Soft tissue sarcomas are often highly heterogeneous tumours with variable components which can include cellular tumour, fat, necrosis and cystic change. Post-treatment changes often cannot be described by standard size criteria e.g. RECIST 1.1, as responding tumours may not change size, or may grow, after radiotherapy.^1,2 In non-resectable tumours, or trials of combined radiotherapy and systemic agents, non-invasive methods are required for assessment of response. Prior knowledge of functional imaging characteristics is required, however, to select appropriate MRI techniques and sequence parameters. Estimates of repeatability inform on sensitivity of imaging metrics to detect post-treatment changes.

To establish pre-treatment estimates of apparent diffusion coefficient (ADC), parameters of the intra-voxel incoherent motion model (IVIM:$$$\;$$$diffusion coefficient$$$\;$$$D,$$$\;$$$fraction$$$\;$$$f,$$$\;$$$fast exponential component$$$\;$$$D*), transverse relaxation rate (R₂*), fat fraction and enhancing fraction (EF) in retroperitoneal sarcomas; assess repeatability of$$$\;$$$ADC,$$$\;$$$D,$$$\;$$$f,$$$\;$$$D*$$$\;$$$and$$$\;$$$R₂*; and assess heterogeneity of these parameters within tumours and between patients, in order to inform protocol development for sarcoma clinical trials, where functional MRI has not been used extensively.

22 patients with retroperitoneal sarcoma were imaged prior to treatment (21 imaged twice), with their written consent, as part of a prospective single-centre study. Tumours were 3 leiomyosarcomas, 17 well-differentiated/dedifferentiated liposarcomas, 1 lipoma and 1 spindle cell sarcoma. Scans were carried out using a 1.5T MR scanner (Aera, Siemens GmbH, Erlangen, Germany) (Figure$$$\;$$$1). Axial T₂-w images, diffusion-weighted images (DWI), Dixon images and pre-contrast T₁-weighted images were acquired from the whole tumour volume. T₂*-w images for estimation of R₂* and an additional DWI series for estimation of IVIM parameters were acquired from a smaller number of slices centred on the central slice of the tumour. Following administration of Gd-based contrast agent (Dotarem, 0.2ml/kg body weight, administered at 2ml/sec using a power-injector; images acquired 4 minutes after injection), post-contrast T₁-w images were acquired for evaluation of EF. 1 patient did not have T₂*-w imaging. 1 patient did not have post-contrast imaging. Dixon images were acquired in 12 patients. In 21 patients, DWI and T₂*-w imaging were repeated after a short break during which the patient left the scanner room and was repositioned for the second scan.

Analysis was carried out using in-house software. Pixel values evaluated at all pixels in the regions of interest (ROIs, Figure 2) were combined to create a volume of interest (VOI) for each quantitative parameter. A threshold in signal intensity was applied to exclude suppressed fat from analysis of DW and T₂*-w images since these sequences employed fat suppression; 2 tumours were excluded from analysis of DW and T₂*-w images as >90% of pixels were excluded. For ADC,$$$\;$$$D,$$$\;$$$f,$$$\;$$$D*$$$\;$$$and$$$\;$$$R₂*, the median, mean, standard deviation, 10^th, 25^th, 75^th and 90^th centiles, skew and kurtosis of pixel values in the VOI were evaluated. Fat fraction, calculated using Dixon fat and water images, was defined as the ratio of the pixel value in the fat image to the sum of the pixel values in fat and water images. EF was defined as the fraction of pixels in the VOI that increased in signal intensity by >5% between pre- and post-contrast T₁-w images.

Bland-Altman plots of untransformed data showed a relationship between the differences in the repeated measurements and their means that was reduced by using the natural logarithm of the data. The Coefficient of Variation (CV) of the log-transformed data was used to describe the repeatability of fitted parameters$$$\;\mathrm{CV}=\sqrt{\mathrm{exp}\left(\Sigma_i d_i^2/2N\right)-1}$$$, where$$$\;d_i\;$$$is the difference between paired measurements for patient$$$\;i$$$ and$$$\;N\;$$$is the number of patients.

Repeatability of ADC summary statistics was excellent (Figure$$$\;$$$3). Repeatability of D was also good, with poorer repeatability of$$$\;$$$f,$$$\;$$$D*$$$\;$$$and$$$\;$$$R₂*, particularly centile estimates. Median estimates of all parameters showed large ranges across the cohort (Figures$$$\;$$$3$$$\;$$$and$$$\;$$$4). The range of each estimated parameter within tumours was also wide, indicated by the large standard deviations (Figure$$$\;$$$3).

The large ranges of estimated parameters indicates wide inter-tumour heterogeneity, possibly reflecting the mixture of sarcoma sub-types included in the study, which is typical of sarcoma trials as this is a rare tumour type. The large standard deviation of parameters within tumours reflects the intra-tumour heterogeneity. The excellent repeatability of ADC statistics indicates high sensitivity to treatment-induced changes, including centile estimates, which may be sensitive to different components of heterogeneous tumours and furthermore have been shown to be earlier indicators of response than mean ADC in some tumours.³ Good quality images can be obtained for estimation of$$$\;$$$ADC,$$$\;$$$D,$$$\;$$$f,$$$\;$$$D*,$$$\;$$$R₂*,$$$\;$$$fat fraction and enhancing fraction in retroperitoneal sarcomas. Repeatability of ADC is excellent, with slightly poorer repeatability of D,$$$\;$$$f,$$$\;$$$D*$$$\;$$$and$$$\;$$$R₂*. Knowledge of the ranges of these parameters within tumours and between subjects informs protocol development for clinical trials.