Quantifying intrafractional gastric motion using auto-segmentation on MRI: Deformation and respiratory-induced displacement compared.

BACKGROUND AND PURPOSE: For accurate pre-operative gastric radiotherapy, intrafractional changes must be taken into account. The aim of this study is to quantify local gastric deformations and compare these deformations with respiratory-induced displacement. MATERIALS AND METHODS: Coronal 2D MRI scans (15-16 min; 120 repetitions of 25-27 interleaved slices) were obtained for 18 healthy volunteers. A deep-learning network was used to auto-segment the stomach. To separate out respiratory-induced displacements, auto-segmentations were rigidly shifted in superior-inferior (SI) direction to align the centre of mass (CoM) within every slice. From these shifted auto-segmentations, 3D iso-probability surfaces (isosurfaces) were established: a reference surface for POcc  = 0.50 and 50 other isosurfaces (from POcc  = 0.01 to 0.99), with POcc indicating the probability of occupation by the stomach. For each point on the reference surface, distances to all isosurfaces were determined and a cumulative Gaussian was fitted to this probability-distance dataset to obtain a standard deviation (SDdeform ) expressing local deformation. For each volunteer, we determined median and 98th percentile of SDdeform over the reference surface and compared these with the respiratory-induced displacement SDresp , that is, the SD of all CoM shifts (paired Wilcoxon signed-rank, α = 0.05). RESULTS: Larger deformations were mostly seen in the antrum and pyloric region. Median SDdeform (range, 2.0-2.9 mm) was smaller than SDresp (2.7-8.8 mm) for each volunteer (p < 0.00001); 98th percentile of SDdeform (3.2-7.3 mm) did not significantly differ from SDresp (p = 0.13). CONCLUSION: Locally, gastric deformations can be large. Overall, however, these deformations are limited compared to respiratory-induced displacement. Therefore, unless respiratory motion is considerably reduced, the need to separately include these deformation uncertainties in the treatment margins may be limited.

A unified model for reconstruction and R2* mapping of accelerated 7T data using the quantitative recurrent inference machine.

Quantitative MRI (qMRI) acquired at the ultra-high field of 7 Tesla has been used in visualizing and analyzing subcortical structures. qMRI relies on the acquisition of multiple images with different scan settings, leading to extended scanning times. Data redundancy and prior information from the relaxometry model can be exploited by deep learning to accelerate the imaging process. We propose the quantitative Recurrent Inference Machine (qRIM), with a unified forward model for joint reconstruction and R2*-mapping from sparse data, embedded in a Recurrent Inference Machine (RIM), an iterative inverse problem-solving network. To study the dependency of the proposed extension of the unified forward model to network architecture, we implemented and compared a quantitative End-to-End Variational Network (qE2EVN). Experiments were performed with high-resolution multi-echo gradient echo data of the brain at 7T of a cohort study covering the entire adult life span. The error in reconstructed R2* from undersampled data relative to reference data significantly decreased for the unified model compared to sequential image reconstruction and parameter fitting using the RIM. With increasing acceleration factor, an increasing reduction in the reconstruction error was observed, pointing to a larger benefit for sparser data. Qualitatively, this was following an observed reduction of image blurriness in R2*-maps. In contrast, when using the U-Net as network architecture, a negative bias in R2* in selected regions of interest was observed. Compressed Sensing rendered accurate, but less precise estimates of R2*. The qE2EVN showed slightly inferior reconstruction quality compared to the qRIM but better quality than the U-Net and Compressed Sensing. Subcortical maturation over age measured by a linearly increasing interquartile range of R2* in the striatum was preserved up to an acceleration factor of 9. With the integrated prior of the unified forward model, the proposed qRIM can exploit the redundancy among repeated measurements and shared information between tasks, facilitating relaxometry in accelerated MRI.

Motion-compensated FDG PET/CT for oesophageal cancer.

PURPOSE: Respiratory-induced motion of oesophageal tumours and lymph nodes can influence positron-emission tomography/computed tomography (PET/CT). The aim was to compare standard three-dimensional (3D) and motion-compensated PET/CT regarding standardized uptake value (SUV), metabolic tumour volume (MTV) and detection of lymph node metastases. METHODS: This prospective observational study (NCT02424864) included 37 newly diagnosed oesophageal cancer patients. Diagnostic PET/CT was reconstructed in 3D and motion-compensated PET/CT. MTVs of the primary tumour were calculated using an automated region-growing algorithm with SUV thresholds of 2.5 (MTV2.5) and ≥ 50% of SUVmax (MTV50%). Blinded for reconstruction method, a nuclear medicine physician assessed all lymph nodes showing 18F­fluorodeoxyglucose uptake for their degree of suspicion. RESULTS: The mean (95% CI) SUVmax of the primary tumour was 13.1 (10.6-15.5) versus 13.0 (10.4-15.6) for 3D and motion-compensated PET/CT, respectively. MTVs were also similar between the two techniques. Bland-Altman analysis showed mean differences between both measurements (95% limits of agreement) of 0.08 (-3.60-3.75), -0.26 (-2.34-1.82), 4.66 (-29.61-38.92) cm3 and -0.95 (-19.9-18.0) cm3 for tumour SUVmax, lymph node SUVmax, MTV2.5 and MTV50%, respectively. Lymph nodes were classified as highly suspicious (30/34 nodes), suspicious (20/22) and dubious (66/59) for metastases on 3D/motion-compensated PET/CT. No additional lymph node metastases were found on motion-compensated PET/CT. SUVmax of the most intense lymph nodes was similar for both scans: mean (95% CI) 6.6 (4.3-8.8) and 6.8 (4.5-9.1) for 3D and motion-compensated, respectively. CONCLUSION: SUVmax of the primary oesophageal tumour and lymph nodes was comparable on 3D and motion-compensated PET/CT. The use of motion-compensated PET/CT did not improve lymph node detection.

First system for fully-automated multi-criterial treatment planning for a high-magnetic field MR-Linac applied to rectal cancer.

Background and purpose: In this study we developed a workflow for fully-automated generation of deliverable IMRT plans for a 1.5 T MR-Linac (MRL) based on contoured CT scans, and we evaluated automated MRL planning for rectal cancer.Methods: The Monte Carlo dose calculation engine used in the clinical MRL TPS (Monaco, Elekta AB, Stockholm, Sweden), suited for high accuracy dose calculations in a 1.5 T magnetic field, was coupled to our in-house developed Erasmus-iCycle optimizer. Clinically deliverable plans for 23 rectal cancer patients were automatically generated in a two-step process, i.e., multi-criterial fluence map optimization with Erasmus-iCycle followed by a conversion into a deliverable IMRT plan in the clinical TPS. Automatically generated plans (AUTOplans) were compared to plans that were manually generated with the clinical TPS (MANplans).Results: With AUTOplanning large reductions in planning time and workload were obtained; 4-6 h mainly hands-on planning for MANplans vs ∼1 h of mainly computer computation time for AUTOplans. For equal target coverage, the bladder and bowel bag Dmean was reduced in the AUTOplans by 1.3 Gy (6.9%) on average with a maximum reduction of 4.5 Gy (23.8%). Dosimetric measurements at the MRL demonstrated clinically acceptable delivery accuracy for the AUTOplans.Conclusions: A system for fully automated multi-criterial planning for a 1.5 T MR-Linac was developed and tested for rectal cancer patients. Automated planning resulted in major reductions in planning workload and time, while plan quality improved. Negative impact of the high magnetic field on the dose distributions could be avoided.

Mapping of sentinel lymph node drainage using SPECT/CT to tailor elective nodal irradiation in head and neck cancer patients (SUSPECT-2): a single-center prospective trial.

BACKGROUND: The majority of patients with head and neck squamous cell carcinoma (HNSCC) receive bilateral elective nodal irradiation (ENI), in order to reduce the risk of regional failure. Bilateral ENI, as compared to unilateral ENI, is associated with higher incidence of acute and late radiation-induced toxicity with subsequent deterioration of quality of life. Increasing evidence that the incidence of contralateral regional failure (cRF) in lateralized HNSCC is very low (< 10%) suggests that it can be justified to treat selected patients unilaterally. This trial aims to minimize the proportion of patients that undergo bilateral ENI, by using lymph drainage mapping by SPECT/CT to select patients with a minimal risk of contralateral nodal failure for unilateral elective nodal irradiation. METHODS: In this one-armed, single-center prospective trial, patients with primary T1-4 N0-2b HNSCC of the oral cavity, oropharynx, larynx (except T1 glottic) or hypopharynx, not extending beyond the midline and planned for primary (chemo) radiotherapy, are eligible. After 99mTc-nanocolloid tracer injection in and around the tumor, lymphatic drainage is visualized using SPECT/CT. In case of contralateral lymph drainage, a contralateral sentinel node procedure is performed on the same day. Patients without contralateral lymph drainage, and patients with contralateral drainage but without pathologic involvement of any removed contralateral sentinel nodes, receive unilateral ENI. Only when tumor cells are found in a contralateral sentinel node the patient will be treated with bilateral ENI. The primary endpoint is cumulative incidence of cRF at 1 and 2 years after treatment. Secondary endpoints are radiation-related toxicity and quality of life. The removed lymph nodes will be studied to determine the prevalence of occult metastatic disease in contralateral sentinel nodes. DISCUSSION: This single-center prospective trial aims to reduce the incidence and duration of radiation-related toxicities and improve quality of life of HNSCC patients, by using lymph drainage mapping by SPECT/CT to select patients with a minimal risk of contralateral nodal failure for unilateral elective nodal irradiation. TRIAL REGISTRATION: ClinicalTrials.gov Identifier: NCT03968679, date of registration: May 30, 2019.

Feasibility of Micro-Computed Tomography Imaging for Direct Assessment of Surgical Resection Margins During Breast-Conserving Surgery.

BACKGROUND: To analyze the feasibility and accuracy of micro-computed tomography (micro-CT) for surgical margin assessment in breast excision specimen. MATERIALS AND METHODS: Two data sets of 30 micro-CT scans were retrospectively evaluated for positive resection margins by four observers in two phases, using pathology as a gold standard. Results of phase 1 were evaluated to define micro-CT evaluation guidelines for phase 2. Interobserver agreement was also assessed (kappa). In addition, a prospective study was conducted in which 40 micro-CT scans were directly acquired, reconstructed, and evaluated for positive resection margins by one observer. A suspect positive resection margin on micro-CT was annotated onto the specimen with ink, enabling local validation by pathology. Main outcome measures were accuracy, sensitivity, specificity, and positive predictive value (PPV). RESULTS: Average accuracy, sensitivity, specificity, and PPV for the four observers were 63%, 38%, 70%, and 22%, respectively, in phase 1 and 72%, 40%, 78%, and 26%, respectively, in phase 2. The interobserver agreement was fair [kappa (range), 0.31 (0.12-0.80) in phase 1 and 0.23 (0-0.43) in phase 2]. In the prospective study 70% of the surgical resection margins were correctly evaluated. Ten specimens were annotated for positive resection margins, which correlated with three positive and three close (<1 mm) margins on pathology. Sensitivity, specificity, and PPV were 38%, 78%, and 30%, respectively. CONCLUSIONS: Micro-CT imaging of breast excision specimen has moderate accuracy and considerable interobserver variation for analysis of surgical resection margins. Especially sensitivity and PPV need to be improved before micro-CT-based margin assessment can be introduced in clinical practice.

Feasibility of MRI-based reference images for image-guided radiotherapy of the pelvis with either cone-beam computed tomography or planar localization images.

PURPOSE: This study introduces methods to conduct image-guided radiotherapy (IGRT) of the pelvis with either cone-beam computed tomography (CBCT) or planar localization images by relying solely on magnetic resonance imaging (MRI)-based reference images. MATERIAL AND METHODS: Feasibility of MRI-based reference images for IGRT was evaluated against kV CBCT (50 scans, 5 prostate cancer patients) and kV & MV planar (5 & 5 image pairs and patients) localization images by comparing the achieved patient position corrections to those obtained by standard CT-based reference images. T1/T2*-weighted in-phase MRI, Hounsfield unit conversion-based heterogeneous pseudo-CT, and bulk pseudo-CT images were applied for reference against localization CBCTs, and patient position corrections were obtained by automatic image registration. IGRT with planar localization images was performed manually by 10 observers using reference digitally reconstructed radiographs (DRRs) reconstructed from the pseudo-CTs and standard CTs. Quality of pseudo-DRRs against CT-DRRs was evaluated with image similarity metrics. RESULTS: The SDs of differences between CBCT-to-MRI and CBCT-to-CT automatic gray-value registrations were ≤1.0 mm & ≤0.8° and ≤2.5 mm & ≤3.6° with 10 cm diameter cubic VOI and prostate-shaped VOI, respectively. The corresponding values for reference heterogeneous pseudo-CT were ≤1.0 mm & ≤0.7° and ≤2.2 mm & ≤3.3°, respectively. Heterogeneous pseudo-CT was the only type of MRI-based reference image working reliably with automatic bone registration (SDs were ≤0.9 mm & ≤0.7°). The differences include possible residual errors from planning CT to MRI registration. The image similarity metrics were significantly (p≤0.01) better in agreement between heterogeneous pseudo-DRRs and CT-DRRs than between bulk pseudo-DRRs and CT-DRRs. The SDs of differences in manual registrations (3D) with planar kV and MV localization images were ≤1.0 mm and ≤1.7 mm, respectively, between heterogeneous pseudo-DRRs and CT-DRRs, and ≤1.4 mm and ≤2.1 mm between bulk pseudo-DRRs and CT-DRRs. CONCLUSION: This study demonstrated that it is feasible to conduct IGRT of the pelvis with MRI-based reference images.

Portal dosimetry in wedged beams.

Portal dosimetry using electronic portal imaging devices (EPIDs) is often applied to verify high-energy photon beam treatments. Due to the change in photon energy spectrum, the resulting dose values are, however, not very accurate in the case of wedged beams if the pixel-to-dose conversion for the situation without wedge is used. A possible solution would be to consider a wedged beam as another photon beam quality requiring separate beam modeling of the dose calculation algorithm. The aim of this study was to investigate a more practical solution: to make aSi EPID-based dosimetry models also applicable for wedged beams without an extra commissioning effort of the parameters of the model. For this purpose two energy-dependent wedge multiplication factors have been introduced to be applied for portal images taken with and without a patient/phantom in the beam. These wedge multiplication factors were derived from EPID and ionization chamber measurements at the EPID level for wedged and nonwedged beams, both with and without a polystyrene slab phantom in the beam. This method was verified for an EPID dosimetry model used for wedged beams at three photon beam energies (6, 10, and 18 MV) by comparing dose values reconstructed in a phantom with data provided by a treatment planning system (TPS), as a function of field size, depth, and off-axis distance. Generally good agreement, within 2%, was observed for depths between dose maximum and 15 cm. Applying the new model to EPID dose measurements performed during ten breast cancer patient treatments with wedged 6 MV photon beams showed that the average isocenter underdosage of 5.3% was reduced to 0.4%. Gamma-evaluation (global 3%/3 mm) of these in vivo data showed an increase in percentage of points with γ ≤ 1 from 60.2% to 87.4%, while γmean reduced from 1.01 to 0.55. It can be concluded that, for wedged beams, the multiplication of EPID pixel values with an energy-dependent correction factor provides good agreement between dose values determined by an EPID and a TPS, indicating the usefulness of such a practical solution.

A deep learning based dynamic arc radiotherapy photon dose engine trained on Monte Carlo dose distributions.

Background and purpose: Despite hardware acceleration, state-of-the-art Monte Carlo (MC) dose engines require considerable computation time to reduce stochastic noise. We developed a deep learning (DL) based dose engine reaching high accuracy at strongly reduced computation times. Materials and methods: Radiotherapy treatment plans and computed tomography scans were collected for 350 treatments in a variety of tumor sites. Dose distributions were computed using a MC dose engine for ∼30,000 separate segments at 6 MV and 10 MV beam energies, both flattened and flattening filter free. For dynamic arcs these explicitly incorporated the leaf, jaw and gantry motions during dose delivery. A neural network was developed, combining two-dimensional convolution and recurrence using 64 hidden channels. Parameters were trained to minimize the mean squared log error loss between the MC computed dose and the model output. Full dose distributions were reconstructed for 100 additional treatment plans. Gamma analyses were performed to assess accuracy. Results: DL dose evaluation was on average 82 times faster than MC computation at a 1 % accuracy setting. In voxels receiving at least 10 % of the maximum dose the overall global gamma pass rate using a 2 % and 2 mm criterion was 99.6 %, while mean local gamma values were accurate within 2 %. In the high dose region over 50 % of maximum the mean local gamma approached a 1 % accuracy. Conclusions: A DL based dose engine was implemented, able to accurately reproduce MC computed dynamic arc radiotherapy dose distributions at high speed.

Dynamic recurrent inference machines for accelerated MRI-guided radiotherapy of the liver.

Recurrent inference machines (RIM), a deep learning model that learns an iterative scheme for reconstructing sparsely sampled MRI, has been shown able to perform well on accelerated 2D and 3D MRI scans, learn from small datasets and generalize well to unseen types of data. Here we propose the dynamic recurrent inference machine (DRIM) for reconstructing sparsely sampled 4D MRI by exploiting correlations between respiratory states. The DRIM was applied to a 4D protocol for MR-guided radiotherapy of liver lesions based on repetitive interleaved coronal 2D multi-slice T2-weighted acquisitions. We demonstrate with an ablation study that the DRIM outperforms the RIM, increasing the SSIM score from about 0.89 to 0.95. The DRIM allowed for an approximately 2.7 times faster scan time than the current clinical protocol with only a slight loss in image sharpness. Correlations between slice locations can also be used, but were found to be of less importance, as were a majority of tested variations in network architecture, as long as the respiratory states are processed by the network. Through cross-validation, the DRIM is also shown to be robust in terms of training data. We further demonstrate a good performance across a large range of subsampling factors, and conclude through an evaluation by a radiation oncologist that reconstructed images of the liver contour and inner structures are of a clinically acceptable standard at acceleration factors 10x and 8x, respectively. Finally, we show that binning the data with respect to respiratory states prior to reconstruction comes at a slight cost to reconstruction quality, but at greater speed of the overall protocol.

Baseline shift corrections towards the heart: External validation of the impact on survival in early-stage NSCLC patients.

PURPOSE: To externally validate Johnson-Hart et al. findings: the association of tumor baseline shifts towards the heart with overall survival (OS) in SBRT for NSCLC. Further analysis included investigating the presence of interfractional heart baseline shifts and the association of OS with heart dose change during treatment. METHODS: Data from 416 SBRT early-stage NSCLC patients was collected. Pearson's correlations (PCCs) between clinical variables and treatment-averaged tumor shifts towards/away from the heart were explored. Validation of published multivariable Cox model was performed. PCCs between heart and tumor baseline shifts were analyzed. Dose accumulation was performed following daily CBCT-to-pCT deformable registration. Maximum heart dose (D0) was computed for planned and accumulated doses. Differences in OS according to shifts towards/away from the heart or D0 increase/decrease were analyzed. Significant D0 differences between patients with D0 increase/decrease and different tumor locations were explored. RESULTS: Tumor shifts towards/away from the heart showed no significant association with OS (p = 0.91). Distance between PTV and heart correlated significantly (PCC = 0.18) with shifts to the heart. Cox model did not validate in our cohort. Heart presented baseline shifts positively correlated with tumor baseline shifts in all three directions (PCC ≥ 0.38; p < 0.001). Counterintuitively, patients experiencing increased D0 during treatment showed significantly better OS (p = 0.0077). Upper-lobe tumor patients with increased D0 had lower D0 than those with decreased D0 (right-upper-lobe p ≤ 0.018). CONCLUSIONS: In our SBRT cohort, the shifts towards the heart were not associated with worse OS. Moderate correlations were found between tumor and heart baseline shifts in each direction. Moreover, the distance between the PTV and the heart showed a significant correlation with shifts to the heart.

Impact of cone-beam computed tomography artifacts on dose calculation accuracy for lung cancer.

BACKGROUND: To implement image-guided adaptive radiotherapy (IGART), many studies investigated dose calculations on cone-beam computed tomography (CBCT). A high HU accuracy is crucial for a high dose calculation accuracy and many imaging sites showed satisfactory results. It has been shown that the dose calculation accuracy for lung cancer lags behind. PURPOSE: To examine why the dose calculation accuracy for lung is insufficient, the relative effects of the field-of-view (FOV), breathing motion, and scatter on dose calculation accuracy were studied. METHODS: A framework was built to simulate CBCT scans for lung cancer patients by forward projecting repeat CT (rCT) scans for two scan geometries: small (SFOV) and medium FOV (MFOV). Breathing motion was modeled by applying a 4D deformation vector field to the mid-position rCT. Scatter was modeled by Monte-Carlo simulations with/without an anti-scatter grid (ASG). Simulated projections were reconstructed using filtered back-projection with/without scatter correction. In case of the SFOV, the CBCT images were patched with the planning CT scan in axial direction. The treatment plan was recalculated on the rCT and simulated CBCT. The mean Hounsfield unit (HU) difference (ΔHUmean), the structural similarity index measure (SSIM), and γ metrics were calculated for the CBCT datasets of various imaging settings. RESULTS: The differences in HU, SSIM and dose calculation accuracy for CBCTs with and without breathing motion were negligible (mean ΔHUmean = 6.4 vs. 13.7, mean SSIM = 0.941 vs. 0.957, mean γ (ref = MFOV) = 0.75). The SFOV resulted in a lower HU (mean ΔHUmean = -9.2 vs. 13.7) and SSIM (mean SSIM = 0.912 vs. 0.957), and therefore in dose differences compared to the MFOV (mean γ = 1.22). Scatter led to considerable discrepancies in all metrics. Adding only the ASG improved the results more than only applying a scatter correction algorithm. Combining ASG and scatter correction algorithm resulted in an even higher dose calculation accuracy. CONCLUSIONS: Scatter and FOV are the main contributors to dose inaccuracies and motion has only a minor effect on dose calculation accuracy. Therefore, utilizing an appropriate scatter correction and FOV is important to achieve sufficient dose calculation accuracy to facilitate IGART for lung.

Intrafraction Motion Management With MR-Guided Radiation Therapy.

High quality radiation therapy requires highly accurate and precise dose delivery. MR-guided radiotherapy (MRgRT), integrating an MRI scanner with a linear accelerator, offers excellent quality images in the treatment room without subjecting patient to ionizing radiation. MRgRT therefore provides a powerful tool for intrafraction motion management. This paper summarizes different sources of intrafraction motion for different disease sites and describes the MR imaging techniques available to visualize and quantify intrafraction motion. It provides an overview of MR guided motion management strategies and of the current technical capabilities of the commercially available MRgRT systems. It describes how these motion management capabilities are currently being used in clinical studies, protocols and provides a future outlook.

On retrospective k-space subsampling schemes for deep MRI reconstruction.

Acquiring fully-sampled MRI k-space data is time-consuming, and collecting accelerated data can reduce the acquisition time. Employing 2D Cartesian-rectilinear subsampling schemes is a conventional approach for accelerated acquisitions; however, this often results in imprecise reconstructions, even with the use of Deep Learning (DL), especially at high acceleration factors. Non-rectilinear or non-Cartesian trajectories can be implemented in MRI scanners as alternative subsampling options. This work investigates the impact of the k-space subsampling scheme on the quality of reconstructed accelerated MRI measurements produced by trained DL models. The Recurrent Variational Network (RecurrentVarNet) was used as the DL-based MRI-reconstruction architecture. Cartesian, fully-sampled multi-coil k-space measurements from three datasets were retrospectively subsampled with different accelerations using eight distinct subsampling schemes: four Cartesian-rectilinear, two Cartesian non-rectilinear, and two non-Cartesian. Experiments were conducted in two frameworks: scheme-specific, where a distinct model was trained and evaluated for each dataset-subsampling scheme pair, and multi-scheme, where for each dataset a single model was trained on data randomly subsampled by any of the eight schemes and evaluated on data subsampled by all schemes. In both frameworks, RecurrentVarNets trained and evaluated on non-rectilinearly subsampled data demonstrated superior performance, particularly for high accelerations. In the multi-scheme setting, reconstruction performance on rectilinearly subsampled data improved when compared to the scheme-specific experiments. Our findings demonstrate the potential for using DL-based methods, trained on non-rectilinearly subsampled measurements, to optimize scan time and image quality.

Stomach Motion and Deformation: Implications for Preoperative Gastric Cancer Radiation Therapy.

PURPOSE: Selection and development of image guided strategies for preoperative gastric radiation therapy requires quantitative knowledge of the various sources of anatomic changes of the stomach. This study aims to investigate the magnitude of interfractional and intrafractional stomach motion and deformation using fiducial markers and 4-dimensional (4D) imaging. METHODS AND MATERIALS: Fourteen patients who underwent preoperative gastric cancer radiation therapy received 2 to 6 fiducial markers distributed throughout the stomach (total of 54 markers) and additional imaging (ie, 1 planning 4D computed tomography [pCT], 20-25 pretreatment 4D cone beam [CB] CTs, 4-5 posttreatment 4D CBCTs). Marker coordinates on all end-exhale (EE) and end-inhale (EI) scans were obtained after a bony anatomy match. Interfractional marker displacements (ie, between EE pCT and all EE CBCTs) were evaluated for 5 anatomic regions (ie, cardia, small curvature, proximal and distal large curvature, and pylorus). Motion was defined as displacement of the center-of-mass of available markers (COMstomach), deformation as the average difference in marker-pair distances. Interfractional (ie, between EE pCT and all EE CBCTs), respiratory (between EE and EI pCT and CBCTs), and pre-post (pre- and posttreatment EE CBCTs) motion and deformation were quantified. RESULTS: The interfractional marker displacement varied per anatomic region and direction, with systematic and random errors ranging from 1.6-8.8 mm and 2.2-8.2 mm, respectively. Respiratory motion varied per patient (median, 3-dimensional [3D] amplitude 5.2-20.0 mm) and day (interquartile range, 0.8-4.2 mm). Regarding COMstomach motion, respiratory motion was larger than interfractional motion (median, 10.9 vs 8.9 mm; P < .0001; Wilcoxon rank-sum), which was larger than pre-post motion (3.6 mm; P < .0001). Interfractional deformations (median, 5.8 mm) were significantly larger than pre-post deformations (2.6 mm; P < .0001), which were larger than respiratory deformation (1.8 mm; P < .0001). CONCLUSIONS: The demonstrated sizable stomach motions and deformations during radiation therapy stress the need for generous nonuniform planning target volume margins for preoperative gastric cancer radiation therapy. These margins can be decreased by daily image guidance and adaptive radiation therapy.

Enhancing the reliability of deep learning-based head and neck tumour segmentation using uncertainty estimation with multi-modal images.

Objective.Deep learning shows promise in autosegmentation of head and neck cancer (HNC) primary tumours (GTV-T) and nodal metastases (GTV-N). However, errors such as including non-tumour regions or missing nodal metastases still occur. Conventional methods often make overconfident predictions, compromising reliability. Incorporating uncertainty estimation, which provides calibrated confidence intervals can address this issue. Our aim was to investigate the efficacy of various uncertainty estimation methods in improving segmentation reliability. We evaluated their confidence levels in voxel predictions and ability to reveal potential segmentation errors.Approach.We retrospectively collected data from 567 HNC patients with diverse cancer sites and multi-modality images (CT, PET, T1-, and T2-weighted MRI) along with their clinical GTV-T/N delineations. Using the nnUNet 3D segmentation pipeline, we compared seven uncertainty estimation methods, evaluating them based on segmentation accuracy (Dice similarity coefficient, DSC), confidence calibration (Expected Calibration Error, ECE), and their ability to reveal segmentation errors (Uncertainty-Error overlap using DSC, UE-DSC).Main results.Evaluated on the hold-out test dataset (n= 97), the median DSC scores for GTV-T and GTV-N segmentation across all uncertainty estimation methods had a narrow range, from 0.73 to 0.76 and 0.78 to 0.80, respectively. In contrast, the median ECE exhibited a wider range, from 0.30 to 0.12 for GTV-T and 0.25 to 0.09 for GTV-N. Similarly, the median UE-DSC also ranged broadly, from 0.21 to 0.38 for GTV-T and 0.22 to 0.36 for GTV-N. A probabilistic network-PhiSeg method consistently demonstrated the best performance in terms of ECE and UE-DSC.Significance.Our study highlights the importance of uncertainty estimation in enhancing the reliability of deep learning for autosegmentation of HNC GTV. The results show that while segmentation accuracy can be similar across methods, their reliability, measured by calibration error and uncertainty-error overlap, varies significantly. Used with visualisation maps, these methods may effectively pinpoint uncertainties and potential errors at the voxel level.

Interobserver variation in tumor delineation of liver metastases using Magnetic Resonance Imaging.

Background and purpose: Magnetic Resonance Imaging (MRI) guided stereotactic body radiotherapy (SBRT) of liver metastases is an upcoming high-precision non-invasive treatment. Interobserver variation (IOV) in tumor delineation, however, remains a relevant uncertainty for planning target volume (PTV) margins. The aims of this study were to quantify IOV in MRI-based delineation of the gross tumor volume (GTV) of liver metastases and to detect patient-specific factors influencing IOV. Materials and methods: A total of 22 patients with liver metastases from three primary tumor origins were selected (colorectal(8), breast(6), lung(8)). Delineation guidelines and planning MRI-scans were provided to eight radiation oncologists who delineated all GTVs. All delineations were centrally peer reviewed to identify outliers not meeting the guidelines. Analyses were performed both in- and excluding outliers. IOV was quantified as the standard deviation (SD) of the perpendicular distance of each observer's delineation towards the median delineation. The correlation of IOV with shape regularity, tumor origin and volume was determined. Results: Including all delineations, average IOV was 1.6 mm (range 0.6-3.3 mm). From 160 delineations, in total fourteen single delineations were marked as outliers after peer review. After excluding outliers, the average IOV was 1.3 mm (range 0.6-2.3 mm). There was no significant correlation between IOV and tumor origin or volume. However, there was a significant correlation between IOV and regularity (Spearman's ρs = -0.66; p = 0.002). Conclusion: MRI-based IOV in tumor delineation of liver metastases was 1.3-1.6 mm, from which PTV margins for IOV can be calculated. Tumor regularity and IOV were significantly correlated, potentially allowing for patient-specific margin calculation.

Patient-reported outcomes after personalised dose-escalation for stage II-III non-small-cell lung cancer patients: Results from the randomised ARTFORCE PET-Boost trial.

BACKGROUND AND PURPOSE: The ultimate challenge in dose-escalation trials lies in finding the balance between benefit and toxicity. We examined patient-reported outcomes (PROs), including health-related quality of life (HRQoL) in patients with locally advanced non-small cell lung cancer (LA-NSCLC), treated with dose-escalated radiotherapy. MATERIALS AND METHODS: The international, randomised, phase 2 ARTFORCE PET-Boost study (NCT01024829) aimed to improve 1-year freedom from local failure rates in patients with stage II-III NSCLC, with a ≥ 4 cm primary tumour. Treatment consisted of an individualised, escalated fraction dose, either to the primary tumour as a whole or to its most FDG-avid subvolume (24 x 3.0-5.4 Gy). Patients received sequential or concurrent chemoradiotherapy, or radiotherapy only. Patients were asked to complete the EORTC QLQ-C30, QLQ-LC13, and the EuroQol-5D at eight timepoints. We assessed the effect of dose-escalation on C30 sum score through mixed-modelling and evaluated clinically meaningful changes for all outcomes. RESULTS: Between Apr-2010 and Sep-2017, 107 patients were randomised; 102 were included in the current analysis. Compliance rates: baseline 86.3%, 3-months 85.3%, 12-months 80.3%; lowest during radiation treatment 35.0%. A linear mixed-effect (LME) model revealed no significant change in overall HRQoL over time, and no significant difference between the two treatment groups. Physical functioning showed a gradual decline in both groups during treatment and at 18-months follow-up, while clinically meaningful worsening of dyspnoea was seen mainly at 3- and 6-months. CONCLUSION: In patients with LA-NSCLC treated with two dose-escalation strategies, the average patient-reported HRQoL remained stable in both groups, despite frequent patient-reported symptoms, including dyspnoea, dysphagia, and fatigue.

A multicentric randomized controlled phase III trial of adaptive and 18F-FDG-PET-guided dose-redistribution in locally advanced head and neck squamous cell carcinoma (ARTFORCE).

BACKGROUND AND PURPOSE: This multicenter randomized phase III trial evaluated whether locoregional control of patients with LAHNSCC could be improved by fluorodeoxyglucose-positron emission tomography (FDG-PET)-guided dose-escalation while minimizing the risk of increasing toxicity using a dose-redistribution and scheduled adaptation strategy. MATERIALS AND METHODS: Patients with T3-4-N0-3-M0 LAHNSCC were randomly assigned (1:1) to either receive a dose distribution ranging from 64-84 Gy/35 fractions with adaptation at the 10thfraction (rRT) or conventional 70 Gy/35 fractions (cRT). Both arms received concurrent three-cycle 100 mg/m2cisplatin. Primary endpoints were 2-year locoregional control (LRC) and toxicity. Primary analysis was based on the intention-to-treat principle. RESULTS: Due to slow accrual, the study was prematurely closed (at 84 %) after randomizing 221 eligible patients between 2012 and 2019 to receive rRT (N = 109) or cRT (N = 112). The 2-year LRC estimate difference of 81 % (95 %CI 74-89 %) vs. 74 % (66-83 %) in the rRT and cRT arm, respectively, was not found statistically significant (HR 0.75, 95 %CI 0.43-1.31,P=.31). Toxicity prevalence and incidence rates were similar between trial arms, with exception for a significant increased grade ≥ 3 pharyngolaryngeal stenoses incidence rate in the rRT arm (0 versus 4 %,P=.05). In post-hoc subgroup analyses, rRT improved LRC for patients with N0-1 disease (HR 0.21, 95 %CI 0.05-0.93) and oropharyngeal cancer (0.31, 0.10-0.95), regardless of HPV. CONCLUSION: Adaptive and dose redistributed radiotherapy enabled dose-escalation with similar toxicity rates compared to conventional radiotherapy. While FDG-PET-guided dose-escalation did overall not lead to significant tumor control or survival improvements, post-hoc results showed improved locoregional control for patients with N0-1 disease or oropharyngeal cancer treated with rRT.

Adaptive and innovative Radiation Treatment FOR improving Cancer treatment outcomE (ARTFORCE); a randomized controlled phase II trial for individualized treatment of head and neck cancer.

BACKGROUND: Failure of locoregional control is the main cause of recurrence in advanced head and neck cancer. This multi-center trial aims to improve outcome in two ways. Firstly, by redistribution of the radiation dose to the metabolically most FDG-PET avid part of the tumour. Hereby, a biologically more effective dose distribution might be achieved while simultaneously sparing normal tissues. Secondly, by improving patient selection. Both cisplatin and Epidermal Growth Factor Receptor (EGFR) antibodies like Cetuximab in combination with Radiotherapy (RT) are effective in enhancing tumour response. However, it is unknown which patients will benefit from either agent in combination with irradiation. We will analyze the predictive value of biological markers and (89)Zr-Cetuximab uptake for treatment outcome of chemoradiation with Cetuximab or cisplatin to improve patient selection. METHODS: ARTFORCE is a randomized phase II trial for 268 patients with a factorial 2 by 2 design: cisplatin versus Cetuximab and standard RT versus redistributed RT. Cisplatin is dosed weekly 40 mg/m(2) for 6 weeks. Cetuximab is dosed 250 mg/m(2) weekly (loading dose 400 mg/m(2)) for 6 weeks. The standard RT regimen consists of elective RT up to 54.25 Gy with a simultaneous integrated boost (SIB) to 70 Gy in 35 fractions in 6 weeks. Redistributed adaptive RT consists of elective RT up to 54.25 Gy with a SIB between 64-80 Gy in 35 fractions in 6 weeks with redistributed dose to the gross tumour volume (GTV) and clinical target volume (CTV), and adaptation of treatment for anatomical changes in the third week of treatment.Patients with locally advanced, biopsy confirmed squamous cell carcinoma of the oropharynx, oral cavity or hypopharynx are eligible.Primary endpoints are: locoregional recurrence free survival at 2 years, correlation of the median (89)Zr-cetuximab uptake and biological markers with treatment specific outcome, and toxicity. Secondary endpoints are quality of life, swallowing function preservation, progression free and overall survival. DISCUSSION: The objective of the ARTFORCE Head and Neck trial is to determine the predictive value of biological markers and (89)Zr-Cetuximab uptake, as it is unknown how to select patients for the appropriate concurrent agent. Also we will determine if adaptive RT and dose redistribution improve locoregional control without increasing toxicity.ClinicalTrials.gov Identifier: NCT01504815.