The positron emission tomography-magnetic resonance (PET/MR) system allows fluorodeoxyglucose (FDG)-PET and a diffusion-weighted image (DWI) to be acquired simultaneously with precise image co-registration. Further, this system enables direct voxel-by-voxel comparison of the standardized uptake value (SUV) and apparent diffusion coefficient (ADC). Our previous study showed that this system performed better diagnostically than a single measurement such as SUV max or ADC minimum¹. The purpose of the present study was to investigate the feasibility of the voxel-by-voxel comparisons of SUVs and ADCs in the evaluation of the treatment effect in soft-tissue tumors.

Six patients with high-grade soft-tissue sarcomas (2 osteosarcomas, 2 pleomorphic sarcomas, 1 synovial sarcoma, and 1 clear cell sarcoma) were included in this study. Four patients underwent chemotherapy and 2 patients underwent heavy-particle radiotherapy. The patients were scanned by using the Ingenuity TF PET/MR (Philips Healthcare, Cleveland, OH) before and after the treatment. After the scout image and a 3D T1-weighted image for attenuation correction were acquired, the participants underwent PET imaging with 3D-ordered subset expectation maximization (3D-OSEM) and time of flight (TOF). The sampling time was 5 minutes per station, and the images were reconstructed with 2³ mm voxels. After PET imaging, a series of diagnostic MR images were obtained. Transverse 2D fat-suppressed T2-weighted images (T2WI) and T1WI were obtained by using a turbo-spin echo sequence to cover the entire tumor volume. A DWI with a spatially selective RF pulse² (zoomed DWI, b = 0 and 800) was obtained after the anatomical images were acquired, and an ADC map was generated. Additional diagnostic imaging was provided in the clinic, as necessary. The total scan time was 50 to 60 min.

The procedures for image processing are shown in Figure 1. First, image registration with rigid transformation was performed between the T2WI, ADC map, and PET image on an IntelliSpace Portal workstation (version 6.0, Philips Healthcare). The T2WI and PET image were then resliced in accordance with the ADC map (5 or 6 mm). Subsequent image processing was performed by using Image J software (version 1.43, NIH, Bethesda, MD). On T2WI, regions of interest (ROIs) were manually drawn along the border of the tumor in all slices. The ROIs were then copied onto the ADC maps and PET images. After extracting the tumor area, in-plane image resolution was interpolated to 4² mm from 2² mm. The ADCs and SUVs were recorded on a voxel-by-voxel basis for all slices, and scatter plots were generated for each tumor.

The tumor volume, SUV peak, and ADC minimum were recorded for each tumor as ‘conventional’ measurements. To quantify the heterogeneity of SUV/ADC, the 95% area of bivariate normal distribution of the 2 parameters was calculated by using our in-house program and a commercial software package (MATLAB 2015a, MathWorks, Natick, MA). The reverse correlation between SUV and ADC was also analyzed by using Pearson’s correlation.

A representative case of osteosarcoma before and after chemotherapy is shown in Figure 2. A remarkable decrease of FDG uptake was observed, but the change in tumor size was minimal. The significant difference between pre- and post-treatment is clearly demonstrated in the scatter plots shown in Figures 3 and 4. A comparison between each measurement for pre- and post-treatment is shown in Figure 5. The correlation coefficient of the SUV and ADC significantly increased after treatment in all 6 sarcomas (-0.52 ± 0.15 vs. -0.16 ± 0.18, P < 0.01 by paired t-test). On the other hand, tumor volume, SUV peak, ADC minimum, and SUV/ADC heterogeneity did not show significant differences between pre- and post-treatment. These results suggest that a change in the correlation between the SUV and ADC could be an early indicator of the treatment effect in soft-tissue sarcoma.Multiparametric voxel-based analysis of SUVs and ADCs with a PET/MR system could be a promising tool for the evaluation of the treatment effect in soft-tissue tumors.