ORCID Profile
0000-0002-8276-9774
Current Organisation
University of New South Wales
Does something not look right? The information on this page has been harvested from data sources that may not be up to date. We continue to work with information providers to improve coverage and quality. To report an issue, use the Feedback Form.
In Research Link Australia (RLA), "Research Topics" refer to ANZSRC FOR and SEO codes. These topics are either sourced from ANZSRC FOR and SEO codes listed in researchers' related grants or generated by a large language model (LLM) based on their publications.
Nanomaterials | Medical Devices | Nanotechnology | Sensor Technology (Chemical aspects)
Diagnostic Methods | Health Status (e.g. Indicators of Well-Being) |
Publisher: JMIR Publications Inc.
Date: 09-04-2018
Publisher: JMIR Publications Inc.
Date: 16-03-2018
Publisher: JMIR Publications Inc.
Date: 03-07-2023
DOI: 10.2196/43154
Abstract: Tuberculosis (TB) was the leading infectious cause of mortality globally prior to COVID-19 and chest radiography has an important role in the detection, and subsequent diagnosis, of patients with this disease. The conventional experts reading has substantial within- and between-observer variability, indicating poor reliability of human readers. Substantial efforts have been made in utilizing various artificial intelligence–based algorithms to address the limitations of human reading of chest radiographs for diagnosing TB. This systematic literature review (SLR) aims to assess the performance of machine learning (ML) and deep learning (DL) in the detection of TB using chest radiography (chest x-ray [CXR]). In conducting and reporting the SLR, we followed the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. A total of 309 records were identified from Scopus, PubMed, and IEEE (Institute of Electrical and Electronics Engineers) databases. We independently screened, reviewed, and assessed all available records and included 47 studies that met the inclusion criteria in this SLR. We also performed the risk of bias assessment using Quality Assessment of Diagnostic Accuracy Studies version 2 (QUADAS-2) and meta-analysis of 10 included studies that provided confusion matrix results. Various CXR data sets have been used in the included studies, with 2 of the most popular ones being Montgomery County (n=29) and Shenzhen (n=36) data sets. DL (n=34) was more commonly used than ML (n=7) in the included studies. Most studies used human radiologist’s report as the reference standard. Support vector machine (n=5), k-nearest neighbors (n=3), and random forest (n=2) were the most popular ML approaches. Meanwhile, convolutional neural networks were the most commonly used DL techniques, with the 4 most popular applications being ResNet-50 (n=11), VGG-16 (n=8), VGG-19 (n=7), and AlexNet (n=6). Four performance metrics were popularly used, namely, accuracy (n=35), area under the curve (AUC n=34), sensitivity (n=27), and specificity (n=23). In terms of the performance results, ML showed higher accuracy (mean ~93.71%) and sensitivity (mean ~92.55%), while on average DL models achieved better AUC (mean ~92.12%) and specificity (mean ~91.54%). Based on data from 10 studies that provided confusion matrix results, we estimated the pooled sensitivity and specificity of ML and DL methods to be 0.9857 (95% CI 0.9477-1.00) and 0.9805 (95% CI 0.9255-1.00), respectively. From the risk of bias assessment, 17 studies were regarded as having unclear risks for the reference standard aspect and 6 studies were regarded as having unclear risks for the flow and timing aspect. Only 2 included studies had built applications based on the proposed solutions. Findings from this SLR confirm the high potential of both ML and DL for TB detection using CXR. Future studies need to pay a close attention on 2 aspects of risk of bias, namely, the reference standard and the flow and timing aspects. PROSPERO CRD42021277155 www.crd.york.ac.uk rospero/display_record.php?RecordID=277155
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
Date: 2022
Publisher: JMIR Publications Inc.
Date: 16-10-2017
Abstract: n a home telemonitoring trial, patient adherence with scheduled vital signs measurements is an important aspect that has not been thoroughly studied and for which data in the literature are limited. Levels of adherence have been reported as varying from approximately 40% to 90%, and in most cases, the adherence rate usually dropped off steadily over time. This drop is more evident in the first few weeks or months after the start. Higher adherence rates have been reported for simple types of monitoring and for shorter periods of intervention. If patients do not follow the intended procedure, poorer results than expected may be achieved. Hence, analyzing factors that can influence patient adherence is of great importance. he goal of the research was to present findings on patient adherence with scheduled vital signs measurements in the recently completed Commonwealth Scientific and Industrial Research Organisation (CSIRO) national trial of home telemonitoring of patients (mean age 70.5 years, SD 9.3 years) with chronic conditions (chronic obstructive pulmonary disease, coronary artery disease, hypertensive diseases, congestive heart failure, diabetes, or asthma) carried out at 5 locations along the east coast of Australia. We investigated the ability of chronically ill patients to carry out a daily schedule of vital signs measurements as part of a chronic disease management care plan over periods exceeding 6 months (302 days, SD 135 days) and explored different levels of adherence for different measurements as a function of age, gender, and supervisory models. n this study, 113 patients forming the test arm of a Before and After Control Intervention (BACI) home telemonitoring trial were analyzed. Patients were required to monitor on a daily basis a range of vital signs determined by their chronic condition and comorbidities. Vital signs included noninvasive blood pressure, pulse oximetry, spirometry, electrocardiogram (ECG), blood glucose level, body temperature, and body weight. Adherence was calculated as the number of days during which at least 1 measurement was taken over all days where measurements were scheduled. Different levels of adherence for different measurements, as a function of age, gender, and supervisory models, were analyzed using linear regression and analysis of covariance for a period of 1 year after the intervention. atients were monitored on average for 302 (SD 135) days, although some continued beyond 12 months. The overall adherence rate for all measurements was 64.1% (range 59.4% to 68.8%). The adherence rates of patients monitored in hospital settings relative to those monitored in community settings were significantly higher for spirometry (69.3%, range 60.4% to 78.2%, versus 41.0%, range 33.1% to 49.0%, P .001), body weight (64.5%, range 55.7% to 73.2%, versus 40.5%, range 32.3% to 48.7%, P .001), and body temperature (66.8%, range 59.7% to 73.9%, versus 55.2%, range 48.4% to 61.9%, P=.03). Adherence with blood glucose measurements (58.1%, range 46.7% to 69.5%, versus 50.2%, range 42.8% to 57.6%, P=.24) was not significantly different overall. Adherence rates for blood pressure (68.5%, range 62.7% to 74.2%, versus 59.7%, range 52.1% to 67.3%, P=.04), ECG (65.6%, range 59.7% to 71.5%, versus 56.5%, range 48.7% to 64.4%, P=.047), and pulse oximetry (67.0%, range 61.4% to 72.7%, versus 56.4%, range 48.6% to 64.1%, P=.02) were significantly higher in males relative to female subjects. No statistical differences were observed between rates of adherence for the younger patient group (70 years and younger) and older patient group (older than 70 years). atients with chronic conditions enrolled in the home telemonitoring trial were able to record their vital signs at home at least once every 2 days over prolonged periods of time. Male patients maintained a higher adherence than female patients over time, and patients supervised by hospital-based care coordinators reported higher levels of adherence with their measurement schedule relative to patients supervised in community settings. This was most noticeable for spirometry. ustralian New Zealand Clinical Trials Registry ACTRN12613000635763 www.anzctr.org.au/Trial/Registration/TrialReview.aspx?id=364030& isReview=true (Archived by WebCite at xPOU3DpR).
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
Date: 12-2021
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
Date: 2021
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
Date: 2020
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
Date: 2020
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
Date: 2023
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
Date: 2019
Publisher: Wiley
Date: 12-02-2018
DOI: 10.1002/RNC.4061
Publisher: JMIR Publications Inc.
Date: 16-10-2017
Publisher: Wiley
Date: 12-10-2018
DOI: 10.1002/RNC.4376
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
Date: 2022
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
Date: 04-2023
Publisher: JMIR Publications Inc.
Date: 18-12-2017
Abstract: easonal variation has an impact on the hospitalization rate of patients with a range of cardiovascular diseases, including myocardial infarction and angina. This paper presents findings on the influence of seasonal variation on the results of a recently completed national trial of home telemonitoring of patients with chronic conditions, carried out at five locations along the east coast of Australia. he aim is to evaluate the effect of the seasonal timing of hospital admission and length of stay on clinical outcome of a home telemonitoring trial involving patients (age: mean 72.2, SD 9.4 years) with chronic conditions (chronic obstructive pulmonary disease coronary artery disease, hypertensive diseases, congestive heart failure, diabetes, or asthma) and to explore methods of minimizing the influence of seasonal variations in the analysis of the effect of at-home telemonitoring on the number of hospital admissions and length of stay (LOS). atients were selected from a hospital list of eligible patients living with a range of chronic conditions. Each test patient was case matched with at least one control patient. A total of 114 test patients and 173 control patients were available in this trial. However, of the 287 patients, we only considered patients who had one or more admissions in the years from 2010 to 2012. Three different groups were analyzed separately because of substantially different climates: (1) Queensland, (2) Australian Capital Territory and Victoria, and (3) Tasmania. Time series data were analyzed using linear regression for a period of 3 years before the intervention to obtain an average seasonal variation pattern. A novel method that can reduce the impact of seasonal variation on the rate of hospitalization and LOS was used in the analysis of the outcome variables of the at-home telemonitoring trial. est patients were monitored for a mean 481 (SD 77) days with 87% (53/61) of patients monitored for more than 12 months. Trends in seasonal variations were obtained from 3 years’ of hospitalization data before intervention for the Queensland, Tasmania, and Australian Capital Territory and Victoria subgroups, respectively. The maximum deviation from baseline trends for LOS was 101.7% (SD 42.2%), 60.6% (SD 36.4%), and 158.3% (SD 68.1%). However, by synchronizing outcomes to the start date of intervention, the impact of seasonal variations was minimized to a maximum of 9.5% (SD 7.7%), thus improving the accuracy of the clinical outcomes reported. easonal variations have a significant effect on the rate of hospital admission and LOS in patients with chronic conditions. However, the impact of seasonal variation on clinical outcomes (rate of admissions, number of hospital admissions, and LOS) of at-home telemonitoring can be attenuated by synchronizing the analysis of outcomes to the commencement dates for the telemonitoring of vital signs. ustralian New Zealand Clinical Trial Registry ACTRN12613000635763 www.anzctr.org.au/Trial/Registration/TrialReview.aspx?id=364030& isReview=true (Archived by WebCite at www.webcitation.org/ 6xLPv9QDb)
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
Date: 04-2021
Publisher: Wiley
Date: 29-08-2019
DOI: 10.1002/RNC.4727
Publisher: JMIR Publications Inc.
Date: 10-2022
Abstract: uberculosis (TB) was the leading infectious cause of mortality globally prior to COVID-19 and chest radiography has an important role in the detection, and subsequent diagnosis, of patients with this disease. The conventional experts reading has substantial within- and between-observer variability, indicating poor reliability of human readers. Substantial efforts have been made in utilizing various artificial intelligence–based algorithms to address the limitations of human reading of chest radiographs for diagnosing TB. his systematic literature review (SLR) aims to assess the performance of machine learning (ML) and deep learning (DL) in the detection of TB using chest radiography (chest x-ray [CXR]). n conducting and reporting the SLR, we followed the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. A total of 309 records were identified from Scopus, PubMed, and IEEE (Institute of Electrical and Electronics Engineers) databases. We independently screened, reviewed, and assessed all available records and included 47 studies that met the inclusion criteria in this SLR. We also performed the risk of bias assessment using Quality Assessment of Diagnostic Accuracy Studies version 2 (QUADAS-2) and meta-analysis of 10 included studies that provided confusion matrix results. arious CXR data sets have been used in the included studies, with 2 of the most popular ones being Montgomery County (n=29) and Shenzhen (n=36) data sets. DL (n=34) was more commonly used than ML (n=7) in the included studies. Most studies used human radiologist’s report as the reference standard. Support vector machine (n=5), k-nearest neighbors (n=3), and random forest (n=2) were the most popular ML approaches. Meanwhile, convolutional neural networks were the most commonly used DL techniques, with the 4 most popular applications being ResNet-50 (n=11), VGG-16 (n=8), VGG-19 (n=7), and AlexNet (n=6). Four performance metrics were popularly used, namely, accuracy (n=35), area under the curve (AUC n=34), sensitivity (n=27), and specificity (n=23). In terms of the performance results, ML showed higher accuracy (mean ~93.71%) and sensitivity (mean ~92.55%), while on average DL models achieved better AUC (mean ~92.12%) and specificity (mean ~91.54%). Based on data from 10 studies that provided confusion matrix results, we estimated the pooled sensitivity and specificity of ML and DL methods to be 0.9857 (95% CI 0.9477-1.00) and 0.9805 (95% CI 0.9255-1.00), respectively. From the risk of bias assessment, 17 studies were regarded as having unclear risks for the reference standard aspect and 6 studies were regarded as having unclear risks for the flow and timing aspect. Only 2 included studies had built applications based on the proposed solutions. indings from this SLR confirm the high potential of both ML and DL for TB detection using CXR. Future studies need to pay a close attention on 2 aspects of risk of bias, namely, the reference standard and the flow and timing aspects. ROSPERO CRD42021277155 www.crd.york.ac.uk rospero/display_record.php?RecordID=277155
Start Date: 08-2022
End Date: 07-2027
Amount: $5,000,000.00
Funder: Australian Research Council
View Funded Activity