Computer-Aided Diagnosis of ER-positive, HER2-Negative Breast Cancer on Dynamic Contrast-Enhanced MRI: Associations With Oncotype DX Recurrence Scores.

Abstract

Purpose: To investigate whether kinetic features derived from computer-aided diagnosis (CAD) and their heterogeneity at dynamic contrast-enhanced MRI are associated with the risk of breast cancer recurrence in patients with ER-positive, HER2-negative invasive breast cancer who have undergone the Oncotype DX assay.

Methods: We retrospectively evaluated 227 women (mean age, 53 years) with newly diagnosed ER-positive, HER2-negative breast cancer who had undergone preoperative MRI and the Oncotype DX assay between 2018 and 2022. The enhancement kinetic parameters and their kinetic heterogeneity for each breast cancer were assessed using a CAD system and correlated with the Oncotype DX RS.

Results: Of the 227 cancers, 61 (26.9%) had low (< 11), 141 (62.1%) had intermediate (11-25), and 25 (11.0%) had high (≥ 26) RS. The high-risk group (RS ≥ 26) had significantly higher mean Angio volume and kinetic heterogeneity on CAD compared to the nonhigh-risk group (RS < 26) (2.52 cm³ vs. 1.42 cm³, P = .023; .78 vs. .62, P = .025). Multivariate regression showed that higher kinetic heterogeneity (odds ratio [OR] 12.57, P = .024), high histologic grade (grade 3; OR 21.30, P< .001), progesterone receptor negativity (OR 21.30, P = .005) and higher Ki-67 (≥ 25%; OR 21.30, P= .006) were associated with a high-risk of recurrence.

Conclusion: Higher kinetic heterogeneity on CAD was associated with an increased risk of recurrence in patients with ER-positive, HER2-negative breast cancer.