Abstract 45P
Background
Gene expression profiling (GEP)-based prognostic signatures are rapidly integrated into clinical decision-making for systemic management of breast cancer patients. However, GEP remains relatively underdeveloped for loco-regional risk assessment. Yet, local-regional recurrence (LRR), especially early after surgery, is associated with poor survival.
Methods
GEP was performed on two independent luminal-like breast cancer cohorts of patients developing early (≤ 5 years after surgery) or late (>5 years) LRR and used, by training and testing approach, to build a gene signature able to intercept women at risk of developing early LRR. The GEP data of two in silico datasets and of a third independent cohort were used to explore its prognostic value.
Results
Analysis of the first two cohorts lead to the identification of three genes, CSTB, CCDC91 and ITGB1, whose expression, derived by principal component analysis, generated a 3-gene signature significantly associated with early LRR in both cohorts (p-value: <0.001 and 0.005, respectively), overcoming the discriminatory capability of age, hormone receptor status and therapy. Remarkably, the integration of the signature with these clinical variables leads to an AUC of 0.878 (95% CI: 0.810-0.945). In in silico datasets we found that the 3-gene classifier retained its association, showing higher values in the early relapsed patients. Moreover, in the third additional cohort, the signature significantly associated with relapse-free survival (HR 2.88; 95% CI 1.11-7.49, p-value: 0.030).
Conclusions
Our 3-gene signature represents a new exploitable tool to aid treatment choice in patients with luminal-like breast cancer at risk of developing early recurrence.
Legal entity responsible for the study
The authors.
Funding
AIRC (Italian Association for Cancer Research), Grants to MPC (AIRC IG #18425 and AIRC IG #24363) and Italian Ministry of Health, “Ricerca corrente” Funds.
Disclosure
All authors have declared no conflicts of interest.