Abstract 2O
Background
Localized breast cancer (LBC) embraces multiple genetic mutagenic events during tumor development. Copy number variant (CNV) and single-nucleotide variant (SNV) generated distinct cellular subclones within the same tumor, leading to intra-tumoral heterogeneity (ITH). Although ITH is a risk factor for therapy-resistance and relapse, its mechanism at the single-cell resolution remains unknown. Here, we study the genetic bases underlying ITH in a cohort of HR+/HER2–, node-positive LBC.
Methods
ScDNA-seq was performed on intact nuclei isolated from frozen tumors using TAPESTRI platform. ITH was assessed using CNV- and SNV-based clusterization (CBC and SBC, respectively). Correlations between ITH and clinical data, including molecular subtypes, relapse and survivorship were examined.
Results
The cohort contains 138 patients, in which 36.7% (n = 51) were luminal A and 62.3% (n = 86) luminal B. Overall, 7.3% (n = 10) and 27.5% (n = 38) of patients developed local and distant relapses with a survival rate of 63.8% (n = 90). Intra-tumor clusters varied from 1 to 4 in CBC and from 1 to 5 in SBC, with 63.8 % patients (n = 88) in CBC and 73.9% patients (n = 102) in SBC having at least 2 clusters. In CBC, specific gene amplifications and deletions occurred mostly in tumors having 2 and 3 clusters, in which deletion events were more prevalent in the 3-cluster tumors and in luminal B than luminal A. We captured CBC-specific genes including AKT1, ATM, KRAS, and PIK3CA, which associated with 2-cluster tumors; and BRCA1, ERBB2, GATA3, IDH1 and NRAS with 3-cluster tumors. In SBC, tumors having more than 2 clusters occurred more frequently in luminal B than luminal A (Wilcoxon test, p = 0.015). Notably, mutational phylogenies in luminal A and B tumors were different, implying SNV-derived ITH may emerge and evolve independently via distinct mutations and clonal patterns.
Conclusions
CNV- and SNV-clustering based ITH highly prevailed in HR+/HER2– LBC patients at the single-cell resolution. Luminal B tumors carried more deletions and more SNV-clusters than luminal A. By uncovering distinct genetic bases leading to CBC and SBC between the 2 subtypes, our findings could potentially contribute to precision medicine as an early diagnosis and prognosis tool.
Editorial acknowledgement
Clinical trial identification
Legal entity responsible for the study
Institut Gustave Roussy.
Funding
Institut Gustave Roussy.
Disclosure
All authors have declared no conflicts of interest.
Resources from the same session
1O - SPICE: Probabilistic reconstruction of copy-number evolution in metastatic cancers
Presenter: Abigail Bunkum
Session: Proffered Paper session 1
Resources:
Abstract
Invited discussant
Presenter: Christos Sotiriou
Session: Proffered Paper session 1
Resources:
Slides
Webcast
Q&A
Presenter: Andres Cervantes
Session: Proffered Paper session 1
Resources:
Webcast
3O - Are targeted therapies effective in the relapsed or metastatic cancer setting? A Cochrane meta-analysis
Presenter: Farasat Kazmi
Session: Proffered Paper session 1
Resources:
Abstract
Slides
Webcast
4O - Ultra-fast deep-learned CNS tumor classification during surgery
Presenter: Bastiaan Tops
Session: Proffered Paper session 1
Resources:
Abstract
5O - Deciphering resistance mechanisms in cancer: Insights from the final report of MATCH-R study with focus on molecular drivers
Presenter: Kristi Beshiri
Session: Proffered Paper session 1
Resources:
Abstract
Slides
Webcast
Invited discussant
Presenter: Emre Kocakavuk
Session: Proffered Paper session 1
Resources:
Slides
Webcast
Q&A
Presenter: Andres Cervantes
Session: Proffered Paper session 1
Resources:
Webcast