Automatic delineation of on-line head-and-neck computed tomography images: toward on-line adaptive radiotherapy

Enhanced IDOL segmentation framework using personalized hyperspace learning IDOL

BACKGROUND

Adaptive radiotherapy (ART) workflows have been increasingly adopted to achieve dose escalation and tissue sparing under shifting anatomic conditions, but the necessity of recontouring and the associated time burden hinders a real-time or online ART workflow. In response to this challenge, approaches to auto-segmentation involving deformable image registration, atlas-based segmentation, and deep learning-based segmentation (DLS) have been developed. Despite the particular promise shown by DLS methods, implementing these approaches in a clinical setting remains a challenge, namely due to the difficulty of curating a data set of sufficient size and quality so as to achieve generalizability in a trained model.

PURPOSE

To address this challenge, we have developed an intentional deep overfit learning (IDOL) framework tailored to the auto-segmentation task. However, certain limitations were identified, particularly the insufficiency of the personalized dataset to effectively overfit the model. In this study, we introduce a personalized hyperspace learning (PHL)-IDOL segmentation framework capable of generating datasets that induce the model to overfit specific patient characteristics for medical image segmentation.

METHODS

The PHL-IDOL model is trained in two stages. In the first, a conventional, general model is trained with a diverse set of patient data (n = 100 patients) consisting of CT images and clinical contours. Following this, the general model is tuned with a data set consisting of two components: (a) selection of a subset of the patient data (m < n) using the similarity metrics (mean square error (MSE), peak signal-to-noise ratio (PSNR), structural similarity index (SSIM), and the universal quality image index (UQI) values); (b) adjust the CT and the clinical contours using a deformed vector generated from the reference patient and the selected patients using (a). After training, the general model, the continual model, the conventional IDOL model, and the proposed PHL-IDOL model were evaluated using the volumetric dice similarity coefficient (VDSC) and the Hausdorff distance 95% (HD95%) computed for 18 structures in 20 test patients.

RESULTS

Implementing the PHL-IDOL framework resulted in improved segmentation performance for each patient. The Dice scores increased from 0.81 ± $ \pm $ 0.05 with the general model, 0.83± 0.04 $ \pm 0.04$ for the continual model, 0.83± 0.04 $ \pm 0.04$ for the conventional IDOL model to an average of 0.87± 0.03 $ \pm 0.03$ with the PHL-IDOL model. Similarly, the Hausdorff distance decreased from 3.06± 0.99 $ \pm 0.99$ with the general model, 2.84± 0.69 $ \pm 0.69$ for the continual model, 2.79± 0.79 $ \pm 0.79$ for the conventional IDOL model and 2.36± 0.52 $ \pm 0.52$ for the PHL-IDOL model. All the standard deviations were decreased by nearly half of the values comparing the general model and the PHL-IDOL model.

CONCLUSION

The PHL-IDOL framework applied to the auto-segmentation task achieves improved performance compared to the general DLS approach, demonstrating the promise of leveraging patient-specific prior information in a task central to online ART workflows.

A comprehensive lung CT landmark pair dataset for evaluating deformable image registration algorithms

PURPOSE

Deformable image registration (DIR) is a key enabling technology in many diagnostic and therapeutic tasks, but often does not meet the required robustness and accuracy for supporting clinical tasks. This is in large part due to a lack of high-quality benchmark datasets by which new DIR algorithms can be evaluated. Our team was supported by the National Institute of Biomedical Imaging and Bioengineering to develop DIR benchmark dataset libraries for multiple anatomical sites, comprising of large numbers of highly accurate landmark pairs on matching blood vessel bifurcations. Here we introduce our lung CT DIR benchmark dataset library, which was developed to improve upon the number and distribution of landmark pairs in current public lung CT benchmark datasets.

ACQUISITION AND VALIDATION METHODS

Thirty CT image pairs were acquired from several publicly available repositories as well as authors' institution with IRB approval. The data processing workflow included multiple steps: (1) The images were denoised. (2) Lungs, airways, and blood vessels were automatically segmented. (3) Bifurcations were directly detected on the skeleton of the segmented vessel tree. (4) Falsely identified bifurcations were filtered out using manually defined rules. (5) A DIR was used to project landmarks detected on the first image onto the second image of the image pair to form landmark pairs. (6) Landmark pairs were manually verified. This workflow resulted in an average of 1262 landmark pairs per image pair. Estimates of the landmark pair target registration error (TRE) using digital phantoms were 0.4 mm ± 0.3 mm.

DATA FORMAT AND USAGE NOTES

The data is published in Zenodo at https://doi.org/10.5281/zenodo.8200423. Instructions for use can be found at https://github.com/deshanyang/Lung-DIR-QA.

POTENTIAL APPLICATIONS

The dataset library generated in this work is the largest of its kind to date and will provide researchers with a new and improved set of ground truth benchmarks for quantitatively validating DIR algorithms within the lung.

Functional information guided adaptive radiation therapy

Introduction

Functional informaton is introduced as the mechanism to adapt cancer therapies uniquely to individual patients based on changes defined by qualified tumor biomarkers.

Methods

To demonstrate the methodology, a tumor volume biomarker model, characterized by a tumor volume reduction rate coefficient, is used to adapt a tumor cell survival bioresponse radiotherapy model in terms of therapeutic radiation dose. Tumor volume, acquired from imaging data, serves as a surrogate measurement for tumor cell death, but the biomarker model derived from this data cannot be used to calculate the radiation dose absorbed by the target tumor. However, functional information does provide a mathematical connection between the tumor volume biomarker model and the tumor cell survival bioresponse model by quantifying both data sets in the units of information, thus creating an analytic conduit from bioresponse to biomarker.

Results

The information guided process for individualized dose adaptations using information values acquired from the tumor cell survival bioresponse model and the tumor volume biomarker model are presented in detailed form by flowchart and tabular data. Clinical data are used to generate a presentation that assists investigator application of the information guided methodology to adaptive cancer therapy research.

Conclusions

Information guided adaptation of bioresponse using surrogate data is extensible across multiple research fields because functional information mathematically connects disparate bioresponse and biomarker data sets. Thus, functional information offers adaptive cancer therapy by mathematically connecting immunotherapy, chemotherapy, and radiotherapy cancer treatment processes to implement individualized treatment plans.

Adaptive Region-Specific Loss for Improved Medical Image Segmentation

Defining the loss function is an important part of neural network design and critically determines the success of deep learning modeling. A significant shortcoming of the conventional loss functions is that they weight all regions in the input image volume equally, despite the fact that the system is known to be heterogeneous (i.e., some regions can achieve high prediction performance more easily than others). Here, we introduce a region-specific loss to lift the implicit assumption of homogeneous weighting for better learning. We divide the entire volume into multiple sub-regions, each with an individualized loss constructed for optimal local performance. Effectively, this scheme imposes higher weightings on the sub-regions that are more difficult to segment, and vice versa. Furthermore, the regional false positive and false negative errors are computed for each input image during a training step and the regional penalty is adjusted accordingly to enhance the overall accuracy of the prediction. Using different public and in-house medical image datasets, we demonstrate that the proposed regionally adaptive loss paradigm outperforms conventional methods in the multi-organ segmentations, without any modification to the neural network architecture or additional data preparation.

Generating synthetic images from cone beam computed tomography using self-attention residual UNet for head and neck radiotherapy

Background and purpose

Accurate CT numbers in Cone Beam CT (CBCT) are crucial for precise dose calculations in adaptive radiotherapy (ART). This study aimed to generate synthetic CT (sCT) from CBCT using deep learning (DL) models in head and neck (HN) radiotherapy.

Materials and methods

A novel DL model, the 'self-attention-residual-UNet' (ResUNet), was developed for accurate sCT generation. ResUNet incorporates a self-attention mechanism in its long skip connections to enhance information transfer between the encoder and decoder. Data from 93 HN patients, each with planning CT (pCT) and first-day CBCT images were used. Model performance was evaluated using two DL approaches (non-adversarial and adversarial training) and two model types (2D axial only vs. 2.5D axial, sagittal, and coronal). ResUNet was compared with the traditional UNet through image quality assessment (Mean Absolute Error (MAE), Peak-Signal-to-Noise Ratio (PSNR), Structural Similarity Index (SSIM)) and dose calculation accuracy evaluation (DVH deviation and gamma evaluation (1 %/1mm)).

Results

Image similarity evaluation results for the 2.5D-ResUNet and 2.5D-UNet models were: MAE: 46±7 HU vs. 51±9 HU, PSNR: 66.6±2.0 dB vs. 65.8±1.8 dB, and SSIM: 0.81±0.04 vs. 0.79±0.05. There were no significant differences in dose calculation accuracy between DL models. Both models demonstrated DVH deviation below 0.5 % and a gamma-pass-rate (1 %/1mm) exceeding 97 %.

Conclusions

ResUNet enhanced CT number accuracy and image quality of sCT and outperformed UNet in sCT generation from CBCT. This method holds promise for generating precise sCT for HN ART.

A clinical evaluation of the performance of five commercial artificial intelligence contouring systems for radiotherapy

Purpose/objectives

Auto-segmentation with artificial intelligence (AI) offers an opportunity to reduce inter- and intra-observer variability in contouring, to improve the quality of contours, as well as to reduce the time taken to conduct this manual task. In this work we benchmark the AI auto-segmentation contours produced by five commercial vendors against a common dataset.

Methods and materials

The organ at risk (OAR) contours generated by five commercial AI auto-segmentation solutions (Mirada (Mir), MVision (MV), Radformation (Rad), RayStation (Ray) and TheraPanacea (Ther)) were compared to manually-drawn expert contours from 20 breast, 20 head and neck, 20 lung and 20 prostate patients. Comparisons were made using geometric similarity metrics including volumetric and surface Dice similarity coefficient (vDSC and sDSC), Hausdorff distance (HD) and Added Path Length (APL). To assess the time saved, the time taken to manually draw the expert contours, as well as the time to correct the AI contours, were recorded.

Results

There are differences in the number of CT contours offered by each AI auto-segmentation solution at the time of the study (Mir 99; MV 143; Rad 83; Ray 67; Ther 86), with all offering contours of some lymph node levels as well as OARs. Averaged across all structures, the median vDSCs were good for all systems and compared favorably with existing literature: Mir 0.82; MV 0.88; Rad 0.86; Ray 0.87; Ther 0.88. All systems offer substantial time savings, ranging between: breast 14-20 mins; head and neck 74-93 mins; lung 20-26 mins; prostate 35-42 mins. The time saved, averaged across all structures, was similar for all systems: Mir 39.8 mins; MV 43.6 mins; Rad 36.6 min; Ray 43.2 mins; Ther 45.2 mins.

Conclusions

All five commercial AI auto-segmentation solutions evaluated in this work offer high quality contours in significantly reduced time compared to manual contouring, and could be used to render the radiotherapy workflow more efficient and standardized.

The geometric and dosimetric accuracy of kilovoltage cone beam computed tomography images for adaptive treatment: a systematic review

Objectives

To provide an overview and meta-analysis of different techniques adopted to accomplish kVCBCT for dose calculation and automated segmentation.

Methods

A systematic review and meta-analysis were performed on eligible studies demonstrating kVCBCT-based dose calculation and automated contouring of different tumor features. Meta-analysis of the performance was accomplished on the reported γ analysis and dice similarity coefficient (DSC) score of both collected results as three subgroups (head and neck, chest, and abdomen).

Results

After the literature scrutinization (n = 1008), 52 papers were recognized for the systematic review. Nine studies of dosimtric studies and eleven studies of geometric analysis were suitable for inclusion in meta-analysis. Using kVCBCT for treatment replanning depends on a method used. Deformable Image Registration (DIR) methods yielded small dosimetric error (≤2%), γ pass rate (≥90%) and DSC (≥0.8). Hounsfield Unit (HU) override and calibration curve-based methods also achieved satisfactory yielded small dosimetric error (≤2%) and γ pass rate ((≥90%), but they are prone to error due to their sensitivity to a vendor-specific variation in kVCBCT image quality.

Conclusions

Large cohorts of patients ought to be undertaken to validate methods achieving low levels of dosimetric and geometric errors. Quality guidelines should be established when reporting on kVCBCT, which include agreed metrics for reporting on the quality of corrected kVCBCT and defines protocols of new site-specific standardized imaging used when obtaining kVCBCT images for adaptive radiotherapy.

Advances in knowledge

This review gives useful knowledge about methods making kVCBCT feasible for kVCBCT-based adaptive radiotherapy, simplifying patient pathway and reducing concomitant imaging dose to the patient.

Segmentation by test-time optimization for CBCT-based adaptive radiation therapy

PURPOSE

Online adaptive radiotherapy (ART) requires accurate and efficient auto-segmentation of target volumes and organs-at-risk (OARs) in mostly cone-beam computed tomography (CBCT) images, which often have severe artifacts and lack soft-tissue contrast, making direct segmentation very challenging. Propagating expert-drawn contours from the pretreatment planning CT through traditional or deep learning (DL)-based deformable image registration (DIR) can achieve improved results in many situations. Typical DL-based DIR models are population based, that is, trained with a dataset for a population of patients, and so they may be affected by the generalizability problem.

METHODS

In this paper, we propose a method called test-time optimization (TTO) to refine a pretrained DL-based DIR population model, first for each individual test patient, and then progressively for each fraction of online ART treatment. Our proposed method is less susceptible to the generalizability problem and thus can improve overall performance of different DL-based DIR models by improving model accuracy, especially for outliers. Our experiments used data from 239 patients with head-and-neck squamous cell carcinoma to test the proposed method. First, we trained a population model with 200 patients and then applied TTO to the remaining 39 test patients by refining the trained population model to obtain 39 individualized models. We compared each of the individualized models with the population model in terms of segmentation accuracy.

RESULTS

The average improvement of the Dice similarity coefficient (DSC) and 95% Hausdorff distance (HD95) of segmentation can be up to 0.04 (5%) and 0.98 mm (25%), respectively, with the individualized models compared to the population model over 17 selected OARs and a target of 39 patients. Although the average improvement may seem mild, we found that the improvement for outlier patients with structures of large anatomical changes is significant. The number of patients with at least 0.05 DSC improvement or 2 mm HD95 improvement by TTO averaged over the 17 selected structures for the state-of-the-art architecture VoxelMorph is 10 out of 39 test patients. By deriving the individualized model using TTO from the pretrained population model, TTO models can be ready in about 1 min. We also generated the adapted fractional models for each of the 39 test patients by progressively refining the individualized models using TTO to CBCT images acquired at later fractions of online ART treatment. When adapting the individualized model to a later fraction of the same patient, the model can be ready in less than a minute with slightly improved accuracy.

CONCLUSIONS

The proposed TTO method is well suited for online ART and can boost segmentation accuracy for DL-based DIR models, especially for outlier patients where the pretrained models fail.

A clinical and time savings evaluation of a deep learning automatic contouring algorithm

Automatic contouring algorithms may streamline clinical workflows by reducing normal organ-at-risk (OAR) contouring time. Here we report the first comprehensive quantitative and qualitative evaluation, along with time savings assessment for a prototype deep learning segmentation algorithm from Siemens Healthineers. The accuracy of contours generated by the prototype were evaluated quantitatively using the Sorensen-Dice coefficient (Dice), Jaccard index (JC), and Hausdorff distance (Haus). Normal pelvic and head and neck OAR contours were evaluated retrospectively comparing the automatic and manual clinical contours in 100 patient cases. Contouring performance outliers were investigated. To quantify the time savings, a certified medical dosimetrist manually contoured de novo and, separately, edited the generated OARs for 10 head and neck and 10 pelvic patients. The automatic, edited, and manually generated contours were visually evaluated and scored by a practicing radiation oncologist on a scale of 1-4, where a higher score indicated better performance. The quantitative comparison revealed high (> 0.8) Dice and JC performance for relatively large organs such as the lungs, brain, femurs, and kidneys. Smaller elongated structures that had relatively low Dice and JC values tended to have low Hausdorff distances. Poor performing outlier cases revealed common anatomical inconsistencies including overestimation of the bladder and incorrect superior-inferior truncation of the spinal cord and femur contours. In all cases, editing contours was faster than manual contouring with an average time saving of 43.4% or 11.8 minutes per patient. The physician scored 240 structures with > 95% of structures receiving a score of 3 or 4. Of the structures reviewed, only 11 structures needed major revision or to be redone entirely. Our results indicate the evaluated auto-contouring solution has the potential to reduce clinical contouring time. The algorithm's performance is promising, but human review and some editing is required prior to clinical use.

3D Lightweight Network for Simultaneous Registration and Segmentation of Organs-at-Risk in CT Images of Head and Neck Cancer

Image-guided radiation therapy (IGRT) is the most effective treatment for head and neck cancer. The successful implementation of IGRT requires accurate delineation of organ-at-risk (OAR) in the computed tomography (CT) images. In routine clinical practice, OARs are manually segmented by oncologists, which is time-consuming, laborious, and subjective. To assist oncologists in OAR contouring, we proposed a three-dimensional (3D) lightweight framework for simultaneous OAR registration and segmentation. The registration network was designed to align a selected OAR template to a new image volume for OAR localization. A region of interest (ROI) selection layer then generated ROIs of OARs from the registration results, which were fed into a multiview segmentation network for accurate OAR segmentation. To improve the performance of registration and segmentation networks, a centre distance loss was designed for the registration network, an ROI classification branch was employed for the segmentation network, and further, context information was incorporated to iteratively promote both networks' performance. The segmentation results were further refined with shape information for final delineation. We evaluated registration and segmentation performances of the proposed framework using three datasets. On the internal dataset, the Dice similarity coefficient (DSC) of registration and segmentation was 69.7% and 79.6%, respectively. In addition, our framework was evaluated on two external datasets and gained satisfactory performance. These results showed that the 3D lightweight framework achieved fast, accurate and robust registration and segmentation of OARs in head and neck cancer. The proposed framework has the potential of assisting oncologists in OAR delineation.

Deformable image registration to assist clinical decision for radiotherapy treatment adaptation for head and neck cancer patients

Head and neck (H&N) cancer patients often present anatomical and geometrical changes in tumors and organs at risk (OARs) during radiotherapy treatment. These changes may result in the need to adapt the existing treatment planning, using an expert's subjective opinion, for offline adaptive radiotherapy and a new treatment planning before each treatment, for online adaptive radiotherapy. In the present study, a fast methodology is proposed to assist in planning adaptation clinical decision using tumor and parotid glands percentage volume changes during treatment. The proposed approach was applied to 40 Η&Ν cases, with one planning Computed Tomography (pCT) image and CBCT scans for 6 weeks of treatment per case. Deformable registration was used for each patient's pCT image alignment to its weekly CBCT. The calculated transformations were used to align each patient's anatomical structures to the weekly anatomy. Clinical target volume (CTV) and parotid gland volume percentage changes were calculated in each case. The accuracy of the achieved image alignment was validated qualitatively and quantitatively. Furthermore, statistical analysis was performed to test if there is a statistically significant correlation between CTV and parotid glands volume percentage changes. Average MDA for CTV and parotid glands between corresponding structures defined by an expert in CBCTs and automatically calculated through registration was 1.4 ± 0.1 mm and 1.5 ± 0.1 mm, respectively. The mean registration time of the first CBCT image registration for 40 cases was lower than 3.4 min. Five patients show more than 20% tumor volume change. Six patients show more than 30% parotid glands volume change. Ten out of 40 patients proposed for planning adaptation. All the statistical tests performed showed no correlation between CTV/parotid glands percentage volume changes. The aim to assist in clinical decision making on a fast and automatic way was achieved using the proposed methodology, thereby reducing workload in clinical practice.

Clinically Applicable Segmentation of Head and Neck Anatomy for Radiotherapy: Deep Learning Algorithm Development and Validation Study

BACKGROUND

Over half a million individuals are diagnosed with head and neck cancer each year globally. Radiotherapy is an important curative treatment for this disease, but it requires manual time to delineate radiosensitive organs at risk. This planning process can delay treatment while also introducing interoperator variability, resulting in downstream radiation dose differences. Although auto-segmentation algorithms offer a potentially time-saving solution, the challenges in defining, quantifying, and achieving expert performance remain.

OBJECTIVE

Adopting a deep learning approach, we aim to demonstrate a 3D U-Net architecture that achieves expert-level performance in delineating 21 distinct head and neck organs at risk commonly segmented in clinical practice.

METHODS

The model was trained on a data set of 663 deidentified computed tomography scans acquired in routine clinical practice and with both segmentations taken from clinical practice and segmentations created by experienced radiographers as part of this research, all in accordance with consensus organ at risk definitions.

RESULTS

We demonstrated the model's clinical applicability by assessing its performance on a test set of 21 computed tomography scans from clinical practice, each with 21 organs at risk segmented by 2 independent experts. We also introduced surface Dice similarity coefficient, a new metric for the comparison of organ delineation, to quantify the deviation between organ at risk surface contours rather than volumes, better reflecting the clinical task of correcting errors in automated organ segmentations. The model's generalizability was then demonstrated on 2 distinct open-source data sets, reflecting different centers and countries to model training.

CONCLUSIONS

Deep learning is an effective and clinically applicable technique for the segmentation of the head and neck anatomy for radiotherapy. With appropriate validation studies and regulatory approvals, this system could improve the efficiency, consistency, and safety of radiotherapy pathways.

Automatic Segmentation of Mandible from Conventional Methods to Deep Learning-A Review

Medical imaging techniques, such as (cone beam) computed tomography and magnetic resonance imaging, have proven to be a valuable component for oral and maxillofacial surgery (OMFS). Accurate segmentation of the mandible from head and neck (H&N) scans is an important step in order to build a personalized 3D digital mandible model for 3D printing and treatment planning of OMFS. Segmented mandible structures are used to effectively visualize the mandible volumes and to evaluate particular mandible properties quantitatively. However, mandible segmentation is always challenging for both clinicians and researchers, due to complex structures and higher attenuation materials, such as teeth (filling) or metal implants that easily lead to high noise and strong artifacts during scanning. Moreover, the size and shape of the mandible vary to a large extent between individuals. Therefore, mandible segmentation is a tedious and time-consuming task and requires adequate training to be performed properly. With the advancement of computer vision approaches, researchers have developed several algorithms to automatically segment the mandible during the last two decades. The objective of this review was to present the available fully (semi)automatic segmentation methods of the mandible published in different scientific articles. This review provides a vivid description of the scientific advancements to clinicians and researchers in this field to help develop novel automatic methods for clinical applications.

A validation framework to assess performance of commercial deformable image registration in lung radiotherapy

INTRODUCTION

Deformable image registration (DIR) can play an important role in the context of adaptive radiotherapy. The AAPM Task Group 132 (TG-132) has described several quantitative measures for DIR error assessment but they can only be accurately defined when there is a ground-truth present in high-contrast regions. This work aims to set out a framework to obtain optimal results for CT-CT lung DIR in clinical setting for a commercially available system by quantifying the DIR performance in both low- and high-contrast regions.

METHODS

Five publicly available thorax datasets were used to assess the DIR quality. A "Ghost fiducial" method was implemented by windowing the contrast in a new feature provided by Varian Velocity v4.1. Target registration error (TRE) of the landmarks and Dice-similarity coefficient of the tumour were calculated at three different contrast settings to assess the algorithm in high- and low-contrast scenarios.

RESULTS

For the original unedited dataset, higher resolution DIR methods showed best performance acceptable within the recommended limit according to TG-132, when actual displacements were less than 10 mm. The relation of the actual displacement of the landmarks and TRE shows the limited capacity of the algorithm to deal with movements larger than 10 mm.

CONCLUSION

This work found the performance of DIR methods and settings available in Varian Velocity v4.1 to be a function of contrast level as well as extent of motion. This highlights the need for multiple metrics to assess different aspects of DIR performance for various applications related to low-contrast and/or high-contrast regions.

A generalized framework for analytic regularization of uniform cubic B-spline displacement fields

Image registration is an inherently ill-posed problem that lacks the constraints needed for a unique mapping between voxels of the two images being registered. As such, one must regularize the registration to achieve physically meaningful transforms. The regularization penalty is usually a function of derivatives of the displacement-vector field and can be calculated either analytically or numerically. The numerical approach, however, is computationally expensive depending on the image size, and therefore a computationally efficient analytical framework has been developed. Using cubic B-splines as the registration transform, we develop a generalized mathematical framework that supports five distinct regularizers: diffusion, curvature, linear elastic, third-order, and total displacement. We validate our approach by comparing each with its numerical counterpart in terms of accuracy. We also provide benchmarking results showing that the analytic solutions run significantly faster-up to two orders of magnitude-than finite differencing based numerical implementations.

Variability in commercially available deformable image registration: A multi-institution analysis using virtual head and neck phantoms

PURPOSE

The purpose of this study was to evaluate the performance of three common deformable image registration (DIR) packages across algorithms and institutions.

METHODS AND MATERIALS

The Deformable Image Registration Evaluation Project (DIREP) provides ten virtual phantoms derived from computed tomography (CT) datasets of head-and-neck cancer patients over a single treatment course. Using the DIREP phantoms, DIR results from 35 institutions were submitted using either Velocity, MIM, or Eclipse. Submitted deformation vector fields (DVFs) were compared to ground-truth DVFs to calculate target registration error (TRE) for six regions of interest (ROIs). Statistical analysis was performed to determine the variability between each DIR software package and the variability of users within each algorithm.

RESULTS

Overall mean TRE was 2.04 ± 0.35 mm for Velocity, 1.10 ± 0.29 mm for MIM, and 2.35 ± 0.15 mm for Eclipse. The MIM mean TRE was significantly different than both Velocity and Eclipse for all ROIs. Velocity and Eclipse mean TREs were not significantly different except for when evaluating the registration of the cord or mandible. Significant differences between institutions were found for the MIM and Velocity platforms. However, these differences could be explained by variations in Velocity DIR parameters and MIM software versions.

CONCLUSIONS

Average TRE was shown to be <3 mm for all three software platforms. However, maximum errors could be larger than 2 cm indicating that care should be exercised when using DIR. While MIM performed statistically better than the other packages, all evaluated algorithms had an average TRE better than the largest voxel dimension. For the phantoms studied here, significant differences between algorithm users were minimal suggesting that the algorithm used may have more impact on DIR accuracy than the particular registration technique employed. A significant difference in TRE was discovered between MIM versions showing that DIR QA should be performed after software upgrades as recommended by TG-132.

Dosimetric influence of deformable image registration uncertainties on propagated structures for online daily adaptive proton therapy of lung cancer patients

PURPOSE

A major burden of introducing an online daily adaptive proton therapy (DAPT) workflow is the time and resources needed to correct the daily propagated contours. In this study, we evaluated the dosimetric impact of neglecting the online correction of the propagated contours in a DAPT workflow.

MATERIAL AND METHODS

For five NSCLC patients with nine repeated deep-inspiration breath-hold CTs, proton therapy plans were optimised on the planning CT to deliver 60 Gy-RBE in 30 fractions. All repeated CTs were registered with six different clinically used deformable image registration (DIR) algorithms to the corresponding planning CT. Structures were propagated rigidly and with each DIR algorithm and reference structures were contoured on each repeated CT. DAPT plans were optimised with the uncorrected, propagated structures (propagated DAPT doses) and on the reference structures (ideal DAPT doses), non-adapted doses were recalculated on all repeated CTs.

RESULTS

Due to anatomical changes occurring during the therapy, the clinical target volume (CTV) coverage of the non-adapted doses reduces on average by 9.7% (V95) compared to an ideal DAPT doses. For the propagated DAPT doses, the CTV coverage was always restored (average differences in the CTV V95 < 1% compared to the ideal DAPT doses). Hotspots were always reduced with any DAPT approach.

CONCLUSION

For the patients presented here, a benefit of online DAPT was shown, even if the daily optimisation is based on propagated structures with some residual uncertainties. However, a careful (offline) structure review is necessary and corrections can be included in an offline adaption.

Comparison of predictive performance for toxicity by accumulative dose of DVH parameter addition and DIR addition for cervical cancer patients

We compared predictive performance between dose volume histogram (DVH) parameter addition and deformable image registration (DIR) addition for gastrointestinal (GI) toxicity in cervical cancer patients. A total of 59 patients receiving brachytherapy and external beam radiotherapy were analyzed retrospectively. The accumulative dose was calculated by three methods: conventional DVH parameter addition, full DIR addition and partial DIR addition. ${D}_{2{cm}^3}$, ${D}_{1{cm}^3}$ and ${D}_{0.1{cm}^3}$ (minimum doses to the most exposed 2 cm3, 1cm3 and 0.1 cm3 of tissue, respectively) of the rectum and sigmoid were calculated by each method. V50, V60 and V70 Gy (volume irradiated over 50, 60 and 70 Gy, respectively) were calculated in full DIR addition. The DVH parameters were compared between toxicity (≥grade1) and non-toxicity groups. The area under the curve (AUC) of the receiver operating characteristic (ROC) curves were compared to evaluate the predictive performance of each method. The differences between toxicity and non-toxicity groups in ${D}_{2{cm}^3}$ were 0.2, 5.7 and 3.1 Gy for the DVH parameter addition, full DIR addition and partial DIR addition, respectively. The AUCs of ${D}_{2{cm}^3}$ were 0.51, 0.67 and 0.57 for DVH parameter addition, full DIR addition and partial DIR addition, respectively. In full DIR addition, the difference in dose between toxicity and non-toxicity was the largest and AUC was the highest. AUCs of V50, V60 and V70 Gy were 0.51, 0.63 and 0.62, respectively, and V60 and V70 were high values close to the value of ${D}_{2{cm}^3}$ of the full DIR addition. Our results suggested that the full DIR addition may have the potential to predict toxicity more accurately than the conventional DVH parameter addition, and that it could be more effective to accumulate to all pelvic irradiation by DIR.

Anthropomorphic lung phantom based validation of in-room proton therapy 4D-CBCT image correction for dose calculation

PURPOSE

Ventilation-induced tumour motion remains a challenge for the accuracy of proton therapy treatments in lung patients. We investigated the feasibility of using a 4D virtual CT (4D-vCT) approach based on deformable image registration (DIR) and motion-aware 4D CBCT reconstruction (MA-ROOSTER) to enable accurate daily proton dose calculation using a gantry-mounted CBCT scanner tailored to proton therapy.

METHODS

Ventilation correlated data of 10 breathing phases were acquired from a porcine ex-vivo functional lung phantom using CT and CBCT. 4D-vCTs were generated by (1) DIR of the mid-position 4D-CT to the mid-position 4D-CBCT (reconstructed with the MA-ROOSTER) using a diffeomorphic Morphons algorithm and (2) subsequent propagation of the obtained mid-position vCT to the individual 4D-CBCT phases. Proton therapy treatment planning was performed to evaluate dose calculation accuracy of the 4D-vCTs. A robust treatment plan delivering a nominal dose of 60Gy was generated on the average intensity image of the 4D-CT for an approximated internal target volume (ITV). Dose distributions were then recalculated on individual phases of the 4D-CT and the 4D-vCT based on the optimized plan. Dose accumulation was performed for 4D-vCT and 4D-CT using DIR of each phase to the mid position, which was chosen as reference. Dose based on the 4D-vCT was then evaluated against the dose calculated on 4D-CT both, phase-by-phase as well as accumulated, by comparing dose volume histogram (DVH) values (D_mean, D_2%, D_98%, D_95%) for the ITV, and by a 3D-gamma index analysis (global, 3%/3mm, 5Gy, 20Gy and 30Gy dose thresholds).

RESULTS

Good agreement was found between the 4D-CT and 4D-vCT-based ITV-DVH curves. The relative differences ((CT-vCT)/CT) between accumulated values of ITV D_mean, D_2%, D_95% and D_98% for the 4D-CT and 4D-vCT-based dose distributions were -0.2%, 0.0%, -0.1% and -0.1%, respectively. Phase specific values varied between -0.5% and 0.2%, -0.2% and 0.5%, -3.5% and 1.5%, and -5.7% and 2.3%. The relative difference of accumulated D_mean over the lungs was 2.3% and D_mean for the phases varied between -5.4% and 5.8%. The gamma pass-rates with 5Gy, 20Gy and 30Gy thresholds for the accumulated doses were 96.7%, 99.6% and 99.9%, respectively. Phase-by-phase comparison yielded pass-rates between 86% and 97%, 88% and 98%, and 94% and 100%.

CONCLUSIONS

Feasibility of the suggested 4D-vCT workflow using proton therapy specific imaging equipment was shown. Results indicate the potential of the method to be applied for daily 4D proton dose estimation.

An Automated Workflow to Improve Efficiency in Radiation Therapy Treatment Planning by Prioritizing Organs at Risk

PURPOSE

Manual delineation (MD) of organs at risk (OAR) is time and labor intensive. Auto-delineation (AD) can reduce the need for MD, but because current algorithms are imperfect, manual review and modification is still typically used. Recognizing that many OARs are sufficiently far from important dose levels that they do not pose a realistic risk, we hypothesize that some OARs can be excluded from MD and manual review with no clinical effect. The purpose of this study was to develop a method that automatically identifies these OARs and enables more efficient workflows that incorporate AD without degrading clinical quality.

METHODS AND MATERIALS

Preliminary dose map estimates were generated for n = 10 patients with head and neck cancers using only prescription and target-volume information. Conservative estimates of clinical OAR objectives were computed using AD structures with spatial expansion buffers to account for potential delineation uncertainties. OARs with estimated dose metrics below clinical tolerances were deemed low priority and excluded from MD and/or manual review. Final plans were then optimized using high-priority MD OARs and low-priority AD OARs and compared with reference plans generated using all MD OARs. Multiple different spatial buffers were used to accommodate different potential delineation uncertainties.

RESULTS

Sixty-seven out of 201 total OARs were identified as low-priority using the proposed methodology, which permitted a 33% reduction in structures requiring manual delineation/review. Plans optimized using low-priority AD OARs without review or modification met all planning objectives that were met when all MD OARs were used, indicating clinical equivalence.

CONCLUSIONS

Prioritizing OARs using estimated dose distributions allowed a substantial reduction in required MD and review without affecting clinically relevant dosimetry.

[Multi-scale 3D convolutional neural network-based segmentation of head and neck organs at risk]

OBJECTIVE

To establish an algorithm based on 3D convolution neural network to segment the organs at risk (OARs) in the head and neck on CT images.

METHODS

We propose an automatic segmentation algorithm of head and neck OARs based on V-Net. To enhance the feature expression ability of the 3D neural network, we combined the squeeze and exception (SE) module with the residual convolution module in V-Net to increase the weight of the features that has greater contributions to the segmentation task. Using a multi-scale strategy, we completed organ segmentation using two cascade models for location and fine segmentation, and the input image was resampled to different resolutions during preprocessing to allow the two models to focus on the extraction of global location information and local detail features respectively.

RESULTS

Our experiments on segmentation of 22 OARs in the head and neck indicated that compared with the existing methods, the proposed method achieved better segmentation accuracy and efficiency, and the average segmentation accuracy was improved by 9%. At the same time, the average test time was reduced from 33.82 s to 2.79 s.

CONCLUSIONS

The 3D convolution neural network based on multi-scale strategy can effectively and efficiently improve the accuracy of organ segmentation and can be potentially used in clinical setting for segmentation of other organs to improve the efficiency of clinical treatment.

Dose-Distribution-Driven PET Image-Based Outcome Prediction (DDD-PIOP): A Deep Learning Study for Oropharyngeal Cancer IMRT Application

Purpose

To develop a deep learning-based AI agent, DDD-PIOP (Dose-Distribution-Driven PET Image Outcome Prediction), for predicting ¹⁸FDG-PET image outcomes of oropharyngeal cancer (OPC) in response to intensity-modulated radiation therapy (IMRT).

Methods

DDD-PIOP uses pre-radiotherapy ¹⁸FDG-PET/CT images and the planned spatial dose distribution as the inputs, and it predicts the ¹⁸FDG-PET image outcomes in response to the planned IMRT delivery. This AI agent centralizes a customized convolutional neural network (CNN) as a deep learning approach, and it incorporates a few designs to enhance prediction accuracy. 66 OPC patients who received IMRT treatment on a sequential boost regime (2 Gy/daily fraction) were studied for DDD-PIOP development. 61 patients were used for AI agent training/validation, and the remaining five were used as independent tests. To evaluate the developed AI agent’s performance, the predicted mean standardized uptake values (SUVs) of gross tumor volume (GTV) and clinical target volume (CTV) were compared with the ground truth values. Overall SUV distribution accuracy was evaluated by gamma test passing rates under different criteria.

Results

The developed DDD-PIOP successfully generated ¹⁸FDG-PET image outcome predictions for five test patients. The predicted mean SUV values of GTV/CTV were 3.50/1.41, which were close to the ground-truth values of 3.57/1.51. In 2D-based gamma tests, the average passing rate was 92.1% using 5%/10 mm criteria, which was improved to 95.9%/93.2% when focusing on GTV/CTV regions. 3D gamma test passing rates were 98.7% using 5%/10 mm criteria, and the corresponding GTV/CTV results were 99.8%/99.4%.

Conclusion

The reported results suggest that the developed AI agent DDD-PIOP successfully predicted ¹⁸FDG-PET image outcomes with high quantitative accuracy. The generated voxel-based image outcome predictions could be used for treatment planning optimization prior to radiation delivery for the best individual-based outcome.

A convolutional neural network for contouring metastatic lymph nodes on diffusion-weighted magnetic resonance images for assessment of radiotherapy response

BACKGROUND AND PURPOSE

Retrieving quantitative parameters from magnetic resonance imaging (MRI), e.g. for early assessment of radiotherapy treatment response, necessitates contouring regions of interest, which is time-consuming and prone to errors. This becomes more pressing for daily imaging on MRI-guided radiotherapy systems. Therefore, we trained a deep convolutional neural network to automatically contour involved lymph nodes on diffusion-weighted (DW) MRI of head and neck cancer (HNC) patients receiving radiotherapy.

MATERIALS AND METHODS

DW-images from 48 HNC patients (18 induction-chemotherapy + chemoradiotherapy; 30 definitive chemoradiotherapy) with 68 involved lymph nodes were obtained on a diagnostic 1.5 T MR-scanner prior to and 2-3 timepoints throughout treatment. A radiation oncologist delineated the lymph nodes on the b = 50 s/mm2 images. A 3D U-net was trained to contour involved lymph nodes. Its performance was evaluated in all 48 patients using 8-fold cross-validation and calculating the Dice similarity coefficient (DSC) and the absolute difference in median apparent diffusion coefficient (ΔADC) between the manual and generated contours. Additionally, the performance was evaluated in an independent dataset of three patients obtained on a 1.5 T MR-Linac.

RESULTS

In the definitive chemoradiotherapy patients (n = 96 patients/lymphnodes/timepoints) the DSC was 0.87 (0.81-0.91) [median (1st-3rd quantiles)] and ΔADC was 1.9% (0.8-3.4%) and both remained stable throughout treatment. The network performed worse in the patients receiving induction-chemotherapy (n = 65), with DSC = 0.80 (0.71-0.87) and ΔADC = 3.3% (1.6-8.0%). The network performed well on the MR-Linac data (n = 8) with DSC = 0.80 (0.75-0.82) and ΔADC = 4.0% (0.6-9.1%).

CONCLUSIONS

We established accurate automatic contouring of involved lymph nodes for HNC patients on diagnostic and MR-Linac DW-images.

Evaluation of an Open Source Registration Package for Automatic Contour Propagation in Online Adaptive Intensity-Modulated Proton Therapy of Prostate Cancer

Objective: Our goal was to investigate the performance of an open source deformable image registration package, elastix, for fast and robust contour propagation in the context of online-adaptive intensity-modulated proton therapy (IMPT) for prostate cancer.

Methods: A planning and 7–10 repeat CT scans were available of 18 prostate cancer patients. Automatic contour propagation of repeat CT scans was performed using elastix and compared with manual delineations in terms of geometric accuracy and runtime. Dosimetric accuracy was quantified by generating IMPT plans using the propagated contours expanded with a 2 mm (prostate) and 3.5 mm margin (seminal vesicles and lymph nodes) and calculating dosimetric coverage based on the manual delineation. A coverage of V_95% ≥ 98% (at least 98% of the target volumes receive at least 95% of the prescribed dose) was considered clinically acceptable.

Results: Contour propagation runtime varied between 3 and 30 s for different registration settings. For the fastest setting, 83 in 93 (89.2%), 73 in 93 (78.5%), and 91 in 93 (97.9%) registrations yielded clinically acceptable dosimetric coverage of the prostate, seminal vesicles, and lymph nodes, respectively. For the prostate, seminal vesicles, and lymph nodes the Dice Similarity Coefficient (DSC) was 0.87 ± 0.05, 0.63 ± 0.18, and 0.89 ± 0.03 and the mean surface distance (MSD) was 1.4 ± 0.5 mm, 2.0 ± 1.2 mm, and 1.5 ± 0.4 mm, respectively.

Conclusion: With a dosimetric success rate of 78.5–97.9%, this software may facilitate online adaptive IMPT of prostate cancer using a fast, free and open implementation.

Improvements in CBCT Image Quality Using a Novel Iterative Reconstruction Algorithm: A Clinical Evaluation

PURPOSE

This study aimed to evaluate the clinical utility of a novel iterative cone beam computed tomography (CBCT) reconstruction algorithm for prostate and head and neck (HN) cancer.

METHODS AND MATERIALS

A total of 10 patients with HN and 10 patients with prostate cancer were analyzed. For each patient, raw CBCT acquisition data were used to reconstruct images with a currently available algorithm (FDK_CBCT) and novel iterative algorithm (Iterative_CBCT). Quantitative contouring variation analysis was performed using structures delineated by several radiation oncologists. For prostate, observers contoured the prostate, proximal 2 cm seminal vesicles, bladder, and rectum. For HN, observers contoured the brain stem, spinal canal, right-left parotid glands, and right-left submandibular glands. Observer contours were combined to form a reference consensus contour using the simultaneous truth and performance level estimation method. All observer contours then were compared with the reference contour to calculate the Dice coefficient, Hausdorff distance, and mean contour distance (prostate contour only). Qualitative image quality analysis was performed using a 5-point scale ranging from 1 (much superior image quality for Iterative_CBCT) to 5 (much inferior image quality for Iterative_CBCT).

RESULTS

The Iterative_CBCT data sets resulted in a prostate contour Dice coefficient improvement of approximately 2.4% (P = .029). The average prostate contour Dice coefficient for the Iterative_CBCT data sets was improved for all patients, with improvements up to approximately 10% for 1 patient. The mean contour distance results indicate an approximate 15% reduction in mean contouring error for all prostate regions. For the parotid contours, Iterative_CBCT data sets resulted in a Hausdorff distance improvement of approximately 2 mm (P < .01) and an approximate 2% improvement in Dice coefficient (P = .03). The Iterative_CBCT data sets were scored as equivalent or of better image quality for 97.3% (prostate) and 90.0% (HN) of the patient data sets.

CONCLUSIONS

Observers noted an improvement in image uniformity, noise level, and overall image quality for Iterative_CBCT data sets. In addition, expert observers displayed an improved ability to consistently delineate soft tissue structures, such as the prostate and parotid glands. Thus, the novel iterative reconstruction algorithm analyzed in this study is capable of improving the visualization for prostate and HN cancer image guided radiation therapy.

On-line dose-guidance to account for inter-fractional motion during proton therapy

Background and purpose

Proton therapy (PT) of extra-cranial tumour sites is challenged by density changes caused by inter-fractional organ motion. In this study we investigate on-line dose-guided PT (DGPT) to account inter-fractional target motion, exemplified by internal motion in the pelvis.

Materials and methods

On-line DGPT involved re-calculating dose distributions with the isocenter shifted up to 15 mm from the position corresponding to conventional soft-tissue based image-guided PT (IGPT). The method was applied to patient models with simulated prostate/seminal vesicle target motion of ±3, ±5 and ±10 mm along the three cardinal axes. Treatment plans were created using either two lateral (gantry angles of 90°/270°) or two lateral oblique fields (gantry angles of 35°/325°). Target coverage and normal tissue doses from DGPT were compared to both soft-tissue and bony anatomy based IGPT.

Results

DGPT improved the dose distributions relative to soft-tissue based IGPT for 39 of 90 simulation scenarios using lateral fields and for 50 of 90 scenarios using lateral oblique fields. The greatest benefits of DGPT were seen for large motion, e.g. a median target coverage improvement of 13% was found for 10 mm anterior motion with lateral fields. DGPT also improved the dose distribution in comparison to bony anatomy IGPT in all cases. The best strategy was often to move the fields back towards the original target position prior to the simulated target motion.

Conclusion

DGPT has the potential to better account for large inter-fractional organ motion in the pelvis than IGPT.

AnatomyNet: Deep learning for fast and fully automated whole-volume segmentation of head and neck anatomy

PURPOSE

Radiation therapy (RT) is a common treatment option for head and neck (HaN) cancer. An important step involved in RT planning is the delineation of organs-at-risks (OARs) based on HaN computed tomography (CT). However, manually delineating OARs is time-consuming as each slice of CT images needs to be individually examined and a typical CT consists of hundreds of slices. Automating OARs segmentation has the benefit of both reducing the time and improving the quality of RT planning. Existing anatomy autosegmentation algorithms use primarily atlas-based methods, which require sophisticated atlas creation and cannot adequately account for anatomy variations among patients. In this work, we propose an end-to-end, atlas-free three-dimensional (3D) convolutional deep learning framework for fast and fully automated whole-volume HaN anatomy segmentation.

METHODS

Our deep learning model, called AnatomyNet, segments OARs from head and neck CT images in an end-to-end fashion, receiving whole-volume HaN CT images as input and generating masks of all OARs of interest in one shot. AnatomyNet is built upon the popular 3D U-net architecture, but extends it in three important ways: (a) a new encoding scheme to allow autosegmentation on whole-volume CT images instead of local patches or subsets of slices, (b) incorporating 3D squeeze-and-excitation residual blocks in encoding layers for better feature representation, and (c) a new loss function combining Dice scores and focal loss to facilitate the training of the neural model. These features are designed to address two main challenges in deep learning-based HaN segmentation: (a) segmenting small anatomies (i.e., optic chiasm and optic nerves) occupying only a few slices, and (b) training with inconsistent data annotations with missing ground truth for some anatomical structures.

RESULTS

We collected 261 HaN CT images to train AnatomyNet and used MICCAI Head and Neck Auto Segmentation Challenge 2015 as a benchmark dataset to evaluate the performance of AnatomyNet. The objective is to segment nine anatomies: brain stem, chiasm, mandible, optic nerve left, optic nerve right, parotid gland left, parotid gland right, submandibular gland left, and submandibular gland right. Compared to previous state-of-the-art results from the MICCAI 2015 competition, AnatomyNet increases Dice similarity coefficient by 3.3% on average. AnatomyNet takes about 0.12 s to fully segment a head and neck CT image of dimension 178 × 302 × 225, significantly faster than previous methods. In addition, the model is able to process whole-volume CT images and delineate all OARs in one pass, requiring little pre- or postprocessing.

CONCLUSION

Deep learning models offer a feasible solution to the problem of delineating OARs from CT images. We demonstrate that our proposed model can improve segmentation accuracy and simplify the autosegmentation pipeline. With this method, it is possible to delineate OARs of a head and neck CT within a fraction of a second.

Early Changes in Serial CBCT-Measured Parotid Gland Biomarkers Predict Chronic Xerostomia After Head and Neck Radiation Therapy

PURPOSE

To determine whether serial cone beam computed tomography (CBCT) images taken during head and neck radiation therapy (HNR) can improve chronic xerostomia prediction.

METHODS AND MATERIALS

In a retrospective analysis, parotid glands (PGs) were delineated on daily kV CBCT images using deformable image registration for 119 HNR patients (60 or 70 Gy in 2 Gy fractions over 6 or 7 weeks). Deformable image registration accuracy for a subset of deformed contours was quantified using the Dice similarity coefficient and mean distance to agreement in comparison with manually drawn contours. Average weekly changes in CBCT-measured mean Hounsfield unit intensity and volume were calculated for each PG relative to week 1. Dose-volume histogram statistics were extracted from each plan, and interactions among dose, volume, and intensity were investigated. Univariable analysis and penalized logistic regression were used to analyze association with observer-rated xerostomia at 1 year after HNR. Models including CBCT delta imaging features were compared with clinical and dose-volume histogram-only models using area under the receiver operating characteristic curve (AUC) for grade ≥1 and grade ≥2 xerostomia prediction.

RESULTS

All patients experienced end-treatment PG volume reduction with mean (range) ipsilateral and contralateral PG shrinkage of 19.6% (0.9%-58.4%) and 17.7% (4.4%-56.3%), respectively. Midtreatment volume change was highly correlated with mean PG dose (r = -0.318, P < 1e^-6). Incidence of grade ≥1 and grade ≥2 xerostomia was 65% and 16%, respectively. For grade ≥1 xerostomia prediction, the delta-imaging model had an AUC of 0.719 (95% confidence interval [CI], 0.603-0.830), compared with 0.709 (95% CI, 0.603-0.815) for the dose/clinical model. For grade ≥2 xerostomia prediction, the dose/clinical model had an AUC of 0.692 (95% CI, 0.615-0.770), and the addition of contralateral PG changes modestly improved predictive performance, with an AUC of 0.776 (0.643-0.912).

CONCLUSIONS

The rate of CBCT-measured PG image feature changes improves prediction over dose alone for chronic xerostomia prediction. Analysis of CBCT images acquired for treatment positioning may provide an inexpensive monitoring system to support toxicity-reducing adaptive radiation therapy.

Feasibility of online IMPT adaptation using fast, automatic and robust dose restoration

Intensity-modulated proton therapy (IMPT) offers excellent dose conformity and healthy tissue sparing, but it can be substantially compromised in the presence of anatomical changes. A major dosimetric effect is caused by density changes, which alter the planned proton range in the patient. Three different methods, which automatically restore an IMPT plan dose on a daily CT image were implemented and compared: (1) simple dose restoration (DR) using optimization objectives of the initial plan, (2) voxel-wise dose restoration (vDR), and (3) isodose volume dose restoration (iDR). Dose restorations were calculated for three different clinical cases, selected to test different capabilities of the restoration methods: large range adaptation, complex dose distributions and robust re-optimization. All dose restorations were obtained in less than 5 min, without manual adjustments of the optimization settings. The evaluation of initial plans on repeated CTs showed large dose distortions, which were substantially reduced after restoration. In general, all dose restoration methods improved DVH-based scores in propagated target volumes and OARs. Analysis of local dose differences showed that, although all dose restorations performed similarly in high dose regions, iDR restored the initial dose with higher precision and accuracy in the whole patient anatomy. Median dose errors decreased from 13.55 Gy in distorted plan to 9.75 Gy (vDR), 6.2 Gy (DR) and 4.3 Gy (iDR). High quality dose restoration is essential to minimize or eventually by-pass the physician approval of the restored plan, as long as dose stability can be assumed. Motion (as well as setup and range uncertainties) can be taken into account by including robust optimization in the dose restoration. Restoring clinically-approved dose distribution on repeated CTs does not require new ROI segmentation and is compatible with an online adaptive workflow.

Molecular Imaging Using PET/CT for Radiation Therapy Planning for Adult Cancers: Current Status and Expanding Applications

Accurate tumor delineation is a priority in radiation therapy (RT). Metabolic imaging has a key and evolving role in target volume selection and delineation. This is especially so for non-small cell lung cancer, squamous cell cancer of the head and neck, and lymphoma, for which positron emission tomography/computed tomography (PET/CT) is complimentary to structural imaging modalities, not only in delineating primary tumors, but also often in revealing previously undiagnosed regional nodal disease. At some sites, PET/CT has been confirmed to enable target size reduction compared with structural imaging alone, with enhanced normal tissue sparing and potentially allowing for dose escalation. These contributions often dramatically affect RT strategies. However, some limitations exist to the use of fluorodeoxyglucose-PET in RT planning, including its relatively poor spatial resolution and partial voluming effects for small tumors. A role is developing for contributions from metabolic imaging to RT planning at other tumor sites and exciting new applications for the use of non-fluorodeoxyglucose metabolic markers for RT planning.

Performance of commercially available deformable image registration platforms for contour propagation using patient-based computational phantoms: A multi-institutional study

PURPOSE

To investigate the performance of various algorithms for deformable image registration (DIR) to propagate regions of interest (ROIs) using multiple commercial platforms.

METHODS AND MATERIALS

Thirteen institutions participated in the study with six commercial platforms: RayStation (RaySearch Laboratories, Stockholm, Sweden), MIM (Cleveland, OH, USA), VelocityAI and Smart Adapt (Varian Medical Systems, Palo Alto, CA, USA), Mirada XD (Mirada Medical Ltd, Oxford, UK), and ABAS (Elekta AB, Stockholm, Sweden). The DIR algorithms were tested on synthetic images generated with the ImSimQA package (Oncology Systems Limited, Shrewsbury, UK) by applying two specific Deformation Vector Fields (DVF) to real patient data-sets. Head-and-neck (HN), thorax, and pelvis sites were included. The accuracy of the algorithms was assessed by comparing the DIR-mapped ROIs from each center with those of reference, using the Dice Similarity Coefficient (DSC) and Mean Distance to Conformity (MDC) metrics. Statistical inference on validation results was carried out in order to identify the prognostic factors of DIR performances.

RESULTS

DVF intensity, anatomic site and participating center were significant prognostic factors of DIR performances. Sub-voxel accuracy was obtained in the HN by all algorithms. Large errors, with MDC ranging up to 6 mm, were observed in low-contrast regions that underwent significant deformation, such as in the pelvis, or large DVF with strong contrast, such as the clinical tumor volume (CTV) in the lung. Under these conditions, the hybrid DIR algorithms performed significantly better than the free-form intensity based algorithms and resulted robust against intercenter variability.

CONCLUSIONS

The performances of the systems proved to be site specific, depending on the DVF type and the platforms and the procedures used at the various centers. The pelvis was the most challenging site for most of the algorithms, which failed to achieve sub-voxel accuracy. Improved reproducibility was observed among the centers using the same hybrid registration algorithm.

Validating Dose Uncertainty Estimates Produced by AUTODIRECT: An Automated Program to Evaluate Deformable Image Registration Accuracy

Deformable image registration is a powerful tool for mapping information, such as radiation therapy dose calculations, from one computed tomography image to another. However, deformable image registration is susceptible to mapping errors. Recently, an automated deformable image registration evaluation of confidence tool was proposed to predict voxel-specific deformable image registration dose mapping errors on a patient-by-patient basis. The purpose of this work is to conduct an extensive analysis of automated deformable image registration evaluation of confidence tool to show its effectiveness in estimating dose mapping errors. The proposed format of automated deformable image registration evaluation of confidence tool utilizes 4 simulated patient deformations (3 B-spline-based deformations and 1 rigid transformation) to predict the uncertainty in a deformable image registration algorithm's performance. This workflow is validated for 2 DIR algorithms (B-spline multipass from Velocity and Plastimatch) with 1 physical and 11 virtual phantoms, which have known ground-truth deformations, and with 3 pairs of real patient lung images, which have several hundred identified landmarks. The true dose mapping error distributions closely followed the Student t distributions predicted by automated deformable image registration evaluation of confidence tool for the validation tests: on average, the automated deformable image registration evaluation of confidence tool-produced confidence levels of 50%, 68%, and 95% contained 48.8%, 66.3%, and 93.8% and 50.1%, 67.6%, and 93.8% of the actual errors from Velocity and Plastimatch, respectively. Despite the sparsity of landmark points, the observed error distribution from the 3 lung patient data sets also followed the expected error distribution. The dose error distributions from automated deformable image registration evaluation of confidence tool also demonstrate good resemblance to the true dose error distributions. Automated deformable image registration evaluation of confidence tool was also found to produce accurate confidence intervals for the dose-volume histograms of the deformed dose.

Comparison of rigid and deformable registration through the respiratory phases of four-dimensional computed tomography image data sets for radiotherapy after breast-conserving surgery

BACKGROUND

The aim of this study was to compare the geometric differences in gross tumor volume (GTV) and surgical clips propagated by rigid image registration (RIR) and deformable image registration (DIR) using a four-dimensional computed tomography (4DCT) image data set for patients treated with boost irradiation or accelerated partial breast irradiation after breast-conserving surgery (BCS).

METHODS

The 4DCT data sets of 44 patients who had undergone BCS were acquired. GTV and selected clips were manually delineated on end-inhalation phase (CT0) and end-exhalation phase (CT50) images of 4DCT data sets. Subsequently, the GTV and selected clips from CT0 images were transformed and propagated to CT50 images using RIR and DIR, respectively. The geometric differences in GTV and surgical clips from DIR were compared with those of RIR.

RESULTS

The mean Dice similarity coefficient (DSC) index was 0.860 ± 0.042 for RIR and 0.870 ± 0.040 for DIR for GTV (P = .000). The three-dimensional distance to the center of mass (COM) of the GTV from RIR was longer than that from DIR (1.22 mm and 1.10 mm, respectively, P = .000). Moreover, in the anterior-posterior direction, displacements from RIR were significantly greater than those from DIR for both GTV (0.70 mm and 0.50 mm, respectively) and selected clips (upper clip, 0.45 mm vs 0.20 mm; inner clip, 0.55 mm vs 0.30 mm; outer clip, 0.40 mm vs 0.20 mm; lower clip, 0.50 mm vs 0.25 mm) (P = .000). However, in the left-right and superior-inferior directions, there were no significant displacement differences between RIR and DIR for GTV and the selected clips (all P > .050).

CONCLUSION

DIR can improve the overlap for GTV registration from CT0 to CT50 images from 4DCT scanning. Furthermore, DIR is superior to RIR in reflecting the displacement of GTV and selected clips in the anterior-posterior direction induced by respiratory movement.

Evaluation of rectum and bladder dose accumulation from external beam radiotherapy and brachytherapy for cervical cancer using two different deformable image registration techniques

We evaluated dose-volume histogram (DVH) parameters based on deformable image registration (DIR) between brachytherapy (BT) and external beam radiotherapy (EBRT) that included a center-shielded (CS) plan. Eleven cervical cancer patients were treated with BT, and their pelvic and CS EBRT were studied. Planning CT images for EBRT and BT (except for the first BT, used as the reference image) were deformed with DIR to reference image. We used two DIR parameter settings: intensity-based and hybrid. Mean Dice similarity coefficients (DSCs) comparing EBRT with the reference for the uterus, rectum and bladder were 0.81, 0.77 and 0.83, respectively, for hybrid DIR and 0.47, 0.37 and 0.42, respectively, for intensity-based DIR (P < 0.05). D1 cm3 for hybrid DIR, intensity-based DIR and DVH addition were 75.1, 81.2 and 78.2 Gy, respectively, for the rectum, whereas they were 93.5, 92.3 and 94.3 Gy, respectively, for the bladder. D2 cm3 for hybrid DIR, intensity-based DIR and DVH addition were 70.1, 74.0 and 71.4 Gy, respectively, for the rectum, whereas they were 85.4, 82.8 and 85.4 Gy, respectively, for the bladder. Overall, hybrid DIR obtained higher DSCs than intensity-based DIR, and there were moderate differences in DVH parameters between the two DIR methods, although the results varied among patients. DIR is only experimental, and extra care should be taken when comparing DIR-based dose values with dose-effect curves established using DVH addition. Also, a true evaluation of DIR-based dose accumulation would require ground truth data (e.g. measurement with physical phantom).

Adherence to ICRU-83 reporting recommendations is inadequate in prostate dosimetry studies

PURPOSE

This study aimed to investigate if the International Commission on Radiation Units and Measurements (ICRU) 83 recommendations for reporting dosimetric endpoints are followed in published prostate studies using modulated techniques.

METHODS AND MATERIALS

Prostate dosimetry studies using inverse planning techniques were identified through a search of PubMed and EMBASE databases. These studies were analyzed to determine if the endpoints reported followed the recommendations outlined in ICRU-83. A data collection form was completed and any alternative methods of reporting were recorded. Results were analyzed using frequencies, percentages, and Fisher exact tests.

RESULTS

The ICRU-83 recommendations were not followed in the majority of studies. For the planning target volume, the dose received by 2% of the volume, the dose received by 98% of the volume, and the dose received by 50% of the volume were reported in 22.9%, 18.8%, and 8.3% of studies, respectively. The adherence to reporting for the clinical target volume was below 5% for all specifications. The mean dose, the dose received by a specified volume, and dose received by 2% of the volume for organs at risk were reported in 47.1%, 83.3%, and 16.7%, respectively. The homogeneity index was used in 14.6% of studies. Conformity was discussed in 45.8% of studies. Confidence intervals were included in 37.5% of studies.

CONCLUSIONS

The reporting recommendations of ICRU-83 were not adhered to in the majority of the dosimetry studies reviewed, highlighting the need for greater diligence for authors and reviewers when publishing planning outcomes for modulated techniques.

A graphical approach to optimizing variable-kernel smoothing parameters for improved deformable registration of CT and cone beam CT images

A systematic method is presented for determining optimal parameters in variable-kernel deformable image registration of cone beam CT and CT images, in order to improve accuracy and convergence for potential use in online adaptive radiotherapy. Assessed conditions included the noise constant (symmetric force demons), the kernel reduction rate, the kernel reduction percentage, and the kernel adjustment criteria. Four such parameters were tested in conjunction with reductions of 5, 10, 15, 20, 30, and 40%. Noise constants ranged from 1.0 to 1.9 for pelvic images in ten prostate cancer patients. A total of 516 tests were performed and assessed using the structural similarity index. Registration accuracy was plotted as a function of iteration number and a least-squares regression line was calculated, which implied an average improvement of 0.0236% per iteration. This baseline was used to determine if a given set of parameters under- or over-performed. The most accurate parameters within this range were applied to contoured images. The mean Dice similarity coefficient was calculated for bladder, prostate, and rectum with mean values of 98.26%, 97.58%, and 96.73%, respectively; corresponding to improvements of 2.3%, 9.8%, and 1.2% over previously reported values for the same organ contours. This graphical approach to registration analysis could aid in determining optimal parameters for Demons-based algorithms. It also establishes expectation values for convergence rates and could serve as an indicator of non-physical warping, which often occurred in cases  >0.6% from the regression line.

Towards ultrasound-guided adaptive radiotherapy for cervical cancer: Evaluation of Elekta's semiautomated uterine segmentation method on 3D ultrasound images

PURPOSE

3D ultrasound (US) images of the uterus may be used to adapt radiotherapy (RT) for cervical cancer patients based on changes in daily anatomy. This requires accurate on-line segmentation of the uterus. The aim of this work was to assess the accuracy of Elekta's "Assisted Gyne Segmentation" (AGS) algorithm in semi-automatically segmenting the uterus on 3D transabdominal ultrasound images by comparison with manual contours.

MATERIALS & METHODS

Nine patients receiving RT for cervical cancer were imaged with the 3D Clarity® transabdominal probe at RT planning, and 1 to 7 times during treatment. Image quality was rated from unusable (0)-excellent (3). Four experts segmented the uterus (defined as the uterine body and cervix) manually and using AGS on images with a ranking > 0. Pairwise analysis between manual contours was evaluated to determine interobserver variability. The accuracy of the AGS method was assessed by measuring its agreement with manual contours via pairwise analysis.

RESULTS

35/44 images acquired (79.5%) received a ranking > 0. For the manual contour variation, the median [interquartile range (IQR)] distance between centroids (DC) was 5.41 [5.0] mm, the Dice similarity coefficient (DSC) was 0.78 [0.11], the mean surface-to-surface distance (MSSD) was 3.20 [1.8] mm, and the uniform margin of 95% (UM95) was 4.04 [5.8] mm. There was no correlation between image quality and manual contour agreement. AGS failed to give a result in 19.3% of cases. For the remaining cases, the level of agreement between AGS contours and manual contours depended on image quality. There were no significant differences between the AGS segmentations and the manual segmentations on the images that received a quality rating of 3. However, the AGS algorithm had significantly worse agreement with manual contours on images with quality ratings of 1 and 2 compared with the corresponding interobserver manual variation. The overall median [IQR] DC, DSC, MSSD, and UM95 between AGS and manual contours was 5.48 [5.45] mm, 0.77 [0.14], 3.62 [2.7] mm, and 5.19 [8.1] mm, respectively.

CONCLUSIONS

The AGS tool was able to represent uterine shape of cervical cancer patients in agreement with manual contouring in cases where the image quality was excellent, but not in cases where image quality was degraded by common artifacts such as shadowing and signal attenuation. The AGS tool should be used with caution for adaptive RT purposes, as it is not reliable in accurately segmenting the uterus on 'good' or 'poor' quality images. The interobserver agreement between manual contours of the uterus drawn on 3D US was consistent with results of similar studies performed on CT and MRI images.

Evaluation of Imaging Dose From Different Image Guided Systems During Head and Neck Radiotherapy: A Phantom Study

This work evaluated and compared the absorbed doses to selected organs in the head and neck region from the three image guided radiotherapy systems: cone-beam computed tomography (CBCT) and kilovoltage (kV) planar imaging using the On-board Imager® (OBI) as well as the ExacTrac® X-ray system, all available on the Varian Novalis TX linear accelerator. The head and neck region of an anthropomorphic phantom was used to simulate patients' head within the imaging field. Nanodots optically stimulated luminescent dosemeters were positioned at selected sites to measure the absorbed doses. CBCT was found to be delivering the highest dose to internal organs while OBI-2D gave the highest doses to the eye lenses. The setting of half-rotation in CBCT effectively reduces the dose to the eye lenses. Daily high-quality CBCT verification was found to increase the secondary cancer risk by 0.79%.

Quality assurance of radiotherapy in the ongoing EORTC 1219-DAHANCA-29 trial for HPV/p16 negative squamous cell carcinoma of the head and neck: Results of the benchmark case procedure

BACKGROUND AND PURPOSE

The phase III EORTC 1219-DAHANCA 29 intergroup trial evaluates the influence of nimorazole in patients with locally advanced head and neck cancer when treated with accelerated radiotherapy (RT) in combination with chemotherapy. This article describes the results of the RT Benchmark Case (BC) performed before patient inclusion.

MATERIALS AND METHODS

The participating centers were asked to perform a 2-step BC, consisting of (1) a delineation and (2) a planning exercise according to the protocol guidelines. Submissions were prospectively centrally reviewed and feedback was given to the submitting centers. Sørensen-Dice similarity index (DSI) and the 95th percentile Hausdorff distance (HD) were retrospectively used to evaluate the agreement between the centers and the expert contours.

RESULTS

Fifty-four submissions (34 delineation and 20 planning exercises) from 19 centers were reviewed. Nine (47%) centers needed to perform the delineation step twice and three (16%) centers 3 times before receiving an approval. An increase in DSI-value and a decrease in HD, in particular for the prophylactic Clinical Target Volume (pCTV), could be found for the resubmitted cases. No unacceptable variations could be found for the planning exercise.

CONCLUSIONS

These BC-results highlight the need for effective and prospective RTQA in clinical trials. Even with clearly defined protocol guidelines, delineation and not planning remain the main reason for unacceptable protocol variations. The introduction of more objective quantitative analysis methods, such as the HD and DSI, in future trials might strengthen the evaluation by experts.

Comprehensive evaluation of ten deformable image registration algorithms for contour propagation between CT and cone-beam CT images in adaptive head & neck radiotherapy

Deformable image registration (DIR) is a critical technic in adaptive radiotherapy (ART) for propagating contours between planning computerized tomography (CT) images and treatment CT/cone-beam CT (CBCT) images to account for organ deformation for treatment re-planning. To validate the ability and accuracy of DIR algorithms in organ at risk (OAR) contour mapping, ten intensity-based DIR strategies, which were classified into four categories—optical flow-based, demons-based, level-set-based and spline-based—were tested on planning CT and fractional CBCT images acquired from twenty-one head & neck (H&N) cancer patients who underwent 6~7-week intensity-modulated radiation therapy (IMRT). Three similarity metrics, i.e., the Dice similarity coefficient (DSC), the percentage error (PE) and the Hausdorff distance (HD), were employed to measure the agreement between the propagated contours and the physician-delineated ground truths of four OARs, including the vertebra (VTB), the vertebral foramen (VF), the parotid gland (PG) and the submandibular gland (SMG). It was found that the evaluated DIRs in this work did not necessarily outperform rigid registration. DIR performed better for bony structures than soft-tissue organs, and the DIR performance tended to vary for different ROIs with different degrees of deformation as the treatment proceeded. Generally, the optical flow-based DIR performed best, while the demons-based DIR usually ranked last except for a modified demons-based DISC used for CT-CBCT DIR. These experimental results suggest that the choice of a specific DIR algorithm depends on the image modality, anatomic site, magnitude of deformation and application. Therefore, careful examinations and modifications are required before accepting the auto-propagated contours, especially for automatic re-planning ART systems.