Advanced multicompartment diffusion model for noninvasive grading of endometrial cancer: comparative analysis with...

Area of Science:

• Radiology
• Oncology
• Medical Imaging

Background:

• Endometrial cancer (EC) grading is crucial for treatment planning.
• Accurate noninvasive grading of EC remains a challenge.
• Diffusion-weighted imaging (DWI) parameters, like apparent diffusion coefficient (ADC), are used but have limitations.

Purpose of the Study:

• To assess the predictive capability of restricted spectrum imaging (RSI) model parameters for endometrial cancer grading.
• To compare the performance of RSI models against traditional mono-exponential model histogram parameters.
• To investigate the utility of RSI in differentiating EC histological grades.

Main Methods:

• Sixty-three EC patients were analyzed using magnetic resonance imaging (MRI).
• Voxel-wise fitting was performed using mono-exponential and 2-4-compartment RSI models.
• Histogram parameters of ADC and RSI components were extracted and statistically analyzed. Diagnostic performance was evaluated using receiver operating characteristic (ROC) curve analysis.

Main Results:

• The four-compartment RSI model was optimal for EC characterization.
• RSI parameters demonstrated significant differences across EC grades (G1/G2, G1/G3, G2/G3).
• Combined RSI parameters achieved high diagnostic performance (AUC up to 0.922), outperforming ADC histogram parameters, especially when tumor size was included.

Conclusions:

• The four-compartment RSI model offers valuable insights into the tumor microenvironment's component weights in EC.
• RSI enhances the noninvasive grading of endometrial cancer lesions.
• This advanced imaging approach shows promise for improved EC management.