Deep Transfer Learning for Breast Cancer Detection in Underserved Regions - Takeaways - MDSpire

Deep Transfer Learning for Breast Cancer Detection in Underserved Regions

  • By

  • Obaid, Mahmoud

  • ODEH, SUHAIL

  • Ashqar, Huthaifa I.

  • Abumwais, Allam

  • Hodrob, Rami

  • June 22, 2026

  • 0 min

Share

  • 1

    Breast cancer is the leading cancer in women in Palestine, accounting for over 34% of cases and 12% of cancer-related deaths.

  • 2

    The study introduces a two-step deep learning method for breast cancer detection in mammograms, targeting low-resource settings.

  • 3

    The first stage utilizes a U-Net architecture with a VGG16 encoder for lesion segmentation, trained on the CBIS-DDSM dataset.

  • 4

    The second stage employs a VGG16 classifier to differentiate benign from malignant tumors, achieving 91% accuracy on the Palestine dataset.

  • 5

    The proposed framework demonstrates feasibility in low-resource environments but requires validation on a larger local dataset.

Original Source(s)

Deep Transfer Learning for Breast Cancer Detection in Underserved Regions

  • by Obaid, Mahmoud , ODEH, SUHAIL , Ashqar, Huthaifa I., Abumwais, Allam , Hodrob, Rami

  • June 22, 2026

Related Content

Antibody-drug conjugates in colorectal cancer: current landscape and future perspectives from clinical trials

Multiparametric MRI-based nomogram integrating clinicopathological factors for predicting HER2 expression status in breast cancer

Durable disease control after multimodal therapy and platinum-based chemotherapy rechallenge in advanced clear cell adenocarcinoma of the urethra with lymph node metastases: a case report