Skip to Main Content

An official website of the United States government

Data-Driven Breast Cancer Research: Machine Learning Applications and Deep Learning Approaches for Early Detection and Prognosis of Breast Cancer

Principal Investigator
Name
Manisha Sharma
Degrees
Ph.D.
Institution
KIET GROUP OF INSTITUTIONS GHAZIABAD
Position Title
Associate Professor
Email
About this CDAS Project
Study
PLCO (Learn more about this study)
Project ID
PLCOI-1758
Initial CDAS Request Approval
Dec 16, 2024
Title
Data-Driven Breast Cancer Research: Machine Learning Applications and Deep Learning Approaches for Early Detection and Prognosis of Breast Cancer
Summary
Breast Cancer is one of the most significant causes of mortality from cancer worldwide improvement of survival rate and treatment methods rely heavily on early detection as well as proper prognosis. The study presents the proposal of developing predictive models of Machine Learning (ML), The study presents the proposal of developing predictive models of Machine Learning (ML), Artificial Neural Networks (ANNs), and Convolutional Neural Networks (CNNs) to improve diagnostic capability for breast cancer cases and predict outcomes. Integration of imaging data like mammograms along with Clinical data would encompass patients' demographic characteristics, nature, and characterization of tumor; the evaluation metrics to evaluate the model are precision, sensitivity, and specificity. The Primary goal will be to achieve an accuracy oriented system for support toward effective clinical decisions that translate to improvement in patient outcome.
Introduction:
The largest percentage of cancer diagnoses in the world is breast cancer. Early detection is a cornerstone in effective treatment. Traditional methods of diagnosis include biopsy and imaging interpretation, which depend on subjective assessment and can therefore vary in outcome. Recent advances in artificial intelligence, particularly in Machine Learning (ML) and deep learning, have shown potential in medical applications, improving accuracy and consistency in diagnostics.
Aims

Research Problem
While several ML and deep learning models have been explored for breast cancer diagnosis, the integration of clinical and imaging data in a unified predictive framework remains underdeveloped. This study aims to fill this gap by combining structured data with image-based insights through ML, ANN, and CNN techniques.
Objectives:
Development and validation of predictive models for breast cancer outcomes using ML, ANN, and CNN. Integration of clinical and imaging data to improve prediction accuracy. Comparison of traditional ML algorithms with ANN and CNN models.
Literature Review:
Numerous studies have applied ML for predicting breast cancer outcomes like recurrence and metastasis. Techniques such as SVMs and Random Forests have been useful with clinical data but are incapable of handling imaging data. Deep learning, especially CNNs, have proven useful in breast cancer detection from mammograms; however, there are not many studies where clinical data is included in the analysis for an all-encompassing model. This research will integrate the power of ML, ANN, and CNN for developing an integrated framework.
Methodology
a. Data Gathering: Clinical and imaging data obtained from cancer registries and hospitals. b. Preprocessing of Data
c. Model Building
d. Evaluation Metrics
e. Tools and Software :Python libraries: TensorFlow, Keras, Scikit-learn, and OpenCV.Hardware: NVIDIA GPU for deep learning computations.
Expected Outcomes: Development of precise predictive models that integrate clinical and imaging data and identification of key predictors of breast cancer outcomes.
Conclusion
This paper intends to utilize advanced ML, ANN, and CNN techniques for accurate prediction of breast cancer outcomes. The proposed models, incorporating clinical and imaging data, are expected to bring a significant improvement over the available diagnostic methods, which in turn will help clinicians take the right decisions and enhance patient care.
Ethical Considerations
This research will also adhere to ethical standards pertaining to human data and biological samples by providing informed consent, ensuring data privacy and adherence to IRB guidelines. References
To be included in this paper: seminal and recent research articles, reviews, and clinical studies on causes of breast cancer published in machine learning peer-reviewed journals.

Collaborators

Dr. Priyanka Rai (KIET GROUP OF INSTUTUTION).
Dr. Akansha Agarwal (KIET GROUP OF INSTUTUTION).