INNOVATION BLOSSOMS IN
AI tool could dramatically decrease
diagnostic times for psoriatic arthritis patients
Researchers in Israel have shown that a new machine-learning tool can speed up the diagnosis of psoriatic arthritis (PsA) by up to 4 years,1 potentially preventing irreversible joint damage and deteriorating function for sufferers.
PsA is a progressive inflammatory condition that effects the joints and connective skin mostly in patients who suffer from psoriasis (a chronic skin disease). The most common symptoms are joint pain and swelling, which can range from mild to severe, but many patients also develop more damaging erosive joint disease and deformities.
The findings from the researchers’ study are being presented today <<Friday>> at the European Academy of Dermatology and Venereology’s (EADV) Spring Symposium in Ljubljana.1
The study retrospectively researched and analysed the medical database of Israel’s second largest health medical organization with over 2.5 million members. PredictAI™ analysed the medical records of over 2000 confirmed PsA patients in order to train the algorithm which was then tested on a separate group of confirmed PsA patients and accurately identified 32-51% of them, one to 4 years prior to a clinician’s diagnosis.1
The researchers developed this algorithm with the aim of shortening the time to diagnosis which takes today an average of 2.5 years from the onset of symptoms.2 32% of patients in the study were identified 4 years prior to the diagnosis and 43% one year before a recorded PsA diagnosis by a clinician. When analysing psoriasis patients’ medical records only, 51% of undiagnosed PsA patients were identified one year prior to first diagnosis
The study’s authors believe it would make the most impact when used in a primary care setting because the symptoms of PsA may be unspecific compared to rheumatoid arthritis and awareness of PsA may be lacking in community medical practice.
It is predominantly dermatologists who treat psoriasis patients and since 10% of those patients may have PsA but not be aware of it, we have an opportunity to ask about joint pain,” said Dr Jonathan Shapiro, Dermatologist, medical advisor to Predicta Med analytics LTD and manager of the tele-dermatology service in Maccabi Healthcare Services in Israel.
“Many psoriasis patients themselves might be unaware they have PsA and will contact a general practitioner or an orthopaedic specialist about joint or back pain – not linking it with their skin condition particularly since the non-specific nature of these symptoms makes it difficult for a clinician to diagnose upon first presentation,” Dr Shapiro explained.
“What PredictAI™ brings is the opportunity to scan large medical databases and use AI methods to search for clues such as complaints of joint pain, orthopaedic specialist consultations, lab results and many other
parameters that can help to identify an undiagnosed PsA patient up to 4 years before first suspicion of PsA and can detect over 50% of these patients.”
The results have been described as “A step towards an improved treatment pathway for patients with this painful condition,” by Professor Dedee Murrell, Professor of Dermatology at the University of New South Wales, Sydney and Chair of the EADV Communications Committee.
“Arthritis due to psoriasis can cause permanent damage to the joints and may present years before any psoriasis in the skin is apparent,” she added.
“Early diagnosis and thus earlier treatment which could prevent pain and permanent joint destruction would be welcomed.
“I would be interested to know if these patients already had joint destruction and a randomised prospective study could be done to determine if earlier diagnosis prevented joint destruction and the future development of other co-morbidities associated with psoriasis.”
Dr Shapiro called on the medical profession to “keep an open mind” on the benefits of decision support tools and should they flag up a potential undiagnosed condition to “consider sending these patients for further investigation.”
The team are planning to continue the research to allow the tool to increase its accuracy and sensitivity.
“Our next step is to check the performance of PredictAI on more databases worldwide which can help validate its results and improve it. We are starting a prospective study which aims at identifying currently undiagnosed PsA patients so we can consider referring them to a rheumatologist for further investigation”
- Shapiro J et al. Development of a machine learning tool for early diagnosis of psoriatic arthritis in a primary care setting: A population based study. Abstract No 629. Presented at EADV Spring Symposium 2022. https://eadvsymposium2022.org/wp-content/uploads/2022/05/eadv_abstracts_book.pdf
- Karmacharya et al. Diagnostic Delay in Psoriatic Arthritis: A Population-based Study. The Journal of Rheumatology, 2021;48(9), pp.1410-1416. Available at: https://www.jrheum.org/content/48/9/1410 [Accessed May 2022].
- Sokoll KB, Helliwell P. Comparison of disability and quality of life in rheumatoid and psoriatic arthritis. The Journal of Rheumatology, 2001;28(8), pp. 1842-6. Available at: https://www.jrheum.org/content/28/8/1842 [Accessed May 2022]
FULL SCIENTIFIC PROGRAMME
ALL SCIENTIFIC NEWS
keep in touch
Stay informed about latest EADV information.
Congresses, Education, JEADV, Membership, Newsletter...
yes, keep me informed
of Dermatology and Venereology
Copyright 2021 EADV
Only trust information from our official website: eadv.org