Ultrasound Detection of Calcinosis and Association with Ulnar Artery Occlusion in Patients with Systemic Sclerosis

Abstract

Objective:

To investigate the ability of ultrasound (US) compared to radiographs in detecting calcinosis in hands/wrists of systemic sclerosis (SSc) patients, and assess US markers of pathologic perfusion.

Methods:

SSc patients were evaluated for calcinosis in the hands/wrists by X-ray and US. Presence or absence of calcinosis was recorded by patient, hand, and anatomical zone; sensitivity and specificity for calcinosis detection by US versus X-ray was determined. Bilateral US vascular measurements of ulnar artery occlusion (UAO) and finger pulp blood flow (FPBF) were obtained. For each hand, associations between markers of pathologic blood flow (UAO, FPBF, and a composite severity score of UAO and FPBF) and presence of calcinosis were assessed using generalized estimating equations.

Results:

Of 43 SSc patients (19 diffuse, 24 limited), 39.5% had calcinosis on X-ray compared to 30.2% on US. Sensitivity and specificity for US was 61% and 95% by zone, 78% and 98% by hand, and 76% and 100% by patient, respectively. UAO was seen in 30% and 28% of left and right hands, respectively; FPBF was absent in ≥1 digit of the left and right hands in 49% and 44%, respectively. UAO was associated with X-ray identified calcinosis by hand (OR 8.08, 95% CI 2.45–26.60, p<0.001), whereas FPBF and the composite severity score were not significant. UAO was associated with calcinosis even in the absence of digital ulcers (OR 33.00, 95% CI 3.39–321.09, p=0.003).

Conclusion:

US was sensitive and highly specific in detecting calcinosis in SSc. UAO was strongly associated with X-ray identified calcinosis.

INTRODUCTION

Systemic sclerosis (SSc) is an autoimmune connective tissue disease characterized by excessive collagen deposition, vascular damage, inflammation, and progressive fibrosis of the skin and visceral organs.¹ Calcinosis cutis, the deposition of calcium in the skin and subcutaneous tissues, is a common and potentially debilitating manifestation of SSc, occurring in approximately 25% of patients.² Calcinosis commonly occurs in the hands with associated pain and functional impairment.

Although the pathogenesis of calcinosis is not fully understood, local trauma, chronic inflammation, and vascular hypoxia have been proposed as potential mechanisms. Calcinosis occurs in tissues that are under chronic stress such as local trauma, or have been damaged secondary to the underlying inflammatory process.³ Macrophages, proinflammatory cytokines, and impairment of calcium-regulating proteins have been implicated.⁴ Furthermore, digital ulcers and ischemia are associated with calcinosis in SSc, suggesting a role of vasculopathy and injury in the pathogenesis of calcinosis.^2,5 In the periphery, destructive and obliterative vasculopathy leads to capillary loss, narrowing, and occlusion of palmar and ulnar arteries resulting in Raynaud’s phenomena, digital ischemia, ulceration, and gangrene, and may contribute to the development of calcinosis and acro-osteolysis.^5-7

Calcinosis involvement is typically evaluated with radiography, bone scintigraphy, computed tomography (CT), and magnetic resonance (MR) imaging, but these techniques can be costly, time consuming, and/or expose patients to radiation. The dramatic vascular and fibrotic changes associated with SSc are particularly amenable to evaluation by ultrasound (US), which has the benefits of being a non-invasive, low cost, non-ionizing, point of care imaging modality. Conveniently, US provides tissue imaging capabilities for vasculopathy-related disease manifestations and has been used to evaluate digital ulcers, calcinosis, and acro-osteolysis,^7–13 making it a promising tool for both research and clinical SSc evaluation and management. Ulnar artery occlusion (UAO) and pathologic finger pulp blood flow (FPBF) are vascular features detectable using US and may be helpful for screening SSc patients at risk for digital vascular complications. UAO is a macrovascular feature occurring in up to 25% of SSc patients and is more frequent in SSc than in the general population. Pathologic FPBF, or a decrease of Doppler signal on the subhypodermal part of finger pulp or on the entire finger pulp, has been reported in up to 51% of SSc patients.⁷

Few studies have used US to evaluate calcinosis and vasculopathy in SSc patients. Freire et al. detected calcinosis in 17/44 (39%) patients using US, with a high sensitivity (89%) compared to radiographs.¹¹ Prior studies in SSc have shown that UAO and pathologic FPBF are associated with new or recurrent digital ulcers.^8–10 Lescoat et al. used US to show concordance between UAO, FPBF, and microvascular disease on capillaroscopy, and that these measures were associated with digital pits, ulcers, acro-osteolysis, and calcinosis.¹⁰

Prior studies were limited because they characterized the presence or absence of calcinosis by patient, and did not specifically evaluate the location of lesions by hand or anatomic zone. Associations of calcinosis with UAO and FPBF also ignored associations within each hand. Our objective was to evaluate the sensitivity and specificity of US to detect calcinosis in the hands and wrists of SSc patients as compared to X-ray by patient, hand, and anatomical zone, and to evaluate the relationship between US markers of pathologic perfusion and the presence of calcinosis.

METHODS

Study population and ethics approval

Between February 2018 and February 2019, consecutive patients meeting 2013 ACR/EULAR classification criteria for SSc at Stanford University were recruited for this cross-sectional study evaluating calcinosis in the hands and wrists by both X-ray and US. This study was approved by Stanford University’s Institutional Review Board and informed consent was obtained from all patients.

Ultrasound evaluation

US measurements were performed using a GE Logiq E system equipped with a 5–12 MHz linear array probe in standard B-mode. The same experienced ultrasonographer (RF) performed all US evaluations and was blinded to all radiographic evaluations. All vascular US measurements were obtained and interpreted prior to US calcinosis evaluation. Measurements of FPBF and UAO were obtained following techniques previously described by Lescoat et al with minor modifications.⁹ US blood flow measurements were made after patients were at rest for 10 minutes in a room with temperatures ranging from 20–23°C. Bilateral US vascular measurements of UAO and FPBF in the 3rd and 4th digits were obtained (Figure 1). Ulnar artery blood flow was measured on the wrist using a transverse view at Guyon’s canal, at the level of the pisiform bone. UAO was defined as a complete lack of blood flow on power doppler US; power doppler was favored over color doppler to maximize sensitivity for blood flow. Loss of FPBF in each hand was defined as a lack of power doppler signal in the subhypodermal finger pulp or entire finger pulp observed in at least 1 of the 2 evaluated fingers per hand. To perform the calcinosis US exam, the dorsal and palmar aspects of the hands and wrists were interrogated by sweeping the probe over the skin surface while submerged in lukewarm water. Calcinosis was identified as hyperechoic lesions with or without shadowing, located in the skin, soft tissue, tendons, peritendinous or periarticular areas (Figure 1). The presence or absence of calcinosis was recorded in each of the 44 anatomical zones as described previously by Chung et al. on behalf of the Scleroderma Clinical Trials Consortium Calcinosis Working Group.¹⁴ The left and right hands were divided into 22 anatomic regions each, which were defined by the underlying osseous structures and included the bone and surrounding soft tissue.

Figure 1. Ultrasound and X-ray findings in systemic sclerosis.

Systemic sclerosis patients with (A) X-ray findings of calcinosis, (B) ultrasound identified calcinosis [red arrowheads] in the longitudinal palmar aspect of the 2^nd finger [blue outline = flexor tendon], (C1-C2) absent and normal FPBF respectively, and (D1-D2) UAO and no UAO, respectively. UAO = ulnar artery occlusion, FPBF = finger pulp blood flow, DIP = distal interphalangeal

Radiographic evaluation

All patients received bilateral hand and wrist x-rays within 3 months of US evaluation. The same trained and experienced reader (MC) evaluated all radiographs according to the study by Chung et al.,¹² and was blinded to the clinical history and US results. The presence or absence of calcinosis was recorded in each of the 44 anatomical zones as described previously. The presence or absence of acro-osteolysis by patient was also recorded.¹⁵

Clinical and Laboratory assessment

Standard clinical parameters were assessed including MRSS and the presence or absence of digital ulcers and pits at the time of the US. Disease manifestations, laboratory parameters, and current medications were assessed by chart review.

Statistical analysis

Sensitivity and specificity for calcinosis detection by US were calculated using radiographs as the reference standard. For each hand, associations between markers of pathologic blood flow (UAO, FPBF, and a composite severity score of UAO and FPBF) and the presence of calcinosis were assessed using generalized estimating equations to estimate the odds ratio and corresponding 95% confidence intervals for the outcome calcinosis adjusting for intra-patient correlation from both left and right hands. Since calcinosis has been associated with digital ulcers, a sensitivity analysis was performed stratifying the cohort into patients with and without a history of digital ulcers. For each patient, the association between UAO and acro-osteolysis was also assessed using generalized estimating equations to estimate the odds ratio and corresponding 95% confidence intervals for the outcome acro-osteolysis. The association between presence of UAO and history of digital ulcers by patient was also assessed; since information on the laterality of digital ulcers was not available, the outcome digital ulcers was evaluated by patient instead of by hand. No pre-calculations of power and sample size were performed considering the exploratory nature of the study. All analyses were conducted using SAS software version 9.4 (SAS Institute Inc., Cary, NC) with the threshold of significance set at P<0.05.

RESULTS

General clinical features

43 SSc patients (19 diffuse and 24 limited) were enrolled and 17 (39.5%) were found to have X-ray evidence of calcinosis in at least 1 hand (10 bilateral, 7 unilateral). Patient characteristics are summarized in Table 1. Most were female (90.7%) and had limited cutaneous SSc (55.8%), with mean age at visit 57.6 ± 11.5 years and mean disease duration from first non-Raynaud symptom 10.5 ± 9.2 years.

Table 1.

Patient Characteristics

	(n = 43)
Age, mean ± SD (years)	57.6 ± 11.5
Female, n (%)	39 (90.7)
Race, n (%)
Caucasian	25 (58.1)
Asian	7 (16.3)
Hispanic	6 (14.0)
Other	5 (11.6)
Diffuse disease, n (%)	19 (44.2)
Time from 1st symptom mean ± SD (years)
Raynaud’s phenomenon	11.6 ± 10.0
Non-Raynaud’s phenomenon	10.5 ± 9.2
Modified Rodnan Skin Score, median (IQR)	7 (3–15)
Autoantibodies, n (%)
Anticentromere	16 (37.2)
Anti-Scl-70	12 (27.9)
Anti-RNA-polymerase III	6 (14.0)
Anti-PM-Scl or anti-U3 RNP	2 (4.7)
Anti-Nuclear antibody	35 (81.4)
Disease features, current or prior, n (%)
Digital Ulcers	20 (46.5)
Gangrene	2 (4.7)
Gastrointestinal involvement	39 (90.7)
Interstitial lung disease	16 (37.2)
Pulmonary hypertension*	8 (18.6)
Arthritis§	29 (67.4)
Myositis	17 (39.5)
Scleroderma renal crisis	0
Telangiectasias	28 (65.1)
Nailfold capillary changes	31 (72.1)
Active Clinical Features, n (%)
Digital ulcers	6 (14.0)
Digital Pits	11 (25.6)
X-ray features on hand and wrist X-ray
Calcinosis	17 (39.5)
Acro-osteolysis	13 (30.2)
Current therapy, n (%)
Calcium channel blocker	23 (53.5)
PDE5 or PDE3 inhibitor	12 (27.9)
SSRI/SNRI	8 (18.6)
ACE inhibitor	3 (7.0)
Endothelin receptor antagonist	1 (2.3)

SD = standard deviation, IQR= interquartile range, anti-Scl-70 = anti-topoisomerase antibody, anti-PM-Scl = anti-polymyositis-scleroderma antibody, anti-U3 RNP = anti-fibrillarin antibody, PDE = phosphodiesterase, ACE = angiotensin converting enzyme.

^*Diagnosed by right heart catheterization

^§reported history of joint pain

Sensitivity, specificity, and feasibility of US-detected calcinosis

US identified 13 (30.2%) patients with calcinosis (9 bilateral, 4 unilateral). Of the 748 zones evaluated, X-rays identified 72 zones (9.6%) with calcinosis while US showed 76 (10.2%). The sensitivity and specificity for US detection of calcinosis versus X-ray was 61% and 95% by zone, 78% and 98% by hand, and 76% and 100% by patient, respectively. On average, the bilateral US calcinosis exam was completed in 7–10 minutes.

Associations of US markers of pathologic perfusion with calcinosis

UAO was observed in 16 patients (7 unilateral, 9 bilateral), and in 30% and 28% of left and right hands, respectively. The absence of FPBF in 3^rd and/or 4^th digits of the hand was seen in 24 patients (8 unilateral, 16 bilateral), and in 49% and 44% of the left and right hands, respectively. Calcinosis by X-ray was detected in 35% and 28% of left and right hands, respectively. Similarly, calcinosis by US was detected in 26% of both left and right hands (Table 2A).

Table 2.

(A) US and X-ray features of calcinosis; (B) Associations between US markers of pathologic blood flow and calcinosis.

(A) US and X-ray features of calcinosis.
	Left hand (n=43)		Right hand (n=43)
UAO, n (%)	13 (30)		12 (28)
FPBF absent* (n%)	21 (49)		19 (44)
Calcinosis by US, n (%)	11 (26)		11 (26)
Calcinosis by X-ray, n (%)	15 (35)		12 (28)
(B) Associations between US markers of pathologic blood flow and calcinosis.
	Unstratified Analysis		Positive digital ulcers		Negative digital ulcers
	OR (95% CI)	p	OR (95% CI)	p	OR (95% CI)	p
UAO	8.08 (2.45–26.60)	<0.001	4.25 (1.17–15.43)	0.03	33.00 (3.39–321.09)	0.003
FPBF	2.66 (0.83–8.52)	0.10	5.11 (0.91–28.65)	0.06	1.39 (0.26–7.35)	0.70
FPBF + UAO severity score	3.23 (0.96–10.92)	0.06	3.66 (0.66–20.13)	0.14	2.78 (0.51–15.10)	0.24

US = ultrasound, UAO = ulnar artery occlusion, FPBF = finger pulp blood flow, OR = odds ratio by generalized estimating equations, CI = confidence interval.

^*FPBF absent if one or both measurements per hand were absent.

Using generalized estimating equations, there was a strong association between UAO and X-ray identified calcinosis (OR 8.08, 95% CI 2.45–26.60, p<0.001). Both FPBF (OR 2.66, 95% CI 0.83–8.52, p=0.10) and the composite severity score (OR 3.23, 95% CI 0.96–10.92, p=0.06) were not significantly associated with calcinosis. When stratified by patients without a history of digital ulcers, UAO was even more strongly associated with X-ray identified calcinosis (OR 33.00, 95% CI 3.39–321.09, p=0.003). In patients with a history of digital ulcers, the association between UAO and calcinosis was attenuated but still significant (OR 4.25, 95% CI 1.17–15.43, p=0.03) (Table 2B).

When analyzing the relationship between UAO and acro-osteolysis by patient, there was an increased but non-significant odds ratio (OR 2.30, 95% CI 0.79–6.72, p=0.13). Similar results were observed between acro-osteolysis and FPBF (OR 2.29, 95% CI 0.71–7.35, p=0.16) or the composite severity index (OR 1.78, 95% CI 0.56–5.71, p=0.33). In contrast, UAO of either hand was strongly associated with digital ulcer history by patient (OR 6.76, 95% CI 1.48–38.21, p=0.01).

DISCUSSION

In this cross-sectional single-center study, we present a novel application of US for detecting calcinosis and vascular blood flow in SSc. This imaging modality allowed non-invasive detection and measurement of the extent of calcinosis abnormalities among patients, their respective hands, and anatomical zones. To our knowledge, this is the first study to report both sensitivity and specificity of US-identified calcinosis compared to radiographs with this degree of anatomical evaluation. We found a strong association between UAO, a marker of decreased peripheral perfusion, and X-ray identified calcinosis in the same hand, providing further support for an association between vascular disease in SSc and the pathogenesis of calcinosis.

In our SSc cohort, the prevalence of calcinosis was 39.5% on radiographs and 30.2% on US, similar to prior cohorts.¹¹ We found US to be highly specific but less sensitive compared to radiographs using all three methodologies when compared by patient, hand, or anatomical zone. The sensitivities of US-identified calcinosis in our study were overall less than those reported by Freire et al.; however, blinding of the ultrasonographer, length of time to complete the study, and other methodologic aspects of this initial report are not available for comparison. Our examiner was blinded to disease subtype and X-ray results, and US vascular measurements were obtained and recorded prior to US calcinosis evaluation for each patient. We also aimed to perform the US exam in a feasible timeframe. Based on our experience, we suspect that slower scanning times would lead to increased sensitivities which may be more appropriate for research purposes. When analyzing US performance versus X-ray by anatomic zone, assigning lesions to anatomic zones when they bordered or spanned two zones may have led to disagreements in zone assignment between X-ray and US. Additionally, while very small punctate calcinosis lesions are conspicous on X-rays, these fine lesions may be missed on dynamic ultrasonography, reducing sensitivity for detection. False negatives on US were equally distributed across anatomic zones. The specificities of US-identified calcinosis were consistently high across all three methodologies in our cohort.

In consensus with Lescoat et al.,⁷ we found a strong association between US-identified UAO and calcinosis in SSc patients, thereby replicating similar results in an independent SSc cohort. Although our OR (OR 8.08, 95% CI 2.45–26.60, p<0.001) was higher than that reported by Lescoat et al. (OR 3.19, 95% CI 1.14–8.93, p=0.025), there was still overlap in the 95% CI and our smaller sample size may have resulted in a wider estimate. Since digital ulcers are associated with calcinosis, we performed a sensitivity analysis stratifying by the presence or absence of digital ulcers to understand their impact on the relationship between US markers of pathologic blood flow and calcinosis. In patients without digital ulcers, UAO was even more strongly correlated with X-ray identified calcinosis (OR 33.00, 95% CI 3.39–321.09, p=0.003), albeit the confidence interval was very large. Collectively, these findings provide powerful evidence linking vasculopathy to many SSc related digital manifestations and may inform future treatment approaches. The association between X-ray identified calcinosis and FPBF or the composite score was not statistically significant, similar to Lescoat et al. Although this result may be due to underpowering in this study, minute-to-minute variability in FPBF due to ambient temperature, medication regimen, stress, and other extrinsic factors may make FPBF a less accurate measure of peripheral perfusion compared to UAO. In addition, FPBF lacks a consensus definition and standardization, which is needed for reliable inter-study comparisons moving forward.

The use of US to detect calcinosis in SSc may have a significant role in both clinical care and research. Continuing advances in US processing algorithms and the recent availability of higher frequency US probes affords greater imaging detail and resolution. US has the potential to objectively characterize and monitor calcinosis in SSc, and thus become a reliable tool that complements clinical evaluation and/or radiographs. Although US was not as sensitive as conventional radiography in detecting calcinosis in SSc patients, it could play a role in screening for macrovascular complications without the risks of ionizing radiation. Additionally, US could potentially be used as a tool to assess for calcinosis and UAO during clinic visits, such that referral for sympathectomy and/or calcinosis removal could be decided upon at the same visit.

Limitations of our study include: ultrasonography is highly operator-dependent and inter-user variability was not studied; this was a single-center study with only one ultrasonographer; there were no follow-up visits to monitor calcinosis over time; and practical scanning times were used for calcinosis detection and more lengthy US examination times were not studied.

Despite these limitations, we have developed a novel method for US detection of calcinosis affecting the hands of SSc patients that is feasible, sensitive, and highly specific compared to radiographs. This novel US method should certainly undergo external validation in future studies, and further work is needed to develop a scoring system to monitor the progression, worsening, or stability of calcinosis over time. We also found a strong relationship between UAO, a marker of decreased peripheral perfusion, and X-ray identified calcinosis, providing further support for an association between vascular disease in SSc and the pathogenesis of calcinosis.

KEY MESSAGES:

• Ultrasound detection of calcinosis in SSc patients was highly specific but less sensitive compared to radiographs.

• A strong relationship between ultrasound detected UAO and X-ray identified calcinosis was found.

• This study supports an association between vascular disease in SSc and the pathogenesis of calcinosis.