Breast Cancer Incidence After a False-Positive Mammography Result

Abstract

Importance: False-positive mammography results are common. However, long-term outcomes after a false-positive result remain unclear.

Objectives: To examine long-term outcomes after a false-positive mammography result and to investigate whether the association of a false-positive mammography result with cancer differs by baseline characteristics, tumor characteristics, and time since the false-positive result.

Design, setting, and participants: This population-based, matched cohort study was conducted in Sweden from January 1, 1991, to March 31, 2020. It included 45 213 women who received a first false-positive mammography result between 1991 and 2017 and 452 130 controls matched on age, calendar year of mammography, and screening history (no previous false-positive result). The study also included 1113 women with a false-positive result and 11 130 matched controls with information on mammographic breast density from the Karolinska Mammography Project for Risk Prediction of Breast Cancer study. Statistical analysis was performed from April 2022 to February 2023.

Exposure: A false-positive mammography result.

Main outcomes and measures: Breast cancer incidence and mortality.

Results: The study cohort included 497 343 women (median age, 52 years [IQR, 42-59 years]). The 20-year cumulative incidence of breast cancer was 11.3% (95% CI, 10.7%-11.9%) among women with a false-positive result vs 7.3% (95% CI, 7.2%-7.5%) among those without, with an adjusted hazard ratio (HR) of 1.61 (95% CI, 1.54-1.68). The corresponding HRs were higher among women aged 60 to 75 years at the examination (HR, 2.02; 95% CI, 1.80-2.26) and those with lower mammographic breast density (HR, 4.65; 95% CI, 2.61-8.29). In addition, breast cancer risk was higher for women who underwent a biopsy at the recall (HR, 1.77; 95% CI, 1.63-1.92) than for those without a biopsy (HR, 1.51; 95% CI, 1.43-1.60). Cancers after a false-positive result were more likely to be detected on the ipsilateral side of the false-positive result (HR, 1.92; 95% CI, 1.81-2.04) and were more common during the first 4 years of follow-up (HR, 2.57; 95% CI, 2.33-2.85 during the first 2 years; HR, 1.93; 95% CI, 1.76-2.12 at >2 to 4 years). No statistical difference was found for different tumor characteristics (except for larger tumor size). Furthermore, associated with the increased risk of breast cancer, women with a false-positive result had an 84% higher rate of breast cancer death than those without (HR, 1.84; 95% CI, 1.57-2.15).

Conclusions and relevance: This study suggests that the risk of developing breast cancer after a false-positive mammography result differs by individual characteristics and follow-up. These findings can be used to develop individualized risk-based breast cancer screening after a false-positive result.

Figure 1.. Cumulative Breast Cancer Incidence Among Women With vs Without a False-Positive Result at Index Mammography

Women were followed up from the date of their next screening after their index mammography.

Figure 2.. Hazard Ratios (HRs) for Breast Cancer After a False-Positive Result, by Tumor Characteristics

HRs and 95% CIs are estimated by stratified Cox proportional hazards regression models. P values are derived from the z score for differences in HRs from baseline categories. Luminal A subtype was defined as ERBB2 negative, estrogen receptor (ER) positive, grade I, or ERBB2 negative, ER positive, grade II, Ki-67 ≤10%, or ERBB2 negative, ER negative, grade II, 11% ≤ Ki-67 ≤ 20%, progesterone receptor (PR) ≥20%; luminal B subtype was defined as ERBB2 negative, ER positive, grade II, 11% ≤ Ki-67 ≤ 20%, PR <20%, or ERBB2 negative, ER positive, grade II, Ki-67 ≥21%, or ERBB2 negative, ER positive, grade III; triple-negative subtype was defined as ERBB2 negative, ER negative, and PR negative.

Figure 3.. Hazard Ratios for Breast Cancer After a False-Positive Result, by Side and by Follow-Up Time

Hazard ratios and 95% CIs are estimated by flexible parametric models, adjusting for age and calendar year of mammography, family history of breast cancer, and educational level. The shaded area indicates 95% CIs.