Arthritis & Rheumatism (Arthritis Care & Research) Vol. 57, No. 3, April 15, 2007, pp 461– 465 DOI 10.1002/art.22612 © 2007, American College of Rheumatology
ORIGINAL ARTICLE
Computed Tomography Measurement of Tophus Volume: Comparison With Physical Measurement NICOLA DALBETH, BARNABY CLARK, KATE GREGORY, GREGORY D. GAMBLE, ANTHONY DOYLE, AND FIONA M. MCQUEEN
Objective. Computed tomography (CT) has high accuracy for tophus detection. This study assessed reliability of CT measurement of tophus volume and compared reproducibility of CT with physical measurement of tophus size. Methods. Forty-seven hand tophi were analyzed in 20 patients with gout. The longest tophus diameter was recorded by 2 independent observers. All patients proceeded to CT scanning of the hands on a Philips Brilliance scanner (0.8-mm slices). Two independent observers measured tophus volume using the Surface Shaded Display 3-dimensional function on the Philips CT workstation. Five patients underwent repeat physical and CT assessments within 1 week (18 observations). Inter- and intraobserver reproducibility were analyzed by limits of agreement and coefficients of variation. Results. Of the 47 lesions identified as tophi on physical examination, 42 (89%) were also identified on CT. The mean (95% confidence interval [95% CI]) difference between observers for physical measurement was 0.45 mm (ⴚ4.07, 4.96) and for CT was 65.2 mm3 (ⴚ293.0, 423.3). The mean (95% CI) difference between visits for physical measurement was ⴚ0.72 mm (ⴚ5.47, 4.03) and for CT was ⴚ13.1 mm3 (ⴚ112.5, 86.3). There was no difference between coefficients of variation for inter- and intraobserver reproducibility for the 2 measurement techniques. For tophi identified by physical and CT assessment, there was good correlation between measurements (r ⴝ 0.91, P < 0.0001). Conclusion. CT assessment of tophus volume is reliable and reproducible. However, physical measurement correlates well with CT and has equivalent reproducibility. These data support the use of physical measurement as a simple and reliable method to assess tophus size. KEY WORDS. Gout; Computed tomography; Tophus.
INTRODUCTION Chronic gout is characterized by the presence of tophi, collections of monosodium urate crystals surrounded by chronic inflammatory cells including giant cells, and connective tissue. Tophi typically occur in both subcutaneous tissues and within affected joints, and may cause pain, cosmetic problems, mechanical obstruction of joint movement, and joint destruction. Therefore, assessment of tophus size is an important outcome in clinical practice and Supported by the Auckland Regional Rheumatology Trust. Dr. Dalbeth’s work was funded by the Health Research Council of New Zealand. Ms Gregory’s work was funded by a Maurice and Phyllis Paykel Summer studentship. Nicola Dalbeth, MD, FRACP, Barnaby Clark, MBChB, Kate Gregory, Gregory D. Gamble, MSc, Anthony Doyle, FRANZCR, Fiona M. McQueen, MD, FRACP: University of Auckland, Auckland, New Zealand. Address correspondence to Nicola Dalbeth, MD, FRACP, Department of Medicine, Faculty of Medical and Health Sciences, University of Auckland, 85 Park Road, Grafton, Auckland, New Zealand. E-mail:
[email protected]. Submitted for publication June 16, 2006; accepted in revised form August 18, 2006.
in studies of chronic gout, particularly in trials of uratelowering therapies. The importance of tophus size assessment as an outcome measure has been emphasized by the inclusion of this outcome in the Outcome Measures in Rheumatology Clinical Trials VII core set for chronic gout therapy (1). A number of different methods of tophus size assessment have been described, including physical measurement, ultrasound assessment, and magnetic resonance imaging (MRI) (2– 6). Previous studies have demonstrated that computed tomography (CT) has excellent accuracy in detecting the presence of tophi in patients with gout (7– 10). However, the reliability of CT to assess tophus size as a potential clinical outcome measure has not been described. Several methods of physical assessment of tophus size have been reported. Perez-Ruiz and colleagues described measurement of the longest diameter of index tophi by calipers in a study of benzbromarone and allopurinol (2). Using this method, their group demonstrated a linear relationship between mean serum uric acid during uratelowering therapy and velocity of tophus diameter reduction. Recently, Schumacher and colleagues have described the reliability of measuring tophus area by physical mea461
462 surement of 2 axes using a tape measure (3). This method was found to have good intra- and interobserver reproducibility. This method has also been used in the Febuxostat versus Allopurinol Controlled Trial (FACT) study, with a recent report demonstrating that febuxostat therapy was superior to allopurinol in achieving tophus regression over a 1-year period (11). To date, no study has compared physical measurement with imaging methods. The goal of the present study was to determine whether CT scanning can reliably assess tophus size. In addition, we compared the reliability of physical measurement with CT assessment.
PATIENTS AND METHODS Twenty patients with chronic gout were recruited from rheumatology outpatient clinics at Auckland District Health Board and Counties Manukau District Health Board, Auckland, New Zealand from November 2005 to February 2006. The inclusion criteria were a diagnosis of gout as defined by the Wallace et al classification criteria for acute gout (12), ability to provide written informed consent, and ability to fully flex and abduct the shoulders (required for positioning in the CT scanner). Women of childbearing age and patients with acute gout flares were excluded from the study. The Northern Regional Ethics Committee approved this study. All patients provided written informed consent before inclusion in the study. For patients with hand tophi, a maximum of 3 tophi were chosen as index tophi by clinical examination. These tophi were subject to both physical and CT size assessment. When ⬎3 tophi were present, those tophi that could be easily measured because of physical location or size were preferentially chosen for analysis. The location of the index tophi was recorded in detail by both written anatomic descriptions and hand diagrams. These records were used to identify the index tophi for blinded independent physical assessments over time and to identify the index tophi for assessment on CT scanning. All 20 patients proceeded to CT scanning of the hands. Five patients returned within 1 week for repeat physical and CT assessment of tophus size. Physical measurement of the longest tophus diameter was recorded by 2 independent observers (KG and ND) on the same day using 150-mm digital Vernier calipers (Qingda Tlead International Company, Qingda, Shandong, China) according to the method described by Perez-Ruiz et al (2). The borders of the longest diameter were marked by pen, and the diameter was recorded. The pen marks were completely removed before the second independent observer measurements. The CT scans of the hands were performed on a Philips Brilliance 16-slice scanner (Philips Medical Systems, Best, The Netherlands). The scanning range was from the fingertips to 2 cm proximal to the radiocarpal joints in an axial plane. The left and right hands were both imaged during the same acquisition. Patients were positioned prone, shoulders in forward flexion, with the forearm and wrists forward of the body in a neutral position. The palms were opposed and separated by a foam pad. All scans were
Dalbeth et al performed with the same image protocol: acquisition at 16 ⫻ 0.75 mm, reconstructed on a bone algorithm, 768pixel matrix, to 0.8-mm slices with a 0.4-mm increment (kilovoltage peak 140, 120 milliampere seconds/slice). Additional reconstructions were done on a soft tissue algorithm, 512-pixel matrix, also to a 0.8-mm slice with a 0.4-mm increment. The images were viewed as 0.8-mm slices on a Philips CT workstation (Philips Medical Systems) and were reconstructed to 3-mm slices for viewing on Picture Archiving Communication System (Impax version 4.5, Agfa Healthcare, Mortsel, Belgium). An average Hounsfield unit measurement of the density of each tophus was generated on the CT workstation using standard software. A representative slice through the center of the tophus was chosen and the region of interest selected by drawing freehand around the tophus margins. Calcification within the tophus was identified by visualization of the tophus through all axial slices and was confirmed by measuring Hounsfield units of the suspected area of calcification. Hounsfield units for areas of calcification measured ⬎250 in all cases. CT tophus volume was assessed by 2 independent observers (ND and BC) using the Surface Shaded Display 3-dimensional function on the Philips CT workstation. This function allows quantitative assessment of predefined tissues. By drawing freehand around the region of interest (the tophus) on the multiple 2-dimensional slices, a 3-dimensional tophus model with an estimated volume was generated (Figure 1). Intraobserver reproducibility was assessed by comparing differences in measurements between 2 visits. Interobserver reproducibility was assessed by comparing differences between 2 observers at the same time point. Intraand interobserver reproducibility were analyzed by intraclass correlation coefficients, the limits of agreement (13), and coefficients of variation (14). The smallest detectable difference was also calculated (15). For the purposes of these analyses the unit of investigation was assumed to be the tophus. Tophi reproducibility within an individual are assumed to be independent of each other.
RESULTS Patient characteristics. Baseline characteristics of the patients are summarized in Table 1. There were 11 (55%) patients with aspirate-proven gout, and the remaining patients had a clinical diagnosis of gout based on the Wallace et al classification criteria (12). The majority of patients (16 [80%] of 20) had tophaceous disease, and 14 (70%) patients had at least 1 hand tophus. A total of 47 index hand tophi were measured, and 9 tophi were analyzed at 2 time points (18 observations). CT analysis of tophus size. Of the 47 index tophi identified on physical examination, 42 (89%) were also identified on CT. In 5 patients, 5 (11%) tophi that were present on physical examination were not identified on CT scanning. Of these 5 patients, 4 (80%) had tophi identified by physical examination and CT at other sites in the hand. Tophi identified by physical measurement alone occurred
Tophus Measurement in Gout
463 compared with those not detected by CT (16.2 mm versus 14.1 mm; P ⫽ 0.71). The median volume of visible tophi measured by CT was 265 mm3 (range 13– 4,934 mm3). The median Hounsfield units of the tophi was 129.1 (range 68.5–222). There was a moderate correlation between tophus size and Hounsfield units (r ⫽ 0.37, P ⫽ 0.02). Calcification was present in 12 (29%) of 42 tophi identified on CT. Calcification occurred more often in large tophi; the mean ⫾ SD tophus volume for tophi with calcification was 1,959.5 ⫾ 1,649.2 mm3 compared with 435.6 ⫾ 746.1 mm3 for those without calcification (P ⫽ 0.0002). For CT measurement of tophus volume, intraclass correlation coefficient for interobserver reproducibility was 0.989 (95% confidence interval [95% CI] 0.981, 0.994) and for intraobserver reproducibility was 1.000 (95% CI 0.999, 1.000). Further analysis of inter- and intraobserver reproducibility for CT measurements is summarized in Table 2. Bland-Altman plots for comparison between CT observations are shown in Figures 2A and 2B. Physical analysis of tophus size. The median longest diameter by physical assessment was 11.0 mm (range 3.5– 59.1 mm). For physical measurement of tophus diameter, intraclass correlation coefficient for interobserver reproducibility was 0.985 (95% CI 0.972, 0.992) and for intraobserver reproducibility was 0.996 (95% CI 0.989, 0.999). Further analysis of inter- and intraobserver reproducibility for physical measurements is summarized in Table 2. Bland-Altman plots for comparison between physical observations are shown in Figures 2C and 2D. Comparison between physical and CT measurements. There was no consistent difference between the coefficients of variation for inter- and intraobserver reproducibility of the 2 measurement techniques (Table 2). There was more variability in CT volume measurement between different observers, but very little variability within a single observer for CT measurement (Figures 2A and 2B). For tophi that were identified by both physical and CT assessment, there was good correlation between physical measurement of the longest diameter and CT measurement of volume (r ⫽ 0.91, P ⬍ 0.0001).
Figure 1. Computed tomography 3-dimensional reconstruction and measurement of tophus volume. A, The index tophus was outlined by drawing freehand in multiple 2-dimensional axial planes. The axial plane is shown in the main panel, and the coronal and sagittal planes in the smaller panels on the right. B, Three-dimensional reconstruction of the scanning area prior to selection of the tophus. The tophus of interest is evident over the dorsum of the hand. In addition, tophi are observed in the region of the second proximal interphalangeal (PIP) joint, fifth PIP joint, and fifth distal interphalangeal joint. C, Three-dimensional reconstruction of the selected tophus with calculated volume.
in various locations (overlying finger pulp, first interphalangeal joint, proximal interphalangeal joints, and second metacarpophalangeal joint), and there was no difference in tophus location in those detected by CT and those not detected by CT. In addition, there was no difference in mean physical diameter between tophi detected by CT
DISCUSSION This study demonstrated that CT assessment of tophus volume is reliable and reproducible. These findings confirm previous work by Gerster and colleagues that CT is a useful method of tophus detection, both within the joint and in extraarticular locations (8 –10). Gerster et al have also suggested that CT has greater specificity for detection of tophi than other forms of advanced imaging and can differentiate tophi from other subcutaneous nodules, such as nodules associated with rheumatoid arthritis and dyslipidemia (7,9,10). The development of advanced 3-dimensional modeling and volume assessment now allows for CT technology to be used to accurately analyze the size of tophaceous nodules. Our results suggest that CT assessment has good reliabil-
464
Dalbeth et al
Table 1. Clinical characteristics of study participants* All patients (n ⴝ 20)
Variable Age, years Male sex, no. (%) Ethnicity, no. (%) Pacific New Zealand Maori New Zealand European/other Total number of tophi Gout disease duration, years Serum uric acid, mg/dl Patients taking allopurinol, no. (%)
Patients returning for second visit (n ⴝ 5)†
57.5 (38–75) 19 (95)
46 (39–65) 5 (100)
6 (30) 5 (25) 9 (45) 10 (0–244) 17 (1–50) 8.13 (4.23–11.52) 17 (85)
1 (20) 4 (80) 10 (0–93) 14 (11–37) 7.62 (6.61–11.52) 4 (80)
* Values are the median (range) unless otherwise indicated. † P ⬎ 0.34 for all comparisons between the entire group and the returning group.
ity compared with other forms of advanced imaging for tophus size assessment, such as MRI or ultrasound (4,6). In addition to its specificity and reliability, CT scanning has the advantage of storing images for serial assessments. This method allows for rapid scanning of numerous tophi; for example, using the protocol developed for this study, both hands (and all tophi within the scanning field) were imaged in ⬍5 minutes. This short scanning time is of particular benefit for patients with joint pain or deformity. The cost of CT is competitive compared with other methods of advanced imaging for monitoring of tophus size; although prices may vary depending on local availability and imaging protocols, the typical price for CT is US$200 – $300, compared with US$100 –$200 for ultrasound and US$400 –$700 for MRI. However, there are several disadvantages to using CT scanning, particularly the use of ionizing radiation, although positioning away from core structures as used in this study significantly reduces the risks related to radiation exposure. Adequate positioning of the patient within the scanner may be difficult in the presence of significant shoulder and elbow disease. Furthermore, although the scans can be rapidly acquired, 3-dimensional volume assessment can be time consuming. In our study, the mean Hounsfield units in the hand tophi were more variable and generally lower than previously reported within knee tophi (8). The underlying reason for this difference is uncertain, but it is conceivable
that the density of the tophus may vary in relation to size, stage of growth, or location. We did demonstrate a relationship between tophus size and Hounsfield units. Furthermore, there was a strong relationship between calcification within the tophus and tophus size. These findings suggest that larger, more established tophi, as may be found in the knee, have greater density. Furthermore, not all tophi that were identified by clinical examination were recognized on CT. We were unable to identify any difference in size or location between the tophi identified on both physical examination and CT compared with those identified by physical examination alone. This disparity may have occurred because these particular tophi had density similar to soft tissue on CT. Alternatively, it is possible that these nodules identified on physical examination were not in fact tophi, but instead represented other soft tissue or bony pathologies. For example, a firm nodule overlying a joint in a patient with gout may be interpreted as a tophus, but alternatively may be due to an osteophyte. We do not believe that these observations limit the validity of CT measurement of tophus size, because the vast majority of tophi identified on physical examination were easily recognizable by CT. A further goal of this study was to compare the reliability of physical measurement with that of CT assessment. Analysis of precision using coefficients of variation and the smallest detectable difference has demonstrated that
Table 2. Summary of inter- and intraobserver reproducibility for physical and CT measurement of tophus size*
Interobserver reproducibility (assessor 1 vs assessor 2) CT volume, mm3 Physical diameter, mm Intraobserver reproducibility (visit 1 vs visit 2) CT volume, mm3 Physical diameter, mm
SDD
ICC, mean (95% CI)
Limits of agreement, mean ⴞ SD
CV, % mean (95% CI)
772.6 ⫾ 1,218.1 16.1 ⫾ 13.3
358.1 4.5
0.989 (0.981, 0.994) 0.985 (0.972, 0.992)
65.2 ⫾ 182.7 0.45 ⫾ 2.30
16.7 (13.8, 21.2) 10.2 (8.5, 12.9)
1,042.9 ⫾ 1,675.8 21.1 ⫾ 18.4
99.4 4.7
Measurement, mean ⴞ SD
1.000 (0.999, 1.000) ⫺13.1 ⫾ 50.7 0.996 (0.989, 0.999) ⫺0.72 ⫾ 2.42
3.4 (2.5, 5.3) 8.1 (6.1, 12.2)
* CT ⫽ computed tomography; SDD ⫽ smallest detectable difference; ICC ⫽ intraclass correlation coefficient; CV ⫽ coefficient of variation; 95% CI ⫽ 95% confidence interval.
Tophus Measurement in Gout
465 Study design. Dalbeth, Clark, Gregory, Doyle, McQueen. Acquisition of data. Dalbeth, Clark, Gregory, Doyle. Analysis and interpretation of data. Dalbeth, Clark, Gamble, Doyle, McQueen. Manuscript preparation. Dalbeth, Clark, Gregory, Doyle, McQueen. Statistical analysis. Dalbeth, Gamble.
REFERENCES
Figure 2. Bland-Altman plots of difference. A, Difference between observers for computed tomography (CT) measurement of tophus volume (mm3). B, Difference between visit 1 and visit 2 for CT measurement of tophus volume (mm3). C, Difference between observers for physical measurement of longest tophus diameter (mm). D, Difference between visit 1 and visit 2 for physical measurement of longest tophus diameter (mm). Solid horizontal lines represent the mean difference; broken horizontal lines represent 2 SDs of difference from the mean.
physical measurement of tophus diameter is also reliable and reproducible. This method has been shown to discriminate between effective treatments over time. Physical assessment of tophus size does have some disadvantages, particularly the inability to store images for later reassessment in a serial manner. Unlike advanced imaging techniques, this method only allows for assessment of tophi located in subcutaneous or superficial locations. However, in addition to reliability and reproducibility, physical measurement has a number of other advantages, particularly time benefits, cost effectiveness, and patient acceptability. In summary, CT assessment of tophus volume is a reliable and reproducible method of analyzing tophus size. However, physical measurement of tophus diameter correlates well with CT measurement of tophus volume and has equivalent reproducibility to CT assessment. These data further support the use of physical measurement as a simple, cost-effective, and reliable method to assess the size of tophi in subcutaneous locations, as an outcome measure for clinical studies, and as a clinical practice method for long-term monitoring of patients with tophaceous gout. AUTHOR CONTRIBUTIONS Dr. Dalbeth 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.
1. Schumacher HR Jr, Edwards LN, Perez-Ruiz F, Becker M, Chen LX, Furst DE, et al. Outcome measures for acute and chronic gout [review]. J Rheumatol 2005;32:2452–5. 2. Perez-Ruiz F, Calabozo M, Pijoan JI, Herrero-Beites AM, Ruibal A. Effect of urate-lowering therapy on the velocity of size reduction of tophi in chronic gout. Arthritis Rheum 2002; 47:356 – 60. 3. Schumacher HR Jr, Becker MA, Palo WA, Streit J, MacDonald PA, Joseph-Ridge N. Tophaceous gout: quantitative evaluation by direct physical measurement. J Rheumatol 2005;32: 2368 –72. 4. Perez Ruiz F, Martin JI, Nolla JM. US reliability for the measurement of deep, periarticular tophi in patients with chronic gout [abstract]. Arthritis Rheum 2005;52 Suppl 9:S106. 5. Perez Ruiz F, Martin JI, Canteli B, Nolla Sole JM. Ultrasonography (US) is as useful as Magnetic Resonance Imaging (MRI) for the evaluation of deep periarticular tophi in chronic gout: a comparative study [abstract]. Arthritis Rheum 2005;52 Suppl 9:S106. 6. Schumacher HR Jr, Becker MA, Edwards NL, Palmer WE, MacDonald PA, Palo W, et al. Magnetic resonance imaging in the quantitative assessment of gouty tophi [published erratum appears in Int J Clin Pract 2006;60:630]. Int J Clin Pract 2006; 60:408 –14. 7. Gerster JC, Landry M, Dufresne L, Meuwly JY. Imaging of tophaceous gout: computed tomography provides specific images compared with magnetic resonance imaging and ultrasonography. Ann Rheum Dis 2002;61:52– 4. 8. Gerster JC, Landry M, Duvoisin B, Rappoport G. Computed tomography of the knee joint as an indicator of intraarticular tophi in gout. Arthritis Rheum 1996;39:1406 –9. 9. Gerster JC, Landry M, Rivier G. Computed tomographic imaging of subcutaneous gouty tophi. Clin Rheumatol 1998;17: 62– 4. 10. Gerster JC, Landry M, Rappoport G, Rivier G, Duvoisin B, Schnyder P. Enthesopathy and tendinopathy in gout: computed tomographic assessment. Ann Rheum Dis 1996;55: 921–3. 11. Becker MA, Schumacher HR Jr, Wortmann RL, MacDonald PA, Eustace D, Palo WA, et al. Febuxostat compared with allopurinol in patients with hyperuricemia and gout. N Engl J Med 2005;353:2450 – 61. 12. Wallace SL, Robinson H, Masi AT, Decker JL, McCarty DJ, Yu TF. Preliminary criteria for the classification of the acute arthritis of primary gout. Arthritis Rheum 1977;20:895–900. 13. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307–10. 14. Vangel MG. Confidence intervals for a normal coefficient of variation. Am Stat 1996;50:21– 6. 15. Bruynesteyn K, Boers M, Kostense P, van der Linden S, van der Heijde D. Deciding on progression of joint damage in paired films of individual patients: smallest detectable difference or change. Ann Rheum Dis 2005;64:179 – 82.
