Objective: To evaluate variations in intra- and interfractional tumour motion, and the effect on internal target volume (ITV) contour accuracy, using deformable image registration of real-time two-dimensional-sagittal cine-mode MRI acquired during lung stereotactic body radiation therapy (SBRT) treatments. Methods: Five lung tumour patients underwent free-breathing SBRT treatments on the ViewRay system, with dose prescribed to a planning target volume (defined as a 3–6 mm expansion of the 4DCT-ITV). Sagittal slice cine-MR images (3.5 × 3.5 mm2 pixels) were acquired through the centre of the tumour at 4 frames per second throughout the treatments (3–4 fractions of 21–32 min). Tumour gross tumour volumes (GTVs) were contoured on the first frame of the MR cine and tracked for the first 20 min of each treatment using offline optical-flow based deformable registration implemented on a GPU cluster. A ground truth ITV (MR-ITV20 min) was formed by taking the union of tracked GTV contours. Pseudo-ITVs were generated from unions of the GTV contours tracked over 10 s segments of image data (MR-ITV10 s). Results: Differences were observed in the magnitude of median tumour displacement between days of treatments. MR-ITV10 s areas were as small as 46% of the MR-ITV20 min. Conclusion: An ITV offers a “snapshot” of breathing motion for the brief period of time the tumour is imaged on a specific day. Real-time MRI over prolonged periods of time and over multiple treatment fractions shows that ITV size varies. Further work is required to investigate the dosimetric effect of these results. Advances in knowledge: Five lung tumour patients underwent free-breathing MRI-guided SBRT treatments, and their tumours tracked using deformable registration of cine-mode MRI. The results indicate that variability of both intra- and interfractional breathing amplitude should be taken into account during planning of lung radiotherapy.

The objective of this study was to explore radiomics features from longitudinal diffusion-weighted MRIs (DWIs) for pathologic treatment effect prediction in patients with localized soft tissue sarcoma (STS) undergoing hypofractionated preoperative radiotherapy (RT). Thirty patients with localized STS treated with preoperative hypofractionated RT were recruited to this longitudinal imaging study. DWIs were acquired at three time points using a 0.35 T MRI-guided radiotherapy system. Treatment effect score (TES) was obtained from the post-surgery pathology as a surrogate of treatment outcome. Patients were divided into two groups based on TES. Response prediction was first performed using a support vector machine (SVM) with only mean apparent diffusion coefficient (ADC) or delta ADC to serve as the benchmark. Radiomics features were then extracted from tumor ADC maps at each of the three time points. Logistic regression and SVM were constructed to predict the TES group using features selected by univariate analysis and sequential forward selection. Classification performance using SVM with features from different time points and with or without delta radiomics were evaluated. Prediction performance using only mean ADC or delta ADC was poor (area under the curve (AUC) < 0.7). For the radiomics study using features from all time points and corresponding delta radiomics, SVM significantly outperformed logistic regression (AUC of 0.91 ± 0.05 v.s. 0.85 ± 0.06). Prediction AUC values using single or multiple time points without delta radiomics were all below 0.74. Including delta radiomics of mid- or post-treatment relative to the baseline drastically boosted the prediction. In this work, an SVM model was built to predict the TES using radiomics features from longitudinal DWI. Based on this study, we found that use of mean ADC, delta ADC, or radiomics features alone was not sufficient for response prediction, and including delta radiomics features of mid- or post-treatment relative to the baseline can optimize the prediction of TES, a pathologic and clinical endpoint.

Objective: To evaluate the feasibility of on-board diffusion-weighted imaging (DWI) with an integrated low-field MRI radiotherapy system to assess responses to neoadjuvant chemoradiation (NAC) in rectal cancer. Methods: A spin echo-based planar imaging diffusion sequence on a 0.35-T MRI radiotherapy system was acquired over the course of NAC. The apparent diffusion coefficients (ADCs) from the tumour regions of interest (ROIs) were calculated. A functional diffusion map (fDM) was created showing a pixelwise ADC analysis of the ROI over the course of treatment. Surgical pathology was correlated with ADC data. Results: Consecutive patients treated on a 0.35-T MRI radiotherapy system were evaluated. Patient A had the worst pathological response to NAC with a tumour regression score of 1 and was the only patient with a negative slope in the change of ADC values over the entire course of NAC, and during both the first and second half of NAC. The fDM from the first half of NAC for Patient A showed discrete dark areas in the tumour ROI, reflecting subregions with decreasing ADC values during NAC. Patient C had the most favourable pathological response to NAC with a Grade 3 response and was the only patient who had an increase in the slope in the change of ADC values from the first to the second half of NAC. Conclusion: DWI using a low-field MRI radiotherapy system for evaluating the responses to NAC is feasible. Advances in knowledge: ADC values obtained using a 0.35-T MRI radiotherapy system over the course of NAC for rectal cancer correlate with pathological responses.

Electron density maps must be accurately estimated to achieve valid dose calculation in MR-only radiotherapy. The goal of this study is to assess whether two deep learning models, the conditional generative adversarial network (cGAN) and the cycle-consistent generative adversarial network (cycleGAN), can generate accurate abdominal synthetic CT (sCT) images from 0.35T MR images for MR-only liver radiotherapy. A retrospective study was performed using CT images and 0.35T MR images of 12 patients with liver (n = 8) and non-liver abdominal (n = 4) cancer. CT images were deformably registered to the corresponding MR images to generate deformed CT (dCT) images for treatment planning. Both cGAN and cycleGAN were trained using MR and dCT transverse slices. Four-fold cross-validation testing was conducted to generate sCT images for all patients. The HU prediction accuracy was evaluated by voxel-wise similarity metric between each dCT and sCT image for all 12 patients. dCT-based and sCT-based dose distributions were compared using gamma and dose-volume histogram (DVH) metric analysis for 8 liver patients. sCT _cycleGAN achieved the average mean absolute error (MAE) of 94.1 HU, while sCT _cGAN achieved 89.8 HU. In both models, the average gamma passing rates within all volumes of interest were higher than 95% using a 2%, 2 mm criterion, and 99% using a 3%, 3 mm criterion. The average differences in the mean dose and DVH metrics were within ±0.6% for the planning target volume and within ±0.15% for evaluated organs in both models. Results : demonstrated that abdominal sCT images generated by both cGAN and cycleGAN achieved accurate dose calculation for 8 liver radiotherapy plans. sCT _cGAN images had smaller average MAE and achieved better dose calculation accuracy than sCTcyleGAN images. More abdominal patients will be enrolled in the future to further evaluate the two models.