High concordance of 70-gene recurrence risk signature and 80-gene molecular subtyping signature between core needle biopsy and surgical resection specimen in early-stage breast cancer

Crozier, J. A.; Barone, J.; Whitworth, P.; Cheong, A.; Maganini, R.; Falk, J.; Wei, J.-P.; Mee, S.; Ruby, J.; Yang, S.; Huynh, Y.; Witteveen, A.; Finn, C.; Corcoran, K.; Dreezen, C.; Dauer, P.; Menicucci, A.; Wang, S.; Tran, A.; Yoder, E.; van der Baan, B.; Audeh, W.; Glas, A.

Introduction: With an increase in neoadjuvant therapy recommendations for most early-stage breast cancer patients due to the COVID-19 pandemic, it has become increasingly imperative to ensure that molecular diagnostic assays provide reliable results from preoperative core needle biopsies. Therefore, the objective of this study was to determine the concordance of MammaPrint results (70-gene signature) and BluePrint results (80-gene signature) between core needle biopsies (CNB) and surgical resection (SR) specimens using prospectively collected matched tissues from patients enrolled in the FLEX trial (NCT03053193). Methods: We analyzed 113 matched CNB and SR tumor specimens from women with early-stage breast cancer enrolled in the FLEX trial. Each patient enrolled in the trial receives a MammaPrint recurrence risk classification test with or without BluePrint molecular subtyping. Concordance of MammaPrint is reported using overall percentage agreement, positive predictive value (PPV, High Risk), negative predictive value (NPV, Low Risk), and Cohens kappa coefficient. Additionally, correlations between sample types are reported using Pearson correlation coefficient. Results: We found good concordance for MammaPrint results between CNB and SR tumor samples (90.3%, k = 0.803), with a 95.1% NPV and 84.6% PPV. There was also a strong correlation of MammaPrint indices between CNB and SR specimens (r = 0.94). In addition to our primary objective, we determined the concordance of BluePrint subtyping in the matched tumor samples, and found excellent concordance (98.2%) and strong correlation of BluePrint scores within each subtype. Conclusion: CNB samples demonstrated overall high concordance with paired SR samples for MammaPrint risk classification, ensuring that physicians are provided with accurate prognostic information for therapy decisions based on testing of core biopsy tissue. Further, BluePrint molecular subtyping also had good concordance between the sample types, outperforming concordance rates based on traditional IHC based classification. Overall, with an increase in neoadjuvant therapy, physicans and patients can be assured that MammaPrint and BluePrint provide reliable results that guide timely and appropriate therapies using preoperative CNB specimens.