The Diagnostic Value of Advanced Tracer Kinetic Models in Evaluating High Grade Gliomas Recurrence and Treatment Response Using Dynamic Contrast-Enhanced MRI

Jianan Zhou ^1,2,3 Zujun Hou ⁴ Xiuqi Guan ⁵ Zhengyang Zhu ^2,3 Wang Han ⁶ Cong Wang ⁷ Wei Luo ⁵ Chuanshuai Tian ^1,2,3 Huiquan Yang ^2,3 Meiping Ye ^2,3 Sixuan Chen ^2,3 Xin Zhang ^1,2,3* Bing Zhang ^1,2,3

• ¹ Department of Radiology, Nanjing Drum Tower Hospital Clinical College of Nanjing Medical University, Nanjing, Liaoning Province, China
• ² Institute of Medical Imaging and Artificial Intelligence, Nanjing University, Nanjing, China
• ³ Medical Imaging Center, Department of Radiology, Nanjing Drum Tower Hospital, Affiliated Hospital of Medical School, Nanjing, China
• ⁴ Second Affiliated Hospital and Yuying Children's Hospital of Wenzhou Medical University, Wenzhou, Zhejiang Province, China
• ⁵ FISCA Healthcare Co., Ltd., Nanjing, China
• ⁶ Nanjing center for Applied Mathematics, Nanjing, China
• ⁷ School of Electronics and Information Engineering, Suzhou Vocational University, Suzhou, China

The purpose of this study was to investigate the diagnostic value of advanced tracer kinetic models (TKMs) in differentiating HGGs recurrence and treatment response.Methods: A total of 52 HGGs were included. DCE images were analyzed using the following TKMs: distributed parameter (DP), tissue homogeneity (TH), Brix's two-compartment (Brix) and extended-Tofts model (ETM), yielding the following parameters: cerebral blood flow (CBF), mean transit time (MTT), plasma volume (Vp), extravascular volume (Ve), vascular permeability (PS) and first-pass extraction ratio (E) in advanced TKMs (DP, TH and Brix); K trans , Ve, Vp and Kep in ETM. Two delineation methods were conducted (routine scans and parameter heat maps). The differences between two MRI scanners were compared. Mann-Whitney U test was used to assess the difference of parameter values. Diagnostic performance was assessed using the method of the receiver operating characteristic (ROC) curves, with the areas under the ROC curves (AUC) to determine the discriminating power of DCE parameters between recurrent tumor group and treatment response group ROC curves. P<0.05 indicates statistical significance.The difference on the normalized kinetic parameter value (with respect to contralateral normal-appearing white matter) between two MRI scanners was statistically insignificant (P＞0.05). MTT and Vp of advanced TKMs were higher in recurrent than in treatment response group (P＜0.05). For ROI delineated on parameter heat maps, MTT(DP) attained the best performance with AUC 0.88, followed by MTT(TH) and Vp (DP, Brix) with AUCs around 0.80 (0.81, 0.80, 0.79 respectively). The best performance in ETM was Vp (AUC = 0.73).MTT (DP, TH), and Vp (DP, Brix) could be potential quantitative imaging biomarkers in distinguishing recurrence and treatment response in HGGs.