Dynomics: A Novel and Promising Approach for Improved Breast Cancer Prognosis Prediction

Area of Science:

• Oncology
• Medical Imaging
• Radiochemistry

Background:

• Traditional breast cancer (BC) imaging like X-rays and MRI have limitations in sensitivity and specificity.
• Positron Emission Tomography (PET) offers improved metabolic activity detection for BC diagnosis and prediction.
• Standard radiomics methods analyze static PET images, potentially missing dynamic information.

Purpose of the Study:

• To introduce and evaluate "dynomics," a novel time-domain radiomics approach for dynamic ¹⁸F-Fluorothymidine (FLT) PET scans in breast cancer.
• To compare the diagnostic and prognostic performance of dynomics against static radiomics and standard PET imaging.
• To assess the utility of dynomics in classifying tumor tissue and predicting response to neoadjuvant chemotherapy.

Main Methods:

• Utilized a public clinical dataset of dynamic ¹⁸F-FLT PET scans from breast cancer patients.
• Extracted radiomic features from both static and dynamic PET images using lesion and reference tissue masks.
• Trained an XGBoost model to classify tumor vs. reference tissue and treatment responders, comparing dynomics, static radiomics, and standard PET data.

Main Results:

• Dynamic and static radiomics significantly outperformed standard PET imaging in tumor tissue classification (94% accuracy).
• Dynomics achieved the highest accuracy (86%) in predicting breast cancer prognosis, surpassing static radiomics and standard PET.
• The XGBoost model demonstrated the effectiveness of dynomics in differentiating tumor characteristics and treatment outcomes.

Conclusions:

• Dynomics represents a significant advancement over conventional static radiomics and standard PET imaging for breast cancer.
• This time-domain radiomics approach provides more precise and reliable quantitative and qualitative metabolic information.
• Dynomics holds promise for enhancing breast cancer diagnosis, prognosis prediction, and guiding improved treatment strategies.