| Author: | Hui, Chi Ho |
| Title: | Development of a clinical management cloud system for scoliosis based on three-dimensional ultrasound imaging |
| Advisors: | Zheng, Yongping (BME) |
| Degree: | Ph.D. |
| Year: | 2023 |
| Subject: | Spine -- Abnormalities Spine -- Imaging Information storage and retrieval systems -- Health services administration Hong Kong Polytechnic University -- Dissertations |
| Department: | Department of Biomedical Engineering |
| Pages: | xiv, 189 pages : color illustrations |
| Language: | English |
| Abstract: | Scoliosis is a common disease among children, which affects two to four percent of the population. Most scoliosis cases are mild, and usually do not require treatment. However, there are chances that the curvature of the spine progresses to a severe angle, and adversely affects the pulmonary functions of the patient, which could be lethal and requires surgical treatment. Adolescent idiopathic scoliosis (AIS) is multi-factorially aetiopathogenic, and the aetiology is not clear, which makes the diagnosis of the disease challenging, where over-treatment and under-treatment could be very undesirable. Surgery for scoliosis should be avoided as the postoperative recovery is a long and excruciating process. Recently, scoliosis screening has been recommended for early detection of the disease. In addition, continuous monitoring and management of scoliosis progression is essential for effective care of the disease, where treatment plan can be designed specifically according to the progression of the patient. In response to the needs of screening, as well as progression monitoring and management, we have designed and developed a web-based clinical management cloud system, and has three main aspects, namely, data collection, data processing, and data analytics. The overall concept is to encourage regular scoliosis examinations. Then, progression data can be obtained by processing the scoliosis examinations of the patients. As a result, the progression data can be further analyzed for the understanding of the disease and predicting the progression of the disease. Conventionally, X-ray examination is the gold standard imaging modality for progression monitoring. However, due to its radiation hazard, it cannot be used frequently. Recently, 3D ultrasound imaging for scoliosis has been demonstrated to be feasible and reliable. A commercially available system, Scolioscan has been available for reliable clinical scoliosis assessment using 3D ultrasound techniques. Based on Scolioscan, we have built Scolioscan Air, which is the first portable three-dimensional ultrasound device for scoliosis assessment. It innovatively uses an optical tracking sensor to detect the spatial position of the ultrasound probe in real-time. Moreover, it is the first 3D ultrasound imaging device that uses active tracking technique to determine the spatial position of the device. The result is greatly improved portability, while maintained comparable accuracy with respect to the original Scolioscan machine. In addition, we have developed Scoliosync, a unique data and image synchronization program to be used with Scolioscan Air. It synchronizes data and images to Scolionet in background, and when internet connection is not available, the new data will be staged locally for later synchronization. This allows Scolioscan Air to be a data collection tools and to be used for school screening and screening in remote area. For the data processing, we have designed and developed Scolionet, it is a web-based clinical management cloud system, which offers a secure central repository for storing patient examination from different Scolioscan machines. The data stored in Scolionet are reconciled to generate the progression data of the patients. Furthermore, various unique functions are implemented in Scolionet to assist research and analysis of the disease, including automatic ultrasound angle measurement, coronal and sagittal angle manual measurement with flexible switching of VPI images. Conventionally, scoliotic angle measurements are manual process, which is tedious and time-consuming. Progression monitoring and prediction is impractical if progression data is relied on manual measurement. In this study, we have attempted to build a machine learning model to predict scoliosis progression with automatic scoliotic angle measurement of Scolioscan VPI images as one of the prediction factors. We have retrieved Scolioscan examinations and X-ray examinations from previous research projects and uploaded to Scolionet for retrospective study and analysis. Moreover, we have selected 176 patient examinations, which have both Scolioscan and X-ray examination performed on the same day. Ultrasound curve angles are obtained via automatic ultrasound measurement and are validated with the X-ray Cobb’s angle. The two measurements show a very high correlation (R2=0.948). These examinations have generated two sets of 88 progression records from the X-ray Cobb’s angle and automatic ultrasound measurement. The two sets of progression records are compared and have demonstrated an accuracy of 86 percent, sensitivity of 72 percent and, specificity of 92 percent. The validated automatic progression records and other prediction factors, including manual sagittal angle measurements are used to train a machine learning decision tree model to predict scoliosis progression. The decision model has demonstrated an accuracy of 78 percent, sensitivity of 48 percent and, specificity of 94 percent, positive predictive value (PPV) of 75 percent and negative predictive value (NPV) of 79 percent. While there are still room for the overall performance of the model to improve, it is the first study that uses a cloud-based data management system for scoliosis progression prediction using machine learning model with automatic ultrasound angle measurement and manual sagittal angle measurement as prediction factors. The result suggests that three-dimensional ultrasound is suitable for progress monitoring and management. Furthermore, Scolionet can also be potentially used with Scolioscan to screen scoliosis and as a treatment outcome measure. |
| Rights: | All rights reserved |
| Access: | open access |
Copyright Undertaking
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item:
https://theses.lib.polyu.edu.hk/handle/200/12554