Testing for peripheral arterial disease in diabetes
File(s)
Author(s)
Normahani, Pasha
Type
Thesis
Abstract
Peripheral arterial disease (PAD) is a major risk factor for cardiovascular disease, foot ulceration and amputation in people with diabetes. Its diagnosis enables optimisation of therapies to manage these risks. Although a variety of bedside tests are available, they have key limitations and there is no agreement as to which is best. Point-of-care duplex ultrasound (podiatry ankle duplex scan; PAD-scan) is a promising and novel modality that allows for detailed interpretation of visual Doppler waveforms captured from blood vessels at the level of the ankle. However, its diagnostic accuracy has not been formally evaluated. In this thesis, I sought to compare the diagnostic accuracy and cost-effectiveness of the PAD-scan to commonly used bedside tests. I also sought to determine the feasibility of machine learning approaches for classification of visual Doppler arterial waveforms.
I conducted a prospective comparative diagnostic accuracy study of bedside tests in patients presenting to two hospital diabetic foot clinics. The reference test was a full lower limb duplex ultrasound. Health economics analysis was performed by constructing a Markov model to estimate health outcomes and costs over 5 years of different testing strategies. Doppler signals were reconstructed from images of Doppler waveforms and labelled using the results of the reference test. The performance of various machine learning approaches were compared for the classification of peripheral arterial disease.
PAD-scan is the most accurate and cost-effective test for the diagnosis of PAD in diabetes. Its adoption may reduce the number of amputations and cardiovascular deaths. Machine learning for the classification of arterial waveforms is feasible and accurate. Future planned work will aim to 1) validate our diagnostic accuracy findings as part of a large multicenter study, 2) incorporate waveform classification into the diabetic foot assessment by developing a fully integrated digital platform 3) evaluate the prognostic value of bedside tests.
I conducted a prospective comparative diagnostic accuracy study of bedside tests in patients presenting to two hospital diabetic foot clinics. The reference test was a full lower limb duplex ultrasound. Health economics analysis was performed by constructing a Markov model to estimate health outcomes and costs over 5 years of different testing strategies. Doppler signals were reconstructed from images of Doppler waveforms and labelled using the results of the reference test. The performance of various machine learning approaches were compared for the classification of peripheral arterial disease.
PAD-scan is the most accurate and cost-effective test for the diagnosis of PAD in diabetes. Its adoption may reduce the number of amputations and cardiovascular deaths. Machine learning for the classification of arterial waveforms is feasible and accurate. Future planned work will aim to 1) validate our diagnostic accuracy findings as part of a large multicenter study, 2) incorporate waveform classification into the diabetic foot assessment by developing a fully integrated digital platform 3) evaluate the prognostic value of bedside tests.
Version
Open Access
Date Issued
2021-07
Date Awarded
2021-11
Copyright Statement
Creative Commons Attribution NonCommercial Licence
License URL
Advisor
Jaffer, Usman
Publisher Department
Department of Surgery & Cancer
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)