Radiomics in Breast Cancer: In-Depth Machine Analysis of MR Images of Metastatic Spine Lesion

Area of Science:

• Radiology
• Medical Imaging
• Oncology

Background:

• Radiomics, using mathematical criteria for image processing, enhances assessment of therapy-associated changes.
• Quantitative and clinical data comparison aids radiologists in optimal decision-making for cancer treatment.

Purpose of the Study:

• To evaluate the capabilities of software operators for detailed analysis of breast cancer metastatic spine lesion images.

Main Methods:

• Analyzed MRI data from three breast cancer patients with spinal metastases.
• Utilized Arzela variation operators for machine analysis of spinal metastases alongside radiologist assessment.
• Examined 12 MRI scans over a 3-month follow-up period.

Main Results:

• Machine analysis identified rapid changes in vertebral image complexity, indicating treatment efficacy (stabilization, improvement, progression).
• Observed positive treatment response indicators like bone object formation and reduced microlevel contrast agent circulation.
• Found a correlation between structural changes and the CA 15-3 cancer marker levels.

Conclusions:

• Machine image analysis algorithms demonstrated high effectiveness in assessing treatment response.
• Results showed a high correlation with radiologist reports and clinical/laboratory data in 9 out of 12 cases.
• Achieved a Pearson correlation coefficient of 0.8 between the classical marker and matrix filter curve.