Deformable Registration Technique Research Articles

Impact of interplay effects on spot scanning proton therapy with motion mitigation techniques for lung cancer: SFUD versus robustly optimized IMPT plans utilizing a four-dimensional dynamic dose simulation tool

BackgroundThe interaction between breathing motion and scanning beams causes interplay effects in spot-scanning proton therapy for lung cancer, resulting in compromised treatment quality. This study investigated the effects and clinical robustness of two types of spot-scanning proton therapy with motion-mitigation techniques for locally advanced non-small cell lung cancer (NSCLC) using a new simulation tool (4DCT-based dose reconstruction).MethodsThree-field single-field uniform dose (SFUD) and robustly optimized intensity-modulated proton therapy (IMPT) plans combined with gating and re-scanning techniques were created using a VQA treatment planning system for 15 patients with locally advanced NSCLC (70 GyRBE/35 fractions). In addition, gating windows of three or five phases around the end-of-expiration phase and two internal gross tumor volumes (iGTVs) were created, and a re-scanning number of four was used. First, the static dose (SD) was calculated using the end-of-expiration computed tomography (CT) images. The four-dimensional dynamic dose (4DDD) was then calculated using the SD plans, 4D-CT images, and the deformable image registration technique on end-of-expiration CT. The target coverage (V98%, V100%), homogeneity index (HI), and conformation number (CN) for the iGTVs and organ-at-risk (OAR) doses were calculated for the SD and 4DDD groups and statistically compared between the SD, 4DDD, SFUD, and IMPT treatment plans using paired t-test.ResultsIn the 3- and 5-phase SFUD, statistically significant differences between the SD and 4DDD groups were observed for V100%, HI, and CN. In addition, statistically significant differences were observed for V98%, V100%, and HI in phases 3 and 5 of IMPT. The mean V98% and V100% in both 3-phase plans were within clinical limits (> 95%) when interplay effects were considered; however, V100% decreased to 89.3% and 94.0% for the 5-phase SFUD and IMPT, respectively. Regarding the significant differences in the deterioration rates of the dose volume histogram (DVH) indices, the 3-phase SFUD plans had lower V98% and CN values and higher V100% values than the IMPT plans. In the 5-phase plans, SFUD had higher deterioration rates for V100% and HI than IMPT.ConclusionsInterplay effects minimally impacted target coverage and OAR doses in SFUD and robustly optimized IMPT with 3-phase gating and re-scanning for locally advanced NSCLC. However, target coverage significantly declined with an increased gating window. Robustly optimized IMPT showed superior resilience to interplay effects, ensuring better target coverage, prescription dose adherence, and homogeneity than SFUD.Trial registration: None.

Evaluating the Necessity of Adaptive RT and the Role of Deformable Image Registration in Lung Cancer with Different Pathologic Classifications.

This study aimed to analyze differential radiotherapy (RT) responses according to the pathological type of lung cancer to see the possibility of applying adaptive radiotherapy (ART). ART planning with resampled-computed tomography was conducted for a total of 30 patients (20 non-small-cell lung cancer patients and 10 small-cell lung cancer patients) using a deformable image registration technique to reveal gross tumor volume (GTV) changes according to the duration of RT. The small-cell lung cancer group demonstrated an average GTV reduction of 20.95% after the first week of initial treatment (p = 0.001), whereas the adenocarcinoma and squamous cell carcinoma groups showed an average volume reduction of 20.47% (p = 0.015) and 12.68% in the second week. The application of ART according to the timing of GTV reduction has been shown to affect changes in radiation dose irradiated to normal tissues. This suggests that ART applications may have to be different depending on pathological differences in lung cancer. Through these results, the present study proposes the possibility of personalized treatment options for individual patients by individualizing ART based on specific radiation responses by pathologic types of lung cancer.

MICDIR: Multi-scale inverse-consistent deformable image registration using UNetMSS with self-constructing graph latent.

Image registration is the process of bringing different images into a common coordinate system - a technique widely used in various applications of computer vision, such as remote sensing, image retrieval, and, most commonly, medical imaging. Deep learning based techniques have been applied successfully to tackle various complex medical image processing problems, including medical image registration. Over the years, several image registration techniques have been proposed using deep learning. Deformable image registration techniques such as Voxelmorph have been successful in capturing finer changes and providing smoother deformations. However, Voxelmorph, as well as ICNet and FIRE, do not explicitly encode global dependencies (i.e. the overall anatomical view of the supplied image) and, therefore, cannot track large deformations. In order to tackle the aforementioned problems, this paper extends the Voxelmorph approach in three different ways. To improve the performance in case of small as well as large deformations, supervision of the model at different resolutions has been integrated using a multi-scale UNet. To support the network to learn and encode the minute structural co-relations of the given image-pairs, a self-constructing graph network (SCGNet) has been used as the latent of the multi-scale UNet - which can improve the learning process of the model and help the model to generalise better. And finally, to make the deformations inverse-consistent, cycle consistency loss has been employed. On the task of registration of brain MRIs, the proposed method achieved significant improvements over ANTs and VoxelMorph, obtaining a Dice score of 0.8013 ± 0.0243 for intramodal and 0.6211 ± 0.0309 for intermodal, while VoxelMorph achieved 0.7747 ± 0.0260 and 0.6071 ± 0.0510, respectively.

A hybrid deformable registration method to generate motion-compensated 3D virtual MRI for fusion with interventional real-time 3D ultrasound.

Ultrasound is often the preferred modality for image-guided therapy or treatment in organs such as liver due to real-time imaging capabilities. However, the reduced conspicuity of tumors in ultrasound images adversely impacts the precision and accuracy of treatment delivery. This problem is compounded by deformable motion due to breathing and other physiological activity. This creates the need for a fusion method to align interventional US with pre-interventional modalities that provide superior soft-tissue contrast (e.g., MRI) to accurately target a structure-of-interest and compensate for liver motion. In this work, we developed a hybrid deformable fusion method to align 3D pre-interventional MRI and 3D interventional US volumes to target the structures-of-interest in liver accurately in real-time. The deformable multimodal fusion method involved an offline alignment of a pre-intervention MRI with a pre-intervention US volume using a traditional registration method, followed by real-time prediction of deformation using a trained deep-learning model between interventional US volumes across different respiratory states. This framework enables motion-compensated MRI-US image fusion in real-time for image-guided treatment. The proposed hybrid deformable registration method was evaluated on three healthy volunteers across the pre-intervention MRI and 20 US volume pairs in the free-breathing respiratory cycle. The mean Euclidean landmark distance of three homologous targets in all three volunteers was less than 3 mm for percutaneous liver procedures. Preliminary results show that clinically acceptable registration accuracies for near real-time, deformable MRI-US fusion can be achieved by our proposed hybrid approach. The proposed combination of traditional and deep-learning deformable registration techniques is thus a promising approach for motion-compensated MRI-US fusion to improve targeting in image-guided liver interventions.

Detailed dosimetric evaluation of inter-fraction and respiratory motion in lung stereotactic body radiation therapy based on daily 4D cone beam CT images

Objective. Periodic respiratory motion and inter-fraction variations are sources of geometric uncertainty in stereotactic body radiation therapy (SBRT) of pulmonary lesions. This study extensively evaluates and validates the separate and combined dosimetric effect of both factors using 4D-CT and daily 4D-cone beam CT (CBCT) dose accumulation scenarios. Approach. A first cohort of twenty early stage or metastatic disease lung cancer patients were retrospectively selected to evaluate each scenario. The planned-dose (3DRef) was optimized on a 3D mid-position CT. To estimate the dosimetric impact of respiratory motion (4DRef), inter-fractional variations (3DAcc) and the combined effect of both factors (4DAcc), three dose accumulation scenarios based on 4D-CT, daily mid-cone beam CT (CBCT) position and 4D-CBCT were implemented via CT-CT/CT-CBCT deformable image registration (DIR) techniques. Each scenario was compared to 3DRef. A separate cohort of ten lung SBRT patients was selected to validate DIR techniques. The distance discordance metric (DDM) was implemented per voxel and per patient for tumor and organs at risk (OARs), and the dosimetric impact for CT-CBCT DIR geometric errors was calculated. Main results. Median and interquartile range (IQR) of the dose difference per voxel were 0.05/2.69 Gy and −0.12/2.68 Gy for and For the IQR was considerably smaller −0.15/0.78 Gy. These findings were confirmed by dose volume histogram parameters calculated in tumor and OARs. For CT-CT/CT-CBCT DIR validation, DDM (95th percentile) was highest for heart (6.26 mm)/spinal cord (8.00 mm), and below 3 mm for tumor and the rest of OARs. The dosimetric impact of CT-CBCT DIR errors was below 2 Gy for tumor and OARs. Significance. The dosimetric impact of inter-fraction variations were shown to dominate those of periodic respiration in SBRT for pulmonary lesions. Therefore, treatment evaluation and dose-effect studies would benefit more from dose accumulation focusing on day-to-day changes then those that focus on respiratory motion.

Correlation between CT Value on Lung Subtraction CT and Radioactive Count on Perfusion Lung Single Photon Emission CT in Chronic Thromboembolic Pulmonary Hypertension.

Lung subtraction CT (LSCT), the subtraction of noncontrast CT from CT pulmonary angiography (CTPA) without spatial misregistration, is easily applicable by utilizing a software-based deformable image registration technique without additional hardware and permits the evaluation of lung perfusion as iodine accumulation, similar to that observed in perfusion lung single photon emission CT (PL-SPECT). The aim of this study was to use LSCT to newly assess the quantitative correlation between the CT value on LSCT and radioactive count on PL-SPECT as a reference and validate the quantification of lung perfusion by measuring the CT value in chronic thromboembolic pulmonary hypertension (CTEPH). We prospectively enrolled 47 consecutive patients with CTEPH undergoing both LSCT and PL-SPECT; we used noncontrast CT, CTPA, and LSCT to measure CT values and PL-SPECT to measure radioactive counts in areas representing three different perfusion classes-no perfusion defect, subsegmental perfusion defect, and segmental perfusion defect; we compared CT values on noncontrast CT, CTPA, and LSCT and radioactive counts on PL-SPECT among the three classes, then assessed the correlation between them. Both the CT values and radioactive counts differed significantly among the three classes (p < 0.01 for all) and showed weak correlation (ρ = 0.38) by noncontrast CT, moderate correlation (ρ = 0.61) by CTPA, and strong correlation (ρ = 0.76) by LSCT. The CT value measurement on LSCT is a novel quantitative approach to assess lung perfusion in CTEPH and only correlates strongly with radioactive count measurement on PL-SPECT.

Contouring lumbosacral plexus nerves with MR neurography and MR/CT deformable registration technique.

It is difficult to contour nerve structures with the naked eye due to poor differentiation between the nerve structures with other soft tissues on CT images. Magnetic resonance neurography (MRN) has the advantage in nerve visualization. The purpose of this study is to identify one MRN sequence to better assist the delineation of the lumbosacral plexus (LSP) nerves to assess the radiation dose to the LSP using the magnetic resonance (MR)/CT deformable coregistration technique. A total of 18 cases of patients with prostate cancer and one volunteer with radiation-induced lumbosacral plexopathy (RILSP) were enrolled. The data of simulation CT images and original treatment plans were collected. Two MRN sequences (Lr_NerveVIEW sequence and Cs_NerveVIEW sequence) were optimized from a published MRN sequence (3D NerveVIEW sequence). The nerve visualization ability of the Lr_NerveVIEW sequence and the Cs_NerveVIEW sequence was evaluated via a four-point nerve visualization score (NVS) scale in the first 10 patients enrolled to determine the better MRN sequence for assisting nerve contouring. Deformable registration was applied to the selected MRN sequence and simulation CT images to get fused MR/CT images, on which the LSP was delineated. The contouring of the LSP did not alter treatment planning. The dosimetric data of the LSP nerve were collected from the dose-volume histogram in the original treatment plans. The data of the maximal dose (Dmax) and the location of the maximal radiation point received by the LSP structures were collected. The Cs_NerveVIEW sequence gained lower NVS scores than the Lr_NerveVIEW sequence (Z=-2.887, p=0.004). The LSP structures were successfully created in 18 patients and one volunteer with MRN (Lr_NerveVIEW)/CT deformable registration techniques, and the LSP structures conformed with the anatomic distribution. In the patient cohort, the percentage of the LSP receiving doses exceeding 50, 55, and 60 Gy was 68% (12/18), 33% (6/18), and 17% (3/18), respectively. For the volunteer with RILSP, the maximum irradiation dose to his LSP nerves was 69 Gy. The Lr_NerveVIEW MRN sequence performed better than the Cs_NerveVIEW sequence in nerve visualization. The dose in the LSP needs to be measured to understand the potential impact on treatment-induced neuropathy.

Accumulated Dose Prediction for Assisting Radiation Treatment in Cervical Cancer

<h3>Purpose/Objective(s)</h3> For cervical cancer treated with combined external beam radiotherapy (EBRT) and brachytherapy (BT), ideally, the treatment optimization and evaluation should take the accumulated dose distribution into account. However, in current clinical practice different treatment parts are linked to different planning images and optimized independently, resulting in multiple dose distributions. In this study, we used a combination of deformable image registration (DIR) and deep learning techniques to predict the accumulated dose for assisting treatment planning and dose estimation. <h3>Materials/Methods</h3> The dose distributions of both EBRT and BT plans from 32 cervical cancer patients were used for this study. Firstly, the BT images were registered to the EBRT images using DIR to obtain the accumulated dose distributions. Then, a deep learning model based on a 101-layers ResNet was trained to predict pixel-wise dose distributions. The input data was combined anatomic maps including the CT images, structure maps and distance maps. The output data was the corresponding accumulated dose maps. We used 5-fold cross-validation to test the performance of the prediction model. The evaluation parameters included the voxel-based mean absolute error (MAE) and dice similarity coefficient (DSC) of isodose volumes for the prescription dose (PD) of EBRT (45/50 Gy) and the summed PD of EBRT and BT (76.25/82 Gy_EQD2). Finally, we randomly selected 10 patients to validate the effect of the dose prediction on treatment planning. For each patient, two EBRT plans were separately designed by a single physicist in the cases of unknowing and knowing the predicted accumulated dose distribution. The resulting two EBRT doses were separately accumulated with BT dose for comparison. <h3>Results</h3> In the comparison between the predicted and original dose distributions, the mean MAE of the whole body, normal tissue, bladder and rectum were 3.73±1.58, 3.27±1.06, 6.55±3.10 and 8.05±4.95 Gy, respectively. The mean DSC of the isodose volumes for 45/50 Gy and 76.25/82 Gy_EQD2 were 0.89±0.02 and 0.69±0.09, respectively. Compared to the planning without knowing the predicted dose, the planning with knowing the predicted dose resulted in a lower D2cc (-1.2±0.80 Gy, p=0.04), V40 (-3.23±2.19%, p=0.002), V50 (-2.19±%1.34, p=0.001) and V60 (-0.79±0.82%, p=0.02) of the rectum. For the bladder, the V40 of were decreased by 1.48±0.58% (p<0.001) but other dosimetric parameters had no significant difference. <h3>Conclusion</h3> The accumulated dose prediction could help to estimate and further decrease the doses to organs at risk. The predicted dose is recommended to be used as a reference during treatment planning in order to pursue a superior accumulated dose distribution for the combined cervical cancer treatment.

PD-0316 Contouring lumbosacral plexus based on MR neurography and MR/CT deformable registration technique

PD-0316 Contouring lumbosacral plexus based on MR neurography and MR/CT deformable registration technique

Changes in Organs at Risk Dose Estimates Due to Anatomic Changes and Registration Method During Head and Neck Radiotherapy and its Relationship to Percent Weight Loss

<h3>Purpose/Objective(s)</h3> Weight loss and anatomic changes are common during radiation therapy for head and neck cancers. The best method for assessing the impact of anatomic changes on dose distribution is unknown. We compared planned doses to organs at risk (OARs) to delivered dose estimates using a cone beam computed tomography (CBCT) from the final week of treatment. <h3>Materials/Methods</h3> Ten patients who underwent bilateral neck radiation and had weekly extended CBCT's available to capture critical OARs were identified. CBCT from the final week was reconstructed to 2mm slices to match the resolution of the CT simulation (CTsim). The CBCT's were registered to the sim using both rigid and deformable registration techniques. The parotid glands, spinal cord, and skin (defined as 3mm from body contour) were contoured on the CBCT's. Mean dose to the parotids, and maximum dose (D0.035cc) to the spinal cord and skin were extracted from the treatment plan based on CTsim, rigid CBCT and deformed CBCT respectively. Independent t-tests were used to determine and compare mean percent (%) dose changes in OAR's from initially planned to delivered dose estimates using the rigid and deformed CBCT's. Linear regression was then performed to identify which set of OAR doses were the best predictors of the variation in % mean weight change during the treatment course. Statistical software was used for the statistical analyses. <h3>Results</h3> Among the 10 patients analyzed, mean % change in mean right parotid dose was -6.99% between planned and rigid CBCT estimated dose and +12.02% between planned and deformed CBCT estimated dose, for a mean total difference of 19.01% (SE 5.61%, p=0.004). These differences were statistically significant. A similar trend was observed for the left parotid gland, with mean % changes of -1.67%, +6% and 7.67% respectively (SE 5.62%, p=0.19). No differences were observed for cord (mean difference 0.19%, SE 1.27%, p = 0.89) or skin (mean difference 0.26%, SE 2.83%, p = 0.92). On regression analysis the delivered dose estimates from the rigid CBCT model were a better predictor of % weight change (R²=0.543) compared to the deformed CBCT (R²=0.463) or CT sim planned doses (R²=0.460) <h3>Conclusion</h3> The data shows a significant difference in estimating dose delivered to the right parotid gland using final week CBCT between using rigid and deformable registration techniques. Change in dose to left parotid gland also trended towards significance and may prove to be meaningful with a larger sample size. There were no significant differences in cord or skin % dose changes. We found that doses collected from the Rigid CBCT were the best predictors of % weight change during the whole course of treatment on regression analysis. Future work will focus on mid treatment dose estimates using a larger sample size, and correlation with clinical outcomes.

Automated Extraction of Anatomical Measurements From Temporal Bone CT Imaging.

Proposed methods of minimally invasive and robot-assisted procedures within the temporal bone require measurements of surgically relevant distances and angles, which often require time-consuming manual segmentation of preoperative imaging. This study aims to describe an automatic segmentation and measurement extraction pipeline of temporal bone cone-beam computed tomography (CT) scans. Descriptive study of temporal bone measurements. Academic institution. A propagation template composed of 16 temporal bone CT scans was formed with relevant anatomical structures and landmarks manually segmented. Next, 52 temporal bone CT scans were autonomously segmented using deformable registration techniques from the Advanced Normalization Tools Python package. Anatomical measurements were extracted via in-house Python algorithms. Extracted measurements were compared to ground truth values from manual segmentations. Paired t test analyses showed no statistical difference between measurements using this pipeline and ground truth measurements from manually segmented images. Mean (SD) malleus manubrium length was 4.39 (0.34) mm. Mean (SD) incus short and long processes were 2.91 (0.18) mm and 3.53 (0.38) mm, respectively. The mean (SD) maximal diameter of the incus long process was 0.74 (0.17) mm. The first and second facial nerve genus had mean (SD) angles of 68.6 (6.7) degrees and 111.1 (5.3) degrees, respectively. The facial recess had a mean (SD) span of 3.21 (0.46) mm. Mean (SD) minimum distance between the external auditory canal and tegmen was 3.79 (1.05) mm. This is the first study to automatically extract relevant temporal bone anatomical measurements from CT scans using segmentation propagation. Measurements from these models can streamline preoperative planning, improve future segmentation techniques, and help develop future image-guided or robot-assisted systems for temporal bone procedures.

3-Dimensional Neurosurgical Planning Using Multilevel CT-MRI Image Registration and Fusion.

Each imaging modalities information is supplemented by medical imaging technologies. The surgical planning platform is improved by fusing these complementary pictures into the same space and conducting 3D volume rendering and depth peeling on fused sequences. The use of 3D visualisation of fused sequences such as CT-MRI of the brain can help locate the point of incision on the cranium and the space lesion. On CT-MRI sequences, image registration is done, followed by fusion, utilising pixel-based and feature-based fusion algorithms, respectively. On the brain CT-MRI image sequences, a unique multilevel registration technique is used in place of the traditional image registration strategy, which has drawbacks. The Multi-resolution registration technique is used for level one registration, and the Bspline Deformable Registration Technique is used for level two registration. The results of the multilevel registration procedure are then used to execute feature-based fusion. This is followed by 3D volume rendering on the fused sequence of CT and registered MRI sequence. The results of fusion are evaluated between the resultants of both level-1 and level-2 registration. The Q4 index and the Correlation Coefficient (CC) are two potential measures for estimating fusion results. Both level-1 & 2 registration approaches yield fusion results. For level-1 & 2 fusion, the average CC measured across all image pairs are 0.71 and 0.85, respectively, while Q4 measured 0.21 and 0.46 for level-1 & 2, respectively. At level-2 registration, both CC and Q4 exhibit an improvement in fusion. Using VTK angle and distance widgets for measuring distances and angles on the 3D model, improving path planning capabilities. The proposed research creates a computer-aided platform for better neurosurgical planning. The multilevel registration method produced promising fusion results and laid the groundwork for enhanced 3D viewing of fused CT-MRI sequences using depth peeling. Distance and angle measurements improve surgical planning capability.

Multi-contrast four-dimensional magnetic resonance imaging (MC-4D-MRI): Development and initial evaluation in liver tumor patients.

To develop a novel multi-contrast four-dimensional magnetic resonance imaging (MC-4D-MRI) technique that expands single image contrast 4D-MRI to a spectrum of native and synthetic image contrasts and to evaluate its feasibility in liver tumor patients. The MC-4D-MRI technique integrates multi-parametric MRI fusion, 4D-MRI, and deformable image registration (DIR) techniques. The fusion technique consists of native MRI as input, image pre-processing, fusion algorithm, adaptation, and fused multi-contrast MRI as output. Four-dimensional deformation vector fields (4D-DVF) were generated from an original T2/T1-w 4D-MRI by deforming end-of-inhalation (EOI) to nine other phase volumes via DIR. The 4D-DVF were applied to multi-contrast MRI to generate a spectrum of 4D-MRI in different image contrasts. The MC-4D-MRI technique was evaluated in five liver tumor patients on tumor contrast-to-noise ratio (CNR), internal target volume (ITV) contouring consistency, diaphragm motion range, and tumor motion trajectory; and in digital anthropomorphic phantoms on 4D-DIR introduced errors in tumor motion range, centroid location, extent, and volume. MC-4D-MRI consisting of 4D-MRIs in native image contrasts (T1-w, T2-w, and T2/T1-w) and synthetic image contrasts, such as tumor-enhanced contrast (TEC) were generated in five liver tumor patients. Patient tumor CNR increased from 2.6±1.8 in the T2/T1-w MRI, to -4.4±2.4, 6.6±3.0, and 9.6±3.9 in the T1-w, T2-w, and TEC MRI, respectively. Patient ITV inter-observer mean Dice similarity coefficient (mDSC) increased from 0.65±0.10 in the original T2/T1-w 4D-MRI, to 0.76±0.14, 0.77±0.12, and 0.86±0.05 in the T1-w, T2-w, and TEC 4D-MRI, respectively. Patient diaphragm motion range absolute differences between the three new 4D-MRIs and original T2/T1-w 4D-MRI were 1.2±1.3, 0.3±0.7, and 0.5±0.5mm, respectively. Patient tumor displacement phase-averaged absolute differences between the three 4D-MRIs and the original 4D-MRI were 0.72±0.33, 0.62±0.54, and 0.74±0.43mm in the superior-inferior (SI) direction, and 0.59±0.36, 0.51±0.30, and 0.50±0.24mm in the anterior-posterior (AP) direction, respectively. In the digital phantoms, phase-averaged absolute tumor centroid shift caused by the 4D-DIR were at or below 0.5mm in SI, AP, and left-right (LR) directions. We developed an MC-4D-MRI technique capable of expanding single image contrast 4D-MRI along a new dimension of image contrast. Initial evaluations in liver tumor patients showed enhancements in image contrast variety, tumor contrast, and ITV contouring consistencies using MC-4D-MRI. The technique might offer new perspectives on the image contrast of MRI and 4D-MRI in MR-guided radiotherapy.

Using CBCT to Monitor Treatment Dose Delivery Accuracy in Head and Neck Proton Radiotherapy

<h3>Purpose/Objective(s)</h3> In comparison with photon radiotherapy, intensity modulated proton therapy (IMPT) has higher sensitivities to anatomical and geometrical changes. In IMPT, variations in tumor size, patient geometry and setup can compromise target coverage or augment dose to nearby organs at risk (OAR). As a result, weekly quality assurance CT scans (QACTs) are commonly used to monitor treatment accuracy and provide indicators for adaptive radiotherapy (ART). This study presents the validation of a tool that applies deformable image registration and image processing techniques to generate accurate noise-free CBCT-based virtual CTs used for proton dose calculation to monitor treatment accuracy in head and neck (HN) patients. <h3>Materials/Methods</h3> A single-institution retrospective IRB-approved analysis was conducted on 25 randomly selected HN patients (15 [58%] males, 10 [42%] females) treated with IMPT at our institution between January 2019 and August 2020. Patients had same-day QACT and CBCTs (total of 194 scans) which were imported into a research version of the treatment planning system. A vendor-developed prototype (RaySearch Laboratories, Sweden) tool was used to correct CBCT artifacts and subsequently create a virtual-CT (vCTs) by deforming the planning CT (pCT). The treatment plan was subsequently recomputed on both the QACT and vCT scans. Differences for dose parameters of clinical target volumes (CTVs) and primary OARs such as oral cavity, larynx, spinal cord, mandible, and parotids were compared between vCT and the reference QACT image sets. <h3>Results</h3> Patient averaged differences between CTV coverage parameters estimated from vCT and QACTs were within a 3% standard deviation; CTV mean doses were 0.4% ± 0.7%, and D_2% and D_95% were 0.8% ± 1.6%, and 0.9% ± 1.7%, respectively. Similarly, good correspondence was observed for CTV V_95% (difference = 0.3% ± 1.0%). High agreement was also observed for OARs mean dose (parotid 0.6% ± 1.0%) and max doses (D_0.01cc): oral cavity (0.9% ± 1.5%), larynx (0.8% ± 1.2%), spinal cord (0.7% ± 1.4%), and mandible (1.0% ± 1.2%). Except for 5 cases where the vCT suffered from uncorrected CBCT artifacts, vCT and QACT estimated changes in target coverage metrics (in percent) with respect to the planning CT were within 3% agreement. Virtual-CTs also accurately indicated scenarios where adaptive radiotherapy (e.g., replan) was initiated based on the QACTs. <h3>Conclusion</h3> For all patients, the dose distributions computed on the vCTs closely matched the reference QACT scans. Our results suggest that CBCT-based dosimetry can be used to evaluate treatment plans, and guide the decision for ART over the treatment course.

Personalized dosimetry of 177Lu-DOTATATE: a comparison of organ- and voxel-level approaches using open-access images

177Lu-DOTATATE (Lutathera®) enables targeted radionuclide therapy of neuroendocrine tumors expressing somatostatin receptor type 2. Though patient-specific dosimetry estimates may be clinically important for predicting absorbed dose-effect relationships, there are multiple relevant dosimetry paradigms which are distinct in terms of clinical effort, numerical output and added-value. This work compares three different approaches for 177Lu-DOTATATE dosimetry, including 1) an organ-level approach based on reference phantom MIRD S-values scaled to patient-specific organ masses (MIRDcalc), 2) an organ-level approach based on Monte Carlo simulation in a patient-specific mesh phantoms (PARaDIM), and 3) a 3D approach based on Monte Carlo simulation in patient-specific voxel phantoms. Method. Serial quantitative SPECT/CT images for two patients receiving 177Lu-DOTATATE therapy were obtained from archive in the Deep Blue database. For each patient, the serial CT images were co-registered to the first time point CT using a deformable registration technique aided by virtual landmarks placed in the kidney pelves and the lesion foci. The co-registered SPECT images were integrated voxel-wise to generate time-integrated activity maps. Lesions, kidneys, liver, spleen, lungs, compact bone, spongiosa, and rest of body were segmented at the first imaging time point and overlaid on co-registered integrated activity maps. The resultant segmentation was used for three purposes: 1) to generate patient-specific phantoms, 2) to determine organ-level time-integrated activities, and 3) to generate dose volume histograms from 3D voxel-based calculations. Results. Mean absorbed doses were computed for lesions and 48 tissues with MIRDcalc software. Mean organ absorbed doses and dose volume histograms were obtained for lesions and 6 tissues with the voxel Monte Carlo approach. Lesion- and organ-level absorbed dose estimates agreed within ±26% for the lesions and ±13% for the critical organs, among the different methods tested. Overall good agreement was observed with the dosimetry estimates from the NETTER-1 trial. Conclusions. For personalized 177Lu-DOTATATE dosimetry, a combined approach was determined to be valuable, which utilized two dose calculation methods supported by a single image processing workflow. In the absence of quantitative imaging limitations, the voxel Monte Carlo method likely provides valuable information to guide treatment by considering absorbed dose non-uniformity in lesions and organs at risk. The patient-scaled reference phantom method also provides valuable information, including absorbed dose estimates for non-segmented organs, and more accurate dose estimates for complex radiosensitive organs including the active marrow.

Deformation analysis of surface and bronchial structures in intraoperative pneumothorax using deformable mesh registration

The positions of nodules can change because of intraoperative lung deflation, and the modeling of pneumothorax-associated deformation remains a challenging issue for intraoperative tumor localization. In this study, we introduce spatial and geometric analysis methods for inflated/deflated lungs and discuss heterogeneity in pneumothorax-associated lung deformation. Contrast-enhanced CT images simulating intraoperative conditions were acquired from live Beagle dogs. The images contain the overall shape of the lungs, including all lobes and internal bronchial structures, and were analyzed to provide a statistical deformation model that could be used as prior knowledge to predict pneumothorax. To address the difficulties of mapping pneumothorax CT images with topological changes and CT intensity shifts, we designed deformable mesh registration techniques for mixed data structures including the lobe surfaces and the bronchial centerlines. Three global-to-local registration steps were performed under the constraint that the deformation was spatially continuous and smooth, while matching visible bronchial tree structures as much as possible. The developed framework achieved stable registration with a Hausdorff distance of less than 1 mm and a target registration error of less than 5 mm, and visualized deformation fields that demonstrate per-lobe contractions and rotations with high variability between subjects. The deformation analysis results show that the strain of lung parenchyma was 35% higher than that of bronchi, and that deformation in the deflated lung is heterogeneous.

Use of deformable image registration techniques to estimate dose to organs at risk following prostate external beam radiation therapy and high-dose-rate brachytherapy.

PurposeThe purpose of this investigation was to examine differences in estimates of accumulated rectal dose when using deformable image registration (DIR) compared with rigid image registration (RIR) methods, and parameter addition methods for combined transrectal ultrasound (TRUS)-based high-dose-rate brachytherapy (HDR-BT) and external beam radiation therapy (EBRT) treatments of prostate cancer.Material and methodsIn this retrospective study, data from 10 patients who had previously received HDR-BT in one 15 Gy fraction, followed by 46 Gy EBRT in twenty-three fractions were used. To estimate total combined dose to the rectum, dose accumulation using both DIR and RIR methods were compared with parameter addition methods, which assume the same region of rectal anatomy receives the maximum dose from both treatment modalities. For both rigid and deformable image registration techniques, the quality of image registration was evaluated through metrics, including mean distance to agreement and dice similarity coefficient of prostate contours. Total D1cc and D2cc for the rectum was calculated and compared using each method.ResultsThe parameter addition methods predicted the highest accumulated dose to the rectum. On average, the predicted D2cc dose was higher than that calculated by the DIR method by 6.59 Gy EQD2 (range, –3.03 to 13.68 Gy EQD2) for partial parameter addition (PPA), and 4.88 Gy EQD2 (range, –3.41 to 11.97 Gy EQD2) for the full parameter addition (FPA) methods. Similarly, RIR predicted higher average doses compared with DIR, with a difference of 3.46 Gy EQD2 (range, –5.50 to 7.90 Gy EQD2). The results showed a significant difference between DIR and parameter addition methods for dose estimation.ConclusionsThis retrospective study demonstrates significant differences in accumulated rectal dose prediction using different image registration methods. Each method has limitations in its application, and when used with real-time HDR-BT dose planning, awareness of these limitations is essential.

Assessment of Ablative Margin After Microwave Ablation for Hepatocellular Carcinoma Using Deep Learning-Based Deformable Image Registration.

Aim: To assess the ablative margin (AM) after microwave ablation (MWA) for hepatocellular carcinoma (HCC) with a deep learning-based deformable image registration (DIR) technique and analyze the relation between the AM and local tumor progression (LTP).Patients and Methods: From November 2012 to April 2019, 141 consecutive patients with single HCC (diameter ≤ 5 cm) who underwent MWA were reviewed. Baseline characteristics were collected to identify the risk factors for the determination of LTP after MWA. Contrast-enhanced magnetic resonance imaging scans were performed within 1 month before and 3 months after treatment. Complete ablation was confirmed for all lesions. The AM was measured based on the margin size between the tumor region and the deformed ablative region. To correct the misalignment, DIR between images before and after ablation was achieved by an unsupervised landmark-constrained convolutional neural network. The patients were classified into two groups according to their AMs: group A (AM ≤ 5 mm) and group B (AM > 5 mm). The cumulative LTP rates were compared between the two groups using Kaplan–Meier curves and the log-rank test. Multivariate analyses were performed on clinicopathological variables to identify factors affecting LTP.Results: After a median follow-up period of 28.9 months, LTP was found in 19 patients. The mean tumor and ablation zone sizes were 2.3 ± 0.9 cm and 3.8 ± 1.2 cm, respectively. The mean minimum ablation margin was 3.4 ± 0.7 mm (range, 0–16 mm). The DIR technique had higher AUC for 2-year LTP without a significant difference compared with the registration assessment without DL (P = 0.325). The 6-, 12-, and 24-month LTP rates were 9.9, 20.6, and 24.8%, respectively, in group A, and 4.0, 8.4, and 8.4%, respectively, in group B. There were significant differences between the two groups (P = 0.011). Multivariate analysis showed that being >65 years of age (P = 0.032, hazard ratio (HR): 2.463, 95% confidence interval (CI), 1.028–6.152) and AM ≤ 5 mm (P = 0.010, HR: 3.195, 95% CI, 1.324–7.752) were independent risk factors for LTP after MWA.Conclusion: The novel technology of unsupervised landmark-constrained convolutional neural network-based DIR is feasible and useful in evaluating the ablative effect of MWA for HCC.

Fully Convolutional Network for Consistent Voxel-Wise Correspondence

In this paper, we propose a fully convolutional network-based dense map from voxels to invertible pair of displacement vector fields regarding a template grid for the consistent voxel-wise correspondence. We parameterize the volumetric mapping using a convolutional network and train it in an unsupervised way by leveraging the spatial transformer to minimize the gap between the warped volumetric image and the template grid. Instead of learning the unidirectional map, we learn the nonlinear mapping functions for both forward and backward transformations. We introduce the combinational inverse constraints for the volumetric one-to-one maps, where the pairwise and triple constraints are utilized to learn the cycle-consistent correspondence maps between volumes. Experiments on both synthetic and clinically captured volumetric cone-beam CT (CBCT) images show that the proposed framework is effective and competitive against state-of-the-art deformable registration techniques.

Accuracy of deformable image registration techniques for alignment of longitudinal cholangiocarcinoma CT images.

Response assessment of radiotherapy for the treatment of intrahepatic cholangiocarcinoma (IHCC) across longitudinal images is challenging due to anatomical changes. Advanced deformable image registration (DIR) techniques are required to correlate corresponding tissues across time. In this study, the accuracy of five commercially available DIR algorithms in four treatment planning systems (TPS) was investigated for the registration of planning images with posttreatment follow-up images for response assessment or re-treatment purposes. Twenty-nine IHCC patients treated with hypofractionated radiotherapy and with pretreatment and posttreatment contrast-enhanced computed tomography (CT) images were analyzed. Liver segmentations were semiautomatically generated on all CTs and the posttreatment CT was then registered to the pretreatment CT using five commercially available algorithms (Demons, B-splines, salient feature-based, anatomically constrained and finite element-based) in four TPSs. This was followed by an in-depth analysis of 10 DIR strategies (plus global and liver-focused rigid registration) in one of the TPSs. Eight of the strategies were variants of the anatomically constrained DIR while the two were based on a finite element-based biomechanical registration. The anatomically constrained techniques were combinations of: (a) initializations with the two rigid registrations; (b) two similarity metrics - correlation coefficient (CC) and mutual information (MI); and (c) with and without a controlling region of interest (ROI) - the liver. The finite element-based techniques were initialized by the two rigid registrations. The accuracy of each registration was evaluated using target registration error (TRE) based on identified vessel bifurcations. The results were statistically analyzed with a one-way analysis of variance (ANOVA) and pairwise comparison tests. Stratified analysis was conducted on the inter-TPS data (plus the liver-focused rigid registration) using treatment volume changes, slice thickness, time between scans, and abnormal lab values as stratifying factors. The complex deformation observed following treatment resulted in average TRE exceeding the image voxel size for all techniques. For the inter-TPS comparison, the Demons algorithm had the lowest TRE, which was significantly superior to all the other algorithms. The respective mean (standard deviation) TRE (in mm) for the Demons, B-splines, salient feature-based, anatomically constrained, and finite element-based algorithms were 4.6 (2.0), 7.4 (2.7), 7.2 (2.6), 6.3 (2.3), and 7.5 (4.0). In the follow-up comparison of the anatomically constrained DIR, the strategy with liver-focused rigid registration initialization, CC as similarity metric and liver as a controlling ROI had the lowest mean TRE - 6.0 (2.0). The maximum TRE for all techniques exceeded 10mm. Selection of DIR strategy was found to be a statistically significant factor for registration accuracy. Tumor volume change had a significant effect on TRE for finite element-based registration and B-splines DIR. Time between scans had a substantial effect on TRE for all registrations but was only significant for liver-focused rigid, finite element-based and salient feature-based DIRs. This study demonstrates the limitations of commercially available DIR techniques in TPSs for alignment of longitudinal images of liver cancer presenting complex anatomical changes including local hypertrophy and fibrosis/necrosis. DIR in this setting should be used with caution and careful evaluation.