The validity and utility of the Cutaneous Disease Area and Severity Index (CDASI) as a clinical outcome instrument in dermatomyositis: a comprehensive review

Abstract

Introduction:

Validated outcome measures are paramount in the assessment of disease progression and evaluation of new therapeutics in clinical trials. Dermatomyositis (DM) is an autoimmune disease that is frequently refractory to current therapies and warrants the development of new treatments. The Cutaneous Disease Area and Severity Index (CDASI) was developed in 2008 in response to a need for a reliable, validated measure of skin disease activity and damage for use in clinical trials and longitudinal monitoring of disease progression.

Methods:

The literature was searched for all studies validating and utilizing the CDASI between 2008 and October 2018 using searches in PubMed. Studies pertaining to validation of the CDASI, correlation with quality of life, use in the evaluation of current therapies and ongoing trials, as well as relationships to biomarkers were included in this review.

Results:

The CDASI was found to have intra- and inter-rater reliability, validity, reproducibility, sensitivity to clinical changes, and ease of use. It has been shown to correlate with quality of life as measured by the Skindex-29 outcome measure. The CDASI activity score has additionally been shown to correlate significantly with IFN-β, a key cytokine in DM pathogenesis.

Conclusion:

The CDASI is a validated measure of dermatomyositis disease and has been shown to be an effective outcome instrument in clinical trials.

INTRODUCTION

Dermatomyositis (DM) is an autoimmune inflammatory myopathy involving the skin, muscle, lungs, heart, and/or esophagus and affecting an estimated 9.63 per 1 million persons.^1–4 Typical skin findings occur in up to 80% of patients with DM and include the heliotrope rash, Gottron’s papules, Gottron’s sign, periungual telangiectasias, and erythema/scale of the neck and back.^1,5–7 DM significantly impacts quality of life (QoL) and is often recalcitrant to standard-of-care therapies.^8,9 Despite this, skin disease has historically been overlooked in DM clinical trials that primarily focus on myositis manifestations.⁹ In 2008, however, the Cutaneous Disease Area and Severity Index (CDASI) was developed by dermatologists with expertise in DM to measure skin disease activity and damage. It has subsequently been modified and validated in juvenile and adult patient clinical and translational studies in the fields of dermatology, rheumatology, and neurology.^10–12It has allowed clinical researchers to assess the efficacy of conventional and new therapeutics, study the course of DM skin disease and its effect on QoL, as well as better understand disease pathogenesis through correlative studies of important biomarkers and the CDASI.

METHODS

We conducted a thorough literature search using PubMed for studies validating and utilizing the CDASI between 2008, when it was first developed, and October 2018. Articles in this review pertain to development and validation, use in therapeutic evaluation, monitoring of disease progression, quality of life, and relationship with biomarkers.

DEVELOPMENT AND VALIDATION OF THE CDASI

Due to the disease’s impact on quality of life and need for new therapeutics, it became evident that a validated outcome instrument for DM was needed. Validated outcome instruments are paramount in clinical trial design and evaluation of therapeutic outcome. In an article discussing the framework and development of clinical outcome measures, Singer et al. argues that the ideal clinical outcome measure is credible, accurate, sensible, comprehensive, sensitive to change across time, feasible, easy to use, and measures clinically relevant outcomes.¹¹

Prior to the origination of the CDASI, a number of cutaneous outcome measures forged the path to its development.¹² The Dermatomyositis Skin Severity Index (DSSI) was developed in 2007 as a skin disease severity measure based on the Psoriasis Area Severity Index (PASI).¹² Scoring is determined by the percentage of body surface area involved on the head, trunk, upper extremity, and lower extremity with a score range on a scale of 0 to 72.¹² The structural framework of the PASI marks a disadvantage of the DSSI as it tends to evaluate lesions similar to each other as found in psoriasis, while the skin lesions in DM are comparatively more diverse. Further, using body surface area as a measure of skin severity represents a challenge in the presence of focal skin lesions and can easily lead to overestimation.¹² In addition, small areas can have major impacts on QoL and require a different approach than simply measuring surface area.

The Cutaneous Assessment Tool (CAT) was also developed in 2007; intended to be a comprehensive tool to capture cutaneous disease activity and damage associated with adult and juvenile IIM, the CAT is an eight-page tool comprised of 21 items including 10 activity, 4 damage, and 7 combined lesions with activity and damage subscore ranges of 0 to 175 and 0 to 33, respectively.¹³ The tool takes 15 minutes to complete and has good activity score inter-rater reliability and moderate damage score_inter-rater reliability.¹³ It was later modified into the CAT-Binary Method (CAT-BM), an abbreviated one-page tool that uses an alternative scoring system of the CAT.^14,15 The activity subscore ranges from 0 to 17 and the damage subscore ranges from 0 to 11 for a maximum score of 28.¹⁴ The activity score is determined by the presence of erythema over 7 anatomical locations along with other lesions characteristic to DM.¹⁴ However, the CAT-BM may not capture changes in disease such as scale or erosion.¹⁴ In addition, partial improvement cannot be measured because the lesion must be completely resolved to capture improvement.

In 2008, dermatologists with an expertise in DM developed the Cutaneous Disease Activity Severity Index (CDASI) in response to a need for a more reliable, validated measure of skin disease severity for use in clinical trials and longitudinal monitoring of disease progression.¹⁰ It is a one-page validated, quantitative outcome measure that grades skin disease activity by the degree of erythema, scale, and presence of erosions or ulceration, and damage by the presence of poikiloderma and/or calcinosis in DM patients.¹⁰ The features selected during the initial development phase included interviews with DM patients. During the CDASI’s preliminary validation, the instrument was compared to the DSSI and the CAT; its validity was assessed by comparing its scoring to visual analogue scales (VAS) via the Global Physician Score (GPhS), Global Patient Score (GPaS), and Global Itch Score (GIS).^{10,12,13,16,17} Klein et al.’s partial validation of the CDASI demonstrated its clinical applicability, as the CDASI correlated more strongly with the global measurements compared to the DSSI and CAT.¹⁰ At that time, the CDASI had four activity measures (erythema, scale, excoriation, and ulceration), and two damage measures (poikiloderma and calcinosis) over 15 anatomical locations.¹⁸ The CDASI tool additionally measures Gottron’s papules, periungual changes, and alopecia as part of activity of DM.¹⁸ A higher score indicates more severe disease.¹²

The original CDASI was refined in 2010 from having four activity and two damage measures to three activity and two damage measures; the activity subscore ranges from 0 to 100 and the damage subscore ranges from 0 to 32.¹⁸ Specific changes included removal of the activity measure ‘excoriation’ as it indirectly measures itch; the word ‘purple’ from the scoring of ‘erythema’ and ‘Gottron’s-Hands’ activity measures as it does not accurately describe dark red; and the addition of ‘erosion’ to the ‘ulceration’ activity measure.¹⁸ A modified scoring system of ‘periungual’ and the ‘Gottron’s-Hand’ damage measure was also established.¹⁸ The modified CDASI nearly perfectly correlates with the original CDASI and has equally good concurrent validity with the physician global measures (PGM)-overall skin (gold standard) and PGM-activity scales.¹⁸ Similar to the original CDASI, the improved CDASI damage subscore poorly correlated with the PGM-overall skin (r_sp=0.15); both the original and modified versions of the CDASI activity and damage subscores correlated with the PGM-activity (r_sp=0.76, r_sp=0.75) and PGM-damage (r_sp=0.90, r_sp=0.90) scales, respectively.¹⁸

In 2012, Goreshi et al. investigated the clinical responsiveness, reliability, and validity of the CDASI and the CAT-BM (Binary Method) in comparison to the Physician Global Assessment (PGA) “gold-standard”.¹⁴ Both the CDASI and CAT-BM demonstrated good construct validity, as they accurately correlated with PGA scales using the VAS and the Likert scales (p<0.0001 for total, activity, and damage scores).¹⁴ The CDASI demonstrated the greatest responsiveness (Standard Response Mean: CDASI 1.25; CAT-BM 0.93; PGA activity 1.03; PGA activity Likert 0.61) and interrater reliability (activity: CDASI 0.75, CAT-BM 0.52, PGA activity 0.72, PGA activity Likert 0.65; damage: CDASI 0.56, CAT-BM 0.34, PGA damage 0.51, PGA damage Likert 0.54; total: CDASI 0.73, CAT-BM 0.43, PGA overall 0.63, PGA overall Likert 0.70) compared to the other instruments. The PGA VAS showed higher intra-rater reliability (activity: CDASI 0.87, CAT-BM 0.71, PGA activity 0.91, PGA activity Likert 0.74; damage: CDASI 0.80, CAT-BM 0.79, PGA damage 0.81, PGA damage Likert 0.71; total: CDASI 0.90, CAT-BM 0.80, PGA overall 0.89, PGA overall Likert 0.88).¹⁴ The CDASI’s intra-rater reliability in activity and total scores was almost perfect; damage intra-rater reliability was good.¹⁴ Additionally, the CDASI was superior to the CAT-BM in capturing extremes of the cutaneous disease spectrum and in treatment responsiveness.¹⁴ The data supported the CDASI as a better clinical tool to evaluate cutaneous severity in the adult DM population.¹⁴

In addition to reflecting clinical change, the CDASI stratifies disease activity by score, serving as an effective assessment tool in DM trials. In an effort to classify cutaneous disease activity by scoring, the University of Pennsylvania (UPenn) and Stanford University evaluated the CDASI and the Physician Global Assessment Visual Analogue Scale (PGA VAS) scores.¹⁹ UPenn reported a CDASI activity score of 19 or less characterized mild disease (sensitivity, 88%; specificity, 85%), whereas Stanford reported a cutoff score of 14 (sensitivity, 78%; specificity, 82%).¹⁹ Activity scores above these cut-off values indicated moderate to severe disease.¹⁹ Using an improvement of 2 cm in VAS scores to distinguish between responding and nonresponding patients, the study determined that a change of 4 points in CDASI score at UPenn or 5 points at Stanford reflect minimal clinically significant change.¹⁹ Although the average CDASI scores were similar between UPenn and Stanford, the average VAS scores between the sites were not, indicating VAS scores are not as reliable. ¹⁹

Due to the systemic nature of DM, the CDASI has been evaluated by other specialists as well. In a recent study assessing intra- and inter-rater reliability of the CDASI, dermatologists, rheumatologists and neurologists individually scored adult DM patient disease using both the CDASI and the PGA.²⁰ Using the intra-class correlation coefficient (ICC) to estimate inter-rater reliability, results showed good inter-rater reliability for CDASI activity scores among dermatologists and rheumatologists (ICC: 0.73 and 0.74, respectively), while inter-rater reliability was moderate for neurologists (ICC: 0.56).²⁰ Inter-rater reliability for CDASI damage scores was excellent across the three specialties (ICC: dermatologists 0.86, rheumatologists 0.82, neurologists 0.84). All physicians showed excellent intra-rater reliability in the activity score (ICC: dermatologists 0.94, rheumatologists 0.88, neurologists 0.82); however, with respect to damage score, neurologists showed moderate intrarater reliability (ICC: 0.66) while dermatologists and rheumatologists showed excellent intra-rater reliability (ICC: 0.92 and 0.91, respectively). Reliability of the PGA was not as strong or consistent, and thus the CDASI was found to be superior in its inter- and intra-rater reliability. This study confirmed the CDASI’s reliability and validity across two other specialties treating adult DM.²⁰

Because the CDASI is only validated in adult DM, Tiao et al. investigated its use by pediatric dermatologists, rheumatologists, and neurologists through comparison with the CAT-BM, the only validated tool for assessing cutaneous activity in juvenile DM (JDM).²¹ While pediatric dermatologists and rheumatologists demonstrated good to moderate inter-rater reliability for CDASI activity (ICC dermatologists 0.52, rheumatologists 0.81) and damage scores (ICC: dermatologists 0.59, rheumatologists 0.59), pediatric neurologists exhibited poor inter-rater reliability (ICC: activity 0.47; damage 0.46).²¹ There was no difference in both mean CDASI activity and damage scores between pediatric dermatologists and rheumatologists but there was a significant difference between pediatric dermatologists and neurologists. Further, pediatric dermatologists and neurologists differed in mean CDASI damage scores, while pediatric dermatologists and rheumatologists reported similar mean CDASI damage scores.²¹ Both instruments demonstrated good construct validity relative to the PGA.²¹ The study concluded that the CDASI activity and damage scores have moderate to excellent reliability in JDM patients among pediatric rheumatologists or dermatologists, but not neurologists. In contrast, the CAT-BM activity scores showed moderate reliability in dermatologists and rheumatologists while damage scores showed poor reliability across the three specialties.²¹ This highlights the reliability of the CDASI in assessing skin disease activity in JDM by rheumatologists and dermatologists.

CORRELATION OF DISEASE ACTIVITY AND QUALITY OF LIFE

The majority of DM patients suffer from cutaneous manifestations, which are known to significantly impair QoL.²² In one of the first studies evaluating QoL in DM, the CDASI significantly correlated with the Dermatology Life Quality Index (DLQI) (r=0.27, p=0.01) and the Skindex-29 subscales of emotions (r=0.46, p=0.002), functioning (r=0.44, p=0.004), and symptoms (r=0.33, p=0.04), QoL instruments specific to skin disease. Results demonstrated more severe skin disease, as measured by the CDASI, is associated with worse QoL. A later study showed that as the CDASI decreases with treatment or improvement of cutaneous disease, QoL improves in DM patients.²² In this prospective study using a 4 point or greater improvement in CDASI activity score as a marker for clinically significant improvement, the change in CDASI activity score correlated with the change in Skindex-29 subscales of emotions (r_sp=0.60, p<0.01), functioning (r_sp=0.60, p<0.01), and symptoms (r_sp=0.67, p<0.001), with responders showing improvement in their Skindex-29 scores relative to nonresponders.²² However, the lowest CDASI score at which the Skindex-29 QoL measure correlated with skin activity was down to a score of 10 for emotions, 8 for functioning, and 7 for symptoms.²³ The CDASI was found to correlate with the DLQI down to a score of 4²³. QoL with regards to these subscales does not improve with cutaneous disease activity below the reported CDASI cut-off values, suggesting that total skin clearance in the treatment of DM is not a meaningful endpoint in the design of clinical trials.²³

Most recently, our group demonstrated a positive correlation between percent change in CDASI scores between visits and a meaningful change in QoL as measured by the Skindex-29 score.²⁴ After correlating the Skindex-29 to an established definition of meaningful change in the DLQI, we found meaningful improvement is 7.86 (p<0.0001) points in the symptoms subscale_and 10.29 (p<0.0001) points in the emotions subscale.²⁴ Our retrospective study showed that for patients with initial CDASI-A scores >14, an approximately 40% change in CDASI-A between the first two visits suggests a meaningful change in the Skindex-29.²⁴ This 40% change is a conservative recommendation, as a 32% (p<0.001) change in the CDASI score relative to the symptoms subscale and a 35% (p<0.001) change in activity score relative to the emotions subscale in DM patients with moderate to severe disease activity was associated with a meaningful change in QoL.²⁴

MONITORING DISEASE PROGRESSION USING THE CDASI

The CDASI was developed in part to longitudinally monitor disease progression. Chansky et al. conducted a retrospective cohort study of 40 DM patients using the CDASI to characterize disease course and disease progression.²⁵ The study classified disease progression as improved, worsened, or stable by calculating an average CDASI improvement rate since baseline visit for an average change score (ACS) in disease activity over time, and classified disease courses as monophasic, polyphasic, or chronic. The authors defined a monophasic course as a cumulative consecutive decrease in CDASI score of at least 4 points without a clinically significant worsening/flare. A course with both a significant improvement and worsening/flare of skin disease was characterized as polyphasic, while a course with a significant worsening/flare without a significant improvement or change in CDASI compared to baseline was deemed as chronic. Using a generated and independently validated cutoff value, disease progression was classified as improved if the ACS decreased by ≥3, and worsened if it increased by >3. Most patients started with moderate-to-severe disease (n=24 [60%]) by baseline CDASI score and improved in disease activity (mean ACS of −8.57) with a polyphasic course (n=19 [79%]).²⁵ The majority of patients with moderate-to-severe disease improved (n=19 [79%]) whereas patients with mild disease tended to remain stable (n=10 [62.5%]) in their disease progression (mean ACS of +0.95).²⁵ In addition to characterizing disease progression, this study highlighted the concept of the “floor effect,” meaning it is less likely to see significant improvement in disease with low baseline CDASI scores compared to higher baseline scores.²⁵ Therefore, in clinical trial design, it is beneficial to include patients with moderate-to-severe disease activity as opposed to mild disease in order to see the effect of a new therapeutic agent.

Another study by Wolstencroft et al. similarly aimed to characterize cutaneous disease course with the primary outcome measure being clinical remission of skin disease as measured by the CDASI at a 3-year follow-up.²⁶ The study presented data from an ongoing, prospective cohort study of patients with adult DM in which the CDASI was used to monitor cutaneous disease progression, and included patients with a baseline CDASI activity score of 12 or greater and 2 or more CDASI scores separated by at least 3 months within the first 3 years of follow-up.²⁶ Clinical remission was defined as achieving a CDASI score of 5 or less.²⁶ Of the 74 patients who met inclusion criteria, only 38% achieved clinical remission.²⁶ Patients with anti-melanoma differentiation-associated protein 5 (MDA5) antibodies were even less likely to achieve clinical remission.²⁶ Utilizing the CDASI further permitted multivariate analysis to identify increased age, dermatomyositis-associated malignancy, and treatment with mycophenolate mofetil as significantly associated with clinical remission.²⁶ In contrast, disease duration at baseline, disease duration prior to systemic therapy, and baseline cutaneous disease activity were not significantly associated with clinical remission.²⁶ Using the CDASI has allowed us to better understand disease progression and factors related to prognosis.

CDASI FOR EVALUATION OF THERAPEUTICS

The CDASI has been useful in evaluating the efficacy of both new and conventional therapeutics. Treatment of cutaneous DM generally follows a stepwise algorithm, despite the literature lacking strong evidence for the use of many of these agents.²⁷ In general, antimalarials are often first line therapy, but if adequate control is not achieved, immunosuppressive agents including methotrexate, mycophenolate mofetil, or azathioprine, are added.²⁷ If both antimalarials and immunosuppressive drugs are ineffective, IVIG can be added and calcineurin inhibitors and/or corticosteroids may be considered if patients remain refractory or have severe disease.²⁷

With the development of the CDASI, several studies have evaluated the efficacy of these conventional treatment options. Anyanwu et al. applied the CDASI to prospectively assess disease activity and outcomes in 41 adult patients with skin-only DM between July 2009 and April 2013.²⁷ A standard algorithm was applied to treatment of the patients. They found that the median final CDASI score for patients with skin-only disease was 13.5 after a median duration of treatment for 24 months, indicating that patients had at least mild disease activity (as measured by the CDASI) at their final visit despite treatment.²⁷_The authors concluded that antimalarials are frequently insufficient in the management of skin-predominant DM and therefore often require second-line agents.²⁷ Another study utilized the CDASI to assess the efficacy of low-dose methotrexate in DM skin lesions.²⁸ In a series of 11 patients, methotrexate significantly reduced DM skin lesions in 8 patients, with the CDASI scores decreasing more than 50% within 2–3 months, supporting the efficacy of the drug.²⁸ More recently, the CDASI was used to assess the utility of the Janus Kinase (JAK)-1/3 inhibitor, tofacitinib, currently approved for the treatment of rheumatoid arthritis but not for cutaneous DM.²⁹ In this study, three patients with severely refractory cutaneous DM were treated with tofacitinib; all three demonstrated clinically significant improvement in disease activity as measured by the CDASI.²⁹ These studies illustrate the value of the CDASI in evaluating the effectiveness of existing treatments.

Additionally, because cutaneous DM is frequently recalcitrant to treatment, newer therapies are warranted and their development must be evaluated by validated outcome measures. A recent 16-week, double-blind, randomized placebo-controlled Phase 2 study investigated the drug Lenabasum in the treatment of refractory, skin predominant DM.³⁰ Lenabasum is a synthetic reverse agonist of the CB2 cannabinoid receptor that decreases cytokine production by skin and peripheral blood mononuclear cells and promotes resolution of the innate immune response.³⁰ The trial included 22 adults with a CDASI activity score >14 with minimal active muscle involvement who had an inadequate response or were unable to tolerate hydroxychloroquine and immunosuppressants.³¹ The CDASI activity score was the primary outcome to assess drug efficacy at baseline and at various time points.³⁰ These studies demonstrate the usefulness of the CDASI outcome measure to assess changes in disease severity in skin-predominant DM clinical trials.

CORRELATION BETWEEN CDASI ACTIVITY SCORE AND BIOMARKERS

In addition to assessing the clinical efficacy of various treatments, the CDASI correlates with DM biomarkers. Previous literature suggested that biomarkers of the type 1 interferon (IFN) pathway are upregulated in the skin of DM patients.³¹ Huard et al. sought to investigate the relationship between biomarkers in this pathway and disease activity as measured by the CDASI in DM.³² In a prospective study enrolling 42 patients, CDASI activity scores were found to significantly correlate with a type 1 IFN composite 10-gene transcript signature (r=0.60, p<0.0001), as well as CXCL-10, a type 1 IFN-inducible protein (r=0.63, p<0.0001), and IFN-β (r=0.54, p=.0003).³² Patients with mild disease activity (CDASI <12) at baseline had low-normal IFN gene signature scores (<3-fold) compared to healthy subjects.³² In contrast, most patients with CDASI activity scores greater than 12 had high gene signatures (>3-fold).³² Moderate-to-severe disease above this cut-off correlated with IFN-β levels (p<0.001).³² The increased IFN gene signature in the blood of DM patients served as a predictive biomarker for moderate-to-severe disease (CDASI >12).³² Results from a more recent study examining the serum of 40 DM patients in Shanghai, China by Chen et al. corroborated the association between disease activity and biomarkers, showing that CDASI activity score significantly correlated with IFN-β (r=0.37, p=0.02) and CXCL-10 (r=0.32, p=.045).³³ Such translational studies allow us to better understand disease pathogenesis by correlating disease activity with various biomarkers.

CONCLUSION

Cutaneous dermatomyositis is an autoimmune disease that is frequently refractory to current therapies and thus warrants the development of new treatments. As such, validated outcome measures must be used to assess the efficacy of potential therapies. The CDASI outcome measure has been validated in six distinct studies in both adult and pediatric populations at two primary institutions, the University of Pennsylvania and Stanford University. It has demonstrated intra- and inter-rater reliability, validity, reproducibility, sensitivity to clinical changes, and ease of use. The CDASI has additionally been shown to correlate with quality of life as measured by the Skindex-29, further highlighting the utility of the CDASI as an outcome measure. Finally, the CDASI activity score has been shown to correlate significantly with IFN-β, a cytokine key in DM pathogenesis. In sum, the CDASI is a validated outcome measure that is useful in evaluating therapies and better understanding disease pathogenesis. There is strong evidence supporting the utility and efficacy of the CDASI as an outcome instrument in the evaluation of patients with dermatomyositis in both the clinical and research settings. Increased awareness and use of this tool to evaluate disease should facilitate the development of future clinical trials investigating conventional and novel therapeutics for dermatomyositis.

ABBREVIATIONS

ACR

American College of Rheumatology

BM

Binary Method

CADM

clinically amyopathic DM; CAT, Cutaneous Assessment Tool

CDASI

Cutaneous Dermatomyositis Disease Area and Severity Index

DLQI

Dermatology Life Quality Index

DM

dermatomyositis

DSSI

Dermatomyositis Skin Severity Index

EULAR

European League Against Rheumatism

GIS

Global Itch Score

GPaS

Global Patient Score

GPhS

Global Physician Score

ICC

intra-class correlation coefficient

IFN

interferon

IL

interleukin

IIM

idiopathic inflammatory myopathy

JAK

Janus kinase

JDM

juvenile dermatomyositis

PGA

physician global assessment

PGM

physician global measures

PM

polymyositis

VAS

visual analogue scales