Supplementary Material for: Comparison of Prosigna® assay and uPA/PAI-1 results in EBC patients

Background: The decision-making process regarding adjuvant treatment in estrogen-receptor-positive (ER+), HER2 negative early-stage breast cancer (EBC) is often not sufficient, when based on established clinicopathological parameters. Additional prognostic tests could facilitate these decisions. This paper compares the well-validated invasion marker uPA/PAI-1 with the Prosigna® assay, which is based on the PAM50 gene signature. Prosigna® also provides a prognostic risk assessment (risk of recurrence, Prosigna®-ROR), taking tumor burden and the intrinsic molecular subtype into account. Method: From 10/2013 to 04/2014, we selected 42 postmenopausal EBC patients (ER+/HER2 ) from the database of the Breast Centre of the University of Munich (LMU). In the context of therapy decision-making, uPA/PAI-1 testing had already been performed on the selected patients' fresh frozen tumor tissue. The patient's data were pseudonymized, and the Prosigna® assay was performed retrospectively using archived Formalin-fixed paraffin-embedded (FFPE) tumor samples. A five-year follow-up was carried out in March 2021. Results: All patients (n=42) had ER+/HER2-negative invasive breast cancers. According to Prosigna® 22 (52.4%) tumors were classified as low-risk, 14 (33.3%) as intermediate-risk and 6 (14.3)% as high-risk. Among the 22 Prosigna® low-risk tumors, 13 (59.1%) were classified uPA/PAI 1 low-risk and 9 (40.9%) uPA/PAI 1 high-risk. The 14 Prosigna® intermediate-risk tumors divided into 6 (42.9%) uPA/PAI-1 low-risk and 8 (57.1%) uPA/PAI 1 high-risk. Among the six patients classified as high-risk using Prosigna® ROR one (16.7 %) was classified as low-risk the other 5 (83.3%) as high-risk using uPA/PAI 1. The comparison of the intrinsic subtypes (Luminal A, Luminal B, basal-like) based on Prosigna® assay with the risk groups of uPA/PAI 1 showed only poor correlation. A five-year follow-up with the endpoints ipsilateral recurrence, contralateral recurrence, distant metastasis, and death could be obtained in 67% (28/42) of the initially included patients. In two of these cases, one of these events occurred. The risk stratification of uPA/PAI 1 and Prosigna® did not match in these tumors. Conclusion: In luminal EBC, there is only limited concordance between risk group classification according to the Prosigna®-ROR, Prosigna®-Luminal subtypes, and risk classification by uPA/PAI 1. Risk classification by uPA/PAI 1 compared to the Prosigna®-assay resulted in a larger high-risk group with a clear recommendation for adjuvant CTX. A larger study population in a prospective setting and more detailed outcome data would be necessary to understand the clinical relevance of the observed discrepancies. Only one evidence-based guideline-recommended prognostic test should be used in a single patient for chemotherapy decision making.