DIAGNOSTIC SENSITIVITY OF BONE SCINTIGRAPHY FOR EQUINE STIFLE DISORDERS SARAH GRAHAM, MAURICIO SOLANO, JAMES SUTHERLAND-SMITH, AMY F SATO , LOUISE MARANDA

Disorders of the stifle are a common cause of lameness in horses yet the accuracy of scintigraphy for diagnosis of stifle conditions is controversial. The aim of retrospective cross-sectional study was to determine the diagnostic sensitivity (Se) of bone scintigraphy in detecting stifle disease and to determine if two orthogonal scintigraphic images improve diagnostic Se. Horses that underwent scintigraphic examination during a two-year period were included. Horses were divided into two groups: group 1 (N = 23) had lameness that was localized to the stifle by intra-articular analgesia and group 2 (N = 182) had lameness that was localized to a different location. Scintigraphic studies (one image or two images) were independently and retrospectively analyzed by two radiologists (R1 and R2). Sensitivity, specificity (Sp) and predictive values (PV), and were calculated for each type of study (one image or two images) and for each radiologist (R1 or R2). The Se to detect stifle disorders varied between radiologists (29.2% and 20.8%). The Sp was 84.5% and 88.3%. When two images were evaluated a decrease in the positive PV for both readers occurred. The Cohen kappa coefficient (κ) between readers was poor when one image (0.084) or two images (0.117) were evaluated. Findings from this study indicated that bone-phase nuclear scintigraphy is reasonably specific but highly insensitive for detecting lameness originating from the stifle in a diverse population of both normal and affected horses. The addition C 2014 American College of a caudal scintigraphic image acquisition did not improve diagnostic sensitivity.  of Veterinary Radiology. Key words: equine lameness, equine stifle, nuclear scintigraphy.

Introduction

from 1986 to present failed to reveal research investigating the sensitivity (Se) and specificity (Sp) of this modality in the horse stifle. The scintigraphic appearance on the lateral image of the normal equine stifle has been described using a quantitative region of interest (ROI) technique.8,9 Although not universally advocated,10 it has been recommended that two orthogonal images of the stifle be obtained in order to improve diagnostic acuity.11 However, it is unknown if adding a second (caudal) image increases the Se or Sp in detecting stifle abnormalities. The aim of this study was to determine the Se and Sp of bone scintigraphy in detecting stifle disease in a mixed population of horses and to determine if adding the caudal image increases that accuracy when compared to lateral images alone. Our hypotheses were that the Se and Sp for detecting various stifle conditions is low, and that providing two orthogonal images would increase the accuracy of this diagnostic modality.

A

CCURACY of nuclear scintigraphy for the diagnosis of equine stifle disorders needs further investigation.1 The majority of available information is limited to case reports studies2–4 focused on traumatic conditions of the area. One standard nuclear medicine textbook5,6 contains a collection of cases demonstrating a subchondral bone cyst, degenerative joint disease, and nonspecific stifle radiopharmaceutical uptake. Another standard nuclear medicine textbook7 devotes a chapter to equine scintigraphy and, while no images of the stifle are presented, there is a discussion of the insensitivity of scintigraphic images for detecting stifle subchondral cyst-like lesions. A PubMed search using the words, equine, nuclear medicine, bone scintigraphy

From the Department of Large Animal Clinical Sciences, College of Veterinary Medicine, University of Florida, 2015 SW 16th Avenue, Gainesville, FL 32608 (Graham); Department of Clinical Sciences, Cummings School of Veterinary Medicine, Tufts University, 200 Westboro Rd, N. Grafton, MA 01536 (Salano, Sutherland-Smith, Sato); Department of Quantitative Health Sciences, University of Massachusetts, Medical School, 55 Lake Ave North, Worcester, MA 01655 (Maranda). Address correspondence and reprint requests to Sarah Graham at the above address. E-mail: [email protected] Received June 18, 2013; accepted for publication April 8, 2014. doi: 10.1111/vru.12184

Vet Radiol Ultrasound, Vol. 56, No. 1, 2015, pp 96–102.

96

VOL. 56, NO. 1

BONE SCINTIGRAPHY FOR EQUINE STIFLE DISORDERS

Materials and Methods Case Selection Medical records were reviewed for horses that underwent bone-phase scintigraphic examinations at Tufts Cummings School of Veterinary Medicine over a 2-year period (2003– 2005). Medical record information included signalment, history, physical, and lameness examination findings, results of intra-articular analgesia, radiographs, ultrasound, arthroscopy, and final diagnosis. Horses were used in the study if a scintigraphic examination of the hindlimbs was performed using a standard technique (described below) and the source of pain causing lameness was localized. Horses were divided into two groups. Group 1 included horses with lameness localized to the stifle and group 2 included horses with lameness localized to a source other than the stifle. Horses were considered to have localized stifle lameness only if there was positive response to intraarticular analgesia of one or more stifle compartments,12 or surgical confirmation of stifle disease.13,14 Horses with clinical signs or diagnostic imaging suggestive of a stifle condition, but not having intra-articular analgesia performed were excluded from the study.

Scintigraphic Technique Each horse received an IV injection of 99m Technetiumhydroxymethylene diphosphonate (Mallinckrodt, St Louis, MO, 10 MBq/kg IV) 2 h prior to imaging. At that time, R horses were also administered furosemide (Salix Intervet, Summit NJ, 1 mg/kg IV). Prior to imaging, all horses were R sedated with detomidine hydrochloride (Dormosedan Phizer, Exton PA, 0.006 mg/kg IV). Fractious horses were R Fort also administered butorphanol tartrate (Torbugesic Dodge, New York NY, 0.1 mg/kg IV). Lateral and caudal images of both stifles were obtained (Fig. 1) as part of a full body or hindlimb scintigraphic study. All scintigrams were acquired using a 55 photomultiplier tube large field of view, rectangular gamma camera fitted with a high resolution general purpose collimator (IS2 Medical Systems, Ottawa, ON). Image processing was performed using standard nuclear medicine software (Segami Corporation, Columbia, MD). Static count-based images were acquired for both lateral (150,000 counts) and caudal (250,000 counts) images of the stifle.

Evaluation of Scintigraphic Studies Scintigraphic images in native interfile format were accessed from the computer database and printed as black dots over a white background on high-quality glossy white paper (Figure 2). The first set of images evaluated included both the caudal and lateral images of both stifles of each horse on a single page. In the second set of images evalu-

97

ated only the lateral images of horses with confirmed stifleassociated lameness and an equal number of randomly assigned normal horses were printed per sheet of paper. All pages were numbered, grouped, and then placed into random order according to a random number generator (http://stattrek.com/Tables/Random.aspx). Two board-certified radiologists (R1 and R2) with experience in routine interpretation of large animal bone scintigrams independently evaluated the scintigraphic images. Details relating to the cases and the study design were not provided. Each image was presented with a corresponding scoring sheet for completion. Radiopharmaceutical uptake was recorded as either normal or abnormal based on subjective image assessment. This assessment relied primarily on the radiologist’s experience gained through their day-today nuclear medicine practice.

Statistical Analysis Each horse was classified as affected (horses with lameness localized to the stifle) and normal (horses with lameness localized to a source other than the stifle). Each joint was considered the unit of analysis and was considered independent from the contralateral stifle in the same animal. This was possible since joint estimations were done in R15 , based on the framework developed by de Leon et al. to evaluate the independence of paired organs.16,17 Comparison of Se, which measures the proportion of actual positives that are correctly identified as positives, Sp, which measures the proportion of negatives that are correctly identified as negatives, and predicted values (PVs) for each stifle and for each radiologist using both one and two images were done using Fisher’s exact test. Positive predictive value (PV) was defined as the proportion of positive test results that are true positives and negative PV was defined as the proportion of negative test results that are truly negative. Associated 95% confidence intervals (95% CIs) were estimated employing the logit back-transformation method, incorporating the so-called Wilson’s continuity correction described by Mercaldo et al.18 All analyses were carried out in SPSS 20.0 (IBM Corporation, Armonk, NY). A value of P < 0.05 was considered significant.

Results Two hundred and five horses were included in the study. The average age was 7.7 years old (median 7 years, range 1–26 years old). There were 58 mares, 128 geldings, 16 stallions, and 3 of unrecorded sex. Breeds included warmblood (71), Thoroughbred (63), Quarter horse (19), Standardbred (12), Morgan (9), Arabian (6), pony (6), draft (5), American Saddlebred (5), American Paint (3), Lipizzaner (2), Paso Fino (1), Appaloosa (1), mixed breed (1), and unknown

98

GRAHAM ET AL.

2015

FIG. 1. For the lateral image (A), the gamma camera was centered in the stifle joint to include distal aspect of the femur, patella, proximal aspect of the tibia and fibula. The image illustrates one of the limitations when imaging the stifle: while every attempt is made to position the camera as close as possible to the skin, the bulge created by the musculature along the lateral proximal aspect of the femur prevents full contact of the detector with the lateral aspect of the joint. This is in spite of the minimal amount of soft tissues present between the lateral aspect of the joint and the detector, In some cases, a lead shield (not pictured here) is placed between the legs to prevent imaging portions of the contralateral stifle. The caudal image (B) included the distal femur and proximal tibia and fibula with the camera angled in a slightly caudomedial to craniolateral direction to mimic the normal orientation of the joint. Similarly to the lateral image, the detector cannot be placed fully against the caudal aspect of the joint as major bony landmarks, such as the ischiatic tuberosity, are protruding against the detector.

breed (1). Twenty-three of the 205 horses were diagnosed with stifle lameness based on a positive response to intraarticular analgesia of the femoropatellar and femorotibial joints. Twenty-two horses exhibit unilateral stifle disease and one horse exhibited bilateral disease for a total of 24 affected stifles. The distribution of final diagnoses per affected stifle, based on followup imaging and surgical findings, was as follows: osteoarthrosis (10/28), subchondral cyst-like lesions (2/28), patellar ligament desmitis (2/28), cranial cruciate ligament desmitis (1/28), synovitis (1/28), and no specific diagnosis (8/28). The values in Table 1 summarize the Se, Sp, and PV per radiologist. For radiologist 1, neither the Se nor the Sp of scintigraphy appeared to change significantly when two images were available for evaluation vs. one image: 29.2%– 29.2% for Se (P = 1.000) and 84.5%–85.5% for Sp (P = 0.813). However, the addition of a second image caused a significant decrease in the positive PV (from 38.9% to 10.4%, P = 0.004), while there was a significant improvement of the negative PV with such an addition (from 79.3%

to 95%, P < 0.001). This pattern was similar for radiologist 2 who had a nonsignificant decrease in Se: 20.8% for two views and 4.2% for one view (P = 0.188). The addition of an orthogonal scintigraphic image worsened Sp nonsignificantly from 93.4% to 88.3% (P = 0.230). This caused a nonsignificant decrease in the positive PV (from 16.7% to 10.0%, P = 1.000), but a significant increase in the negative PV (from 75.5% to 94.7%, P < 0.001).

Discussion The findings of this study show that bone-phase scintigraphy of the stifle has poor Se, which is between 4.2% and 29.2% with one image and 20.8% to 29.2% for two images, for diagnosing abnormality associated with lameness localized to the stifle in a mixed population of horses. However, it has good Sp, between 85.5% and 93.4% with one image and 84.5% to 88.3% for two images. Barr et al.2 reported that the Se of bone scintigraphy in diagnosis meniscal and cruciate

VOL. 56, NO. 1

BONE SCINTIGRAPHY FOR EQUINE STIFLE DISORDERS

99

FIG. 2. Sample set of images presented to the readers for two-image analysis. Top images are craniocaudal images. Bottom images are lateral images. One-image analysis included only lateral images. Note the variation in the pattern of radiopharmaceutical uptake between images. The left-sided images depict better the mayor anatomical landmarks than the right-sided scintigrams. Possible reasons for these variations include, distance from the detector, tissue thickness, motion of the patient, and radiopharmaceutical uptake. These factors further limit the already subjective assessment performed by the readers.

ligament injuries was high (100% for both) and that the Sp of nuclear scintigraphy for identifying these conditions was poor at 25% and 14%, respectively. It is possible that these results are in contrast to ours because the number of cases reported by Barr et al. was small, only affected horses were included and may have been subject to interpretation bias19 as the clinicians were responsible for both examination of the cases and interpretation of the scintigrams. Our results of low Se and high Sp for scintigraphy of the equine stifle are also different from bone scintigraphy of

other regions of the horse, such as the head20 or the distal limbs21,22 where nuclear scintigraphic imaging has been highly sensitive but nonspecific. There are several possible explanations for these conflicting results. The first is that, unlike in other commonly imaged regions, the stifle has a large amount of muscle surrounding it, especially on the caudal aspect of the joint. The increased attenuation of the gamma rays hitting the detector increases the difficulty in discerning subtle differences in count density and there is also a longer target to detector distance, which decreases

100

GRAHAM ET AL.

2015

TABLE 1. Diagnostic sensitivity values calculated for each radiologist when assessing scintigrams for stifle joints in 205 horses Radiologist 1 One Image % [95% CI] Se Sp PV+ PV−

Se Sp PV+ PV−

29.2 [11.0–47.4] 85.5 [77.6–93.4] 38.9 [16.4–61.4] 79.3 [70.5–88.0] One Image % [95% CI] 4.17[−0.38–12.2] 93.4 [87.8–99.0] 16.7[−13.2–46.5] 75.5 [66.8–84.2]

SE 0.093 0.040 0.115 0.045 Radiologist 2 SE 0.041 0.028 0.152 0.044

Two Images % [95% CI]

SE

29.2 [11.0–47.4] 84.5 [80.8–88.1] 10.4 [3.1–17.8] 95.0 [92.7–97.3]

0.093 0.018 0.037 0.012

Two Images % [95% CI] 20.83 [4.6–37.1] 88.3 [85.1–91.5] 10.0 [1.7–18.3] 94.7 [92.4–97.0]

SE 0.083 0.016 0.042 0.012

PV, predictive value; Se, sensitivity; Sp, specificity.

the final resolution of the image. While the lateral aspect of the joint has little soft tissue coverage, the muscles just proximal to the joint space, limit full contact of the camera against the lateral aspect of the joint even when the camera is placed fully against the skin of the horse (Fig. 1). Second, due to the convex shape of the hamstring musculature, the slightly flexed and abducted position of the joint and the fact that major anatomical protuberances are partially if not fully obscured by the soft tissues, correct positioning of the detector is prone to a large amount of variation as was shown recently.23 Third, the distal end of the femur and proximal end of the tibia are thick anatomical landmarks that exhibit the largest amount of radiopharmaceutical uptake. On the lateral images, if there are areas of abnormal activity medially, this thickness further increases the target to detector distance. Moreover, subtle abnormal uptake medially will be most likely obscured by the normal high uptake of the laterally located structures. These factors frequently result in indistinct scintigrams of poor resolution that make the already subjective interpretation of this area more difficult. Fourth, while boney lesions in the stifle are not uncommon,24 the frequency and type of lesions are likely related to study population. The fraction of horses diagnosed with osteoarthrosis was 3% in Jeffcott et al.24 and 0% in Barr et al.2 compared with 41.7% in this study. Therefore, as expected, unless the stifle injury involves a traumatic2 or severe bone lesion3 , bone phase scintigraphy may not yield diagnostic results. Images in the present study were acquired using countbased static image acquisition protocols rather than dynamically acquired protocols with motion correction algorithms. At the authors’ institution, dynamic acquisition of images with motion correction is not applied automatically to the images unless it is perceived that the animal has swayed too much or if the statically acquired image is deemed to have motion blur. Advantages of the statically acquired image protocols include shorter acquisition time and less time spent in postacquisition image processing by the operator. The advantage of a dynamically acquired im-

age includes increased count density, as multiple frames are summed to form a final image of higher resolution and the frames, which are outside a reference point in the image are eliminated.25 Therefore, acquiring images of the stifle in dynamic mode may help overcome the gamma ray attenuation by the muscles in the area of the stifle by increasing the count density. On the other hand, several factors can decrease the quality of a motion-corrected image. A motion correction protocol aligns multiple images in a dynamic series to a point of reference on each of the images. This point of reference is an area with higher counts, usually a large boney protuberance. The type of motion correction varies with the software used. At the authors’ institution, the software (Mirage processing software version 5.407 sp 19e (2005) automatically eliminates images in which their point of reference does not align within certain distance. After this automatic correction is made, the operator can also manually eliminate additional frames. If the degree of motion by the patient is excessive, too many frames are eliminated resulting in an image with adequate count density but of poor resolution. If the amount of gamma rays detected is low, many frames are eliminated since the software cannot effectively identify the point of reference. This can result in an image of low count density and poor resolution. In the area of the stifle, in particular the caudal images, the authors speculate a larger distance from the detector will affect the final image regardless of acquisition mode (static or dynamically acquired). As different manufacturers use different motion correction solutions comparing the quality of static vs. dynamically acquired images was beyond the scope of the present work. The hypothesis that two orthogonal images would increase diagnostic accuracy of scintigraphy of the stifle was refuted. Neither the Se nor the Sp for detecting conditions of the stifle was improved by using two images for one radiologist. The other radiologist saw a modest improvement in Se but deterioration in Sp when two images were provided. When the quality judged as image resolution and the count density, subjectively judged by the degree of darkness, of

VOL. 56, NO. 1

BONE SCINTIGRAPHY FOR EQUINE STIFLE DISORDERS

lateral images, were compared to caudal images, the caudal images were found to exhibit less count density and were of lower resolution, hence of a lower diagnostic quality. This is probably due to the same reasons as listed above, including greater detector to target distance, muscle attenuation of the gamma rays, and inability to position the camera fully against the area of interest. Admittedly, these comparisons were subjective and hence remain anecdotal. A drop in positive PV was observed for both radiologists when a second image was added indicating that Se of the modality to detect lesions of the stifle was reduced. The reason for this reduction remains unknown but we speculate that adding a poor quality image may increase uncertainty and cause the reader to interpret the study as normal. The gold standard19 used to identify stifle-associated lameness in this study was the use of intra-articular analgesia.12 In spite the fact that a positive response to intra-articular analgesia was an inclusion criteria, response to such analgesia can be variable depending on the type of lesion.26,27 In this study, lameness examinations were performed by experienced clinicians who were board-certified by the American College of Veterinary Surgeons at the time of examination. However, studies have shown that subjective evaluation of lameness can be inaccurate28,29 and that clinicians can be biased towards assessing improvement after diagnostic analgesia.30 Use of an objective lameness evaluation system, such as an inertial sensor system,31 had it been available at the time, may have improved precision of the inclusion data. Objective ROI image analysis may allow for more subtle differences in radioisotope uptake to be appreciated when compared with subjective image assessment.32 Additionally, the use of relative uptake ratios (RUR) has been used to avoid error due to variations in the amount of radiopharmaceutical administered, distance from the camera, effective half-life, asymmetrical limb uptake, and other factors that may affect overall counts. Similar to subjective assessments, however, manually defined ROI and RUR allows for operator induced error and bias. In this study, the readers were presented with printed images, rather than images on a computer screen to facilitate assessment of 208 cases in a relatively short period of time. Any potential variable introduced by subjective windowing of each image to the readers likening or differences in monitor quality was also eliminated. Inclusion of ROR and RUR was therefore not possible with our study design. Having all images side by side in a single glossy high quality printed page also

101

eliminated the distractions of searching the case in the picture archiving system database and the time spent placing each stifle side by side on a computer screen. Previous studies have indicated that subjective methods have been shown to correlate highly with semiquantitative techniques.33,34 Subjective assessment is more commonly used and may in fact be more sensitive to focal increase in uptake since ROI can be subject to an averaging effect.35 In the current work, the subjective assessment relayed primarily on the radiologist’s experience gained through their day to day nuclear medicine practice. This experience takes into account variation of uptake among different horses, variation of uptake between legs in the same horses, and even pattern of uptake in young vs. old animals. Such variables are arguably difficult to take into account by ROI assessments. The subjective method of evaluation used in this study resulted in poor agreement between blinded observers. It was never the authors’ intent to compare the performance of the radiologists. Instead, their Cohen kappa coefficients were calculated (κ), to measure the agreement between their readings. For one scintigraphy image, the overall calculated κ was 0.084, and for two scintigraphy images, the overall calculated κ was 0.117. Fleiss’s guidelines for the evaluation of the kappa statistic36 state that any value below 0.40 should be considered poor agreement. Many factors may affect a subjective evaluation among them: use of a gray scale instead of a color palette33 , different environment during reading the examinations37 , and even processing errors created by nuclear medicine postprocessing techniques.38 For the purposes of this study, readers were presented with the same gray scale images with exactly the same image acquisition parameters. Though it remains unknown why poor agreement was detected, it is not surprising to the authors as anecdotally such difference of opinion is not uncommon amongst radiologists, and between clinical and imaging specialists, when assessing bone scintigrams during clinical practice. Bone-phase scintigraphic examination is frequently performed in horses with hindlimb lameness. Stifle problems are a common cause of lameness in the hindlimb in horses of various disciplines. As suspected from previous work4,7 , the study presented here further confirms that bone-phase nuclear scintigraphy is specific but highly insensitive for the detection of lameness causing conditions of the stifle joint in a diverse population of horses. The addition of a caudal image did not improve diagnostic accuracy in this study.

REFERENCES 1. Martinelli MJ, Rantanen N. The role of select imaging studies in the lameness examination. Proceedings of the 48th Annual Convention of the American Association of Equine Practitioners, Orlando, FL, 2002;4–8. 2. Barr ED, Pinchbeck GL, Clegg PD, Singer ER. Accuracy of diagnostic techniques used in investigation of stifle lameness in horses—40 cases. Equine Vet J 2006;18:326–331.

3. O’Rielly JL, Bertone AL, Genovese RL. Treatment of a chronic comminuted fracture of the fibula in a horse. J Am Vet Med Assoc 1998;212: 396–398. 4. Squire KRE, Fessler JF, Cantwell HD, Widmer WR. Enlarging bilateral femoral condylar bone cysts without scintigraphic uptake in a yearling foal. Vet Radiol Ultrasound 1992;33:109–113.

102

GRAHAM ET AL.

5. Dyson S. The Sports Horse. In: Dyson SJ, Pilsworth RC, Twardock AR, Martinelli MJ (eds): Equine Scintigraphy, 1st ed., Chapter 4. Suffolk, UK: UK Equine Veterinary Journal LTD, 2003;221–222. 6. Martinelli M, Arthur R. The American Thorougbred. In Dyson SJ, Pilsworth RC, Twardock AR, Martinelli MJ (eds): Equine Scintigraphy, 1st ed., Chapter 3. Suffolk, UK: UK Equine Veterinary Journal LTD 2003;186. 7. Graham JP, Roberts G. Equine Skeletal Scintigraphy, In: Daniel GB, Berry CR (eds): Textbook of veterinary nuclear medicine, 2nd ed., Chapter 9. Harrisburg, PA: American College of Veterinary Radiology, 2006;176. 8. Dyson S, McNie K, Weekes J, Murray R. Scintigraphic evaluation of the stifle in normal horses and horses with forelimb lameness. Vet Radiol Ultrasound 2007;48:378–382. 9. Hieber N, Lauk H, Ueltschi G. Zur Auswertung szintigraphischer Aufnahmen des Kniegelenks beim Pferd. Pferdeheilkunde 2000;16:568–578 (in German). 10. Hoskinson JJ. Equine nuclear scintigraphy. Indications, uses, and techniques. Vet Clin North Am Equine Pract 2001;17:63–74. 11. Dyson SJ, Martinelli MJ. Image Description and Interpretation in Musculoskeletal Scintigraphy. In Dyson SJ, Pilsworth RC, Twardock AR, Martinelli MJ (eds): Equine Scintigraphy, 1st ed. Chapter 9. Suffolk, UK: UK Equine Veterinary Journal LTD, 2003;85–87. 12. Walmsley JP. Diagnosis and treatment of ligamentous and meniscal injuries in the equine stifle. Vet Clin North Am Equine Pract 2005;21: 651–672. 13. McIlwraith CW, Nixon AJ, Wright IM, Boening KJ. Diagnostic and surgical arthroscopy in the horse, 3rd ed. St. Louis, MO: Mosby, 2005. 14. Cohen JM, Richardson DW, McKnight AL, Ross MW, Boston RC. Long-term outcome in 44 horses with stifle lameness after arthroscopic exploration and debridement. Vet Surg 2009;38:543–551. 15. R Core Team (2012). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3–900051–07–0, Available at: http://www.R-project.org. 16. de Leon AR, Guo M, Rudnisky CJ, Singh GA. likelihood approach to estimating sensitivity and specificity for binocular data: application in ophtalmology. Stat Med 2007;26:3300–3314. 17. de Leon AR, Soo A, Bonzo DC, Rudnisky CJ. Joint estimation of diagnostic accuracy measures for paired organs—application in ophtalmology. Biom. J. 2009;51:837–850. 18. Mercaldo ND, Lau KF, Zhou XH. Confidence intervals for predictive values with an emphasis to case-control studies. Stat Med 2007;26: 2170–2183. 19. Archer DC, Boswell JC, Voute LC, Clegg PD. Skeletal scintigraphy in the horse: current indications and validity as a diagnostic test. Vet J 2007;173:31–44. 20. Archer DC, Blake CL, Singer ER, et al. Scintigraphic appearance of selected diseases of the equine head. Equine Vet Ed 2003;15:305–313. 21. O’Sullivan CB, Lumsden JM. Stress fractures of the tibia and humerus in Thoroughbred racehorses: 99 cases (1992–2000). J Am Vet Med Assoc 2003;222:491–498.

2015

22. Daniel AJ, Judy CE, Rick MC, Saveraid TC, Herthel DJ. Comparison of radiography, nuclear scintigraphy, and magnetic resonance imaging for detection of specific conditions of the distal tarsal bones of horses: 20 cases (2006–2010). J Am Vet Med Assoc 2012;240:1109–1114. 23. Grapperon-Mathis M, Ley C, Berger M, et al. Evaluation of a positioning method for equine stifle scintigrams. Acta Vet Scand 2012;54:1–7. 24. Jeffcott LB and Kold SE. Stifle lameness in the horse: a survey of 86 referred cases. Equine Vet J 1982;14:31–39. 25. Yamaguchi T, Endo Y, Nambo Y, Sato F, Sasaki N, Yamada K. Evaluation of motion correction processing in equine bone scintigraphy by Scheff´e’s method of paired comparisons J Vet Med Sci 2013;75:369–371. 26. Dyson SJ. Lameness Associated with Stifle and Pelvic Regions. Proceedings of the 48th Annual Convention of the American Association of Equine Practitioners; Orlando, FL, 2002; 387–411. 27. Walmsley JP. The Stifle Cahpter 47. In Ross MW, Dyson SJ, (eds): Diagnosis and management of lameness in the horse, 2nd ed. St. Louis: Saunders, 2010;456. 28. Fuller CJ, Bladon BM, Driver AJ, Barr ARS. The intra- and interassessor reliability of measurement of functional outcome by lameness scoring in horses. Vet J 2006;171:281–286. 29. Keegan KG, Dent EV, Wilson DA, et al. Repeatability of subjective evaluation of lameness in horses. Equine Vet J 2010;42:92–97. 30. Arkell M, Archer RM, Guitian FJ, May SA. Evidence of bias affecting the interpretation of the results of local anaesthetic nerve blocks when assessing lameness in horses. Vet Rec 2006;159:346–349. 31. McCracken MJ, Kramer J, Keegan KG, et al. Comparison of an inertial sensor system of lameness quantification with subjective lameness evaluation. Equine Vet J 2012;44:652–656. 32. Davenport-Goodall CL, Ross MW. Scintigraphic abnormalities of the pelvic region in horses examined because of lameness or poor performance: 128 cases (1993–2000). J Am Vet Med Assoc 2004;224:88–95. 33. Erichsen C, Eksell P, Widstrom C, Roethlisberger-Holm K, Johnston C, Lord P. Scintigraphic evaluation of the thoracic spine in the asymptomatic riding horse. Vet Radiol Ultrasound 2003;44:330–338. 34. Kawcak CE, McIlwraith CW, Norrdin RW, Park RD, Steyn PS. Clinical effects of exercise on subchondral bone of carpal and metacarpophalangeal joints in horses. Am J Vet Res 2000;61:1252–1258. 35. Uhlhorn H, Eksell P, Sandgren B, Carlsten J. Sclerosis of the third carpal bone. A prospective study of its significance in a group of young standardbred trotters. Acta Vet Scand 2000;41:51–61. 36. Fleiss JL. Statistical methods for rates and proportions, 2nd ed. New York: John Wiley, 1981;38–46. 37. Brennan PC, McEntee M, Evanoff M, Phillips P, O’Connor WT, Manning DJ. Ambient lighting: effect of illumination on softcopy viewing of radiographs of the wrist. Am J Roentgenol 2007;188: 177–180. 38. Lee A, Forstrom FA, Dunn WL, O’Connor MK, Decklever TD, Hardtman TJ, Howarth DM. Technical pitfalls in image acquisition, processing and display. Semin Nucl Med 1996;26:278–294.

Diagnostic sensitivity of bone scintigraphy for equine stifle disorders.

Disorders of the stifle are a common cause of lameness in horses yet the accuracy of scintigraphy for diagnosis of stifle conditions is controversial...
681KB Sizes 5 Downloads 4 Views