Investigating the clinical use of structured light plethysmography to assess lung function in children with neuromuscular disorders.

BACKGROUND: Children and young people with neuromuscular disorders (NMD), such as Duchenne Muscular Dystrophy (DMD), develop progressive respiratory muscles weakness and pulmonary restriction. Pulmonary function monitoring of the decline in lung function allows for timely intervention with cough assist techniques and nocturnal non-invasive ventilation (NIV). NMD may find the measurement of lung function difficult using current techniques. Structured Light Plethysmography (SLP) has been proposed as a novel, non-contact, self-calibrating, non-invasive method of assessing lung function. The overarching aim of this study was to investigate the use of SLP as a novel method for monitoring respiratory function in children with neuromuscular disease. METHODS: SLP thoraco-abdominal (TA) displacement was correlated with forced vital capacity measurements recorded by spirometry and the repeatability of the measurements with both methods examined. SLP tidal breathing parameters were investigated to assess the range and repeatability of regional right and left side TA displacement and rib cage and abdominal wall displacement. RESULTS: The comparison of the FVC measured with SLP and with spirometry, while having good correlation (R = 0.78) had poor measurement agreement (95% limits of agreement: -1.2 to 1.2L) The mean relative contribution of right and left TA displacement in healthy controls was 50:50 with a narrow range. Repeatability of this measure with SLP was found to be good in healthy controls and moderate in NMD children with/without scoliosis but with a wider range. The majority of the control group displayed a predominant rib cage displacement during tidal breathing and those who displayed predominant abdominal wall displacement showed displacement of both regions close to 50:50 with similar results for the rib cage and abdomen. In comparison, children with NMD have a more variable contribution for all of these parameters. In addition, SLP was able to detect a reduction in abdominal contribution to TA displacement with age in the DMD group and detect paradoxical breathing in children with NMD. Using SLP tracings during tidal breathing we were able to identify three specific patterns of breathing amongst healthy individuals and in children with NMD. CONCLUSIONS: SLP is a novel method for measuring lung function that requires limited patient cooperation and may be especially useful in children with neuromuscular disorders. Measuring the relative contributions of the right and left chest wall and chest versus abdominal movements allows a more detailed assessment.

Segmentation of breast MR images using a generalised 2D mathematical model with inflation and deflation forces of active contours.

In medical computer aided diagnosis systems, image segmentation is one of the major pre-processing steps used to ensure only the region of interest, such as the breast region, will be processed in subsequent steps. Nevertheless, breast segmentation is a difficult task due to low contrast and inhomogeneity, especially when estimating the chest wall in magnetic resonance (MR) images. In fact, the chest wall comprises fat, skin, muscles, and the thoracic skeleton, which can misguide automatic methods when attempting to estimate its location. The objective of the study is to develop a fully automated method for breast and pectoral muscle boundary estimation in MR images. Firstly, we develop a 2D breast mathematical model based on 30 MRI slices (from a patient) and identify important landmarks to obtain a model for the general shape of the breast in an axial plane. Subsequently, we use Otsu's thresholding approach and Canny edge detection to estimate the breast boundary. The active contour method is then employed using both inflation and deflation forces to estimate the pectoral muscle boundary by taking account of information obtained from the proposed 2D model. Finally, the estimated boundary is smoothed using a median filter to remove outliers. Our two datasets contain 60 patients in total and the proposed method is evaluated based on 59 patients (one patient is used to develop the 2D breast model). On the first dataset (9 patients) the proposed method achieved Jaccard = 81.1% ±6.1 % and dice coefficient= 89.4% ±4.1 % and on the second dataset (50 patients) Jaccard = 84.9% ±5.8 % and dice coefficient = 92.3% ±3.6 %. These results are qualitatively comparable with the existing methods in the literature.

Fully automated breast boundary and pectoral muscle segmentation in mammograms.

Breast and pectoral muscle segmentation is an essential pre-processing step for the subsequent processes in computer aided diagnosis (CAD) systems. Estimating the breast and pectoral boundaries is a difficult task especially in mammograms due to artifacts, homogeneity between the pectoral and breast regions, and low contrast along the skin-air boundary. In this paper, a breast boundary and pectoral muscle segmentation method in mammograms is proposed. For breast boundary estimation, we determine the initial breast boundary via thresholding and employ Active Contour Models without edges to search for the actual boundary. A post-processing technique is proposed to correct the overestimated boundary caused by artifacts. The pectoral muscle boundary is estimated using Canny edge detection and a pre-processing technique is proposed to remove noisy edges. Subsequently, we identify five edge features to find the edge that has the highest probability of being the initial pectoral contour and search for the actual boundary via contour growing. The segmentation results for the proposed method are compared with manual segmentations using 322, 208 and 100mammograms from the Mammographic Image Analysis Society (MIAS), INBreast and Breast Cancer Digital Repository (BCDR) databases, respectively. Experimental results show that the breast boundary and pectoral muscle estimation methods achieved dice similarity coefficients of 98.8% and 97.8% (MIAS), 98.9% and 89.6% (INBreast) and 99.2% and 91.9% (BCDR), respectively.

Current issues with standards in the measurement and documentation of human skeletal anatomy.

Digital modeling of human anatomy has become increasingly important and relies on well-documented quantitative anatomy literature. This type of documentation is common for the spine and pelvis; however, significant issues exist due to the lack of standardization in measurement and technique. Existing literature on quantitative anatomy for the spine and pelvis of white adults (aged 18-65 years, separated into decadal categories) was reviewed from the disciplines of anatomy, manipulative therapy, anthropometrics, occupational ergonomics, biomechanics and forensic science. The data were unified into a single normative model of the sub-axial spine. Two-dimensional orthographic drawings were produced from the 590 individual measurements identified, which informed the development of a 3D digital model. A similar review of full range of motion data was conducted as a meta-analysis and the results were applied to the existing model, providing an inter-connected, articulated digital spine. During these data analysis processes several inconsistencies were observed accompanied by an evidential lack of standardization with measurement and recording of data. These have been categorized as: anatomical terminology; scaling of measurements; measurement methodology, dimension and anatomical reference positions; global coordinate systems. There is inconsistency in anatomical terminology where independent researchers use the same terms to describe different aspects of anatomy or different terms for the same anatomy. Published standards exist for measurement methods of the human body regarding spatial interaction, anthropometric databases, automotive applications, clothing industries and for computer manikins, but none exists for skeletal anatomy. Presentation of measurements often lacks formal structure in clinical publications, seldom providing geometric reference points, therefore making digital reconstruction difficult. Published quantitative data does not follow existing international published standards relating to engineering drawing and visual communication. Large variations are also evident in standards or guidelines used for global coordinate systems across biomechanics, ergonomics, software systems and 3D software applications. This paper identifies where established good practice exists and suggests additional recommendations, informing an improved communication protocol, to assist reconstruction of skeletal anatomy using 3D digital modeling.

Computed tomography characterisation of additive manufacturing materials.

Additive manufacturing, covering processes frequently referred to as rapid prototyping and rapid manufacturing, provides new opportunities in the manufacture of highly complex and custom-fitting medical devices and products. Whilst many medical applications of AM have been explored and physical properties of the resulting parts have been studied, the characterisation of AM materials in computed tomography has not been explored. The aim of this study was to determine the CT number of commonly used AM materials. There are many potential applications of the information resulting from this study in the design and manufacture of wearable medical devices, implants, prostheses and medical imaging test phantoms. A selection of 19 AM material samples were CT scanned and the resultant images analysed to ascertain the materials' CT number and appearance in the images. It was found that some AM materials have CT numbers very similar to human tissues, FDM, SLA and SLS produce samples that appear uniform on CT images and that 3D printed materials show a variation in internal structure.

Medical rapid prototyping technologies: state of the art and current limitations for application in oral and maxillofacial surgery.

PURPOSE: We describe state-of-the-art software and hardware requirements for the manufacture of high quality medical models manufactured using medical rapid prototyping. The limitations of medical models, the source of artefacts, and their physical appearance are illustrated along with remedies for their removal. MATERIALS AND METHODS: Medical models were built using predominantly stereolithography and fused deposition modeling at both institutions over a period of 6 years. A combined total of 350 models have been produced for a range of maxillofacial, neurosurgical, and orthopedic applications. Stereolithography, fused deposition modeling, computerized numerical milling, and other technologies are described along with computer software requirements. RESULTS: A range of unwanted artefacts that create distortions on medical models have been identified. These include data import, computed tomography gantry distortion, metal, motion, surface roughness due to support structure removal or surface modeling, and image data thresholding. The source of the artefact has been related to the patient, imaging modality performance, or the modeling technology. Discussion as to the significance of the artefacts on clinical use is provided. CONCLUSIONS: It is recommended that models of human anatomy generated by medical rapid prototyping are subject to rigorous quality assurance at all stages of the manufacturing process. Clinicians should be aware of potential areas for inaccuracies within models and review the source images in cases where model integrity is in doubt.

Virtual surgical planning and CAD/CAM in the treatment of cranial defects.

The purpose of this work was to enhance the clinical outcome in the neurosurgical treatment of cranial defects and fibrous dysplasia. Cranial defects require repair using a variety of materials to protect the brain and provide a good cosmetic outcome for the patient. Virtual neurosurgery and CAD/CAM techniques have been employed to increase the implant 3-dimensional accuracy. The source data was 3D Computed Tomography scans. The CT scans were visualised using surface shading and rotation with volume rotation and re-sizing being used to review the complete data set. The volume data was manipulated using 3D image editing and surface modelling. Physical models of the patients' skulls were created using computed numerical milling which provided a physically accurate template for the production of a cranial implant. This has led to improved fitting for titanium plates requiring less theatre time and in the case of fibrous dysplasia a change in the operative technique from a two stage operation to one stage.

Increasing face validity of a vascular interventional training system.

BACKGROUND: Many aspects of medical training take place on real patients in a live environment thus incurring risk. Apart from the obvious risks to patients there is the issue of X-ray exposure to both staff and trainees. Image quality used during interventional procedures is low to ensure minimum X-ray radiation dose. A virtual interventional system may be used to simulate the interventional cardiology training environment therefore reducing overall risk. The purpose of this project was to determine the appropriate image quality settings on a virtual training system to accurately simulate interventional cardiology and increase the face validity of the overall system. METHODS: Real image data was obtained from patients undergoing diagnostic interventional cardiology examinations in the Royal Victoria Hospital, Belfast. The images were obtained during catheter placement and contrast enhancement around the region of the heart. The diagnostic views were left anterior oblique, right anterior oblique, left lateral, caudal and the spider view. These five views were simulated on a virtual interventional training system (VIST, Mentice Medical Simulation AB, Sweden) by adjusting x-ray tube and table position at 20 different image quality settings. Signal to noise ratio and image contrast were measured for each of these views using soft tissue and vertebra. The results from the real image data and the VIST image data were compared. Calibration curves were compiled to determine the appropriate VIST image quality setting for each procedure. RESULTS: Average SNR ranged from 2.94 to 74.0 and IC ranged from 0.009 to 0.61 on the real image data. This indicates the very wide range of image quality encountered in interventional cardiology. Calibration curves for each view with the corresponding range of real patient image quality were produced. These calibration curves enabled the appropriate VIST image quality setting to be determined and therefore simulate the real examination image quality as closely as possible. CONCLUSIONS: The VIST system has been calibrated in terms of image quality for interventional cardiology examinations. This has increased the face validity of the system.

Assessment of VIE image quality using helical CT angiography: in vitro phantom study.

The aim of this study is to quantify the effects of helical CT acquisitions parameters on the magnitude of three-dimensional stair-step artefacts, visualization of renal ostium and morphologies of suprarenal stents observed using virtual intravascular endoscopy. This was performed in a phantom of the human abdominal aorta with a stent graft in situ. Stair-step artefacts were quantified by measuring the standard deviation of signal intensity on surface shaded images and the influence of these artefacts on the visualization of arterial ostia and stent morphologies were assessed by three radiologists. The methodology may be used to optimise the CT system performance for helical CT angiography in aortic stent grafting.