In many sub-types of soft tissue sarcomas, post-treatment changes often cannot be described by standard size criteria e.g. RECIST 1.1, as responding tumours may not shrink, or may grow, after radiotherapy.^1,2 Functional imaging may provide a non-invasive assessment of response in non-resectable disease or in the development of non-surgical treatments, including combined radiotherapy and systemic agents. Several such clinical trials are emerging, however response assessment is further challenged by marked inter- and intra-tumour heterogeneity, which is common in soft tissue sarcoma trials.¹ Functional imaging enables assessment of the whole tumour volume and provides scope for investigation of heterogeneity of response. To develop quantitative MRI techniques for assessment of response to radiotherapy in retroperitoneal sarcomas.

9 patients with retroperitoneal sarcoma (2 leiomyosarcomas, 6 liposarcomas, 1 spindle cell sarcoma) were imaged before treatment and 2-4 weeks after radiotherapy (50.4Gy in 28 fractions), with their informed consent, as part of a prospective single-centre study. Axial T₂-w images, diffusion-weighted images (DWI; b-values 50,600,900smm^-2), Dixon images and T₁-weighted images (3D FLASH, 17°) were acquired from the whole tumour volume using a 1.5T MR scanner (Aera, Siemens GmbH, Erlangen, Germany). 4-minutes after administration of Gd-based contrast agent (Dotarem, 0.2ml/kg body weight, administered at 2ml/sec), post-contrast T₁-w images were acquired for evaluation of enhancing fraction (EF). 1 patient did not have post-contrast imaging. Dixon images were acquired in 5 patients. In baseline exams, DWI was repeated after a short break during which the patient left the scanner room and was repositioned for the second scan.

Regions of interest were drawn around the whole tumour on all slices of T₂-w images by a consultant radiologist using in-house software and transferred to each functional imaging series. In evaluation of apparent diffusion coefficient (ADC), a threshold in signal intensity was applied to exclude suppressed fat pixels from analysis since DWI employed fat suppression. Median estimates of ADC in the tumour-volume were evaluated; the mean of the results from the first and second scans were used in pre-radiotherapy estimates. Fat fraction, defined as the ratio of the pixel value in the Dixon fat image to the sum of pixel values in fat and water images, was calculated and the median value estimated for the tumour-volume. EF was defined as the fraction of enhancing pixels in the tumour-volume, where 'enhancing' was defined as$$$\;\left(\mathrm{S}_{post}-\mathrm{S}_{pre}\right)/\mathrm{S}_{pre}>5\%$$$, where$$$\;\mathrm{S}_{pre}\;$$$and$$$\;\mathrm{S}_{post}\;$$$represent signal in pre- and post-contrast T₁-weighted images respectively.

The cumulative distribution function (CDF) of pixel ADC estimates in the tumour-volume was used to visualise post-radiotherapy changes in the whole tumour (empirical CDF, Matlab 2014a). Thresholded ADC maps were used depict intra-tumour heterogeneity of post-radiotherapy changes in the central slice. Thresholds, defined by the 5^th and 95^th centiles of baseline ADC estimates (combining all pixels in the two baseline volumes), were applied to pre- and post-radiotherapy ADC maps, as described previously.³ Thresholded enhancement maps were used to depict enhancing and non-enhancing regions.

Whilst some tumours exhibited large increases in median ADC or EF post-radiotherapy (Figure 1a-b), the cohort changes were not significant (p>0.05, paired-sample t-test). There were no clear changes in fat fraction (Figure 1c).

CDF plots for one tumour (Figure 2) show not only shift in median ADC to higher values post-radiotherapy but also marked intra-tumour heterogeneity of ADC estimates at baseline and post-radiotherapy.

Thresholded ADC maps (Figure 3) show an example of intra-tumour heterogeneity of post-radiotherapy changes, with increase in ADC in the most restricted regions (shown by the disappearance of the red regions) and emergence of localised areas of high ADC. Thresholded enhancement maps of the same tumour (Figure 4) show enhancement post-radiotherapy in the posterior part of the tumour, whereas the anterior non-enhancing region appears larger post-radiotherapy. The tumour shown in this example increased in volume from 600cm³ at baseline to 760cm³ post-radiotherapy.

Marked intra-tumour heterogeneity of response to radiotherapy, which is clearly depicted in parametric images, may be due to the heterogeneous nature of soft tissue sarcomas, which can include cellular tumour, fat, necrosis and cystic components at baseline. Whole-tumour metrics and examination of cohort effects may be insufficient to characterise response. Functional maps of heterogeneity may also have applications for radiotherapy dose escalation to more aggressive areas.

The marked inter-tumour heterogeneity in baseline estimates of median ADC, fat fraction and EF may reflect the mixture of sarcoma sub-types, which were included in this study as retroperitoneal sarcoma is a rare tumour type. However, this is an accurate reflection of sarcoma sub-types included in clinical trials.^1,2

Intra-tumour heterogeneity is clearly depicted in parametric images and reveals localised post-radiotherapy changes in ADC and enhancement.