Developing a clinical-pathologic model to predict genomic risk of recurrence in patients with hormone receptor positive, human epidermal growth factor receptor-2 negative, node negative breast cancer. (2021)

• Record Type: Journal Article
• Authors: Batra, Atul; Nixon, Nancy A.; Roldan-Urgoiti, Gloria; Hannouf, Malek B; Abedin, Tasnima; Hugh, Judith; King, Karen; Bigras, Gilbert; Steed, Tanner; Lupichuk, Sasha
• Is Part Of: Cancer treatment and research communications. Number 28(2021)
• Publication Date: 2021
• Subjects: Node negative breast cancer -- Gene expression profiling -- Clinical-pathologic model -- OncotypeDX -- Prosigna
• DOI: 10.1016/j.ctarc.2021.100401
• Languages: English
• ISSNs: 2468-2942

Abstract

Highlights: Common gene expression profile (GEP) tests include Prosigna and Oncotypedx. GEP predicts benefit of chemotherapy in hormone receptor positive breast cancer. Our clinical-pathologic model includes tumour size, grade and progesterone receptor status. Model specificity was 98.4%. Use of clinical-pathologic model could reduce the number of patients requiring costly GEP testing. Abstract: Introduction: Patients with hormone receptor (HR)-positive, human epidermal growth factor receptor-2 (HER2)-negative, node negative (NN) breast cancer may be offered a gene expression profiling (GEP) test to determine recurrence risk and benefit of adjuvant chemotherapy. We developed a clinical-pathologic (CP) model to predict genomic recurrence risk and examined its performance characteristics. Methods: Patients diagnosed with HR-positive, HER2-negative, NN breast cancer with a tumour size < 30 mm and who underwent a GEP test [OncotypeDX or Prosigna] in Alberta from October 2017 through March 2019 were identified. Patients were classified as low or high genomic risk. Multivariable logistic regression analysis was performed to examine the associations of CP factors with genomic risk. A CP model was developed using coefficients of regression and sensitivity analyses were performed. Results: A total of 366 patients were eligible (135 were tested using OncotypeDX and 231 with Prosigna). Of these, 64 (17.5%) patients were classified as high genomic risk. On multivariable logistic regression, tumour size > 20 mm (odds ratio [OR], 3.58; 95% confidence interval [CI], 1.84–6.98; P <0.001), low expression of progesterone receptor (OR, 3.46; 95% CI, 1.76–6.82; P <0.001), and histological grade III (OR, 7.24; 95% CI, 3.82–13.70; P <0.001) predicted high genomic risk. A CP model using these variables was developed to provide a score of 0–4. A CP cut-point of 0, identified 56% of genomic low risk patients with a specificity of 98.4%. Conclusions: A CP model could be used to narrow the population of breast cancer patients undergoing GEP testing.