Abstract Background: Previous researches found that false-positive mammogram results during breast cancer screening are related to age, family history of breast cancer, body mass index, and breast density. In Brazil, there is limited data available on possible factors that may influence the sensitivity of mammography and, consequently, the overall quality of the breast cancer screening program. Aims: This study aims to conduct a sensitivity analysis comparing the results of screening mammograms with diagnosis of breast cancer, taking into account factors that could impact this sensitivity. Methods: All mammogram and biopsy results in the state of São Paulo in 2013 were obtained from the Brazilian Breast Cancer Information System (SISMAMA) database. We obtained information on age (< 50, 50-59, 60-69, or 70+ years), type of lesion (solid, cyst, or calcifications), lesion size (< 10mm, 11-20mm, 21-50mm, or >50mm), lesion boundaries (defined or undefined), lesion characteristics (regular or irregular), and breast density (fatty/predominant fatty or dense/predominant dense). In addition, we evaluated the hormonal therapy usage (yes or no), radiotherapy treatment (yes or no), ethnicity (caucasian, asian, or black), education (illiterate, completed elementary school, completed high school, or completed higher education), skin type (retracted, thickened, or normal), and type of mammogram (simple or computerized). Our analysis focused on mammograms classified as BI-RADS 4, BI-RADS 5, and those with inconclusive results, which subsequently led to a biopsy for diagnosis. We conduct statistical analysis using SPSS v.29 software. To assess the agreement between mammograms and biopsy results (sensitivity), we calculated the proportion of true positive results. We compared the sensitivity across different strata of variables using the Chi-Square test and the V-Cramer measure. Results: A total of 6,064 women with BI-RADS 4, BI-RADS 5, or inconclusive mammogram results underwent biopsy procedures for breast cancer diagnosis. Among these cases, 3,414 (56.3%) were false positives, 1,765 (29.1%) were true positives, and 885 (14.6%) were inconclusive results. The sensitivity of mammographic screening was not significantly associated with ethnicity, education, type of mammogram, skin type, or previous radiotherapy. However, a higher proportion of false positive results was linked to high breast density (71.2%), irregular lesions with defined boundaries (87.5%), particularly calcifications (68.7%) smaller than 10 mm (67.5 %). True positive results, on the other hand, were associated with solid (87.1%) and irregular lesions with undefined boundaries (82.1 %), often larger than 50 mm (57.2%) in fatty breasts (82.7%). A notable proportion of inconclusive results was associated with the presence of dense/predominant dense breasts (88.4%) and irregular lesions with undefined boundaries (87%), particularly solid lesions (89.8%) ranging from 21 to 50 mm in size (60.8%). False positive and inconclusive results were more prevalent among women under 50 years old (79.8%) and those using hormonal therapy (54.8%) compared to true positives. This study has certain limitations, including the availability of data related to lesion density and morphology, as well as body mass index in the SISMAMA database, all of which can potentially impact mammography sensitivity. Furthermore, a notable limitation was a substantial number of missing values for ethnicity and education in the SISMAMA dataset. Conclusion: We identified associations between false-positive and inconclusive results on screening mammograms and factors such as age, breast density, lesion size, lesion types, and hormonal therapy. These results provide valuable information about potential factors that may affect the accuracy of mammography interpretations. Therefore, they must be taken into consideration when planning a breast cancer screening program. Citation Format: Alice Câmara, Lise Cury, Victor Filho. Factors Affecting the Mammography Sensitivity and the Quality of the Breast Cancer Screening Program in the State of São Paulo, Brazil [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO2-29-03.
Read full abstract