagasakaEmail author View ORCID ID profile,
Hitoshi Tsuda,
Yasuyuki Ishii,
Masaru Sakamoto,
Yoshihiro Kikuchi,
Shiho Fukui,
Yuko Miyagawa,
Haruko Hiraike,
Takayuki Kobayashi,
Takayuki Kinoshita,
Yae Kanai,
Tatsuhiro Shibata,
Issei Imoto,
Johji Inazawa,
Osamu Matsubara and
Takuya Ayabe

BMC Cancer201919:521

https://doi.org/10.1186/s12885-019-5737-7

Received: 14 January 2019
Accepted: 21 May 2019
Published: 30 May 2019

Open Peer Review reports

Abstract

Background

Patients with lymph node metastasis-negative (pN0) invasive breast cancer have favorable outcomes following initial treatment. However, false negatives which occur during routine histologic examination of lymph nodes are reported to underestimate the clinical stage of disease. To identify a high-risk group in pN0 invasive breast cancer, we examined copy number alterations (CNAs) of 800 cancer-related genes.

Methods

Using array-based comparative genomic hybridization (CGH) in 51 pN0 cases (19 relapsed and 32 non-relapsed cases), the positivities of specific gene CNAs in the relapsed and non-relapsed groups were compared. An unsupervised hierarchical cluster analysis was then performed to identify case groups that were correlated with patient outcomes.

Results

The cluster analysis identified three distinct clusters of cases: groups 1, 2, and 3. The major component was triple-negative cases (69%, 9 of 13) in group 1, luminal B-like (57%, 13 of 23) and HER2-overexpressing (26%, 6 of 23) subtypes in group 2, and luminal A-like subtype (60%, 9 of 15) in group 3. Among all 51 cases, those in group 1 showed significantly worse overall survival (OS) than group 2 (p = 0.014), and 5q15 loss was correlated with worse OS (p = 0.017). Among 19 relapsed cases, both OS and relapse-free survival (RFS) rates were significantly lower in group 1 than in group 2 (p = 0.0083 and 0.0018, respectively), and 5q15 loss, 12p13.31 gain, and absence of 16p13.3 gain were significantly correlated with worse OS and RFS (p = 0.019 and 0.0027, respectively).

Conclusions

As the target genes in these loci, NR2F1 (5q15), TNFRSF1A (12p13.31), and ABCA3 (16p13.3) were examined. 5q15 loss, 12p13.31 gain, and absence of 16q13.3 gain were potential indicators of high-risk recurrence and aggressive clinical behavior of pN0 invasive breast cancers.