To improve breast cancer detection in mammograms through a hybrid deep learning model that enhances tumor segmentation and classification.
Key Findings:
Achieved accuracy of 93.8%, sensitivity of 94.1%, specificity of 92.7%, precision of 91.2%, F1 score of 92.6%, Dice coefficient of 0.89, and IoU of 0.82 on CBIS-DDSM and DDSM datasets.
Quantum-SpinalNet enhances segmentation and classification precision, supporting clinical diagnostic workflows.
Interpretation:
The Quantum-SpinalNet framework demonstrates significant improvements in mammographic breast cancer detection, addressing challenges in tumor segmentation and classification.
Limitations:
Dependence on large, annotated datasets for training remains a challenge.
Potential variability in algorithm performance based on imaging methods and population profiles.
Conclusion:
Quantum-SpinalNet offers a robust and interpretable solution for breast cancer detection in mammograms, with the potential to enhance clinical diagnostic accuracy.
This twice-monthly newsletter highlights recently published research where Dana-Farber faculty are listed as first or senior authors. The information is pulled from PubMed and this issue notes papers published from April 16 - 30.
