SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES

Objective

To prospectively compare an indocyanine green (ICG)–enhanced optical imaging system with contrast-enhanced magnetic resonance imaging (MRI) for the detection of synovitis in the hands of patients with rheumatologic disorders.

Methods

Forty-five patients (30 women [67%], mean ± SD age 52.6 ± 13.4 years) in whom there was a clinical suspicion of an inflammatory arthropathy were examined with a commercially available device for ICG-enhanced optical imaging as well as by contrast-enhanced 3T MRI as the standard of reference. Three independent readers graded the degree of synovitis in the carpal, metacarpophalangeal, proximal interphalangeal, and distal interphalangeal joints of both hands (1,350 joints), using a 4-point ordinate scale (0 = no synovitis, 1 = mild, 2 = moderate, 3 = severe). Statistical analyses were performed using a logistic generalized estimating equation approach. Agreement of optical imaging ratings made by the different readers was estimated with a weighted kappa coefficient.

Results

When MRI was used as the standard of reference, optical imaging showed a sensitivity of 39.6% (95% confidence interval [95% CI] 31.1–48.7%), a specificity of 85.2% (95% CI 79.5–89.5%), and accuracy of 67.0% (95% CI 61.4–72.1%) for the detection of synovitis in patients with arthritis. Diagnostic accuracy was especially limited in the setting of mild synovitis, while it was substantially better in patients with severely inflamed joints. Moderate interreader and intrareader agreement was observed.

Conclusion

The evaluated ICG-enhanced optical imaging system showed limitations for the detection of inflamed joints of the hand in comparison with MRI.

Rheumatologic disorders are a heterogeneous group of diseases with a high prevalence of ∼22% (1). Even with the advent of remarkable novel therapeutics, inflammatory arthritis continues to be a common cause of disability and is still associated with substantial activity limitation, reduced quality of life, and high health care costs (1). Rheumatoid arthritis (RA) is the most common chronic inflammatory joint disease, with a prevalence of 0.5–1.0% (2, 3). The present strategy of early and aggressive treatment is based on recognition of a “window of opportunity” in early disease, during which initiation of effective therapy may substantially improve short-term and long-term outcomes (4–6). This knowledge has led to a growing need and search for sensitive and specific tools for the early diagnosis of inflammatory arthritides such as RA (4–6).

Conventional radiography is still widely used for detecting bone damage and monitoring disease progression (7). The marked disadvantage of this method is its lack of sensitivity for detecting early erosive changes (7). Ultrasonography and magnetic resonance imaging (MRI) are both very sensitive and specific methods for the detection of early changes in inflammatory arthritis (8, 9). However, ultrasonography is time-consuming, operator dependent, and may not detect early arthritis (8–10). MRI provides high sensitivity for the diagnosis of early synovitis and joint effusions but is limited by long examination times and high costs (11).

Optical imaging is a relatively new, noninvasive, and nonionizing imaging modality with fast image acquisition times (12, 13). The major drawback of optical imaging is the limited tissue penetration of light; however, because inflammatory arthropathies typically affect the small joints of the hands and feet, this is not necessarily a significant limitation for our application.

Various approaches for detecting experimental arthritis using optical imaging have proven to be suitable (12–19). Early hyperemia of inflamed joints could be diagnosed by recording scattering and absorption patterns of light transmitted through inflamed finger joints. In a clinical study, this laser-based technique was shown to provide information about the inflammation status of finger joints with sensitivity and specificity of 80% and 89%, respectively (20); however, imaging was limited to a single finger, and image processing times were long. Among fluorescent dyes, the nonspecific, small molecular fluorescent substance indocyanine green (ICG) appears to be the most promising for this application, because it has been shown to enhance imaging of inflamed joints (14, 15), and, more importantly, it is approved by the US Food and Drug Administration (FDA).

The results of previous studies showed that optical imaging allows, in principle, the detection of inflamed tissue in animal models as well as in humans (12–19). The purpose of this study was to prospectively compare an ICG-enhanced optical imaging system with contrast-enhanced 3T MRI for the detection of synovitis in the hands of patients with rheumatologic disorders. To the best of our knowledge, this is the first larger-scale investigation of the performance of a clinical ICG-enhanced optical imaging system for imaging of synovitis in arthritic hands in comparison with 3T MRI.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES

This study was approved by the local ethics committee at our institution (Technische Universität München, Germany) prior to commencement of the study and was conducted according to the principles of the Declaration of Helsinki, such that all patients provided written informed consent before participating.

Patients.

A total of 45 patients in whom there was a clinical suspicion of an inflammatory joint disease were included in this prospective study. The patients were recruited from January 2010 to October 2010 through the Department of Rheumatology at our hospital. Inclusion criteria for the study population were >1 tender and/or swollen joint among the carpal, metacarpophalangeal (MCP), proximal interphalangeal (PIP), and distal interphalangeal (DIP) joints of both hands, symptom duration between 6 weeks and 24 months, and willingness to participate. Exclusion criteria were pregnancy; renal failure; known allergy against iodine, ICG, or gadolinium; and other contraindications to MRI such as the presence of metallic fragments in the body, the presence of magnetically activated implanted devices, or claustrophobia. Physical examination, optical imaging, and MRI were performed within 1 week of inclusion in the study.

Clinical examination.

The demographic and clinical characteristics of the patients were recorded; these included sex, age, height, weight, disease duration, duration of morning stiffness, joint deformities of any kind, and ongoing systemic infectious disease. Physical examination of each joint was performed by bimanual palpation, by 2 rheumatologists (KT and PM) with longstanding clinical experience in rheumatology. The numbers of swollen and tender joints were recorded. Additionally, all joints were clinically scored for synovitis (0 = no synovitis, as defined by the absence of both tenderness and swelling or the absence of either tenderness or swelling in the joint, and 1 = synovitis, as defined by swelling and tenderness in the joint) by 2 rheumatologists (KT and PM) in consensus and the results were compared with MRI as the standard of reference. The serum C-reactive protein (CRP) level and erythrocyte sedimentation rate (ESR) were measured in all patients. At the end of the clinical examination, the following clinical diagnoses were made by the rheumatologists (KT and PM): RA, undifferentiated arthritis, osteoarthritis (OA), psoriatic arthritis (PsA), spondylarthritis (SpA), and connective tissue disorder.

Optical imaging contrast media.

ICG is a nontargeted hydrophilic anionic near-infrared fluorescent dye that has been approved for a variety of clinical applications (21–23). It has been used for bedside measurement of blood volume (24), tumor detection (25–28), and, more recently, to visualize joint inflammation in preclinical (14, 15) and clinical settings (16). The absorption and emission maximum wavelengths for ICG are ∼805 nm and ∼830 nm, respectively (29). Within seconds after intravenous injection, ICG reversibly binds to up to 98% of all plasma proteins without extravasating. The protein-bound compound is taken up by hepatocytes and excreted via bile fluids without entering the enterohepatic circle (30). ICG (Pulsion) dissolved in sterile water (total volume 1–2 ml) was injected as an intravenous bolus at a dose of 0.1 mg/kg body weight (time to inject 3 seconds).

Optical imaging system.

We used a commercially available, near-infrared fluorescence imaging system (Xiralite X4; Mivenion). This optical imaging system allows for quick sequential capture of the fluorescence intensity signal from the human hand after excitation by light emitting diode (LED). Patients are positioned with their hands placed inside the device behind a screen, to prevent the entrance of ambient light. Data acquisition with the optical imaging system was performed as follows: continuous illumination with LEDs at a wavelength of 740 nm situated overhead was started. ICG was injected 10 seconds after the start of imaging; simultaneously, a cooled CCD camera behind optical filters of ≤800 nm captured images at 1 frame per second for a total time of 6 minutes.

Magnetic resonance imaging.

MRI was performed with a Verio 3T MR scanner (Siemens) using a Flex Large flexible surface coil (Siemens). Patients underwent scanning while in the prone position, with the hands outstretched in praying position in order to scan both hands at once. The following MRI protocol was used.

  • 1
    Coronal fat-saturated intermediate-weighted turbo spin-echo (TSE) (repetition time [TR]/echo time [TE] 4,880/38 msec, echo train length [ETL] 7, resolution 0.49 × 0.49 × 2.5 mm3, field of view [FOV] 220 × 220 mm2, slice 34 mm, gap 0.25 mm, scan time 3 minutes and 56 seconds).
  • 2
    Coronal T1-weighted TSE (TR/TE 790/12 msec, ETL 3, resolution 0.43 × 0.43 × 2.5 mm3, FOV 220 × 220 mm2, slice 34 mm, gap 0.25 mm, scan time 3 minutes and 19 seconds).
  • 3
    Consecutive dynamic scans (n = 139) with a coronal T1-weighted 3-dimensional (3-D) radio frequency–spoiled gradient echo (TR/TE 3.83/1.34 msec, flip angle 25°, resolution 1.0 × 0.86 × 2.5 mm3, FOV 185 × 220 × 45 mm3, scan time 5 minutes and 39 seconds). For better temporal resolution (2.4 seconds), sparse sampling of outer K-space regions was applied (TWIST [time-resolved angiography with stochastic trajectories]; fully sampled central region 33%, sampling of outer region 50%). Automated injection of 0.02 ml/kg gadopentetate dimeglumine (Magnograf; Schering) was started after the third acquisition, at 2 ml/second. At the end of the MR protocol after the TWIST sequence, postcontrast images were acquired.
  • 4
    Transverse T1-weighted fat-saturated TSE (TR/TE 1,160/13 msec, ETL 3, resolution 0.39 × 0.39 × 3 mm3, FOV 139 × 150 mm2, slice 36 mm, gap 0.6 mm, scan time 4 minutes and 42 seconds).
  • 5
    Coronal T1-weighted fat-saturated TSE (TR/TE 1,070/12 msec, ETL 3, resolution 0.43 × 0.43 × 2.5 mm3, FOV 220 × 220 mm2, slice 34 mm, gap 0.25 mm, scan time 4 minutes and 29 seconds).

All sequences were acquired with parallel imaging (GRAPPA [generalized autocalibrating partially parallel acquisitions], acceleration factor 2).

Image evaluation.

Thirty joints in both hands (carpal, MCP, PIP, and DIP joints) of 45 patients (1,350 evaluations per reader) were scored. For optical imaging, joints were scored independently by 3 radiologists (FD, CS, and SiW, each of whom had 6–9 months of optical imaging experience and additional optical imaging evaluation training by the manufacturer). For MRI, joints were scored by 3 radiologists in consensus (RM [4 years of MRI experience], SaW [9 years of MRI experience], and KW [16 years of MRI experience]). A repeated reading of the optical images 4 weeks after the first reading was performed by one reader (CS), with rearrangement of the image sets for intrareader agreement estimation. The readers performed scoring of optical imaging and MRI without knowledge of the clinical assessment and were blinded to the patient name and the results of MRI or optical imaging.

Images obtained by optical imaging and MRI were evaluated in random order using the semiquantitative assessment system suggested by the OMERACT (Outcome Measures in Rheumatology) MRI group (31). Using this method, synovitis is scored on a scale of 0–3 in each joint (0 = no synovitis, 1 = mild arthritis, 2 = moderate arthritis, 3 = severe arthritis). On MRI, synovitis was defined based on the OMERACT RAMRIS (Rheumatoid Arthritis Magnetic Resonance Imaging Scoring) definition (31). On optical imaging, synovitis was defined as an area of focal hyperperfusion related to a joint. For the evaluation of optical imaging, the software provided by the supplier was used (XiraView Software, version 3.5c; Mivenion). For image demonstration (Figures 1–3), we present maximum-intensity projections of the stack of 360 optical images, which were obtained using the software ImageJA Version 1.43h (National Institutes of Health). Bone erosions, bone edema, and tenosynovitis identified by MRI were additionally scored using the OMERACT RAMRIS and PsAMRIS (psoriatic arthritis magnetic resonance imaging scoring) systems (31, 32). Using this method, bone edema was scored on a scale of 0–3 according to the volume of edema compared with the assessed bone volume, and bone erosion was scored on a scale of 0–10 based on the proportion of eroded bone compared with the assessed bone volume in each individual bone (31). Tenosynovitis was scored on a scale of 0–3 in 10 different compartments of the extensor and flexor tendon areas (32).

thumbnail image

Figure 1. Indocyanin green–enhanced optical image (maximum-intensity projection) (a) and magnetic resonance images (d–f) of the right hand of a 34-year-old woman who had pain in the metacarpophalangeal (MCP) and proximal interphalangeal (PIP) joints of both hands for 4 weeks and a clinical picture of seropositive polyarthritis (rheumatoid factor positive, anti–citrullinated protein antibody negative). The patient's C-reactive protein level was 7.3 mg/dl, and the erythrocyte sedimentation rate (1 hour/2 hours) was 51/91 mm. The wrists, MCP joints, and PIP joints of both hands were tender. Contrast-enhanced coronal (b) and transverse (d–f) T1-weighted fat-saturated turbo spin-echo images and coronal maximum-intensity projection of dynamic TWIST (time-resolved angiography with stochastic trajectories) sequence (c) images are shown. Optical imaging shows focal hyperperfusion, indicating synovitis, in the second and third distal MCP joints, the first interphalangeal (IP) joint, and the second through fifth PIP and distal interphalangeal (DIP) joints. Magnetic resonance imaging (MRI) confirmed moderate-to-severe synovitis in the MCP (e), IP/PIP (d), and DIP (b and c) joints. In addition, MRI detected moderate synovitis in the carpal joints (f) that was not seen using optical imaging.

Download figure to PowerPoint

thumbnail image

Figure 2. Indocyanin green–enhanced optical image (maximum-intensity projection) (a) and magnetic resonance images (d–f) of the left hand of a 47-year-old woman with seropositive rheumatoid arthritis (rheumatoid factor positive and anti–citrullinated protein antibody positive). The patient's C-reactive protein level was 0.3 mg/dl, and the erythrocyte sedimentation rate (1 hour/2 hours) was 21/41 mm. The most involved joint was the wrist of the left hand. Contrast-enhanced coronal (b) and transverse (d–f) T1-weighted fat-saturated turbo spin-echo images and coronal maximum-intensity projection of the dynamic TWIST sequence (c) images are shown. Optical imaging shows mild hyperperfusion in the carpus and in the second through fourth PIP joints. MRI revealed severe erosive synovitis in the carpus (b, c, and f) and mild synovitis in the MCP (b and e) and first through fifth IP/PIP (d) joints. See Figure 1 for definitions.

Download figure to PowerPoint

thumbnail image

Figure 3. Indocyanin green–enhanced optical image (maximum-intensity projection) (a) and magnetic resonance images (d–f) of the left hand of a 46-year-old woman who had longstanding seronegative monarthritis of the third distal MCP joint of the left hand; she was not receiving systemic therapy. The patient's C-reactive protein level was 0.2 mg/dl, and the erythrocyte sedimentation rate (1 hour/2 hours) was 11/30 mm. She reported tenderness and swelling in the third distal MCP joint of her left hand. Contrast-enhanced coronal (b) and transverse (df) T1-weighted fat-saturated turbo spin-echo and coronal maximum-intensity projection of the dynamic TWIST sequence (c) images are shown. Optical imaging shows focal hyperperfusion in the third distal MCP joint and mild hyperperfusion in the second through fourth PIP joints. MRI confirmed the presence of severe synovitis in the third distal MCP joint; additionally, mild synovitis could be detected in the carpal (f) and second through fourth MCP and PIP joints. See Figure 1 for definitions.

Download figure to PowerPoint

Statistical analysis.

True-positive, true-negative, false-positive, and false-negative values for detection of synovitis were determined for all optical images as well as for clinical examinations, using MRI as the diagnostic standard of reference. Diagnostic measures were estimated using a logistic generalized estimating equation (GEE) approach (33) to account for correlated observations caused by investigation of several joints in the same patients and by repeated assessment of the same images by multiple readers. In the GEE models, the patient was used as the subject variable, and joints, hand (left/right), and readers were used as within-subject variables. An exchangeable correlation structure was assumed for all models. For the estimation of sensitivity and specificity, binary ratings (0 = negative/1–3 = positive) were considered as the dependent variable, and the value of the gold standard (not diseased/diseased) was considered as the independent variable (34). For assessment of the predictive values, the variables were used vice versa. Accuracy was estimated in a model without covariates, using the variable “agreement between rating and gold standard” as the dependent binary variable. The 95% confidence intervals (95% CIs) for relevant measures considering the estimated correlation structure are presented. The method was used analogously for a separate estimation of the diagnostic measures for each joint. Due to the smaller sample size, separate estimations of sensitivity and specificity for each joint assessed by each reader were conducted, using the method proposed by Drake and Levine (35), considering the correlation between the left and right hands in each patient.

To evaluate agreement of optical imaging ratings made by different readers, a weighted kappa coefficient (36) based on the ordinal ratings (scale 0–3) was estimated for each pair of readers. The mean, minimum, and maximum of the pairwise kappa coefficients and the proportions of concordant dichotomized ratings (not diseased = 0/diseased ≥1) were presented. Intrareader agreement was estimated analogously using repeated ratings made by the same investigator (CS). Kappa values were estimated as a measure of agreement between the optical imaging findings for each reader and clinical investigation, accounting for the amount of agreement expected by chance. Statistical analyses were performed using R program statistical software (R Foundation) with the additional libraries “vcd” (37) and “gee” (38).

RESULTS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES

Patients and clinical examination.

We examined a total of 1,350 individual joints in 45 patients (30 women [67%], mean ± SD age 52.6 ± 13.4 years). No adverse events were observed in these 45 patients after injection of ICG or gadopentetate dimeglumine. On average, patients had 3.3 (95% CI 2.2–4.5) swollen and 7.5 (95% CI 5.5–9.5) tender joints. Among the patients, 25 (56%) had RA, 9 (20%) had undifferentiated arthritis, 4 (9%) had OA, 3 (7%) had PsA, 3 (7%) had SpA, and 1 (2%) had a connective tissue disorder. The mean serum CRP level was 2.4 mg/dl (95% CI 1.5–3.2 mg/dl), and the mean ESR (1 hour/2 hours) was 20/34 mm (95% CI 14/27–26/42).

When compared with MRI as the gold standard, clinical examination displayed a sensitivity of 27.4% (95% CI 20.4–35.7%), a specificity of 93.5% (95% CI 87.4–96.8%), a positive predictive value of 79.2% (95% CI 64.2–89.1%), a negative predictive value of 61.5% (95% CI 55.9–66.8%), and accuracy of 63.3% (95% CI 58.2–68.1%) for the detection of joint inflammation. The probability of concordant findings by optical imaging and clinical examination was estimated to be 71.5% (95% CI 66.8–75.7%). Kappa values of 0.12, 0.15, and 0.19 were observed, indicating low agreement between optical imaging findings and clinical examination.

Qualitative optical imaging results.

After injection of ICG, an almost immediate increase in signal intensity was observed, which peaked ∼60 seconds postinjection and then slowly declined. After a 6-minute observation period, the signal intensity was ∼30% of the maximum intensity. Inflamed joints appeared with a 10–100-fold greater signal intensity compared with the surrounding, normal joints (Figures 1–3). Consistent hyperintensity was visualized in the distal phalanx around the area of the nailbed as well as in the first DIP joint.

Interreader and intrareader agreement.

The proportions of concordant optical imaging ratings in a pairwise comparison between readers were 79.0%, 80.8%, and 83.0%. The mean weighted kappa coefficient for pairwise assessment of interreader agreement was 0.473. The minimum and maximum estimated kappa values between pairs of readers were 0.411 and 0.554, respectively. For the carpal and MCP joints, agreement was better (mean kappa values 0.332–0.656) than that for the PIP and DIP joints (mean kappa values 0.096–0.537). For intrareader agreement of repeated measurement by one reader (CS), a kappa coefficient of 0.507 was estimated, and concordance was observed in 86.4% of the ratings. Thus, according to kappa values (39) and proportions of concordant ratings (40), interreader and intrareader agreement was moderate.

Optical imaging versus MRI.

Of the 1,350 joints of 45 patients evaluated, MRI showed 608 inflamed joints (45.0%), among which 83.8% were mildly inflamed, 14.1% were moderately inflamed, and 2.1% were severely inflamed. With optical imaging, an average of 356 joints (23.3%) with synovitis were observed, among which 59.6% were mildly inflamed, 24.0% were moderately inflamed, and 16.3% were severely inflamed. Using MRI as the standard of reference, optical imaging had a sensitivity of 39.6% (95% CI 31.1–48.7%), a specificity of 85.2% (95% CI 79.5–89.5%), a positive predictive value of 68.7% (95% CI 57.7–77.9%), a negative predictive value of 63.4% (95% CI 56.8–69.4%), and accuracy of 67.0% (95% CI 61.4–72.1%) for the detection of synovitis in the hands of patients with arthritis (Table 1 and Figure 4). For optical imaging in total, there were 777 true-positive, 1,936 true-negative, 290 false-positive, and 1,047 false-negative ratings among 4,050 evaluations by 3 readers, using MRI as the standard of reference.

Table 1. Sensitivity, specificity, predictive values, and accuracy of optical imaging compared with MRI*
JointNo. pos./no. neg.Sensitivity (95% CI)Specificity (95% CI)PPV (95% CI)NPV (95% CI)Accuracy (95% CI)
  • *

    Optical imaging was used to detect joint inflammation in both hands of 45 patients (1,350 joints). 95% CI = 95% confidence interval; PPV = positive predictive value; NPV = negative predictive value; MCP1 = first metacarpophalangeal joint; IP1 = first interphalangeal joint; PIP2 = second proximal interphalangeal joint; DIP2 = second distal interphalangeal joint.

  • Number of joints scored as inflamed (positive) or as not inflamed (negative) by magnetic resonance imaging (MRI).

Carpal85/540.8 (29.9–52.6)60.3 (40.2–77.4)94.3 (85.0–97.9)5.5 (1.8–15.2)44.8 (34.0–56.2)
MCP184/625.1 (15.9–37.2)59.0 (35.9–78.7)89.2 (73.9–96.0)5.9 (1.9–16.6)31.1 (21.2–43.0)
MCP282/838.0 (27.5–49.7)72.8 (29.9–94.4)92.3 (75.7–97.9)9.5 (3.7–22.2)41.5 (31.3–52.5)
MCP378/1241.8 (30.6–54.0)97.6 (29.0–100)95.6 (76.8–99.3)18.0 (9.2–32.1)45.2 (34.2–56.7)
MCP471/1925.5 (15.7–38.7)96.2 (76.1–99.5)95.8 (72.0–99.5)24.9 (15.2–38.1)41.9 (31.0–53.5)
MCP570/2033.4 (23.2–45.5)92.6 (70.9–98.5)92.3 (76.2–97.8)27.2 (17.1–40.4)47.0 (36.3–58.0)
IP126/6448.5 (33.1–64.2)91.8 (86.4–95.2)32.3 (19.0–51.0)72.4 (58.9–82.8)80.0 (71.2–86.6)
PIP229/6171.7 (54.5–84.3)81.2 (72.2–87.7)36.5 (23.8–51.4)71.2 (57.6–81.8)79.3 (71.2–85.5)
PIP329/6161.2 (42.7–77.0)81.5 (73.1–87.8)41.0 (27.2–56.4)73.2 (60.6–83.0)78.1 (70.0–84.6)
PIP423/6776.5 (57.8–88.6)83.0 (73.5–89.7)35.8 (21.9–52.7)81.8 (69.1–90.0)82.2 (74.6–87.9)
PIP519/7179.2 (54.4–92.4)82.8 (72.9–89.6)21.7 (12.1–35.8)79.5 (65.9–88.6)84.1 (75.8–89.9)
DIP23/8740.7 (29.1–53.5)82.1 (73.5–88.4)6.8 (2.2–18.8)98.2 (94.5–99.5)83.0 (75.0–88.8)
DIP32/8853.6 (46.3–60.7)88.2 (81.1–92.9)8.2 (2.0–28.5)99.3 (96.4–99.9)88.5 (81.8–93.0)
DIP44/8652.0 (20.7–81.8)89.9 (83.3–93.9)13.9 (3.4–42.4)97.7 (92.3–99.3)89.3 (82.9–93.4)
DIP53/8754.7 (18.8–86.3)89.4 (83.3–93.4)9.3 (1.6–38.9)98.2 (93.2–99.5)88.9 (82.6–93.1)
thumbnail image

Figure 4. Sensitivity and specificity of optical imaging for the detection of joint inflammation, using magnetic resonance imaging as the standard of reference. Each panel shows the data obtained by an individual reader. Sensitivity and specificity values are presented only for joints for which sufficient data were available (for sensitivity, ≥10 diseased joints; for specificity, 10 healthy joints). C = carpal; MCP1 = first metacarpophalangeal joint; IP1 = first interphalangeal joint; PIP2 = second proximal interphalangeal joint; DIP2 = second distal interphalangeal joint; 95% CI = 95% confidence interval.

Download figure to PowerPoint

Subgroup analysis.

Carpal/MCP versus PIP/DIP joints.

When the carpal and MCP joints and the PIP and DIP joints were analyzed separately, the sensitivity and specificity of optical imaging were 35.3% and 93.3%, respectively, for the carpal and MCP joints and 60.4% and 85.2%, respectively, for the PIP and DIP joints.

Diagnostic ability of optical imaging for the detection of moderate-to-severe lesions.

In order to determine the diagnostic ability of the optical imaging system for the detection of moderate-to-severe lesions as scored on MRI, 2 subgroups were established according to the following disease classification: joints with an MRI synovitis score of ≥2 (n = 98) or an MRI synovitis score of 3 (n = 11) were classified as inflamed. When comparing the subgroup “MRI synovitis score ≥2” (n = 98) with the group with optical imaging scores of ≥2, sensitivity decreased slightly to 31.4% (95% CI 15.0–54.4%), while specificity and accuracy improved to 91.3% (95% CI 86.8–94.4%) and 88.1% (95% CI 82.9–91.9%), respectively (Table 2). When comparing the subgroup “MRI synovitis score = 3” (n = 11) with the group with optical imaging scores of 3, the sensitivity, specificity, and accuracy improved to 67.0% (95% CI 25.1–92.5%), 96.2% (95% CI 93.8–97.7%), and 95.9% (95% CI 93.5–97.5%), respectively (Table 2).

Table 2. Sensitivity and specificity of OI for detecting inflammation compared with MRI, according to synovitis scores*
 No. pos./ no. neg.Sensitivity (95% CI)Specificity (95% CI)
  • *

    95% CI = 95% confidence interval; OI = optical imaging.

  • Number of joints scored as inflamed (positive) or as not inflamed (negative) by magnetic resonance imaging (MRI).

MRI score ≥1608/42  
 OI score ≥1 39.6 (31.1–48.7)85.2 (79.5–89.5)
 OI score ≥2 13.8 (10.8–17.6)92.5 (87.6–95.5)
 OI score ≥3 5.6 (3.6–8.5)97.0 (94.0–98.5)
MRI score ≥298/1,252  
 OI score ≥1 50.0 (34.7–65.4)75.6 (70.1–80.4)
 OI score ≥2 31.4 (15.0–54.4)91.3 (86.8–94.4)
 OI score ≥3 13.4 (4.5–33.5)96.7 (93.9–98.2)
MRI score = 311/1,339  
 OI score ≥1 89.6 (57.9–98.2)74.2 (68.7–79.0)
 OI score ≥2 83.1 (44.9–96.7)90.0 (86.0–92.9)
 OI score ≥3 67.0 (25.1–92.5)96.2 (93.8–97.7)
Patients with RA.

In the 750 joints of 25 patients with a diagnosis of RA, optical imaging had a sensitivity of 42.5% (95% CI 30.6–55.3%), a specificity of 83.1% (95% CI 73.2–89.8%), a positive predictive value of 69.6% (95% CI 53.3–82.2%), a negative predictive value of 61.7% (95% CI 52.4–70.3%), and accuracy of 67.1% (95% CI 59.4–74.0%) for the detection of synovitis.

MRI of bone erosions, bone edema, and tenosynovitis.

The mean bone erosion score (possible total score 300) was 8.78 (95% CI 6.0–11.5) and ranged from a minimum score of 0 to a maximum score of 35. The mean bone edema score (possible total score 90) was 0.96 (95% CI 0.36–1.55) and ranged from a minimum score of 0 to a maximum score of 10. The mean tenosynovitis score (possible total score 30) was 4.5 (95% CI 3.2–5.9) and ranged from a minimum score of 0 to a maximum score of 19.

DISCUSSION

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES

In this study, we demonstrated that the optical imaging system detected synovitis in the hands of patients with rheumatologic disorders with a sensitivity of 39.6% (95% CI 31.1–48.7%) and a specificity of 85.2% (95% CI 79.5–89.5%) compared with MRI. When analyzing only severe synovitis, optical imaging detected synovitis with an improved sensitivity of 67.0% (95% CI 25.1–92.5%) and specificity of 96.2% (95% CI 93.8–97.7%). Thus, although optical imaging was able to detect severe synovitis in a limited number of joints, it was unable to reliably detect joints with mild-to-moderate synovitis. This is certainly a limitation of this technique, which might be overcome in the future by further technical development. Furthermore, we are aware that by using MRI as the standard, we are comparing optical imaging with an established imaging modality associated with decades of clinical use and optimization of imaging techniques and contrast agents. Increasing experience with optical imaging may lead to improvements in interpretation and thus sensitivity. Adding optical imaging would appear to increase the very low sensitivity of physical examination (27.4% [95% CI 20.4–35.7%) for the detection of inflamed joints; the low accuracy of clinical evaluation was most probably attributable to the available study population, with its many mildly inflamed joints.

The results for interreader and intrareader agreement suggest that the basic interpretation of optical images is moderately consistent among readers with different medical experience. Further training and standardization of grading schemes might improve intergroup reproducibility for assessing synovitis, e.g., for multicenter studies using optical imaging.

Our results are in contrast with those observed by Werner et al in a study comparing the diagnostic ability of the Xiralite X4 optical imaging system to detect synovitis with that of clinical examination, articular ultrasonography, and 1.5T MRI (41). That study showed high agreement rates for optical imaging and clinical examination, MRI, and ultrasonography. With MRI used as the standard of reference, the optical imaging system showed a sensitivity of 76% and a specificity of 54% using different analysis parameters (41). Our results do not support such a high correlation. It is conceivable that our use of high-field MRI (3T versus 1.5T in the study by Werner et al) resulted in improved detection of less severe joint inflammation not visible on 1.5T MRI. However, when our findings for severe disease were compared with those reported by Werner et al, the results were consistent.

For this study, we used ICG as a contrast agent for optical imaging. It is among the few fluorescent dyes with FDA approval, albeit for nonimaging applications, and thus is currently the only choice for use in human trials. Nevertheless, the nonspecific dye has numerous limitations, because it is fluorescent only in its free and unbound state, but ∼98% of it is bound to plasma proteins within seconds after injection. Other nonapproved dyes have been used for the detection of arthritis in preclinical models (42). In particular, SIDAG is a carbocyanine dye similar to ICG but with highly hydrophilic properties. Only ∼10% of the dye binds to plasma proteins, while the remaining 90% potentially is able to extravasate into tissue. SIDAG is eliminated via the kidneys and was shown to provide improved enhancement of arthritic joints in an animal model when compared with ICG (15). Additionally, targeted contrast agents may provide more specific enhancement of diseased joints. In a mouse model of preclinical arthritis, a near-infrared fluorescence-labeled folate probe was shown to preferentially accumulate in activated macrophages in arthritic joints (12). These contrast agents are not FDA approved; however, the use of more specialized contrast agents may aid in improving detection of early joint inflammation.

In addition to optimization of the contrast agent, improvement of sensitivity will be facilitated by enhancing the design of imaging devices and using novel imaging methods. The optical imaging device we used obtained images of the hands from above and thus did not include information for the lateral or palmar aspects of inflamed joints. A commercially available 3-D optical tomography system has shown the response to disease-modifying antirheumatic drug treatment in an animal model of arthritis (43), and it would seem promising that the use of data sets obtained by 3-D imaging would increase sensitivity in the clinical setting as well. Novel fluorescence imaging techniques such as fluorescence lifetime imaging and spectral imaging, if applied in the macroscopic setting, may assist in better detecting fluorescent contrast agent and thus inflammation (44, 45).

Optical imaging maintains several advantages when compared with other standard imaging modalities. MRI can provide highly detailed anatomic imaging but requires long examination times and is costly. Our data-acquisition protocol obtained optical imaging data over the course of 6 minutes, using a small-sized machine. Ultrasonography has shown accuracy similar to that of MRI in detecting synovitis, tendon pathologies, and bony erosions in finger joints and can be performed in the office setting (10). However, ultrasonography typically displays interoperator variability, and, if many joints are to be assessed, examination times become lengthy (46). Like MRI and ultrasonography, optical imaging does not use ionizing radiation and displays device acquisition and maintenance costs comparable with those of lower-tier ultrasonography devices.

We acknowledge the limitations of our study, which assessed the sensitivity and specificity of the optical imaging system at one time point. However, the significance of imaging in RA involves not only diagnosis but also the monitoring of disease progression and therapy. Although our results are still suboptimal in the case of mild-to-moderate joint inflammation, the diagnostic accuracy for assessing severe lesions is acceptable compared with MRI. Because these lesions would probably be chosen as marker lesions for response evaluations, optical imaging could be helpful for serial examinations to assess the response to targeted treatment with antirheumatic drugs. Thus, future studies will need to assess the value of the optical imaging system for long-term monitoring of therapy. Another limitation, which is a fundamental difficulty in the design and interpretation of imaging studies, is the absence of a “true” gold standard such as biopsy and histology for the definitive diagnosis of synovitis. The concern with using MRI as the standard of reference might be the potential for false-positive or false-negative interpretations. However, MRI is an established imaging method, and, if the images are evaluated by experienced readers, it offers high sensitivity and specificity for the detection of synovitis (8, 9). Nevertheless, future studies will have to be undertaken in healthy volunteers in order to exclude false-positive or false-negative interpretations of results obtained by both MRI and optical imaging. Finally, the study was limited by the inhomogeneous distribution of many mildly inflamed joints and comparably low numbers of moderately and severely inflamed joints.

In summary, our data demonstrated that the evaluated optical imaging system had limitations in terms of the ability to detect synovitis in patients with arthritis of the hands in comparison with 3T MRI as the standard of reference. Further optimizations will be required before this imaging modality can be recommended as a diagnostic test for arthritis of the joints of the hand.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Meier had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Meier, Thürmel, Moog, Noël, Ahari, Sievert, Dorn, Waldt, Haller, Woertler, Rummeny.

Acquisition of data. Meier, Thürmel, Moog, Noël, Ahari, Sievert, Dorn, Ganter.

Analysis and interpretation of data. Meier, Thürmel, Moog, Noël, Ahari, Sievert, Dorn, Waldt, Schaeffeler, Golovko, Haller, Weckbach, Woertler.

REFERENCES

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES
  • 1
    Prevalence of doctor-diagnosed arthritis and arthritis-attributable activity limitation: United States, 2007-2009. MMWR Morb Mortal Wkly Rep 2010; 59: 12615.
  • 2
    Gabriel SE. The epidemiology of rheumatoid arthritis. Rheum Dis Clin North Am 2001; 27: 26981.
  • 3
    Silman AJ, Pearson JE. Epidemiology and genetics of rheumatoid arthritis. Arthritis Res 2002; 4 Suppl 3: S26572.
  • 4
    Kim JM, Weisman MH. When does rheumatoid arthritis begin and why do we need to know? [review]. Arthritis Rheum 2000; 43: 47384.
  • 5
    Quinn MA, Emery P. Window of opportunity in early rheumatoid arthritis: possibility of altering the disease process with early intervention. Clin Exp Rheumatol 2003; 21: S1547.
  • 6
    Wilke WS, Sweeney TJ, Calabrese LH. Early, aggressive therapy for rheumatoid arthritis: concerns, descriptions, and estimate of outcome. Semin Arthritis Rheum 1993; 23: 2641.
  • 7
    Backhaus M, Kamradt T, Sandrock D, Loreck D, Fritz J, Wolf KJ, et al. Arthritis of the finger joints: a comprehensive approach comparing conventional radiography, scintigraphy, ultrasound, and contrast-enhanced magnetic resonance imaging. Arthritis Rheum 1999; 42: 123245.
  • 8
    Terslev L, Torp-Pedersen S, Savnik A, von der Recke P, Qvistgaard E, Danneskiold-Samsoe B, et al. Doppler ultrasound and magnetic resonance imaging of synovial inflammation of the hand in rheumatoid arthritis: a comparative study. Arthritis Rheum 2003; 48: 243441.
  • 9
    Szkudlarek M, Klarlund M, Narvestad E, Court-Payen M, Strandberg C, Jensen KE, et al. Ultrasonography of the metacarpophalangeal and proximal interphalangeal joints in rheumatoid arthritis: a comparison with magnetic resonance imaging, conventional radiography and clinical examination. Arthritis Res Ther 2006; 8: R52.
  • 10
    Brown AK. Using ultrasonography to facilitate best practice in diagnosis and management of RA. Nat Rev Rheumatol 2009; 5: 698706.
  • 11
    Wakefield RJ, O'Connor PJ, Conaghan PG, McGonagle D, Hensor EM, Gibbon WW, et al. Finger tendon disease in untreated early rheumatoid arthritis: a comparison of ultrasound and magnetic resonance imaging. Arthritis Rheum 2007; 57: 115864.
  • 12
    Chen WT, Mahmood U, Weissleder R, Tung CH. Arthritis imaging using a near-infrared fluorescence folate-targeted probe. Arthritis Res Ther 2005; 7: R3107.
  • 13
    Wunder A, Tung CH, Muller-Ladner U, Weissleder R, Mahmood U. In vivo imaging of protease activity in arthritis: a novel approach for monitoring treatment response. Arthritis Rheum 2004; 50: 245965.
  • 14
    Meier R, Krug C, Golovko D, Boddington S, Piontek G, Rudelius M, et al. Indocyanine green–enhanced imaging of antigen-induced arthritis with an integrated optical imaging/radiography system. Arthritis Rheum 2010; 62: 23227.
  • 15
    Fischer T, Gemeinhardt I, Wagner S, Stieglitz DV, Schnorr J, Hermann KG, et al. Assessment of unspecific near-infrared dyes in laser-induced fluorescence imaging of experimental arthritis. Acad Radiol 2006; 13: 413.
  • 16
    Fischer T, Ebert B, Voigt J, Macdonald R, Schneider U, Thomas A, et al. Detection of rheumatoid arthritis using non-specific contrast enhanced fluorescence imaging. Acad Radiol 2010; 17: 37581.
  • 17
    Hansch A, Frey O, Hilger I, Sauner D, Haas M, Schmidt D, et al. Diagnosis of arthritis using near-infrared fluorochrome Cy5.5. Invest Radiol 2004; 39: 62632.
  • 18
    Hansch A, Frey O, Sauner D, Hilger I, Haas M, Malich A, et al. In vivo imaging of experimental arthritis with near-infrared fluorescence. Arthritis Rheum 2004; 50: 9617.
  • 19
    Simon GH, Daldrup-Link HE, Kau J, Metz S, Schlegel J, Piontek G, et al. Optical imaging of experimental arthritis using allogeneic leukocytes labeled with a near-infrared fluorescent probe. Eur J Nucl Med Mol Imaging 2006; 33: 9981006.
  • 20
    Scheel AK, Krause A, Mesecke-von Rheinbaben I, Metzger G, Rost H, Tresp V, et al. Assessment of proximal finger joint inflammation in patients with rheumatoid arthritis, using a novel laser-based imaging technique. Arthritis Rheum 2002; 46: 117784.
  • 21
    Ando F, Sasano K, Ohba N, Hirose H, Yasui O. Anatomic and visual outcomes after indocyanine green-assisted peeling of the retinal internal limiting membrane in idiopathic macular hole surgery. Am J Ophthalmol 2004; 137: 60914.
  • 22
    Brancato R, Trabucchi G. Fluorescein and indocyanine green angiography in vascular chorioretinal diseases. Semin Ophthalmol 1998; 13: 18998.
  • 23
    El Desoky A, Seifalian AM, Cope M, Delpy DT, Davidson BR. Experimental study of liver dysfunction evaluated by direct indocyanine green clearance using near infrared spectroscopy. Br J Surg 1999; 86: 100511.
  • 24
    Haruna M, Kumon K, Yahagi N, Watanabe Y, Ishida Y, Kobayashi N, et al. Blood volume measurement at the bedside using ICG pulse spectrophotometry. Anesthesiology 1998; 89: 13228.
  • 25
    Gurfinkel M, Thompson AB, Ralston W, Troy TL, Moore AL, Moore TA, et al. Pharmacokinetics of ICG and HPPH-car for the detection of normal and tumor tissue using fluorescence, near-infrared reflectance imaging: a case study. Photochem Photobiol 2000; 72: 94102.
  • 26
    Haglund MM, Berger MS, Hochman DW. Enhanced optical imaging of human gliomas and tumor margins. Neurosurgery 1996; 38: 30817.
  • 27
    Ntziachristos V, Yodh AG, Schnall M, Chance B. Concurrent MRI and diffuse optical tomography of breast after indocyanine green enhancement. Proc Natl Acad Sci U S A 2000; 97: 276772.
  • 28
    Reynolds JS, Troy TL, Mayer RH, Thompson AB, Waters DJ, Cornell KK, et al. Imaging of spontaneous canine mammary tumors using fluorescent contrast agents. Photochem Photobiol 1999; 70: 8794.
  • 29
    Mordon S, Devoisselle JM, Soulie-Begu S, Desmettre T. Indocyanine green: physicochemical factors affecting its fluorescence in vivo. Microvasc Res 1998; 55: 14652.
  • 30
    Meijer DK, Weert B, Vermeer GA. Pharmacokinetics of biliary excretion in man. VI. Indocyanine green. Eur J Clin Pharmacol 1988; 35: 295303.
  • 31
    Ostergaard M, Peterfy C, Conaghan P, McQueen F, Bird P, Ejbjerg B, et al. OMERACT Rheumatoid Arthritis Magnetic Resonance Imaging Studies: core set of MRI acquisitions, joint pathology definitions, and the OMERACT RA-MRI scoring system. J Rheumatol 2003; 30: 13856.
  • 32
    Ostergaard M, McQueen F, Wiell C, Bird P, Boyesen P, Ejbjerg B, et al. The OMERACT psoriatic arthritis magnetic resonance imaging scoring system (PsAMRIS): definitions of key pathologies, suggested MRI sequences, and preliminary scoring system for PsA hands. J Rheumatol 2009; 36: 181624.
  • 33
    Liang KY, Zeger SL. Longitudinal data analysis using generalized linear models. Biometrika 1986; 73: 1322.
  • 34
    Smith P, Hadgu A. Sensitivity and specificity for correlated observations. Stat Med 1992; 11: 15039.
  • 35
    Drake C, Levine RA. Sensitivity, specificity and other diagnostic measures with multiple sites per unit. Contemp Clin Trials 2005; 26: 2529.
  • 36
    Fleiss JL, Levin B, Paik MC. Statistical methods for rates and proportions. 3rd ed. New York: John Wiley & Sons; 2003.
  • 37
    Meyer D, Zeileis A, Hornik K. vcd: Visualizing categorical data. R package version 1. 2-9; 2010.
  • 38
    Carey VJ. Ported to R by Ripley B. gee: Generalized estimation equation solver. R package version 4.13-16, 2010.
  • 39
    Byrt T, Bishop J, Carlin JB. Bias, prevalence and kappa. J Clin Epidemiol 1993; 46: 4239.
  • 40
    Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977; 33: 15974.
  • 41
    Werner SG, Langer HE, Ohrndorf S, Bahner M, Schott P, Schwenke C, et al. Inflammation assessment in patients with arthritis using a novel in vivo fluorescence optical imaging technology. Ann Rheum Dis 2012; 71: 50410.
  • 42
    Golovko D, Meier R, Rummeny E, Daldrup-Link H. Optical imaging of rheumatoid arthritis. Int J Clin Rheumatol 2011; 6: 6775.
  • 43
    Peterson JD, Labranche TP, Vasquez KO, Kossodo S, Melton M, Rader R, et al. Optical tomographic imaging discriminates between disease-modifying anti-rheumatic drug (DMARD) and non-DMARD efficacy in collagen antibody-induced arthritis. Arthritis Res Ther 2010; 12: R105.
  • 44
    Berezin MY, Achilefu S. Fluorescence lifetime measurements and biological imaging. Chem Rev 2010; 110: 264184.
  • 45
    Levenson RM. Spectral imaging and pathology: seeing more. Lab Med 2004; 35: 24451.
  • 46
    Delle SA, Riente L, Bombardieri S. Limits and perspectives of ultrasound in the diagnosis and management of rheumatic diseases. Mod Rheumatol 2008; 18: 12531.