Direct Correlation of Multiple Diffusion-Weighted Imaging Models of Soft Tissue Sarcoma and Quantitative
Jinge Li1, Kai Zhang1, Minting Zheng2
1Department of Radiology, The Second Affiliated Hospital of Dalian Medical University, Dalian, China (J.L., K.Z., Y.Z., K.Z., S.W.).
Rationale And Objectives:
To explore the correlation between the multiple diffusion-weighted imaging (DWI) model parameters and the quantitative histopathological features of soft tissue sarcoma (STS) utilizing a three-step controlled imaging-pathology co-registration protocol.
Materials And Methods:
This prospective study received approval from the institutional ethics committee and enrolled 69 patients diagnosed with soft tissue sarcoma (STS). All patients underwent 3.0 T MR scanning and multiple DWI model parameters were acquired. A three-step controlled imaging-pathology co-registration protocol was employed to guarantee the correspondence between MRI slices and pathological slices. The correlations between multiple DWI model parameters and histopathological features were analyzed to investigate the application value of multiple DWI model parameters in the characterization of STS tissues. Spearman analysis, Kruskal-Wallis H test, Mann-Whitney U test, and receiver operating characteristic (ROC) curve analysis were conducted.
Results:
The multiple DWI model parameters showed varying degrees of correlation with nuclear fraction (r = -0.563 to 0.465), cellular fraction (r = -0.650 to 0.477) and stromal fraction (r = -0.480 to 0.652), respectively. Among them, the f、MDDKI、MK、MDDTI were independently associated with stromal fractions with a combined AUC of 0.890 in distinguishing between the stroma - rich and stroma - poor groups. Besides, MDDTI and DDC predicted fibroblastic and undetermined differentiation STS with AUC of 0.798 and 0.713, respectively.
Conclusion:
By directly linking multiple DWI model parameters to quantitative histopathological features through rigorous spatial co-registration, this study provides biological interpretability for diffusion metrics and supports their potential role as noninvasive biomarkers in STS.
