[OA046] Visualizing sites of increased cellularity and high permeability in soft tissue sarcomas.

Purpose Radiological evaluation of tumor aggressiveness is very frequently based on diffusion and/or perfusion imaging findings and conclusions are used to guide biopsy. The present work describes a post processing process that highlights areas of increased cellularity (low ADC) and also increased vascular transendothelial permeability (high K trans ), two of the most significant markers of malignancy, at the early stage of imaging. Preliminary results tested on 5 patients with soft tissue sarcomas are presented. Methods Quantitative ADC maps were generated from MR data by pixelwise mono-exponential fitting of multi-b (8b, 0–1500) DW images with custom-built tools written in Python. Similarly, K trans map was calculated based on the Extended Tofts Model from T1-w GRE data (temporal resolution 7 s, 45 time points). All pixel values assigned as tumor volume (3D ROI) were used as input for the initial whole tumor ADC and K trans histograms. As a next step, only pixels with values lower than the mean of ADC histogram and K trans values greater than the mean K trans were located and visualized in order to examine if a voxel cluster with adequate size is discriminated after thresholding. Patient population (5 male, mean age 60) comprised 3 dedifferentiated liposarcomas and 2 pleomorphic liposarcomas. Results Whole tumor pixel percentages with ADC value below the mean ADC value of the same ROI (used as histogram threshold) for the five high grade liposarcomas were: 58.7, 82.5, 52.1, 65.7, and 78.8%. Ktrans pixel percentages above the mean K trans (threshold) were: 11.8, 23.4, 4.3, 62.9, 39.9% respectively. The percentage of pixels meeting both criteria for low ADC and high K trans were: 4.2, 14.2, 0.6, 20.1, 15.3% respectively. Conclusion For 4 out of 5 patients visualization of a pixel cluster with adequate size that can be proposed as suitable site for preoperative needle biopsy was possible. Further appropriate ADC/ K trans thresholding can be used to increase disease conspicuity under certain criteria, depending on the special characteristics of each tumor subtype. [ABSTRACT FROM AUTHOR]