Gareth J Price

Automated delineation of radiotherapy volumes: are we going in the right direction?

Rapid and accurate delineation of target volumes and multiple organs at risk, within the enduring International Commission on Radiation Units and Measurement framework, is now hugely important in radiotherapy, owing to the rapid proliferation of intensity-modulated radiotherapy and the advent of four-dimensional image-guided adaption. Nevertheless, delineation is still generally clinically performed with little if any machine assistance, even though it is both time-consuming and prone to interobserver variation. Currently available segmentation tools include those based on image greyscale interrogation, statistical shape modelling and body atlas-based methods. However, all too often these are not able to match the accuracy of the expert clinician, which remains the universally acknowledged gold standard. In this article we suggest that current methods are fundamentally limited by their lack of ability to incorporate essential human clinical decision-making into the underlying models. Hybrid techniques that utilise prior knowledge, make sophisticated use of greyscale information and allow clinical expertise to be integrated are needed. This may require a change in focus from automated segmentation to machine-assisted delineation. Similarly, new metrics of image quality reflecting fitness for purpose would be extremely valuable. We conclude that methods need to be developed to take account of the clinicians expertise and honed visual processing capabilities as much as the underlying, clinically meaningful information content of the image data being interrogated. We illustrate our observations and suggestions through our own experiences with two software tools developed as part of research council-funded projects.

Reduction of motion artefacts in on-board cone beam CT by warping of projection images.

OBJECTIVE We describe the development and testing of a motion correction method for flat panel imager-based cone beam CT (CBCT) based on warping of projection images. METHODS Markers within or on the surface of the patient were tracked and their mean three-dimensional (3D) position calculated. The two-dimensional (2D) cone beam projection images were then warped before reconstruction to place each marker at the projection from its mean 3D position. The motion correction method was tested using simulated cone beam projection images of a deforming virtual phantom, real CBCT images of a moving breast phantom and clinical CBCT images of a patient with breast cancer and another with pancreatic cancer undergoing radiotherapy. RESULTS In phantom studies, the method was shown to greatly reduce motion artefacts in the locality of the radiotherapy target and allowed the true surface shape to be accurately recovered. The breast phantom motion-compensated surface was within 1 mm of the true surface shape for 90% of surface points and greater than 2 mm from the true surface at only 2% of points. Clinical CBCT images showed improved image quality in the locality of the radiotherapy target after motion correction. CONCLUSION The proposed method is effective in reducing motion artefacts in CBCT images.

A method to calculate coverage probability from uncertainties in radiotherapy via a statistical shape model.

In this paper we describe a technique that may be used to model the geometric uncertainties that accrue during the radiotherapy process. Using data from in-treatment cone beam CT scans, we simultaneously analyse non-uniform observer delineation variability and organ motion together with patient set-up errors via the creation of a point distribution model (PDM). We introduce a novel method of generating a coverage probability matrix, that may be used to determine treatment margins and calculate uncertainties in dose, from this statistical shape model. The technique does not assume rigid body motion and can extrapolate shape variability in a statistically meaningful manner. In order to construct the PDM, we generate corresponding surface points over a set of delineations. Correspondences are established at a set of points in parameter space on spherically parameterized and canonical aligned outlines. The method is demonstrated using rectal delineations from serially acquired in-treatment cone beam CT image volumes of a prostate patient (44 image volumes total), each delineated by a minimum of two observers (maximum six). Two PDMs are constructed, one with set-up errors included and one without. We test the normality assumptions of the PDMs and find the distributions to be Gaussian in nature. The rectal PDM variability is in general agreement with data in the literature. The two resultant coverage probability matrices show differences as expected.

Real-time optical measurement of the dynamic body surface for use in guided radiotherapy

Optical measurements are increasingly used in radiotherapy. In this paper we present, in detail, the design and implementation of a multi-channel optical system optimized for fast, high spatial resolution, dynamic body surface measurement in guided therapy. We include all algorithmic modifications and calibration procedures required to create a robust, practical system for clinical use. Comprehensive static and dynamic phantom validation measurements in the radiotherapy treatment room show: conformance with simultaneously measured cone beam CT data to within 1 mm over 62% ± 8% of the surface and 2 mm over 90% ± 3%; agreement with the measured radius of a precision geometrical phantom to within 1 mm; and true real-time performance with image capture through to surface display at 23 Hz. An example patient dataset is additionally included, indicating similar performance in the clinic.

An analysis of breast motion using high-frequency, dense surface points captured by an optical sensor during radiotherapy treatment delivery

Patient motion is an important factor affecting the quality of external beam radiotherapy in breast patients. We analyse the motion of a dense set of surface points on breast patients throughout their treatment schedule to assess the magnitude and stability of motion, in particular, with respect to breast volume. We use an optical sensor to measure the surface motion of 13 breast cancer patients. Patients were divided into two cohorts dependent upon breast volume. Measurements were made during radiotherapy treatment beam delivery for an average of 12 fractions per patient (total 158 datasets). The motion of each surface point is parameterized in terms of its period, amplitude and relative phase. Inter-comparison of the motion parameters across treatment schedules and between patients is made through the creation of corresponding regions on the breast surfaces. The motion period is spatially uniform and is similar in both patient groups (mean 4 s), with the small volume cohort exhibiting greater inter-fraction period variability. The mean motion amplitude is also similar in both groups with a range between 2 mm and 4 mm and an inter-fraction variability generally less than 1 mm. There is a phase lag of up to 0.4 s across the breast, led by the sternum. Breast patient motion is reasonably stable between and during treatment fractions, with the large volume cohort exhibiting greater repeatability than the small volume one.

Phase unwrapping algorithms for use in a true real-time optical body sensor system for use during radiotherapy

An evaluation of the suitability of eight existing phase unwrapping algorithms to be used in a real-time optical body surface sensor based on Fourier fringe profilometry is presented. The algorithms are assessed on both the robustness of the results they give and their speed of execution. The algorithms are evaluated using four sets of real human body surface data, each containing five-hundred frames, obtained from patients undergoing radiotherapy, where fringe discontinuity is significant. We also present modifications to an existing algorithm, noncontinuous quality-guided path algorithm (NCQUAL), in order to decrease its execution time by a factor of 4 to make it suitable for use in a real-time system. The results obtained from the modified algorithm are compared with those of the existing algorithms. Three suitable algorithms were identified: two-stage noncontinuous quality-guided path algorithm (TSNCQUAL)-the modified algorithm presented here-for online processing and Flynns minimum discontinuity algorithm (FLYNN) and preconditioned conjugate gradient method (PCG) algorithms for enhanced accuracy in off-line processing.

Developing and Validating a Survival Prediction Model for NSCLC Patients Through Distributed Learning Across 3 Countries

Purpose Tools for survival prediction for non-small cell lung cancer (NSCLC) patients treated with chemoradiation or radiation therapy are of limited quality. In this work, we developed a predictive model of survival at 2 years. The model is based on a large volume of historical patient data and serves as a proof of concept to demonstrate the distributed learning approach. Methods and Materials Clinical data from 698 lung cancer patients, treated with curative intent with chemoradiation or radiation therapy alone, were collected and stored at 2 different cancer institutes (559 patients at Maastro clinic (Netherlands) and 139 at Michigan university [United States]). The model was further validated on 196 patients originating from The Christie (United Kingdon). A Bayesian network model was adapted for distributed learning (the animation can be viewed at https://www.youtube.com/watch?v=ZDJFOxpwqEA). Two-year posttreatment survival was chosen as the endpoint. The Maastro clinic cohort data are publicly available at https://www.cancerdata.org/publication/developing-and-validating-survival-prediction-model-nsclc-patients-through-distributed, and the developed models can be found at www.predictcancer.org. Results Variables included in the final model were T and N category, age, performance status, and total tumor dose. The model has an area under the curve (AUC) of 0.66 on the external validation set and an AUC of 0.62 on a 5-fold cross validation. A model based on the T and N category performed with an AUC of 0.47 on the validation set, significantly worse than our model (P<.001). Learning the model in a centralized or distributed fashion yields a minor difference on the probabilities of the conditional probability tables (0.6%); the discriminative performance of the models on the validation set is similar (P=.26). Conclusions Distributed learning from federated databases allows learning of predictive models on data originating from multiple institutions while avoiding many of the data-sharing barriers. We believe that distributed learning is the future of sharing data in health care.

Early clinical evaluation of a novel three-dimensional structure delineation software tool (SCULPTER) for radiotherapy treatment planning

Modern radiotherapy treatment planning (RTP) necessitates increased delineation of target volumes and organs at risk. Conventional manual delineation is a laborious, time-consuming and subjective process. It is prone to inconsistency and variability, but has the potential to be improved using automated segmentation algorithms. We carried out a pilot clinical evaluation of SCULPTER (Structure Creation Using Limited Point Topology Evidence in Radiotherapy) - a novel prototype software tool designed to improve structure delineation for RTP. Anonymized MR and CT image datasets from patients who underwent radiotherapy for bladder or prostate cancer were studied. An experienced radiation oncologist used manual and SCULPTER-assisted methods to create clinically acceptable organ delineations. SCULPTER was also tested by four other RTP professionals. Resulting contours were compared by qualitative inspection and quantitatively by using the volumes of the structures delineated and the time taken for completion. The SCULPTER tool was easy to apply to both MR and CT images and diverse anatomical sites. SCULPTER delineations closely reproduced manual contours with no significant volume differences detected, but SCULPTER delineations were significantly quicker (p<0.05) in most cases. In conclusion, clinical application of SCULPTER resulted in rapid and simple organ delineations with equivalent accuracy to manual methods, demonstrating proof-of-principle of the SCULPTER system and supporting its potential utility in RTP.

Self-management of patient body position, pose, and motion using wide-field, real-time optical measurement feedback: results of a volunteer study.

PURPOSE We present the results of a clinical feasibility study, performed in 10 healthy volunteers undergoing a simulated treatment over 3 sessions, to investigate the use of a wide-field visual feedback technique intended to help patients control their pose while reducing motion during radiation therapy treatment. METHODS AND MATERIALS An optical surface sensor is used to capture wide-area measurements of a subjects body surface with visualizations of these data displayed back to them in real time. In this study we hypothesize that this active feedback mechanism will enable patients to control their motion and help them maintain their setup pose and position. A capability hierarchy of 3 different level-of-detail abstractions of the measured surface data is systematically compared. RESULTS Use of the device enabled volunteers to increase their conformance to a reference surface, as measured by decreased variability across their body surfaces. The use of visual feedback also enabled volunteers to reduce their respiratory motion amplitude to 1.7 ± 0.6 mm compared with 2.7 ± 1.4 mm without visual feedback. CONCLUSIONS The use of live feedback of their optically measured body surfaces enabled a set of volunteers to better manage their pose and motion when compared with free breathing. The method is suitable to be taken forward to patient studies.

Next generation optical surface sensing for real-time measurement in radiotherapy

With the introduction of intensive new treatments such as hypo-fractionation and proton beam therapy, localization of the tumor target volume and tracking of points across the skin entrance surface have become critically important. Optical metrology has been used to monitor the patients bulk position and motion throughout treatment. However systems have not been capable of high temporal and spatial resolution whole-surface topology measurement. We describe the implementation of such a system based on Fourier profilometry. Its algorithm is split into four separate processing stages, including spatial phase determination: descriptions of each stage are given along with the modifications made to increase performance. The optimized system is capable of processing 23 frames per second (fps), with each frame providing 512×512 measured points. The data density, accuracy and performance of the system are an order of magnitude improvement on commercially available clinical systems. We show that this performance permits genuinely real-time measurement of a patient, live during both setup and radiation treatment delivery. It is also fast enough to provide smooth dynamic visualizations of motion at all points on the wraparound body surface for radiotherapy staff and intuitive, direct feed-back to patients.