Introduction to matrix-based method for analyzing hybrid multidimensional prostate MRI data.

Area of Science:

• Radiology and Medical Imaging
• Biophysics
• Computational Biology

Background:

• Prostate cancer (PCa) diagnosis relies on imaging techniques, but differentiating cancerous from normal tissue can be challenging.
• Hybrid multidimensional MRI (HM-MRI) offers rich data but requires advanced analytical methods.
• Current MRI analysis methods like apparent diffusion coefficient (ADC) and T2 mapping have limitations in clearly delineating PCa.

Purpose of the Study:

• To introduce and validate a new matrix-based approach for analyzing HM-MRI data.
• To assess the utility of calculated eigenvalues and their ratios in distinguishing prostate cancer from normal tissue.
• To evaluate the potential clinical applicability of this novel HM-MRI analysis method.

Main Methods:

• Linearization of HM-MRI data by taking the natural logarithm of signal intensity.
• Construction of a hybrid symmetric matrix for each pixel by multiplying the pixel's matrix by its transpose.
• Calculation of eigenvalues from the hybrid symmetric matrix and definition of an eigenvalue ratio (λr) for quantitative comparison.

Main Results:

• Eigenvalue ratio maps clearly visualized prostate cancer regions, differing significantly from standard ADC and T2 maps.
• Prostate cancer tissue exhibited significantly larger eigenvalue ratios (λr) compared to normal prostate tissue (p < 0.001).
• The new method showed significantly smaller ADC and T2 values in PCa compared to normal tissue.

Conclusions:

• The matrix-based analysis of HM-MRI data provides novel, clinically relevant information for prostate cancer detection.
• Eigenvalue ratios derived from HM-MRI data effectively differentiate prostate cancer from normal tissue.
• This eigenvalue-based method is user-friendly and readily implementable in clinical practice for PCa identification and staging.