A Signature That May Be Predictive of Early Versus Late Recurrence After Radiation Treatment for Breast Cancer That May Inform the Biology of Early, Aggressive Recurrences

Abstract

Purpose: Unmet clinical needs in breast cancer (BC) management include the identification of patients at high risk of local failure despite adjuvant radiation and an understanding of the biology of these recurrences. We previously reported a radiation response signature and here extend those studies to identify a signature predictive of recurrence timing (before or after 3 years).

Methods and materials: Two independent patient cohorts were used. The training cohort included 119 patients with in-breast tumor recurrence (343 total), and the validation testing cohort had 16 patients with recurrences (112 total). All patients received radiation treatment after breast-conserving surgery. Initial feature selection used Spearman rank correlation, and a linear model was trained and locked before testing and validation. Cox regression was used for univariate and multivariable analyses (UVA and MVA, respectively). Biologically related concepts were identified using gene set enrichment analysis.

Results: Spearman correlation identified 485 genes whose expression was significantly associated with recurrence time (early vs late). Feature reduction further refined the list to 41 genes retained within the signature. In training, the correlation of score to recurrence time was 0.85 (P value < 1.3 × 10^-31) with an area under the curve (AUC) of 0.91. Application of this early versus late signature to an independent BC testing and validation set accurately identified patients with early versus late recurrences (Spearman correlation = 0.75, P value = .001, AUC = 0.92, sensitivity = 0.75, specificity = 1.0, positive predictive value = 1.0, and negative predictive value = 0.8). Unique associations of breast cancer intrinsic subtype to timing of local recurrence were identified. In UVA and MVA the early versus late recurrence signature remained the most significant factor associated with recurrence. Gene set enrichment analysis identified proliferation and epidermal growth factor receptor concepts associated with early recurrences and luminal and ER-signaling pathways associated with late recurrences. Knockdown of genes associated with the early and late recurrences demonstrated novel effects on proliferation and clonogenic survival, respectively.

Conclusions: We report a breast cancer gene signature that may identify patients unlikely to respond to adjuvant radiation and may be used to predict timing of recurrences with implications for potential treatment intensification and duration of follow-up for women with breast cancer treated with radiation.