Research on application of multiparametric MRI to predict FNCLCC grading and ki67 expression in soft tissue sarcoma...

To investigate an in vivo biopsy method based on CT-MRI fusion image registration and evaluate its value for improving the accuracy of preoperative core needle biopsies(CNBs) histological grading in soft tissue sarcoma (STS)and determining the correlation between multiparametric MRI(mp MRI) and both STS French Federation of Cancer Center (FNCLCC) grading and the Ki-67 labeling index (LI).

32 STS patients were enrolled prospectively and underwent 3.0T MRI, Diffusion weighted imaging(DWI) and Hydrogen proton magnetic resonance spectroscopy (¹H MRS) examination, then underwent a preoperative CT guided CNBs that was confirmed by subsequent surgery. We used a novel image fusion registration method to ensure the biopsy sampling point was located in a high cell density area and the MRI-Region of Interest (ROI) was compatible with the biopsy specimen. We evaluated the accuracy of biopsy pathology diagnoses compared to that of surgery, and compared the mp MRI parameters between FNCLCC low- and high-grade/Ki67 low- and high-expression groups. The statistical analyses included the intraclass correlation coefficient(ICC), the Wilcoxon rank-sum test, Receiver Operating Characteristic (ROC) curves, and Spearman's rank correlation.

The pathological diagnosis accuracy of the biopsy specimens was 100%, the subtype diagnosis accuracy was 87.5%, while the grading diagnosis accuracy was 93.8% compared with that of the surgical specimens. The FNCLCC high-grade group and Ki67 high-expression group had lower Apparent diffusion coefficient(ADC) values and higher Cholin/Creatin (Cho/Cr) ratios. The area under the curve(AUC) of ADCmin, ADCmean values and Cho/Cr ratios to discriminate between low- and high-grade groups were 0.956,0.969,0.917, low and high expression groups were 0.929,0.957,0.943; The ADC values of STSs correlated negatively while the Cho/Cr ratios correlated positively with FNCLCC grading (r=-0.782,-0.814, 0.758,p<0.001) and Ki-67 LI (r=-0.853, -0.902,0.710,p<0.001).

The CT-MRI fusion image registration method may improve the accuracy of pre treatment STS CNBs pathology diagnoses and can be used for precise imaging-pathology control studies. ADC values and Cho/Cr ratios may therefore serve as a valuable supplement to STS histopathological grading and correlate with Ki67 expression.