ASPO Abstracts

Mammography's Blind Spot: Getting a grip on the 30% of all breast cancers that are missed by routine screening

Authors: Zhang Z, Irvin VL, Simon MS, Chlebowski RT, Luoh SW, Shadyab AH, Krok-Schoen JL, Tabung FK, Qi L, Edwards BJ, Stefanick ML, Schedin P, Jindal S

Category: Early Detection & Risk Prediction
Conference Year: 2020

Abstract Body:
Purpose of the study: Interval breast cancers (IBC) present symptomatically after a non-suspicious mammographic screen and before the next scheduled mammographic screen. The objective of this study was to investigate the aggressive attributes of IBC by length of inter-screening period compared to screen-detected breast cancers (SBC). Methods: This study utilized data from the Women's Health Initiative (WHI). WHI enrolled 68,132 postmenopausal women (50-79 years) in randomized controlled clinical trials where mammographic screening was protocol mandated. Using a novel algorithm to accurately identify IBC and SBC cases, T-test, χ2 test and Fine-Gray competing risk models were conducted to compare the demographics, tumor characteristics and the breast cancer-specific mortality of IBC vs. SBC.Results: Of 5,455 breast cancer incidences, 3,019 (55.3%) were compliant to protocol screening and comprise our study population. Among these 3,019 patients, 1,050 were IBC cases including 324 diagnosed within 1 year IBC1yr) and 726 diagnosed between 1-2.5 years (IBC1-2.5yr) following recent mammogram, as well as 1,969 SBC cases. Compared to SBC, IBC had significantly larger tumor size, higher clinical stage and more lymph node involvement but no significant difference in molecular subtypes. For breast cancer-specific mortality, IBC1yr had a higher risk compared to SBC (HR=1.69; 95% CI=1.20-2.37) after controlling for molecular subtype, histology, waist-to-hip ratio, menopausal hormone clinical trial arm and dietary modification trial arm. This significance persisted after further controlling for other confounding factors such as age, race/ethnicity etc. (HR=1.64; 95%CI=1.14-2.34). In fact, the elevated risk of death for IBC1yr was associated with increased tumor size and lymph node involvement, as differences were no longer statistically significant after further controlling for these tumor phenotypes (HR=1.24; 95% CI=0.87-1.77) (IBC1-2.5yr: HR=1.09; 95% CI=0.79-1.51).Conclusions: IBC diagnosed within 1 year after the last mammogram compared to SBC have worse survival, possibly due to a more aggressive tumor phenotype comprising larger tumor size and lymph node involvement. Further studies are warranted to confirm these findings and identify improved screening measures.

Keywords: breast cancer, interval, mammography