1417 - CanAssist-breast -an immunohistochemistry based test for risk of recurrence prediction for early stage breast cancer patients: a cost-effective and accurate solution for Asia

Background

Current molecular risk stratification tests have helped clinicians to optimize chemotherapy for hormone receptor positive, HER2 negative early stage breast cancer patients leading to huge savings in treatment costs and improved quality of life. However, current tests are not impactful in the Asian subcontinent due to the extreme cost-sensitivity of the market. Aim of this study was to develop and validate a simple, cost-effective and accurate test based on insightful biomarkers involved in tumor recurrence using robust statistical methodology to stratify patients based on risk of recurrence.

Methods

A retrospective cohort of 300 patients, was used to develop ‘CanAssist-Breast’(CAB)- an immunohistochemistry based test comprising 5 biomarkers plus three clinical parameters (Tumor size, node status and grade) using machine learning based algorithm. Retrospective clinical validation on 850+ cases was performed and Kaplan Meier survival analysis, multivariate analysis was performed to assess robustness of the test.

Results

CanAssist-Breast classifies patients into ‘low or high’ risk of recurrence based on recurrence score on a scale of 1-100 with a cut off at 15.5. Clinical validation of CAB showed distant metastasis-free survival (DMFS) was significantly different between low- (DMFS: 95%) and high-risk (DMFS 80%) groups in the validation cohort treated with hormone therapy alone (n = 195) and in the entire validation cohort of 857 patients as well. In multivariate analysis, CAB risk score was the most significant independent predictor of distant recurrence with a hazard ratio of 4.25 (P = 0.009). Patients stratified as high-risk by CAB have 19% chemotherapy benefit. We also show that CAB can further identify discrete low- and high-risk sub-groups within IHC4 intermediate risk group and also in a node and age independent manner.

Conclusions

To our knowledge, CAB is the first machine learning based prognostic risk of recurrence prediction classifier using a combination of unique biomarkers and clinicopathological parameters. We believe that CAB enables accurate treatment planning in early stage HR+/HER2- breast cancer patients in low-resource settings.

Editorial acknowledgement

Clinical trial identification

Legal entity responsible for the study

OncoStem Diagnostics.

Funding

Privately funded by venture capitalist.

Disclosure

M. Bakre: Founder and CEO: OncoStem Diagnostics; co-inventor of the patent; Stock: OncoStem Diagnostics.