Aims:
Assessment of 'risk of recurrence' in ER+ breast cancer patients based on clinical parameters and existing hormone receptor signaling pathway and/or proliferation based biomarkers is insufficient, leading to treatment of majority of patients with chemotherapy. First generation risk identification tests like OncotypeDx and Mammaprint are not impactful in India and SE Asia as are largely prognostic with limited chemotherapy-predictivity and are prohibitively expensive. A cost-effective 'predictive' test which will accurately estimate the 'risk of recurrence' for a 'broader' (node - & +) set of breast cancer patients in low resource settings is urgently required.
Materials and Methods:
Using a retrospective training cohort of 300 node– and node+ patients, we developed 'CanAssist-Breast'- a Morphometric Immunohistochemistry based test comprising 5 biomarkers plus three clinical parameters (Tumor size, grade and node status) to arrive at 'CanAssist-Breast Score'. The risk stratification model was developed using cutting edge support vector based machine learning technology. CanAssist-Breast Score stratifies patients into an all actionable 'low or high' risk for recurrence, with no intermediate zone. CanAssist-Breast biomarkers include cancer stem cell markers, Cadherins, and ATP transporter proteins - all critical players in the various steps of chemotherapy resistance leading to metastasis.
Results:
We validated CanAssist-Breast in accordance with EGAPP recommendations which require that prognostic tests be validated both analytically and clinically prior to being utilized in patients. Analytical validation experiments were performed to assess 'variation' in the outcome prediction due to critical IHC variables. We tested inter-pathologists, sample, operator and laboratory site variation and found high concordance in the outcome predictions across all variables, confirming the robustness and reproducibility of the test.
Extended clinical validation on 1000+ pre and post-menopausal cases shows NPV of 95%. The majority of patients in 'low risk' had Stage 2, Grade 2/3 disease over Stage 1, Grade 1 disease, demonstrating that CanAssist-Breast reclassifies patients who would be considered high risk clinically.
In a head-to-head pilot study of 100 patients with Oncotype Dx, CanAssist-Breast test had about 80% concordance with Oncotype in the 'low risk' category. Importantly, CanAssist-Breast correctly stratified few recurred cases as 'high risk' which were called 'low risk' by Oncotype Dx and thus were not treated with chemotherapy.
Conclusion:
In conclusion, we have developed a robust, accurate and low-cost prognostic test to predict risk of recurrence and enable optimal treatment planning in patients with early stage Breast Cancer.
Citation Format: SP S, Bakre MM, Ramkumar C, Basavaraj C, Attuluri A, Madhav L, Prakash C, Naidu N, Malpani S. Development and validation of a broad-based second generation multi marker “Morphometric IHC” test for optimal treatment planning of stage 1 and 2 breast cancer patients in low resource settings [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr P3-08-10.
