|
|
Adres - Aravind Diabetic Retinopathy Examination Software
Software for reading and grading Diabetic Retinopathy through tele-ophthalmology: ADRES 3.0.
Introduction:
Diabetic retinopathy results from microvascular end-organ damage to the retina, and is found to occur in 60%1 of people twenty years after onset of type 2 diabetes mellitus; sight threatening retinopathy is found in 2.0% of persons who had diabetes for less than five years, and in 15.5% of persons who have diabetes for 15 or more years. Identifying these patients so as to prevent their going blind is primarily through opportunistic case detection2 in a country like India, since a routine screening programme even if only for known diabetics, is not a feasible option.
The current gold standard for diagnosing the condition is through fundus photography; comparing the fundus pathology with that seen on stereoscopic viewing of seven standard photographs for grading retinopathy or through flourescein angiographic diagnosis. However, even in countries where facilities for close monitoring of diabetics are available, there is no consensus on preferred practice with regard to cost-effective, valid methods of screening for diabetic retinopathy.3 Telemedicine has resulted in new possibilities such as capturing images at a remote location and transmitting them to a skilled grader at a specialist center. While reducing costs and maintaining good quality care,4 this intensifies the need to validate digital images as a tool for diagnosing diabetic retinopathy and following its progression.
This paper describes a telemedicine-based system for grading diabetic retinopathy using digital imaging, and its validity relative to the current standard practice of clinical examination by ophthalmologists.
Design:
A prospective, comparative observational study was conducted on a series of patients with type 2 diabetes mellitus who presented at the retina clinic of a tertiary care center in India. The participants were consecutively recruited from the out-patient clinic between May and June 2006 by the clinic coordinator, and assigned to the clinician who was attending to out-patients on the day.
Each patient underwent a routine detailed comprehensive ophthalmic examination by one of three trained retina specialists, and digital fundus photographs were taken by a technician. The same three clinicians subsequently graded the digital images using the manual software ‘Aravind Diabetic Retinopathy Screening (ADRES 3.0)’ reading and grading system, being masked to the clinical grading assigned to individual patients. The study compares clinical assessments as the current practice gold standard, with manual grading of the five 45 degree photographs generated by digital imaging of the fundus.
|
