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

Prediction of Oncotype DX Recurrence Score Based on Systematic Evaluation of Ki-67 Scores in Hormone Receptor-Positive Early Breast Cancer

PURPOSE

Oncotype DX (ODX) predicts the risk of recurrence and benefits of adding chemotherapy for patients with estrogen receptor positive (ER+)/human epidermal growth factor receptor 2 negative (HER2-) early-stage breast cancer. We aimed to develop a simplified scoring system using readily available clinicopathological parameters to predict a high-risk ODX recurrence score (RS) while minimizing reproducibility issues regarding Ki-67 index evaluation methods.

METHODS

We enrolled 300 patients with ER+/HER2- early breast cancer, for whom ODX RS data were available in the test set. Using the QuPath image analysis platform, we systematically evaluated the average, hotspot, and hottest spot Ki-67 scores in the test set. Logistic regression analyses were conducted to establish a predictive scoring system for high-risk ODX RS. An independent validation set comprising 117 patients over different periods was established.

RESULTS

Factors such as age ≤ 50 years, invasive ductal carcinoma tumor type, histologic grade 2 or 3, tumor necrosis, progesterone receptor negativity, and a high Roche-analyzed Ki-67 score (> 20) were associated with high-risk ODX RS. These variables were incorporated into our scoring system. The area under the curve of the scoring system was 0.8057. When applied to both the test and validation sets with a cutoff value of 3, the sensitivity of our scoring system was 92%.

CONCLUSION

We successfully developed a scoring system based on the systematic evaluation of Ki-67 scoring methods. We believe that our user-friendly predictive scoring system for high risk ODX RS could help clinicians in identifying patients who may or may require additional ODX testing.

Prediction of Oncotype DX Recurrence Score Using Clinicopathological Variables in Estrogen Receptor-Positive/Human Epidermal Growth Factor Receptor 2-Negative Breast Cancer

PURPOSE

Oncotype DX (ODX) is a well-validated multigene assay that is increasingly used in Korean clinical practice. This study aimed to develop a clinicopathological prediction (CPP) model for the ODX recurrence scores (RSs).

METHODS

A total of 297 patients (study group, n = 175; external validation group, n = 122) with estrogen receptor-positive, human epidermal growth factor receptor 2 (HER2)-negative, T1-3N0-1M0 breast cancer, and available ODX test results were included in the study. Risk categorization as determined by ODX RSs concurred with the TAILORx study (low-risk, RS ≤ 25; high-risk, RS > 25). Univariate and multivariate logistic regression analyses were used to assess the relationships between clinicopathological variables and risk stratified by the ODX RSs. A CPP model was constructed based on regression coefficients (β values) for clinicopathological variables significant by multivariate regression analysis.

RESULTS

Progesterone receptor (PR) negativity, high Ki-67 index, and nuclear grade (NG) 3 independently predicted high-risk RS, and these variables were used to construct the CPP model. The C-index, which represented the discriminatory ability of our CPP model for predicting a high-risk RS, was 0.915 (95% confidence interval [CI], 0.859-0.971). When the CPP model was applied to the external validation group, the C-index was 0.926 (95% CI, 0.873-0.978).

CONCLUSION

Our CPP model based on PR, Ki-67 index, and NG could aid in the selection of patients with breast cancer requiring an ODX test.

Use of a supervised machine learning model to predict Oncotype DX risk category in node-positive patients older than 50 years of age

PURPOSE

The use of the Oncotype DX recurrence score (RS) to predict chemotherapy benefit in patients with hormone receptor-positive/HER2 negative (HR+/HER2-) breast cancer has recently expanded to include postmenopausal patients with N1 disease. RS availability is limited in resource-poor settings, however, prompting the development of statistical models that predict RS using clinicopathologic features. We sought to assess the performance of our supervised machine learning model in a cohort of patients > 50 years of age with N1 disease.

METHODS

We identified patients > 50 years of age with pT1-2N1 HR+/HER2- breast cancer and applied the statistical model previously developed in a node-negative cohort, which uses age, pathologic tumor size, histology, progesterone receptor expression, lymphovascular invasion, and tumor grade to predict RS. We measured the model's ability to predict RS risk category (low: RS ≤ 25; high: RS > 25).

RESULTS

Our cohort included 401 patients, 60.6% of whom had macrometastases, with a median of 1 positive node. The majority of patients had a low-risk observed RS (85.8%). For predicting RS category, the model had specificity of 97.3%, sensitivity of 31.8%, a negative predictive value of 87.9%, and a positive predictive value of 70.0%.

CONCLUSION

Our model, developed in a cohort of node-negative patients, was highly specific for identifying cN1 patients > 50 years of age with a low RS who could safely avoid chemotherapy. The use of this model for identifying patients in whom genomic testing is unnecessary would help decrease the cost burden in resource-poor settings as reliance on RS for adjuvant treatment recommendations increases.