Two prediction models that may identify progression in early diffuse cutaneous systemic sclerosis (dcSSc) have been developed for clinical practice and randomized controlled trials, researchers reported.
The first predictive model, which included skin thickening measured by the modified Rodnan skin score (mRSS), disease duration, and the interaction between the two, had an accuracy of 60.9%, an area under the curve (AUC) of 0.666 and a positive predictive value (PPV) of 33.8%. By adding a variable for anti-RNA polymerase III (Pol3) positivity, the model reached an accuracy of 71%, AUC of 0.711, and PPV of 41%, according to Ariane L. Herrick, MD, of the University of Manchester in England, and colleagues.
“Among patients with early dcSSc, those with shorter disease duration and lower mRSS are most likely to be ‘progressors’ with a trade-off between the two factors and patients who are Pol3+ have the highest mRSS peaks and tend to reach peak mRSS earliest, providing a valuable message for clinicians that patients with short disease duration and Pol3+ must be especially closely monitored,” they wrote in the Annals of the Rheumatic Diseases.
While the model incorporating Pol3 more accurately identified high-risk patients, it may be too restrictive, they noted, adding that “both models were more flexible (for a given skin score) and more accurate than a ’22 mRSS’ cut-off model and may offer advantages for cohort enrichment in clinical trials to ensure that the most informative patients are included.”
Patients with dcSSc have high morbidity and mortality, which is associated with the degree of severity of skin fibrosis and thickening measured by the mRSS. The mRSS gauges skin thickness at 17 sites on a scale of 0 to 3, for a maximum score of 51. The skin score tends to progress rapidly in the first 3 to 5 years of dcSSc, then plateau and fall.
In the prospective European Scleroderma Observational Study (ESOS), researchers studied 326 adult patients with early dcSSc from 50 centers and 19 countries. Of the total sample, 65 patients started on methotrexate, 118 on mycophenolate mofetil, 87 on cyclophosphamide, and 56 had no immunosuppressant.
The primary outcome measure was the mRSS, which was recorded every 3 months. Because progression status did not significantly differ among treatment groups, the researchers analyzed mRSS trajectories regardless of treatment protocol. “Progressors” were defined as patients who experienced a 5-unit and 25% increase in mRSS over 12 months (±3 months).
In the study, 22.5% of early dcSSc patients progressed and 77.5% did not. Progression status could not be assessed in 33 patients.
At baseline, progressors had shorter disease duration (median 8.1 vs 12.6 months, P=0.001) and lower mRSS (median 19 vs 21 units, P=0.030) than nonprogressors.
Pol3+ patients had higher skin scores at baseline than patients in other autoantibody groups (P=0.003), despite similar disease durations (P=0.593). They seemed more likely to progress than other groups: 29.2% of Pol3+ patients were progressors, compared with 11.9% in the no-autoantibody group (P=0.105). Pol3+ patients had the highest peaks of all autoantibody groups, with a median peak of 35 units (P=0.001).
As a single predictor for progression, mRSS performed poorly with an AUC of 0.588. Duration of skin thickening performed better, with an AUC of 0.634. A model that combined mRSS, disease duration and the interaction between the two improved those univariate performances, with an AUC of 0.666. That model, known as model A, had 73.4% sensitivity, 57.2% specificity, and accurately predicted 60.9% of cases.
Adding an indicator variable for Pol3 positivity (model B) improved results. This yielded an AUC of 0.711, 60.4% sensitivity, 74.2% specificity, and accurately predicted 71% of cases.
“The key finding here was the development of a predictive model for mRSS (disease) progression which had an accuracy of 60.9%, achieved by recognizing that the initial skin score is a poor predictor of progression on its own and that prediction is improved by simultaneously accounting for disease duration,” the authors wrote. “By including autoantibodies in this analysis, the model improved and reached an accuracy of 71.0%.”
Model A can help inform randomized controlled trials, they noted, while model B can help identify patients at higher risk for progression in clinical settings. Using model B to inform patient selection into clinical trials may be too restrictive and could produce a sample that does not reflect the overall dcSSc population, they cautioned.
They added that by measuring skin scores every 3 months, they were able to capture peaks in mRSS that might not have been seen otherwise. “Had we only recorded baseline and 12-month data (two observations), 53% of our cases of progression would have been missed,” they wrote.
The study has several limitations, including the fact that, for some patients, peak mRSS may have occurred before they entered the study. Among patients whose progression status was unknown, 36.4% died, which might have induced bias.
Also, the researchers have not externally validated their models, adding that this will be an important step before they are used widely.
ESOS was funded by a grant from the European League Against Rheumatism Orphan Disease Programme, and Scleroderma, and Raynaud’s U.K.
Herrick and co-authors disclosed relationships with Actelion, Apricus, BMS, Celgene, Bayer, Boehringer Ingelheim, JB Therapeutics, Sanofi-Aventis, Novartis, Array BioPharma, Active Biotech, Galapagos, Medac, Pfizer, Anamar RuiYi, 4D Science, Biogenidec, BMS, EpiPharm, Ergonex, espeRare Foundation, Genentech/Roche, GSK, Inventiva, Lilly, Medac, Medimmune, Pharmacyclics, Serodapharm, Sinoxa ,UCB, Richter Gedeon, Inventiva, and Merck-Serono.