Developing a reliable way to predict which patients with psoriasis will go on to develop psoriatic arthritis (PsA) has been a research priority for years, but to date no silver bullet has materialized. Continuing investigations and promising solutions could help clinicians predict who will develop PsA or intercept the disease before it can fully manifest. And in the time-strapped dermatologist’s office, tried-and-true measures can help detect and address PsA as early as possible.

Currently, it’s known that up to 30% of people with psoriasis go on to develop PsA, with a reported median time of 2.5 years from symptom onset to diagnosis.

“There are a number of prediction models out there, and there are definitely recognized risk factors. Things like family history, obesity, severe psoriasis, and location of psoriasis all can increase your risk,” PsA prevention researcher Laura Coates, MBChB, PhD, of the Nuffield Department of Orthopaedics, Rheumatology, and Musculoskeletal Sciences at the University of Oxford in Oxford, England, told Medscape Medical News. “But the prediction models mostly haven’t been validated, and they don’t really give you a number. So they might say you are at slightly higher risk, but they’re not sophisticated enough to really define that risk very well.”

Factors such as nail dystrophy, scalp lesions, and intragluteal or perianal psoriasis have long been identified as increasing the likelihood of PsA. The key barrier to prediction lies not in a lack of knowledge but rather in the complex nature of PsA itself. 

In joint comments emailed to Medscape Medical News, co-presidents of the Group for Research and Assessment of Psoriasis and Psoriatic Arthritis Arthur Kavanaugh, MD, rheumatologist, professor of medicine, and director of the Center for Innovative Therapy at University of California San Diego Health, and Joseph Merola, MD, MMSc, rheumatologist-dermatologist, the Mary Kay Distinguished Chair in Dermatology, and professor of medicine in the Division of Rheumatic Diseases at University of Texas Southwestern Medical Center in Dallas, said, “It is not possible to predict with certainty which individual patient with psoriasis will develop PsA.”

“One of the greatest challenges is that PsA is an extremely heterogeneous disease,” Merola and Kavanaugh explained. “Patients may develop peripheral arthritis, axial arthritis, enthesitis, dactylitis, or various combinations of these manifestations. The challenge is therefore not only predicting whether someone will develop PsA but also determining which disease domains may be involved, what type of PsA may develop, and when it may become clinically apparent.”

Some researchers are increasingly adopting the concept of disease interception. A January review published in The Lancet Rheumatology defined interception as a form of prevention, “identifying and addressing early or subclinical manifestations to limit or halt progression.”

“Perhaps the most exciting direction is disease interception,” Merola and Kavanaugh said. “Studies are beginning to examine whether patients can be identified as being at sufficiently high risk for PsA to allow intervention before clinically apparent arthritis develops…. This represents an important evolution in the field: moving from identifying factors associated with PsA to prospectively identifying high-risk patients and ultimately determining whether the development of the disease can be delayed or prevented.”

Existing Tools for Screening or Early Diagnosis

The five-question Psoriasis Epidemiology Screening Tool (PEST) and the Toronto Psoriatic Arthritis Screen are two examples of popular PsA screening questionnaires, but neither is designed to predict who may develop PsA. 

“Standard patient-reported screening questionnaires (eg, PEST) frequently trigger false positives due to osteoarthritis, fibromyalgia, or mechanical joint pain,” Jose Scher, MD, Steere Abramson Professor of Medicine, director of the NYU Langone’s Judith and Stewart Colton Center for Autoimmunity, and director of the NYU Langone Health Psoriatic Arthritis Center, at NYU Grossman School of Medicine in New York City, told Medscape Medical News.

Arguably the most promising and well-known prediction model is the Psoriatic Arthritis Risk Estimation Tool (PRESTO-PsA), a risk calculator that estimates the percentage risk of developing PsA at 1 or 5 years using data on patient demographics, psoriasis features and treatments, comorbidities, and musculoskeletal symptoms. Validation efforts are ongoing.

“PRESTO is the first prediction model that predicts the risk of developing PsA within 1 to 5 years among psoriasis patients,” PRESTO co-creator Lihi Eder, MD, PhD, associate professor of medicine, director of the psoriatic arthritis program, and co-director of the cardio-rheumatology program at the University of Toronto and Women’s College Hospital in Toronto, Ontario, Canada, told Medscape Medical News. “It was developed using a prospective cohort, which is the ideal study design for the development of such tools. It allows clinicians to identify patients with psoriasis who are at high risk of progressing to PsA within a defined timeframe.”

An even newer ultrasonographic solution to diagnose early PsA is the Diagnostic Ultrasound Enthesitis Tool (DUET), which Eder also helped to develop. First unveiled in 2025, the tool involves the ultrasound assessment of 16 entheseal sites, with images reviewed to derive a consensus score using both inflammatory lesions and structural lesions. In the first clinical study of the tool, the DUET score was significantly higher among patients with early PsA than among those in the control group or patients with only psoriasis. 

“The score defines which entheseal sites should be scanned and how sonographic abnormalities should be scored to help distinguish PsA from other rheumatic conditions,” Eder said. She noted that the tool is ready to be used by clinicians, with detailed study results expected to be released “soon.”

Building Larger Prediction Models

Some researchers are taking an all-of-the-above approach when seeking new prediction or interception tools.

“There is growing interest in combining clinical, imaging, and biomarker data through larger prediction models and potentially machine learning,” Merola and Kavanaugh said in their joint response. “Ultimately, the goal is to provide an individualized risk estimate for a patient with psoriasis.”

A study published in September in The Lancet Rheumatology sought to create prediction models for patients with new-onset psoriasis. The first set of models, designed to identify patients for rheumatologist referral, included five risk factors: arthralgia, fatigue, psoriasis phenotype, high-sensitivity C-reactive protein (hs-CRP), and psoriasis disease activity. The second set of models, designed for patients with subclinical PsA, included up to nine risk factors: recurrent uveitis, number of sites with self-reported joint pain, tender joint count, triglycerides, axial and peripheral entheseal pain, the visual analog scale for self-rated health, hs-CRP, and the HLA-B27 blood biomarker. 

Of the 628 participants who reported no previous PsA diagnosis, 83 (13%) were diagnosed with PsA by a rheumatologist. According to the researchers, the models showed “good discrimination,” with optimism-adjusted area under the curve ranging from 0.76 to 0.84. Pain, HLA-B27, and systemic inflammation were the strongest predictors for future PsA.

The authors stressed that the algorithms should be externally validated before implementation in clinical practice.

“We have developed models for dermatologist and GP [general practitioner] perspectives that inform the decision on rheumatologist referral and for risk prediction in patients who have been examined by a rheumatologist but could not be diagnosed with PsA,” study first author Axel Svedbom, PhD, researcher in the Division of Dermatology and Venereology at the Karolinska Institutet, Stockholm, Sweden, told Medscape Medical News. “These two settings differ in terms of available data, relevant alternatives, and actions that can be taken. Therefore, they require different models. Earlier models do not make this distinction or only cater to one of the decision problems.”

A major initiative working to identify and validate PsA biomarkers and risk-prediction models is the HIPPOCRATES program in Europe. The HIPPOCRATES Prospective Observational Study (HPOS) was developed to identify demographic, clinical, and biological predictors for the development of PsA through semi-structured questionnaires completed by adult participants with psoriasis but with no prior PsA diagnosis. The primary outcome is the new onset of PsA confirmed by a diagnosis from their doctor. 

According to Coates, the study’s lead investigator, HPOS has recruited more than 9000 participants to date from about 10 countries across Europe. Researchers aim to release findings next year.

Coates and her colleagues decided to conduct the prospective study because available data were not “concrete enough to base our prediction models on,” said Coates. “We’re just working on putting the data together to try and look at validating existing predictive models.”

Mining Longitudinal Electronic Medical Records Data

Further down the road, electronic health records and AI could play a greater role in prediction efforts. In 2023, a Taiwanese research team published a machine-learning model to predict the short-term risk for PsA using longitudinal electronic medical records data. The study included 443 patients who developed PsA after a diagnosis of psoriasis and 1772 with psoriasis who did not develop PsA. A neural network analyzed 2.5 years of time-series medical records data, including International Classification of Diseases, Ninth Revision, Clinical Modification codes for each visit and prescription medications obtained on each date, to predict PsA onset within the ensuing 6 months. The model achieved an area under the receiver operating characteristic curve of 0.70, with 80% sensitivity, 60% specificity, and a 93% negative predictive value.

“You probably need a tool that says, ‘Hey, this person has symptoms that could be suggestive of inflammatory arthritis, and that warrant an investigation,’” Dylan McGagh, BMBCh, academic clinical fellow with the Nuffield Department of Orthopaedics, Rheumatology, and Musculoskeletal Sciences at the University of Oxford, and first author of The Lancet Rheumatology review on PsA interception, told Medscape Medical News. “So it’s potentially enriching that initial referral pool. The model happens in the background, and then it gives a pop-up for that physician using electronic records to say, ‘Hey, this person has these factors.’”

Active and Passive Collection of Patients’ Activity, Signs, and Symptoms

Smartphones or watches also could play a role in prediction. A 2024 proof-of-concept study showed that a smartphone app known as Psorcast achieved “significant clinical validity” in remotely measuring and tracking signs and symptoms of psoriasis and PsA using smartphone sensors and image data.

“A smartphone app could periodically collect brief patient-reported information about symptoms such as pain, stiffness, and fatigue, and could also ask people to perform short movement tasks to assess aspects of upper- or lower-limb function,” McGagh said. “Wearable devices such as smartwatches could complement this by passively measuring mobility and physical activity in everyday life, potentially capturing changes in how much someone moves, patterns of activity and inactivity, and sleep.”

Can Early Biologic Treatment Stop the Transition?

The continuing advancement of biologic therapies may have implications for intercepting PsA. One study currently underway, PAMPA, is a phase 4, randomized, double-blind, placebo-controlled interventional trial across US and Canadian sites evaluating whether early intervention with the interleukin (IL)-23 inhibitor guselkumab can prevent the transition from psoriasis to PsA. The ongoing trial aims to enroll 350 participants with a preexisting psoriasis diagnosis and features associated with increased PsA risk. The trial’s investigators, including Scher, aim to randomly assign 200 patients to receive guselkumab 100 mg or placebo for 24 weeks, after which the placebo arm will switch to open-label guselkumab 100 mg, and the other arm already taking guselkumab will continue to receive it open-label. The investigators also aim to enroll an observational control group of 150 patients who receive standard-of-care treatment. 

“[The goal is to] determine if early guselkumab reduces subclinical joint and entheseal inflammation at week 24 on musculoskeletal power Doppler ultrasound (MSK-PDUS) compared to nonbiologic standard of care,” Scher said. “[A second goal is to] evaluate whether early IL-23 blockade delays or prevents full conversion to clinical PsA at 2 years.”

Taking Action in the Clinic

The actual work of prediction or interception falls most heavily on the time-strapped dermatologist. Just as some researchers are taking a combined approach to predicting PsA, the best option for practicing dermatologists today may sometimes be one that incorporates several measures.

Asking simple questions can help uncover telltale signs and symptoms without adding major burdens on each relevant appointment.

“In a busy dermatology clinic, it’s very hard to spend a lot of time,” Cheryl Rosen, MD, head of the Division of Dermatology at Toronto Western Hospital and University Health Network Hospital in Toronto, Ontario, Canada, as well as a consultant dermatologist for Weeneebayko Area Health Authority, told Medscape Medical News. “This is one of the main difficulties that dermatologists have. So simple questions about morning stiffness, aches, or pains in any joints can work. How long does morning stiffness last?”

According to Kevin Sharghi, MD, assistant professor of dermatology at Johns Hopkins School of Medicine in Baltimore, there is no substitute for a thorough physical exam.

“The best practice is to make psoriatic arthritis screening a routine part of every psoriasis visit, because symptoms can be subtle and patients may not realize they are connected to their skin disease,” Sharghi told Medscape Medical News. “I ask about the most common symptoms: joint pain, swelling, and morning stiffness, or swollen fingers or toes. On exam, I look for tender or swollen joints, dactylitis, inflammation where tendons attach to bone, and I examine the nails for changes such as pitting or lifting. A validated questionnaire can be helpful, but it does not replace a direct conversation and physical exam. If anything is concerning, early referral to a rheumatologist is essential.”

Early diagnosis, rather than prediction per se, remains the more common strategy in the clinic today. But the ongoing research continues to put new and more powerful tools in the hands of clinicians.

“The exciting next step is moving beyond earlier diagnosis toward prediction and prevention,” Merola and Kavanaugh said. “If we can more accurately identify which patients are truly at high risk and recognize the earliest stages of transition, we may eventually be able to intervene before significant musculoskeletal disease develops.”

Coates reported receiving grants/research support from AbbVie, Amgen, Janssen, and UCB; working as a paid consultant for AbbVie, Amgen, Bristol Myers Squibb, Eli Lilly, Enlivex, Janssen, Moonlake, Novartis, Oruka, Pfizer, Proximi-T, Sitryx, Takeda, and UCB; and being paid as a speaker for AbbVie, Amgen, Eli Lilly, Janssen, Novartis, Pfizer, and UCB. 

Kavanaugh reported financial support for consulting and/or clinical research from Amgen, AbbVie, Bristol Myers Squibb, Eli Lilly, Janssen, Pfizer, Takeda, UCB, and Novartis. Merola is a consultant and/or investigator for Amgen, AstraZeneca, Biogen, Boehringer Ingelheim, Bristol Myers Squibb, AbbVie, Eli Lilly, Moonlake, Novartis, Janssen, Oruka, UCB, Sanofi, Regeneron, Sun Pharma, Galderma, and Pfizer. 

Scher reported financial support from UCB, Novartis, Johnson & Johnson, Bristol Myers Squibb, and Oruka. Eder reported financial support from Bristol Myers Squibb, AbbVie, Johnson & Johnson, Eli Lilly, Amgen, Celltrion, Novartis, Pfizer, Fresenius Kabi, Novo Nordisk, and UCB.

Svedbom reported receiving grants from AbbVie, Janssen, and Eli Lilly; consultancy fees from AbbVie, Almirall SA, Bristol Myers Squibb, Eli Lilly, Icon, and Quantify Research; and speakers’ fees from AbbVie, Almirall SA, Janssen, and UCB. McGagh reported receiving a research grant from Novartis. Rosen reported financial support from AbbVie, Novartis, UCB, and Pfizer. Sharghi reported having no disclosures.

Scott Harris has covered rheumatology and dermatology for more than a decade. He lives near Washington, DC.