Diffusion weighted imaging (DWI) for adaptive avoidance of functional parotid gland tissue in patients receiving radiotherapy (RT) for head and neck cancer

Abstract

Purpose: Despite current efforts to use highly conformal techniques to deliver maximal tumour coverage whilst sparing organs at risk, a considerable number of head and neck cancer (HNC) patients receiving RT suffer from significant late toxicities such as xerostomia. DWI is a modified MR sequence which identifies water diffusion characteristics from which tissue properties can be inferred1. We hypothesize that DWI and apparent diffusion coefficient (ADC) maps can distinguish areas of different functionality within parotid glands (PGs). As literature supports a dose-volume effect relationship for PG toxicity2, if we are able to track serial changes in PG cellularity in response to RT and where this change occurs, this information may be useful to guide PG sparing to best preserve salivary function. Methods: Multi-parametric MR sequences were acquired for 10patients with newly diagnosed HNC receiving RT at timepoints before, during (week2,3,5,6) and 4 and 12 weeks post-RT. Using MIM imaging software, PG ipsilateral and contralateral to patient's primary were contoured on T2-MR with DWI overlaid. Corresponding ADC maps were extracted. Volumed PGs were sectioned according to ADC maps: 1) high-cellular- volume (HCV) ≤1000 mm2/s, 2) moderate-cellular-volume (MCV) 1000-2500 mm2/s and 3) low-cellular-volume (LCV) ≥2500 mm2/s. The volume (mL) of each 'section' was recorded for each scan. Results: LCVmean of ipsilateral and contralateral PGs increased in volume from DWI1 (pre-treatment) to DWI7 (12 weeks post-RT); 0.12 ml±0.15 to 1.47 ml±0.87 (p ≤ 0.05) and 0.15 ml±0.21 to 0.69 ml±0.44 (p ≤ 0.05) respectively. HCVmean of ipsilateral and contralateral PGs decreased in volume from DWI1 to DWI7; 29.91 ml±7.45 to 8 ml±3.94 (p ≤ 0.05) and 28.15 ml±7.91 to 9.86 ml±4.82 (p ≤ 0.05) respectively. By using ADC maps as a surrogate for change in cellularity, whereby LCV represents fibrotic/inflamed tissue, MCV mixed tissue and HCV functional tissue; our findings suggest there is a trend towards increasing volume of low cellularity (fibrotic) tissue during- and post- RT and decreasing volume of high cellularity (functional) tissue. This supports current literature which describes an increase in ADC maps in irradiated PG tissue over time3. MCV mean of ipsilateral and contralateral PGs from DWI1 to DWI7 did not record a significant change; 20.78 ± 10.59 to 28.10 ± 7.26 and 20.35 ± 10.24 to 26.6 ± 4.66 respectively (p ≥ 0.05). Conclusion: In many cases it is not possible to spare whole PG without compromising target volume coverage. Our study proposes that DWI can identify HCVs of PGs which we suggest should be spared to minimise xerostomia. DWI may therefore play a role in functional image-guided dose-painting IMRT by providing an extra dimension for improving the therapeutic ratio.L618976849 <br />