Development of an Integrated Ultrasound Predictive Model and Framework for Breast Cancer Diagnosis in Zambia

Background: Developed countries bear a higher burden of breast cancer compared to developing ones. However, they have higher survival rates (80-90%) compared to developing countries (57%). In Zambia, this is compounded by the lack of context-specific integrated predictive model and framework. Aim: To develop an integrated breast cancer ultrasound predictive model and framework suitable for clinical use in low-resource settings. Methods: A quantitative analytical cross-sectional study design was used. Participants were selected through systematic random sampling and the breast ultrasound features were documented using the modified BI-RADS Atlas fifth edition. Histology results for the same patients were documented using the modified RCP reporting proformas. Finally, the ultrasound and histology findings were compared for each patient. Results: The findings of this study indicate that productive age, multi-parity, marital status (married), and employment status (employed) are significantly associated with breast cancer (P<0.05). The strength for the associations were 0.6930 and 0.7872 for reproductive age and multi-parity respectively, whereas the strength for the associations of marital and employment status with breast cancer were 0.4455 and 0.4624 respectively. Irregular shape, vertical orientation, hypoechogenicity, complex echopattern, irregular or spiculated margin contours, compression of surrounding tissue, and absence of hyperechogenic spots or a hyperechoic halo or thin capsule were found to be associated with breast cancer (P<0.05). The strength of the association of the preceding ultrasound features with breast cancer was 0.4953, 0.3712, 0.7989, 0.4722, 0.4783, 0.3527 and 0.4540 respectively. Additionally, the study revealed that breast cancer lesions in stage 2 were the most prevalent, and ductal carcinomas were the most common histological type. Conclusion: In countries with limited resources with limited access to imaging diagnostic tools, ...........