Diagnostic Algorithm for Accurate detection of breast carcinoma on ultrasound

EXECUTIVE SUMMARY Introduction Breast cancer remains the most common form of cancer among women. Due to its high incidence, technology that improves detection rates needs to be developed. While routine mammography remains the gold standard for breast cancer detection, considerable research is being done to alter breast cancer detection methods and diagnostic processes. The increased interest in ultrasound as a diagnostic tool for breast cancer detection has led to rapid developments in the application of computer-aided diagnosis (CAD) for breast ultrasound. Aim This study aimed to develop and test targeted diagnostic segmentation algorithms utilising imaging software for the ultrasound-based diagnosis of suspicious mass lesions of the breast. Materials and method The first tier of the study was a retrospective, cross-sectional study with a population of 1000 ultrasound images. This included images from Parklane Women’s Imaging Centre, performed between January 2017 and December 2018. All malignant lesions as detected on ultrasound and subsequently confirmed with histology as lobular carcinoma (LC) and ductal carcinoma (DC) were included. The second tier of the study was a prospective case-controlled study with a population estimate of 100 ultrasound images of all suspicious mass lesions detected on ultrasound performed in a radiology department from April 2020 to March 2021. The breast ultrasound images used in the study were samples of diagnostic cases obtained during routine clinical care at the radiology department. The final tier compared the histological output to the MIPARTM software image segmentation analysis to determine if a clinically valid algorithm had been developed. Conclusion The use of MIPARTM software for the segmentation and morphological assessment of breast cancer masses depicted on ultrasound is limited. The proposed research study promoted the further development of CAD software for breast ultrasound. Thesis (PhD (Clinical Anatomy))--University of Pretoria, 2021.Technology Innovation Agency TIAAnatomyPhD (Clinical Anatomy)Unrestricte