Abstract PS4-01-09: How does proteomics-based prognostic testing compare to genomic tests and online clinical risk prediction tools in early breast cancer patients?

Abstract Objective: ∼70% of HR+/HER2- early breast cancer have low risk of breast cancer recurrence. Hence, prognostication in these patients to assess chemotherapy benefit is a huge value add. Online tools such as Nottingham prognostic index (NPI), PREDICT etc are often used as they are quick and free. However use of genomic prognostic tests like Oncotype DX (ODX), Mammaprint (MP) etc is increasing. CanAssist Breast (CAB) is a proteomics-based prognostic test that uses an AI-based algorithm to segregate EBC patients as ‘low or high’ risk for recurrence. CAB is validated in global studies and clinically used on ∼10,000 patients to date in the Indian subcontinent, UAE, Turkey, Iran, and Saudi Arabia. Comparative analysis of prognostic tests is important to evaluate the relative performance of different prognostic tests in a population and assess how a new test performs in relation to established, validated ones. We have compared CAB with NPI, PREDICT, ODX, and MP, and here we showcase the results of those comparative studies. Methods: A patient cohort of 1474 from Europe, India and US was used to compare CAB with NPI and PREDICT. NPI risk groups were categorized into three prognostic groups: good (GPG-NPI index ≤ 3.4), moderate (MPG 3.41-5.4), and poor (PPG > 5.4). Patients with chemotherapy benefit of < 2% were classified as low risk and ≥ 2% high risk by PREDICT. CAB uses a cut-off of 15.5 to stratify patients into low risk (≤15.5) and high risk (>15.5) categories. Agreement between CAB and NPI/PREDICT risk groups were assessed by kappa coefficient. Retrospective comparison of risk stratification by CAB with ODX and MP was done with 109 (US and India) and 43 (EU) patients, and prospectively with a total of 116 Turkish patients- 58 patients in each group. Accuracy/ negative predictive value was calculated using MedCalc. Concordance of CAB with ODX or MP was calculated using the overall percentage agreement. Results: Risk proportions generated by all tests were: CAB low:high 74:26; NPI good:moderate:poor prognostic group- 38:55:7; PREDICT low:high 63:37. 65% of NPI-MPG patients were called low risk by CAB. From PREDICT high risk patients, CAB segregated 51% as low risk, thus preventing over-treatment in these patients. Overall, there was a fair agreement between CAB and NPI [κ=0.31(0.278-0.346)] / PREDICT [κ=0.398 (0.35-0.446)], with a concordance of 97% / 88% between CAB and NPI/PREDICT low risk categories. In cohorts with mostly T1N0 patients, NPI and PREDICT segregated more as low risk compared to CAB, suggesting that T1N0 patients with aggressive biology are missed by online tools but not by CAB. Comparison of CAB with ODX retrospectively (n=109) and prospectively (n=58) showed similar low:high risk proportions as 83:17 and 79:21 for CAB; 90:10 and 83:17 for ODX. An overall concordance of 75% and 65%, and low risk concordance of 82% and 77% was observed between both the tests in retrospective and prospective studies. Retrospective (n=43) and prospective (n=58) comparison of CAB with MP showed similar low:high risk proportions of 65:35 and 66:34 for CAB; 56:46 and 52:48 for MP. Overall concordance between the two tests was 75% in the retrospective study and 62% in the prospective study, while low risk concordance was 83% and 77%, respectively. Conclusion: CAB provided unbiased risk stratification across cohorts of various geographies with minimal impact by clinical parameters, whether compared with online tools and tumor based prognostic tests. CAB is useful for all EBC patients and specifically in NPI-MPG and PREDICT high risk patients for making accurate decisions on chemotherapy use. CAB shows good concordance with ODX and MP in low risk categories. This, coupled with high accuracy comparable with ODX, shows that CAB is an excellent, cost-effective, and quick alternative. Citation Format: M. Bakre, T. Durgekar, S. BA, P. Shrivastava, G. Basaran, G. Ozge, T. Korkmaz. How does proteomics-based prognostic testing compare to genomic tests and online clinical risk prediction tools in early breast cancer patients? [abstract]. In: Proceedings of the San Antonio Breast Cancer Symposium 2025; 2025 Dec 9-12; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2026;32(4 Suppl):Abstract nr PS4-01-09.