Automated Cluster-based Quantitative Analysis of Ultrafast DCE MRI for Differential Breast DCIS Grading

Area of Science:

• Radiology and Medical Imaging
• Oncology
• Biostatistics

Background:

• Ductal carcinoma in situ (DCIS) grading is crucial for treatment decisions.
• Differentiating low-grade from high-grade DCIS using imaging remains challenging.
• Ultrafast dynamic contrast-enhanced (DCE) MRI offers detailed kinetic information.

Purpose of the Study:

• To evaluate the efficacy of k-means clustering (KMC) analysis on ultrafast DCE MRI kinetic parameters for DCIS grading.
• To determine if KMC can identify aggressive DCIS lesions.
• To assess the performance of KMC-derived parameters in differentiating DCIS grades.

Main Methods:

• Retrospective analysis of ultrafast DCE MRI data from 57 patients with DCIS.
• KMC applied to classify breast parenchyma and lesion voxels into five clusters based on kinetic parameters.
• Statistical analysis including Kruskal-Wallis and ROC curve analysis to compare KMC parameters across DCIS grades.

Main Results:

• The kinetic parameter alpha was significantly associated with DCIS grades (P = .04).
• KMC-derived parameters, including alpha, demonstrated areas under the ROC curve > 0.70 for differentiating DCIS grades.
• Alpha achieved the highest AUC (0.77) for identifying low-grade DCIS.

Conclusions:

• KMC analysis of ultrafast DCE MRI kinetic parameters can effectively differentiate low-grade from intermediate- to high-grade DCIS.
• The parameter alpha derived from KMC shows the most promise for identifying low-grade DCIS.
• This approach may aid in better risk stratification and treatment planning for DCIS patients.