Abstract A035: Subtype conversion and misclassification of ipsilateral second breast cancer

Abstract

Abstract Breast cancer is a heterogeneous disease defined by receptor status (estrogen, progesterone, or Her2; ER, PR, Her2) and intrinsic subtype. Tumor evolution may be reflected by shifts in intrinsic subtype, as women with metastatic tumors may display metastases with different subtypes than their primary tumors. Subtype conversion may introduce the possibility for additional treatments or reflect treatment resistance and a need for new therapeutic regimens. Roughly 30% of distant breast cancer metastases show conversion of one or more receptor, but the frequency of such conversion in ipsilateral in-breast recurrences is not well characterized. There are also uncertainties in the assays used to classify tumors, and the extent to which misclassification plays a role in observed conversion is unclear. The aims of this study are to describe the frequency of conversions in ipsilateral breast tumors as a function of time since diagnosis and to determine whether conversions are explained by misclassification. We identified 77 and 922 women with ipsilateral breast tumors in the Carolina Breast Cancer Study (CBCS) and SEER18, respectively. All CBCS tumors were classified by a pathologist as recurrences; SEER18 tumors were included based on laterality because recurrence status was unknown. We cross-tabulated primary and secondary tumor phenotypes based on receptor statuses (CBCS) or intrinsic subtype (SEER18) and calculated the frequency of conversion within categories defined by time between diagnoses (early: <2 years, late: >5 years). We also simulated expected frequencies of change based on a range of test accuracies to assess how misclassification influences estimates. In CBCS, second ipsilateral tumor pairs changed receptor status 52% of the time. Overall, 40% of ER+ tumors, 55% of PR+ tumors, and 50% of Her2+ tumors lost positivity. Early recurrences were more likely to lose ER or PR positivity (85% ER/83% PR loss) than late recurrences (19% ER/50% PR loss) but were less likely to lose Her2 positivity (33% vs 66%). In SEER18, a smaller fraction (16%) of tumors converted to different intrinsic subtypes, but rates of conversion varied by subtype. While only 9% of luminal (Lum) A primary tumors changed subtype, 47% of LumB, 50% of Her2, and 38% of triple negative primary tumors converted. More late-recurring (vs early-recurring) LumA tumors changed subtype (12% vs 7%), whereas early-recurring triple negative tumors changed subtype more than late-recurring triple negative tumors (43% vs 31%). Simulations suggested that accuracy would need to be very low (<60%) to account for changes of these magnitudes. Differences between CBCS and SEER18 may reflect differences in the source populations for the two studies. Ipsilateral breast cancers frequently display time-dependent phenotypic differences from primary tumors, and in some cases these differences may be due to misclassification. Future analyses should consider changes in RNA-based measures of subtype, genomic assessment of recurrences, and how frequency of conversion varies by demographic characteristics. Citation Format: Sarah C. Van Alsten, Melissa A Troester, Colin Begg. Subtype conversion and misclassification of ipsilateral second breast cancer [abstract]. In: Proceedings of the AACR Special Conference on the Evolutionary Dynamics in Carcinogenesis and Response to Therapy; 2022 Mar 14-17. Philadelphia (PA): AACR; Cancer Res 2022;82(10 Suppl):Abstract nr A035.