Associations of Breast Cancer Method of Detection With Tumor Characteristics, Treatments, and Overall Survival: A Propensity-Score-Matched Analysis.

Abstract

BACKGROUND. A contributor to variable breast cancer screening guidelines has been limited research assessing associations of the method of cancer detection with cancer outcomes. OBJECTIVE. The purpose of this study is to compare tumor characteristics, treatments, and survival outcomes between patients with breast cancer detected by mammographic screening versus by clinical symptoms. METHODS. This retrospective study included patients with biopsy-proven breast cancer from January 1, 2010, to January 1, 2020. For each patient, the method of detection was classified as screening-detected (i.e., screening mammography in the absence of symptoms) or clinically detected (i.e., presentation with symptoms leading to diagnostic imaging). Propensity-score matching was performed between the screening-detected and clinically detected cohorts by age, race, ethnicity, insurance status, and year of diagnosis. Tumor characteristics, treatments, and overall survival rates were compared between cohorts. Overall survival was compared between cohorts using Kaplan-Meier curves, incorporating a 0.5-year lead time bias adjustment in the screening-detected cohort. RESULTS. Before propensity-score matching, the analysis included 1460 patients with breast cancer (screening-detected cohort, 932 patients; clinically detected cohort, 528 patients). After propensity-score matching, each cohort included 507 patients (mean age in the screening-detected and clinically diagnosed cohorts of 61.2 and 62.0 years, respectively). Tumor size measured 1-10 mm in 42.0% of screening-detected cancers versus in 13.0% of clinically detected cancers (p < .001). Frequencies of noninvasive disease, localized disease, regional disease, and distant metastatic disease were 35.5%, 54.0%, 9.9%, and 0.0% in the screening-detected cohort versus 8.9%, 47.3%, 34.1%, and 7.7% in the clinically detected cohort, respectively (p < .001). Screening-detected and clinically detected cohorts showed significant differences in frequencies of surgical therapies (lumpectomy, 66.1% vs 39.3%; mastectomy, 22.1% vs 34.3%; p < .001), radiation therapy (50.1% vs 41.2%; p = .006), and chemotherapy (15.0% vs 40.2%; p < .001). The screening-detected in comparison with the clinically detected cohort showed significantly greater 5-year survival rate (94.4% vs 79.6%; p < .001) and 10-year survival rate (82.7% vs 66.1%; p < .001). Kaplan-Meier curves indicated significantly greater overall survival in the screening-detected cohort (p < .001), with a progressively increasing difference in survival with increasing time since diagnosis. CONCLUSION. Screening-detected cancers in comparison with clinically detected cancers were associated with smaller size, earlier stage, less-invasive therapies, and improved overall survival. CLINICAL IMPACT. The findings provide strong evidence supporting mammographic screening.