Following tumour microenvironment after Neoadjuvant radiotherapy with IVIM perfusion analysis.

Purpose or Objective
Neoadjuvant radiotherapy (NeoRT) improves tumor local control and facilitates tumor resection in many cancers. The timing between the end of the NeoRT and surgery is driven by the occurrence of side effects or the tumor downsizing. Some clinical studies demonstrated that the timing of surgery and the RT schedule influence tumor dissemination and subsequently patient overall survival. Previously, we developed a pre-clinical model demonstrating an impact of NeoRT schedule and the timing of surgery on metastatic spreading (Leroi et al. Oncotarget 2015). Here, we evaluate the impact of NeoRT on the tumor microenvironment by functional MRI (fMRI). We aim to identify non-invasive markers allowing to determine the best timing to perform surgery and avoiding tumor spreading.
Material and Methods
Based on our NeoRT model, MDA-MB 231 and 4T1 cells were implanted in the flank of SCID and BalbC mice, respectively. We locally irradiated tumors with 2x5Gy and then surgically removed at different time points after RT. Diffusion Weighted (DW) -MRI was performed every 2 days between RT and surgery. For each tumors we acquired 8 slices of 1 mm thickness and 0.5 mm gap with an 'in plane voxel resolution” of 0.5 mm. For DW-MRI, we performed FSEMS (Fast Spin Echo MultiSlice) sequences, with 9 different B-value (from 40 to 1000) and B0, in the 3 main directions. We performed IVIM (IntraVoxel Incoherent Motion) analysis to obtain information on intravascular diffusion, related to perfusion (F: perfusion factor) and subsequently tumor vessels perfusion.
Results
With the MDA-MB 231, we observed a significant peak of F at day 6 after irradiation, this increasing is about 60% of the basal value (n=6, p<0,05). Moreover, D* parameters (also related to perfusion) increase at the same time. The other parameters of the DW-MRI, ADC and D presented no modification. We observed similar results with 4T1 cells, where F increased at day 3 (about 55%, n=10, p<0,05) then returned to initial level. The difference in timing for the peak of F (day 6 vs day 3) could be related to the difference in tumor growth according to the cell line (four weeks for MDA-MB 231 cells vs one week for 4T1cells). We performed surgery at the time of the F parameter peak in the MDA-MB 231 and we observed a decrease of the metastasic burden compared to surgery performed at day 4 or day 11(absolute number of metastasis 23 VS 1 VS 8 with n=4).
Conclusion
For the first time, we demonstrate the feasibility of repetitive fMRI imaging in preclinical models after NeoRT. With these models, we show a significant difference in perfusion-related parameters (D* and F) at a specific time point depending of the tumor cells. These modifications are correlated to a decrease of metastasis spreading related to the surgery procedure. These results open new perspectives in the personalized medicine and MRI guided surgery timing after NeoRT.