BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:Computer-Aided Diagnosis of Breast Cancer via Mammography
DTSTART;VALUE=DATE-TIME:20260518T122500Z
DTEND;VALUE=DATE-TIME:20260518T123300Z
DTSTAMP;VALUE=DATE-TIME:20260624T105808Z
UID:indico-contribution-3806@indico.tlabs.ac.za
DESCRIPTION:Speakers: Andrew Lucio Mbewe (Researcher)\n## Advanced Nuclear
  Science and Technology Techniques (ANSTT6) Workshop\n*18–22 May 2026 
 — iThemba LABS\, Cape Town*\n\n# Computer-Aided Diagnosis of Breast Canc
 er via Mammography\n\n**Andrew Lucio Mbewe and Dr. Maluba Vernon Chisapi**
  \n*University of Zambia\, School of Natural and Applied Sciences\, Lusaka
 \, Zambia*\n\n---\n\n### Abstract\nZambia faces a critical shortage of rad
 iologists and limited mammography infrastructure. This has resulted in sig
 nificant diagnostic delays and in turn to high mortality rates due to late
 -stage breast cancer presentation [1]. This study aimed to develop and val
 idate a computer-aided diagnosis (CAD) system utilizing the YOLOv11 deep l
 earning architecture to automate the detection and classification of breas
 t cancer lesions in mammograms [2]. A quantitative research design was emp
 loyed\, using a dataset of 4\,060 anonymized mammograms collected from Mai
 na Soko Hospital in Lusaka\, Zambia. The model was developed using a progr
 essive training strategy\, incorporating curriculum learning [3] and utili
 zed both manual and model-assisted annotation to identify masses\, calcifi
 cations\, and architectural distortions. The CAD system achieved robust pe
 rformance results\, reaching an accuracy of **71.4%**\, precision of **72.
 2%**\, recall of **70.8%**\, and an F1-score of **71.1%**. While the progr
 essive training strategy successfully improved detection of underrepresent
 ed lesions like architectural distortions\, the model faced challenges wit
 h small lesions and false positive results. These findings demonstrate tha
 t deep learning-based CAD systems can enhance radiological workflows in re
 source-limited environments through fast\, automated screening.\n\n**Keywo
 rds:** Breast Cancer\, YOLOv11\, Computer-Aided Diagnosis (CAD)\, Mammogra
 phy\, Zambia.  \n**Category:** Radiation and Health Physics\n\n---\n\n### 
 References\n1. F. Bray et al.\, “Global cancer statistics 2022: GLOBOCAN
  estimates of incidence and mortality worldwide for 36 cancers in 185 coun
 tries\,” *CA Cancer J Clin*\, vol. 74\, no. 3\, pp. 229–263\, 2024.\n2
 . T. Abd El-Hafeez\, M. A. Shams\, and N. E. Farrag\, “Optimizing YOLOv1
 1 for automated classification of breast cancer histopathology images\,”
  *Scientific Reports*\, vol. 15\, p. 1234\, 2025.\n3. Y. Bengio\, J. Loura
 dour\, R. Collobert\, and J. Weston\, “Curriculum learning\,” in *Proc
 . of the 26th Annual Int. Conf. on Machine Learning*\, pp. 41–48\, 2009.
 \n\nhttps://indico.tlabs.ac.za/event/139/contributions/3806/
LOCATION:NRF-iThemba LABS\, Old Faure Road\, Cape Town Auditorium
URL:https://indico.tlabs.ac.za/event/139/contributions/3806/
END:VEVENT
END:VCALENDAR