Methods of Assessing Nailfold Capillaroscopy Compared to Video Capillaroscopy in Patients with Systemic Sclerosis—A Critical Review of the Literature

Abstract Introduction: Nailfolds of patients with systemic sclerosis (SSc) provide an opportunity to directly visualize microvascular remodeling in SSc. Nailfold video capillaroscopy (NVC) remains the gold standard for assessing nailfold capillaroscopy (NFC). However, access to NVC is limited by expense and expertise. This review aims to synthesize current research on other NFC devices compared to NVC. Methods: The literature search included the primary research of adult patients with SSc as defined by the 2013 ACR/EULAR criteria. Methods of assessing NFC included stereomicroscopy/wide-field microscopy, ophthalmoscopy, dermatoscopy, smartphone devices, and digital USB microscopy. Primary outcomes included both qualitative (normal vs. abnormal nailfolds, overall pattern recognition, presence/absence of giant capillaries, hemorrhages, and abnormal morphology) and quantitative (capillary density and dimension) measures. Results: The search yielded 471 studies, of which 9 were included. Five studies compared NVC to dermatoscopy, two compared it to widefield/stereomicroscopy, one to smartphone attachments, and one to USB microscopy. In dermatoscopy studies, NVC had a higher percentage of images that were interpretable (63–77% vs. 100%), classifiable (70% vs. 84%), or gradable (70% vs. 79.3%) across three studies. Dermatoscopy had a lower sensitivity (60.2% vs. 81.6%) and higher specificity (92.5% vs. 84.6%) compared to NVC. One stereomicroscopy study found a significant difference between methods in capillary density in limited cutaneous SSc, while another found correlations in all parameters between stereomicroscopy and NVC. One smartphone lens had good agreement with NVC on abnormal capillary morphology and density. USB microscopy was able to differentiate between SSc and healthy controls using mean capillary width but not by capillary density. Discussion: A dermatoscope may serve as a more portable and affordable screening tool to identify a normal “scleroderma pattern”, and images that need further corroboration by NVC. NFC parameters reported are heterogenous and the standardization of these parameters is important, especially in non-gold-standard devices.


Introduction Key Messages
Systemic sclerosis (SSc) is a poorly understood autoimmune connective-tissue disease characterized by fibrosis of the skin and/or internal organs, immune dysregulation, and microvasculopathy. The nailfolds of patients with SSc provide an opportunity to directly visualize microvascular remodeling in SSc. This was first visualized by Maricq and colleagues using widefield microscopy in the 1970s [1]. Since then, abnormal nailfold capillaries has been proposed as a method to distinguish primary Raynaud's phenomenon (RP) from RP due to SSc [2].
Longitudinally, in patients with RP who meet the classification criteria for "early" SSc with no other evidence of connective tissue disease, abnormalities in images via nailfold capillaroscopy (NFC) together with autoantibodies were found to be highly predictive of developing definite disease [2,3]. Furthermore, a normal pattern on NFC combined with the absence of SSc-specific auto-antibodies rules out the presence of SSc [4]. Efforts have been combined internationally to standardize NFC [5]. Patients with SSc typically exhibit a "scleroderma-spectrum pattern" characterized by a loss of capillaries, enlarged capillary loops, capillary hemorrhage, and the disruption of capillary appearance [6]. As a result of NFC's diagnostic value, it has been included in the 2013 American College of Rheumatology (ACR)/European League Against Rheumatism (EULAR) classification criteria for SSc [7].
In addition to its diagnostic value, emerging evidence suggests that NFC parameters can be used in prognostication and disease monitoring in SSc. These parameters include the presence of giant capillaries, microhemorrhages, decrease in capillary density, and presence of avascular areas or disorganized angiogenesis [8] (as illustrated in Figure 1). Microhemorrhages and giant capillaries (>50 mm in apical diameter) have been associated with the active phase of disease, with a stronger correlation seen with microhemorrhages than with giant capillaries [9]. An increase in capillary apical width and decrease in capillary density were also found in patients with SSc-pulmonary arterial hypertension (PAH) [10]-although this may not be only present in PAH associated with SSc. A decreased capillary density on NFC represents the proliferative microvasculopathy that is well described in patients with PAH. Intriguingly, capillary density is inversely associated with presence of SSc-interstitial lung disease (ILD), the longitudinal progression of SSc-ILD, and the presence and severity of SSc-PAH by right-heart catheterization [11,12]. Additionally, some have found capillary loss in 'late' SSC to be associated with the increased risk of digital skin ulcers [13,14]. This may stem from decreased capillary loss and a subsequent chronic hypoxic state systemically-particularly in acral tissues such as the digits [8]. Further to that, "late" NFC changes representing disorganized angiogenesis as a consequence of a chronic hypoxic state may also be detectable [8]. This disorganized form of angiogenesis prospectively correlates with the development of SSc-PAH [12]. Numerous methods have been described to assess nailfold capillaries including the stereomicroscope, dermatoscope, ophthalmoscope, smartphone devices, digital USB microscopes, and the current gold-standard device, the nailfold videocapillaroscope (NVC) ( Figure 2, Table 1). Nailfold videocapillaroscopy combines a microscope with a digital video camera with the magnification ranging between 50× and 500×. The nailfold images are then stored and analyzed on specific parameters in each single frame separately and adjacent images can be combined to visualize the entire nailfold by software [15]. In addition, a capillary score that is comprised of features associated with SSc-related disease activity (e.g., giant/enlarged capillaries and microhemorrhages) and vascular remodeling (e.g., capillary disorganization, loss, and dropout) is generated after the examination [8]. Hence, NVC has become the accepted as the "gold-standard". However, access to NVC is mostly restricted to centers with special interests due to expense and required expertise when assessing NFC visually. In fact, according to a survey of SSc specialists in mostly academic hospitals in the USA found that while 91% of experts assess NFC "always" or "most of the time", 64% use a dermatoscope or ophthalmoscope for evaluation, and only 7% use NVC [16]. Despite growing interest in the utility of NFC in SSc, there are no systematic reviews that compare the different devices against the gold standard, NVC. This systematic review synthesizes evidence comparing different NFC devices to NVC in assessing NFC in patients with SSc.

Materials and Methods
The review included primary research from observational studies with a comparison group. Studies on adults (age 18 or older) with SSc were included. Systemic sclerosis was defined by the 2013 ACR/EULAR classification criteria. Methods of assessing NFC included but were not limited to stereomicroscopy/wide-field microscopy, ophthalmoscopy, dermatoscopy, smartphone devices, and digital USB microscopy [5]. Nailfold video capillaroscopy was defined as a microscope with a digital camera with a magnification that ranges between 50× to 500×. Studies that included NVC and another form of capillaroscopy assessment were included. The primary outcomes of interest included qualitative (normal vs. abnormal nailfolds, overall pattern recognition (scleroderma pattern), or quantitative (capillary density, capillaries with abnormal morphology, capillary dimension, and presence/absence of hemorrhage) outcomes captured by different devices. There were no limitations on publication date or publication type. The search was limited to the English language.
The literature search included articles that were published until 23 May 2022 in MEDLINE, EMBASE, and Web of Science. A combination of the following keywords was used including: capillaroscopy, USB, smartphone, stereomicroscopy, dermatoscopy, nailfold video capillaroscopy, ophthalmoscopy, and SSc. The search terms were developed based on equipment names used in a previously published review by Smith et al. [5]. The search strategy is included in Supplementary Materials. Title and abstract screening (level 1), full text (level 2) screening, and data abstraction was completed by one reviewer (ZM) then confirmed by a second reviewer (MSO). Abstracted data included authors, year of publication, country of publication, sample size, eligibility criteria, NFC assessment device, NVC device, assessors, outcomes, and main results. Covidence was used to eliminate duplicates and screen records. Given the heterogeneity of outcomes reported and the results of included studies, the results were synthesized narratively.

Results
The search resulted in 138, 330, and 3 studies from MEDLINE, EMBASE, and Web of Science, respectively. After duplicates were removed, 457 titles and abstracts were screened. Fourteen abstracts underwent full-text screening. Nine studies met the final inclusion criteria. Five studies compared NVC to dermatoscopy, two compared it with widefield/stereomicroscopy, one with smartphone attachments, and one with USB microscopy.

Videocapillaroscopy vs. Dermatoscopy
Five of nine included studies compared dermatoscopy to NVC (Table 2) [17][18][19][20][21]. The sample size across studies ranged from 20 to 170 participants. Two included studies were from the United States, two were from the United Kingdom, and one from Turkey. The dermatoscopes used, videocapillaroscopes used, and outcomes measured were heterogenous across studies. Radic and colleagues in the United States compared two dermatoscopes including a non-contact polarised HEINE Delta 20 T dermatoscope connected to a digital camera with 10-16× magnification (D1), as well as DermLite DL3 dermatoscope with 10× lens with a DermLite connection kit for iPad mini 4 (D2) in a group of 100 patients with RP. This group developed an algorithm based on the presence or absence of scleroderma patterns (reduced capillary density, capillary enlargement, hemorrhages, and capillary morphology), and categorized images to normal, definite scleroderma, non-specific, and non-interpretable groups. According to the algorithm, non-specific and non-interpretable groups underwent further evaluation with NVC. In this group, the images considered normal or definite scleroderma by dermatoscope were consistent with NVC classifications. There was a higher percentage of non-interpretable images with dermatoscopes (37% using D1, 23% using D2), compared to non-interpretable images with NVC (0%). From the developed algorithm, 50% of dermatoscopic images in this group were non-specific or non-interpretable, which required further imaging using NVC. All scleroderma patterns by NVC were found to be either a non-specific or definite scleroderma pattern by dermatoscope [17]. Hughes et al. included 32 patients (8 controls, 3 primary RP, and 21 patients with SSc spectrum disorders) [18]. Their group used a semi-quantitative scale based on the severity of abnormalities ranging from 0 (normal) to 3 (grossly abnormal). Similarly in this group, there was a significant difference in classibility between NVC vs. dermatoscope (84% vs. 70%, p < 0.001). Additionally, severity scores were significantly higher in NVC (1.69, 95% CI 1.44-2.94) than in dermatoscopy 1.26 (1.04-1.49). The correlation between two techniques on classibility was 0.4 (95% CI 0.3-0.49) by the same rater, and for severity was 0.65 (95% CI 0.55-0.73) [18]. Dinsdale et al. had the largest sample size of 170 patients (99 SSc patients, and 71 controls) who underwent dermatoscopy from 10 expert observers from seven centers. Nailfolds were assessed as either ungradeable (due to the extreme severity of the abnormality or poor image quality) or gradeable. Gradeable nailfolds then were classified as normal, non-specific, early, active, or late SSc. Gradeability was 70.9% using dermatoscopy and 79.3% using NVC. With dermatoscopy, sensitivity was 60.2% and specificity was 92.5% whereas sensitivity was 81.6% and specificity was 84.6% with NVC to identify disease status of the study participant (non-SSc or SSc) [19]. Dogan et al. compared the DermLite ProGen with a digital camera and found moderate agreement (k = 0.52) between the dermatoscope and NVC in classifying patients with SSc (n = 39) into early-phase, late-phase, and active-phase disease groups [20]. In an abstract presented at ACR, Stever and colleagues assessed SSc patients using a DermLite dermatoscope compared to NVC (Optilia) with two trainees and two experts [21]. Of those 20 patients, the dermatoscope was able to recognize abnormal capillaroscopy in 13 of 20 patients (65%) [21]. Hence, in contrast to NVC, NFC interobserver variability may be a prominent feature.    (3), and active phase (4) Cohen's kappa agreement between groups is k = 0.52.

Stereomicroscopy vs. Nailfold Videocapillaroscopy
Two included studies compared stereomicroscopy to NVC [22,23] (Table 3). Widlt and colleagues measured capillary density (loops/mm) using two stereomicroscopy methods (direct counting using stereozoom microscope and stereozoom equipped with DeltaPix camera) with NVC in 62 patients (41 Ssc and 21 controls). The median capillary density was found to be 4.3, 5.4, and 6.1 loops/mm in limited cutaneous SSc (lcSSC) and 4.5, 5.0, and 6.3 in diffuse cutaneous SSc (dcSSC) when measured by direct counting, stereozoom with digital camera, and with NVC, respectively. Within patients with lcSSC, the capillary density was significantly different between the three methods [22]. In Brazil, a study with a larger group of 252 patients (101 with SSc, 61 with RP associated with undifferentiated connective tissue disease, 37 primary RP, and 52 healthy controls) found that widefield stereomicroscopy under 10-25× magnification correlated with NVC in all parameters including number of capillaries (r = 0.874), enlarged capillaries (r = 0.902), giant capillaries (r = 0.882), microhemorrhages (r = 0.601), and avascular score (r = 0.814) [23]. The areaunder-ROC-curve analysis showed widefield NFC and the videocapillaroscopy showed similar performances in discriminating between SSc patients and healthy controls [23].

Smartphone Attachments
One abstract published by an Australian group compared three iPhone microscopic lenses (Olloclip ×30 magnification, Nightstar ×60 magnification, and GoMicro ×60 magnification) compared to widefield binocular nailfold capillaroscopy (×20-80 magnification) and Capiscope NVC (×100-300 magnification) in 10 patients with SSc (Table 3) [24]. This found that the Nightstar lens was superior to other smartphone lenses with good agreement on the assessment of abnormal capillary morphology and density compared to widefield binocular NFC and NVC [24].

USB Microscopy
One pilot study recruited 20 patients with lcSSc and 20 healthy controls (Table 2). Nailfold capillaroscopy was completed using XCSOURCE TE389, a USB microscope [25]. This group was able to obtain high-quality NFC images using USB capillaroscopy but the morphology in some frames was obscured. The parameters assessed included mean capillary width and capillary density. USB microscopy was able to differentiate between patients with SSc and healthy controls using mean capillary width, with an area under the ROC of 0.81 (SE 0.120) with USB microscopy compared to that of NVC (0.81, SE 0.095). USB microscopy was unable to discriminate between SSc and healthy controls based on capillary density [25].

Discussion
This review has summarized existing evidence comparing methods of assessing NFC to the current gold standard, NVC. Most of the current evidence comparing the utility of other NFC devices to NVC focused on dermatoscopy. Of these studies, evidence is consistent that a higher proportion of images acquired by dermatoscopes were noninterpretable or non-classifiable compared to those acquired by NVC [17][18][19]. Compared to NVC, the dermatoscope was less sensitive but more specific in detecting abnormalities in classifiable images [18]. In terms of scleroderma pattern subclassifications (early, active, and late), there is moderate agreement between dermatoscopy and NVC found in a small group of patients. This may be explained by the lower magnification of dermatoscopes and the fact that they are unable to evaluate capillaries in sufficient detail compared to NVC. When comparing a stereomicroscope with NVC, one study found a significant difference in capillary density between stereomicroscopy and NVC [22]. However, a study using a larger sample found that widefield stereomicroscopy strongly correlated with all parameters related to scleroderma pattern including capillary density, enlarged capillaries, giant capillaries, and avascular areas with excellent intra-inter-observer reliability for both techniques with experienced observers [23]. The search yielded only one study comparing other techniques such as smartphone attachments and USB microscopy [24]. The limited data show that smartphones microscopy had good agreement on abnormal capillary morphology and density compared to NVC. USB microscopy was able to differentiate patients with SSc from the control group using capillary width measurements, but not based on capillary density [24]. Nailfold capillaroscopy is essential to the assessment of SSc. However, the goldstandard method, NVC, is not always available and can be limited by cost as well as its laborious process [5]. The results of this review suggest that dermatoscope and stereomicroscopy may be useful alternatives to NVC. Despite lower magnification, dermatoscope may serve as a screening tool to identify normal, "scleroderma pattern", and images need further corroboration by NVC (either non-interpretable or non-specific). However, one may expect that a high proportion of patients will require NVC corroboration as 50% of dermatoscopy images required further NVC [17]. Additionally, dermatoscopy may have some ability to differentiate between scleroderma patterns (early, active, late) based on limited evidence. Dermatoscopes have the advantage of portability as well as being affordable [5]. These advantages are reflected in a survey of SSc experts in the USA in 2019, which indicated that 64% of clinicians use a dermatoscope or ophthalmoscope, with only 7% using NVC [16]. Additionally, training curriculums have been developed to use dermatoscopy in rheumatology fellowship programs [26]. On the other hand, stereomicroscopy showed a strong correlation in all parameters related to scleroderma pattern compared to NVC, with the ability to discriminate between SSc patients from controls by experienced observers. Though not portable, stereomicroscopy has the advantage of being able to visualize the whole nailfold microvasculature for the prompt localization of abnormalities and is the more cost-effective compared to NVC [5]. However, stereomicroscopy may not be able to capture high quality NFC images in patients with SSc with finger-flexion ankylosis. Current evidence is scarce but promising in using smartphone microscopy, and USB microscopy to assess NFC in patients with SSc, though further research is needed. Other potential tools that are similar include epiluminescence microscopes (e.g., Dino-Lite US), as these devices generate high-quality digital images compared, albeit, at a lower magnification compared to the NFC gold standards, but with a fraction of the cost [27]. However, their specificity and/or sensitivity have not been directly compared.
On the other hand, as reflected in the articles included in this review, NFC parameters measured are heterogenous and not consistently reported. These parameters assessed by NVC have important clinical implications. For example, NVC-detected capillary density is a reliable predictor of overall disease progression, the occurrence of digital ulcers, the progression of pulmonary disease, and skin fibrosis [11,12,28]. Additionally, capillary density can be used to monitor the effects of treatment [29]. Using NVC, the inter-rater and intra-rater reliability for NVC is excellent across specific capillaroscopy parameters including capillary-loop dimensions, widened and giant capillaries, and capillary density [12]. The reliability of these specific quantitative metrics in other devices is not well studied. While one of the included studies reported excellent interobserver and intraobserver agreement for widefield microscopy to classify images by specific capillaroscopic parameters (capillary density, enlarged capillary loops, microhemorrhages, giant capillary loops, avascular score) [23], the reliability of dermatoscopes to study these metrics may be limited by its magnification and are not yet well described [30,31]. This is the first review to synthesize current evidence comparing NVC to other methods for assessing NFC to our knowledge. This review has captured studies from the USA and Europe, which may have captured differences in practice patterns [16,32]. Additionally, this review included all publication types, which also included conference abstracts that added to available evidence.
The limitations of this review included its language limitation as well as limitations on the search strategy's inclusion of different equipment terminology. Another limitation is the inclusion of only patients with SSc, when there is evidence to suggest NFC changes in other connective-tissue disorders [33,34]. The results from this review may not extrapolate to other settings as it is not clear in included studies how much training/expertise evaluators had when assessing nailfolds with different methods. In 2020, Smith and colleagues proposed the standardization of the assessment of NFC by NVC in SSc [5]. The emergence of portable and low-cost imaging techniques makes NFC assessment in SSc more accessi-ble. More research is needed to evaluate performance characteristics using standardized parameters on NVC in larger studies.
Author Contributions: Z.M. and M.O. reviewed the studies and generated the first version of the manuscript; D.J.M., R.G. and J.W.C.T. read, edited, and provided valuable feedback. We confirm that neither the manuscript nor any parts of its content are currently under consideration for or published in another journal. All authors have read and agreed to the published version of the manuscript.