Radiographic Analysis of Trochlear Notch Sclerosis in the Diagnosis of Osteoarthritis Secondary to Medial Coronoid Disease

Home , Trochlear notch

Radiographic analysis of trochlear notch sclerosis in the diagnosis of osteoarthritis secondary to medial coronoid disease

D. Draffan1; I. Carrera1; S. Carmichael1; J. Heller1; G. Hammond2 1Institute of Comparative Medicine Faculty of Veterinary Medicine, University of Glasgow, UK; 2Hospital for Small Animals, Easter Bush Veterinary Centre, Roslin, Midlothian, UK

1) Increased density of the bone (e.g. en- Summary Introduction largement of the trabeculae, reduced Trochlear notch sclerosis (TNS) as assessed by radiography has been shown to be increased in elbow dysplasia inter-trabecular bone marrow spaces and (ED) associated medial coronoid process disease (MCD). Medial coronoid process disease (MCD) is subchondral compaction); The aims of this study were to investigate whether two a term that represents a spectrum of pathol- 2) Superimposed osteophytes, and peri- defined radiographic tests evaluating TNS would in- ogies which affect the medial coronoid pro- osteal/endosteal new bone; crease the sensitivity of detecting osteoarthritis second- cess, including fragmentation of the medial 3) Superimposed soft tissue mineralisation. ary to MCD, and to assess whether there was a cor- coronoid process (1, 2). MCD is one of the relation between increasing TNS with severity of MCD. abnormalities seen in elbow dysplasia (ED); Radiographic evidence of TNS secondary to Sixty-one dogs with MCD (121 elbows) were selected the other lesions regularly included are unu- MCD has been described (25) and the inter- from the imaging database. The controls were nine ca- nited anconeal process, and osteochondro- national elbow working group (IEWG) in- davers (18 disease-free elbows). Standard Inter- sis or osteochondritis dissecans of the medi- clude the use of TNS as part of their grading national Elbow Working Group radiographs and CT al humeral condyle (3). Some authors also system for elbow arthrosis. The purpose of scans were taken of each elbow. Plain radiographs were analysed using various assessments: osteophyte include elbow joint incongruency (4, 5). this study was to evaluate the reliability of grade (0–3), coronoid grade (0–3), TNS descriptive MCD is the most common manifestation of qualitative and quantitative radiographic grade (0–3) and TNS ratio. The TNS ratio was calcu- ED (6–9), and, according to many studies, it TNS assessment that could be easily trans- lated from the depth of ulnar sclerosis at the level of is the most frequent cause of thoracic limb ferred to general practice in order to in- the disto-cranial margin of the humeral condyles di- lameness in young, large breed dogs crease the sensitivity of the radiographic di- vided by the cranio-caudal ulna depth. The TNS de- (10–18). ED is the most common cause of agnosis of arthrosis secondary to MCD. The scriptive assessment and ratio both increased the over- elbow osteoarthritis (19). It is therefore of primary hypothesis was that radiographic all sensitivity of diagnosing osteoarthritis secondary to paramount importance that this disease is TNS tests increase the sensitivity of detect- MCD above the other radiographic tests. The sensitivity diagnosed in general practice in order to ing MCD. The second hypothesis was that of the TNS ratio at <0.3 mm was 91%-96% and the allow appropriate therapy, specialist referral increasing TNS correlates with increasing TNS descriptive assessment was 77%-96%. Radio- (if required) and client education to be com- severity of MCD as evaluated by CT. graphic TNS significantly increased with increasing severity of coronoid disease grade as evaluated by CT menced as early as possible. Radiographic p<0.01. The finding that TNS increases the sensitivity signs of MCD are well documented and are of diagnosing osteoarthritis secondary to MCD is valu- based on the early signs of osteoarthritis able to those that have neither a CT scanner nor arth- (13, 20). It is known that the radiographic Materials and methods roscopy readily available. The assessments described in diagnosis of MCD is difficult, with poor this study are easy to apply and do not require any sensitivity compared to advanced imaging In this retrospective study the clinical data- sophisticated technology in order to detect sclerosis. modalities, such as computed tomography base at the University of Glasgow Small (CT) and magnetic resonance imaging Animal Hospital was reviewed from Keywords (MRI) (20–23), but radiology still remains 2004–2007 in order to identify dogs that had Trochlear notch sclerosis, medial coronoid disease, the primary imaging modality in general undergone both CT and radiographic evalu- elbow dysplasia practice. Trochlear notch sclerosis (TNS) is ation of the elbow joint. Additional recorded a radiological term which describes in- clinical information included: age, breed, Vet Comp Orthop Traumatol 2009; 22: 7–15 creased bone radio-opacity in the region of sex, weight, duration of lameness, and the Prepublished online: December 10, 2008 the ulnar trochlear notch, although the term results of arthroscopy, if performed. The in- doi:10.3415/VCOT-07–12–0110 sclerosis means a hardening. It is thought to clusion criteria for each dog were; 1) tho- be due to three different pathological pro- racic limb lameness localised to the elbow cesses (24): joint, 2) diagnostic quality radiographs Received: December 8, 2007 Accepted: May 13, 2008 Vet Comp Orthop Traumatol 1/2009

Table 1 numbers were covered and every elbow was Assessment Grade Description Descriptions of the assess- assigned a number. Each observer analysed ments made for CT and 0 normal coronoid as detected by CT radiographic analyses. the radiographs independently and one 1 altered attenuation week later they analysed the CT scans. CT Coronoid disease grade 2 fissured coronoid For radiographic analyses, osteophyte grade, medial coronoid process disease 3 fragmented coronoid grade and TNS sclerosis grade were assessed 0 no osteophytes on a categorical scale (Table 1). Continuous 1 largest osteophyte <2mm Radiographic osteophyte grade measures of TNS ratio were obtained using 2 largest osteophyte 2–5mm the medio-lateral radiographic projection 3 largest osteophyte >5mm (Fig. 1A, B) and the following standardised 0 normal coronoid measurements and calculations: the depth of sclerosis was measured at the level of the Radiographic medial coronoid 1 mild change in coronoid shape cranio-distal margin of the humeral condyle disease grade 2 marked change in coronoid shape in millimetres (A) and the cranio-caudal 3 fragmented coronoid ulnar depth was measured in millimetres 0 no sclerosis (outer cortex-outer cortex) (B) at the same 1 mild sclerosis, trabecular pattern still level (Fig. 1A, B) 90° to the caudal ulnar Trochlear notch sclerosis (TNS), easily seen shaft. The overall TNS ratio of sclerosis to medio-lateral projection, qualitative 2 moderate sclerosis, trabecular pattern ulnar depth was then calculated as A/B. (descriptive) grade slightly unclear Computed tomograms were assessed for 3 severe sclerosis, trabecular pattern the grade of MCD on a categorical scale cannot be seen (Table 1). A group consensus was also obtained using a combination of the assessments of the CT scans for all of the observers for MCD. By simple majority, the con- (standard IEWG views) and CT scans taken coronoid process. One millimetre sequen- sensus CT coronoid disease grade was used within one week of each other, and 3) evi- tial slices with a slice thickness of 1.2 mm, as the criteria for the diagnosis of MCD to dence of MCD on CT scan in at least one (slice overlap 0.2 mm), were obtained. The which all radiographic assessments were re- elbow. images were reviewed by all of the observ- lated. This categorisation was termed ‘con- Cases with incomplete medical records ers using a bone window (window width sensus’ and is treated as the gold standard or any other pathology within the elbow 3500 Hounsfield units, window level 500 for the diagnosis of presence and extent of were excluded. Elbows demonstrating Hounsfield units). MCD in this paper. MCD with concurrent incongruency were A control population of nine adult cross included. breed canine cadavers, (euthanatised for hu- Standard IEWG radiographic views of mane reasons unrelated to the elbow), with each elbow were obtained (neutral and flex- the same median weight (28 kg) and age Statistical analysis ed mediolateral and extended craniocau- (one year) as the disease population, were Initial graphical analyses of the data were dal), using a rare earth cassette (Quanta de- selected. The forelimbs were collected by undertaken for each radiographic measure- tail; DuPont) and high detail film (Cronex shoulder disarticulation, wrapped in saline- ment and for each observer. Kruskal-Wallis 10T; Agfar). Table top exposures were per- soaked towels and stored at –20°C until the analyses were used in order to compare the formed without the use of a grid and settings time of examination. Before examination, TNS ratio results between levels of consen- were adjusted depending on tissue thick- the limbs were thawed for 24 hours at room sus for each observer and follow-up compari- ness, according to pre-existing exposure temperature remaining in the saline-soaked sons were undertaken using Mann-Whitney charts. CT scans were obtained using a third towels. Standard IEWG radiographs and a tests, adjusted for multiple comparisons. generation scanner (Excel 2400 elite; Els- CT scan were taken of each elbow; gross The sensitivity of each imaging test was cint). The dogs were sedated and placed in post-mortem examination confirmed the calculated (dichotomising the results of the sternal recumbency, the forelimbs extended lack of elbow pathology. categorical data into presence (grade 1, 2 or with the elbows perpendicular to the CT Three observers, one board certified ra- 3) or absence (grade 0) of each radiographic gantry; the head was pulled caudally in diologist, one resident radiologist and one feature and using varying cut-off values for order to avoid interference with the beam resident surgeon, assessed each radiograph TNS ratio) for the diagnosis of coronoid dis- (26). The region to be scanned included the and CT. The observers were blinded to the ease for each observer. Simple two by two entire elbow articulation from the anconeus identity of the cases and whether or not they tables and standard calculations were used to the radioulnar articulation distal to the were diseased. All of the names and case for sensitivity calculations.

Vet Comp Orthop Traumatol 1/2009

Downloaded from www.vcot-online.com on 2011-11-23 | ID: 1000542141 | IP: 137.166.68.75 For personal or educational use only. No other uses without permission. All rights reserved. 9 Radiographic analysis of trochlear notch sclerosis

Inter-observer agreement was assessed using weighted Cohen’s Kappa statistics for the categorical outcomes (radiographic osteophyte grade, MCD grade, TNS description and CT coronoid disease grade) and intraclass correlation coefficients for the continuous measure of TNS ratio. Statistical analyses were performed using Minitab® Statistical Software version 15 and StatsDi- rect Statistical Software; (StatsDirect Ltd., 2005, Coventry, UK).

Results A) Sixty-one patients met the inclusion criteria of the diseased population. These 61 patients had 121 elbows examined and of these, 12 were found to be non-diseased on CT examination. The median age of the dogs was 12 months (range: 5–108 months), and the median weight was 28 kg (range: 18–46 kg). Male dogs were mildly more affected than fe- male dogs (52% and 48%, respectively). The median duration of lameness was two months (range: 1 week to 48 months), and 80.3% (49/61) of the cases were affected bilaterally. Labrador Retrievers were over-represented (49%, [30/61]), and a variety of other breeds Fig. 1 were also identified (Table 2). Of the 121 el- A) Line A: extent of scler- bows, 66 arthroscopies were performed, all osis. B) Line B: ulnar depth 90° to caudal ulnar B) of which were positive for MCD. Only one shaft. elbow had arthroscopic MCD which was not detected by CT (Table 2). There was moderate to substantial inter- observer agreement with most assessments Kruskal-Wallis tests revealed a differ- (p<0.01), for categorical data Kappa was ence in the median values of TNS ratio be- Discussion between 0.40 – 0.72 and intraclass cor- tween consensus categories for all of the ob- relation coefficients for continuous data servers (p<0.001 for each observer), and The purpose of this study was to evaluate the were between 0.46 – 0.61. Therefore, the follow-up comparisons showed that there usefulness and reliability of qualitative and comparison of consensus and radiographic were significant differences between cat- quantitative radiographic TNS assessments assessments (osteophyte grade, radio- egories 0 and 1 for observers 1 (p=0.018) that could be easily transferred to general graphic coronoid disease grade, TNS de- and 2 (p=0.006) and approaching signifi- practice in order to increase the sensitivity of scriptive grade and TNS ratio) of elbows are cance for observer 3 (p=0.053). Significant the radiographic diagnosis of osteoarthritis only displayed for observer one (Figs. 2–6). differences were also present between cat- secondary to MCD. This study showed that Only observer three for TNS descriptive as- egories 0 and 2, and 0 and 3, respectively, for our radiographic TNS assessments increased sessment had slight to moderate agreement all of the observers (p<0.001). the sensitivity of detecting MCD, and that in- with the other observers Kappa 0.1 – 0.2. The results for calculations of sensitivity creasing TNS correlated to increasing grade Fig. 5 represents TNS ratio scores for each for each radiographic assessment are pres- of MCD as evaluated by CT. observer compared with a dichotomised ented in Table 3. The TNS description and It has previously been shown that radi- consensus outcome, showing that TNS ratio ratio with a cut off <0.3 mm had a higher ography is inferior to CT at detecting MCD, increased in MCD positive elbows. sensitivity than the osteophyte or coronoid particularly as it mainly relies on arthrosis disease assessments. secondary to the disease. Carpenter and

Vet Comp Orthop Traumatol 1/2009

others (1993) showed that CT had an accu- TNS is to be used for the assessment of os- elbow further, for example by CT or arthro- racy of 86.7% compared to radiography, teoarthritis (20). scopic assessment. 56.7% (20). The current study does not re- The radiographic TNS descriptive as- A limitation of our study is the lack of fute that CT is a superior imaging technique sessment was highly sensitive (77–96%) in ability to obtain estimates for accuracy, spe- for the diagnosis of MCD. CT has many ad- detecting osteoarthrosis secondary to MCD. cificity and predictive values for the use of vantages over plain imaging, including lack Grading systems have been developed TNS in the diagnosis of MCD. The popu- of superimposition of bony structures, which now include TNS (such as the IEWG lation used in the current retrospective study multiplanar viewing, the ability to recon- and Lang's scoring systems), but to our was biased due to the lack of availability of struct images, enhanced contrast, improved knowledge there are not any estimates of the CT scans of animals with other manifes- definition of the coronoid process and de- sensitivity of these systems, neither as sub- tations of elbow disease (other than incon- gree of MCD and the ability to detect other jective assessments nor in combination with gruency). For example, observer three had a lesions which may be the cause of elbow the other radiographic signs of ED (28). Our higher sensitivity but may have had a lower pain/lameness, such as incomplete ossifica- results concur with those of Burton and specificity at detecting TNS due to inexperi- tion of the humeral condyle (27). The pur- others (2007) who demonstrated that there ence and likely over-interpretation of the pose of this study was to aid practitioners was a statistically significant relationship radiograph, this may occur in practice. who do not have access to a CT scanner or between trochlear notch radiopacity and Intra-observer analysis of TNS would have arthroscopy in the diagnosis of arthritis sec- MCD using quantitative digital analysis improved the validation of our results. ondary to MCD, and to highlight that TNS is (25). However TNS is not specific to MCD. Further studies in these areas would be an important radiological sign that is highly It is a radiographic sign indicative of in- beneficial. Despite this, our high sensitivity sensitive for the diagnosis of arthritis sec- creased bone hardening, which can occur in results show that TNS is a useful screening ondary to MCD, particularly when com- many pathological conditions, including the test and may be used as a test to rule out bined with the other radiographic signs. other manifestations of elbow dysplasia, in- MCD. In the authors’ opinion, while false In an attempt to make the TNS descrip- complete ossification of the humeral con- positive results may occur, it is preferrable tive evaluation quantitative and to measure dyle, trauma, abnormal loading of the limb to over diagnose and pursue further diag- the extent of sclerosis, we describe a simple and neoplasia. Therefore, TNS must be used nostics, in order to enable the commence- technique of measuring the TNS ratio. This simply as an indicator to investigate the ment of therapy in a timely manner, than to ratio was measured in the coronoid region of wait for more traditionally relied on radio- the trochlear since Burton et al. (2007) had graphic signs of osteoarthritis to be de- described this area as having the greatest tected, such as osteophytes. Table 2 Numbers of breeds affected with medial coron- radiopacity (25). As hypothesised, the TNS oid disease (MCD). The radiograph coronoid assessment had ratio was significantly greater in diseased a poor sensitivity (58% – 63%), which is dogs and, similar to the descriptive assess- comparable to other studies with ranges re- Breed of dog No. of dogs ment, increased the sensitivity of diagnos- with MCD ported between 10–62% (20, 29, 30). It is ing arthritis secondary to MCD. Of particu- difficult to interpret the medial coronoid on Labrador Retriever 30 lar note was that diseased elbows without a radiograph due to superimposition of the any secondary osteophytes were more likely Rottweiler 5 radial head and humeral condyle. The scores to be detected using the TNS description Golden Retriever 4 may have been improved if oblique views and/or ratio assessments, than would other- Staffordshire Bull Terrier 3 had been included, such as the distomedial- wise have been the case. The ratio number of Border Collie 2 proximo-lateral oblique (Di35M-PrLO) < 0.3 mm should not be taken as definitive Hungarian Vizla 2 (29). This was not available to us due to the as there are many variables which may in- retrospective nature of this study, and this is fluence this, such as patient positioning, Airdale Terrier 2 not a standard view used at our institute. radiographic quality and observer measur- Flat-Coated Retriever 2 The osteophyte assessment concurred ing ability. For example, a measurement German Shepherd Dog 2 with previous studies showing it was only variation of 1 mm may make the difference Springer Spaniel 2 sensitive as an indicator of late signs of os- between whether the elbow is categorised as Bernese Mountain Dog 1 teoarthritis (28, 31). Observer three had diseased or not and this is of particular im- Boxer 1 poor osteophyte assessment sensitivity and portance if there are not any other radio- this is likely to be due to observer inexperi- graphic signs indicative of disease. TNS as- English Bulldog 1 ence. sessment is affected by radiographic quality, Doberman 1 As the TNS ratio or descriptive grade in- (exposure, beam centering, patient position, Newfoundland 1 creased so did the grade of coronoid disease. collimation and film/screen system used), Toy Poodle 1 This was a general trend, and was more pro- hence it is of paramount importance to en- South African Boerbel 1 nounced using the descriptive assessment sure high quality imaging if evaluation of than the ratio. This is likely because the ex-

Vet Comp Orthop Traumatol 1/2009

Downloaded from www.vcot-online.com on 2011-11-23 | ID: 1000542141 | IP: 137.166.68.75 For personal or educational use only. No other uses without permission. All rights reserved. 11 Radiographic analysis of trochlear notch sclerosis

tent of sclerosis (ratio) may not increase to a degree that can be measured, whereas there may be an increase of detectable radiopacity. We therefore recommend that the descriptive and ratio assessments are used to- gether. It is not possible to use the TNS ratio and descriptive assessments in order to de- termine whether a coronoid is fragmented, fissured or has decreased subchondral density since there is too great an overlap between the groups to give definitive cut off points. Further diagnostics are required in order to assess the severity of coronoid disease. In this study, CT was used to confirm the presence or absence of MCD. Ideally, a combination of arthroscopic diagnosis and CT should be used to minimise misdiagno- sis (32, 33). However, it has been shown that the absence of CT or arthroscopic abnormalities does not rule out elbow lesions IEWG (33). The collection of tissue at the time of Obs1 3 arthroscopy for histopathology would be 2 ideal (2), but the associated morbidity is un- 1 acceptable. Arthroscopy was not performed 0 in all of the cases for a number of reasons, some of which being owner choice, finan- cial concerns or the fact that lameness was responding to conservative management. Gross post mortem examination was per- Fig. 2 Chart of radiographic osteophyte grade versus computed tomography (CT) consensus grade for observer one, where formed on all of the normal dog cadavers; 0= normal coronoid as detected by CT; 1= altered attenuation; 2= fissured coronoid; and 3=fragmented coronoid. they were rejected if any signs of elbow pathology were detected. It is accepted by the authors that histopathology would have been optimal since Danielson et al. (2006) proved that fissuring occurs in the subchondral bone before gross fibrillation of articu- lar cartilage develops (2, 34, 35). We assessed the grade of CT diagnosed MCD by group consensus of the appearance of the medial coronoid process on CT transverse image. Individual variation has been shown in grading MCD (5). The observers did not have any difficulty interpreting whether the coronoid was fragmented (group 3) or not, but there was discussion when differentiating between hypoattenu- Fig. 3 ation (group 1) and fissuring (group 2). Chart of radiographic medial coronoid disease Tromblee et al. (2007) concluded that the grade versus computed to- diagnostic certainty for altered subchondral mography (CT) consensus bone density was primarily influenced by grade for observer one, CDG window setting and was optimized at a where 0= normal coron- Obs1 oid as detected by CT; 1= 3 width of 3,500 HU and window level of 500 2 altered attenuation; 2= 1 HU (5). Whereas structural medial coronoid fissured coronoid; and 0 process abnormalities were influenced most 3=fragmented coronoid.

Vet Comp Orthop Traumatol 1/2009

by image plane, the transverse plane being the most favourable. In our study, we fol- lowed those recommendations in an attempt to maximize our sensitivity in detecting severity of MCD. There was a general agreement amongst observers for the radiographic assessments, though from the sensitivity results it can be clearly seen that observer 3 had a greater sensitivity for TNS. This is likely due to in- experience and over-interpretation of the degree of trochlear notch radiopacity. Inter- observer variation also reflects the observer

noise inherent in the decision-making process, for example, if a fragmented medial coronoid process or osteophytes are ob- served or the contralateral limb has signs of osteoarthritis, then the observer may have greater certainty that the elbow is diseased TNS and this may bias their judgment when in- Obs1 specting the degree of TNS (36). A recent 3 2 study by Tromblee and others also found 1 this, whereby observers had a greater diag- 0 nostic certainty for normal elbows than for dysplastic elbows (5). The median age of our diseased groups of 12 months and median weight of 28 kg Fig. 4 Chart of radiographic TNS descriptive grade versus computed tomography (CT) consensus grade for observer one, was similar to other studies (37, 38). The where 0= normal coronoid as detected by CT; 1= altered attenuation; 2= fissured coronoid; and 3=fragmented coronoid. over-representation of Labrador Retrievers is supported by many studies (10–12, 16, 32, 38, 39). The lower number of breeds, such as German Shepherd Dogs, Rottweilers, Springer Spaniels, Bernese Mountain Dogs, and Mastiffs, is likely to re- flect a lower overall number of these breeds in the referred dog population but may also be due to our exclusion criteria. For example, most Springer Spaniels that had MCD also had incomplete ossification of the humeral condyle. Many studies show that male dogs are clinically more affected than females and our study supported this

albeit not as strongly as other studies (1, 30, 40–43), which again may be due to our strin- gent exclusion criteria. In our diseased dog population, 80.3% were bilaterally affected, a finding which is similar to previous reports (19, 44, 45). In our clinic, both elbows are imaged regard- less of whether the dog is clinically affected bilaterally or unilaterally. This is to help pre- Fig. 5 Boxplot of radiographic trochlear notch sclerosis (TNS) ratio versus dichotomised computed tomography (CT) con- vent diseased but clinically ‘silent’ elbows sensus for each observer (Obs), where 0= normal coronoid as detected by CT; and 1= medial coronoid disease positive. The whiskers indicate the range of values, the box contains the 25th and 75th centiles and the line within the box indicates the medi- from being missed. However, it is still possi- an. Outliers are indicated with an asterix. ble that a normal CT is seen with a histologi-

Vet Comp Orthop Traumatol 1/2009

Downloaded from www.vcot-online.com on 2011-11-23 | ID: 1000542141 | IP: 137.166.68.75 For personal or educational use only. No other uses without permission. All rights reserved. 13 Radiographic analysis of trochlear notch sclerosis

cally abnormal coronoid (5). This may have reduced the sensitivity results. Further assessment of all of the elbows by arthroscopy would have allowed a more detailed examination of the coronoid process and may have improved the identification of diseased elbows (9). There were 12 contralateral elbows within the clinical cases which did not have CT evidence of MCD yet in some cases the observers assessed these elbows radio- graphically as having TNS sclerosis (Fig. 4). As previously mentioned, this may have been because of bias due to the observation of more convincing evidence of osteoarthritis in the opposite elbow (39), or for a patho- physiological reason that cannot be determined from this study. Trochlear notch radiopacity is not a di- rect measurement of bone density itself. Our study may have been further validated by Fig. 6 Boxplot of radiographic trochlear notch sclerosis (TNS) ratio versus computed tomography (CT) consensus grade for each observer (Obs), where 0= normal coronoid as detected by CT; 1= altered attenuation; 2= fissured coronoid; and using computed tomographic osteoabsorp- 3=fragmented coronoid. The whiskers indicate the range of values, the box contains the 25th and 75th centiles and the line tionmetry (46–48), or dual-energy x-ray ab- within the box indicates the median. Outliers are indicated with an asterix. sorptionmetry (49). The reason for the increased radiopacity with MCD and why this is increased with increasing severity of disease is unknown. In humans, increasing Table 3 Sensitivity results for each observer for the radiographic and CT assessments for the diagnosis of osteoarthritis sec- bone mineral density has been found in os- ondary to medial coronoid disease. 95% confidence intervals are in brackets. Trochlear notch sclerosis (TNS) sensitivities have been shown for cut off points of 0.3 mm-0.6 mm. Sensitivities for diagnosis of CT coronoid disease assessment versus consen- teoarthritis of the knee(50) and it may be sus are also shown. that this is also true with osteoarthritis of the elbow in dogs. Pathological humeroulnar in- Imaging assessment Sensitivity (true positive proportion) congruency may also play a role in this due 95% confidence intervals are below in brackets to eccentric loading of the joint (25). Con- Observer 1 Observer 2 Observer 3 versely, other studies have indicated that 0.72 0.74 0.43 Radiographic osteophyte assessment humeroulnar incongruency may be a nor- (0.63–0.80) (0.65–0.82) (0.33–0.53) mal finding in both humans (51) and Radiographic medial coronoid disease 0.63 0.63 0.58 dogs(52). assessment (0.54–0.72) (0.54–0.72) (0.48–0.67) In summary, this study demonstrates that 0.89 0.77 0.96 Radiographic TNS description assessment radiographic assessments of TNS increases (0.82–0.94) (0.69–0.85) (0.91–0.99) the sensitivity of diagnosing arthritis sec- 0.96 0.91 0.93 Radiographic TNS ratio cut off <0.3 mm ondary to MCD. This technique is likely to (0.91–0.98) (0.85–0.96) (0.86–0.97) be valuable to veterinarians who have 0.71 0.55 0.72 Radiographic TNS ratio cut off <0.4 mm neither a CT scanner nor arthroscopy (0.62–0.79) (0.45–0.64) (0.62–0.80) readily available. Specificity has been un- 0.24 0.19 0.34 Radiographic TNS ratio cut off <0.5 mm able to be elucidated from this study; there- (0.18–0.33) (0.12–0.28) (0.26–0.45) fore further diagnostics are required to rule 0.02 0.04 0.12 Radiographic TNS ratio cut off <0.6 mm out other causes of elbow disease. The as- (0.003–0.06) (0.01–0.09) (0.07–0.2) sessments described in this study are easy to Radiographic TNS description + ratio 0.97 0.91 0.96 apply in general practice and should be used <0.3 mm combined in parallel (0.92–0.99) (0.84–0.96) (0.90–0.99) as a screening test for diagnosing arthritis Radiographicosteophyte/coronoid/TNS 0.98 0.93 0.96 secondary to MCD. description/TNS sclerosis ratio <0.3 mm (0.94–0.99) (0.87–0.97) 0.91–0.99) combined in parallel Acknowledgements CT coronoid disease assessment versus 0.89 0.79 0.72 The authors acknowledge and thank Professors D. consensus (0.82 – 0.94) (0.70– 0.86) (0.63–0.81) Bennett and M. Sullivan for proofing the paper.

Vet Comp Orthop Traumatol 1/2009

Thanks are extended to Professor E. M. Scott and Dr. tigations into hereditary disposition. Tijdschr sessment of joint spaces in the canine elbow. J D. Mellor for statistic assistance; as well as to the im- Diergeneeskd 1988; 113(Suppl 1): 49S-51S. Small Anim Pract 2006; 47: 66–74. aging department for their expertise and technical 17. Grondalen J, Rorvik AM. Arthrosis in the elbow 33. Moores AP, Benigni L, Lamb CR. Computed to- support. joint of young rapidly growing dogs. IV. Ununited mography versus arthroscopy for the detection of anconeal process. A follow up investigation of op- canine elbow dysplasia lesions. Veterinary Ortho- erated dogs. Nord Vet Med 1980; 32: 212–218. paedic Society Congress; March 8–15, 2008; Big 18. Morgan JP, Wind A, Davidson AP. Bone dyspla- Sky, Montana; 2008. p. 51. References sias in the Labrador retriever: A radiographic 34. Wolschrijn CF, Weijs WA. Development of the study. J Am Anim Hosp Assoc 1999; 35: subchondral bone layer of the medial coronoid 1. Gemmill TJ, Clements DN. Fragmented coronoid 332–340. process of the canine ulna. Anat Rec A Discov process in the dog: is there a role for incongruen- 19. Olsson SE. The early diagnosis of fragmented cor- Mol Cell Evol Biol 2005; 284: 439–445. cy? J Small Anim Pract 2007; 48: 361–368. onoid process and osteochondritis dissecans of the 35. Wolschrijn CF, Weijs WA. Development of the 2. Danielson KC, Fitzpatrick N, Muir P et al. Histo- canine elbow joint. J Am Anim Hosp Assoc 1983; trabecular structure within the ulnar medial cor- morphometry of fragmented medial coronoid pro- 19: 616–626. onoid process of young dogs. Anat Rec A Discov cess in dogs: a comparison of affected and normal 20. Carpenter LG, Schwarz PD, Lowry JE et al. Com- Mol Cell Evol Biol 2004; 278: 514–519. coronoid processes. Vet Surg 2006; 35: 501–509. parison of radiologic imaging techniques for diag- 36. Meyer-Lindenberg A, Langhann A, Fehr M et al. 3. Schulz KS, Krotscheck U. Canine elbow dyspla- nosis of fragmented medial coronoid process of Prevalence of fragmented medial coronoid pro- sia. In: Slatter D, editor. A textbook of small ani- rd the cubital joint in dogs. J Am Vet Med Assoc cess of the ulna in lame adult dogs. Vet Rec 2002; mal surgery. 3 ed. Philadelphia: W.B. Saunders 1993; 203: 78–83. 151: 230–234. 2003; pp 1927–1952. 21. Snaps FR, Balligand MH, Saunders JH et al. Com- 37. Meyer-Lindenberg A, Fehr M, Nolte I. Co-exist- 4. Samoy Y, Van Ryssen B, Gielen I et al. Review of parison of radiography, magnetic resonance im- ence of ununited anconeal process and frag- the literature – Elbow incongruity in the dog. Vet aging, and surgical findings in dogs with elbow mented medial coronoid process of the ulna in the Comp Orthop Traumatol 2006; 19: 1–8. dysplasia. Am J Vet Res 1997; 58: 1367–1370. dog. J Small Anim Pract 2006; 47: 61–65. 5. Tromblee TC, Jones JC, Bahr AM et al. Effect of 22. Snaps FR, Park RD, Saunders JH et al. Magnetic 38. Gemmill TJ, Mellor DJ, Clements DN et al. Evalu- computed tomography display window and image resonance arthrography of the cubital joint in dogs ation of elbow incongruency using reconstructed plane on diagnostic certainty for characteristics of affected with fragmented medial coronoid pro- CT in dogs suffering fragmented coronoid pro- dysplastic elbow joints in dogs. Am J Vet Res cesses. A J Vet Res 1999; 60: 190–193. cess. J Small Anim Pract 2005; 46: 327–333. 2007; 68: 858–871. 23. Snaps FR, Saunders JH, Park RD et al. Compari- 39. LaFond E, Breur GJ, Austin CC. Breed suscepti- 6. Wind AP. Elbow incongruity and developmental son of spin echo, gradient echo and fat saturation bility for developmental orthopedic diseases in elbow diseases in the dog .1. J Am Anim Hosp magnetic resonance imaging sequences for im- dogs. J Am Anim Hosp Assoc 2002; 38: 467–477. Assoc 1986; 22: 711–724. aging the canine elbow. Vet Radiol Ultrasound 40. D.Bonagura J, editor. Kirk’s Current Veterinary 7. Wind AP, Packard ME. Elbow incongruity and de- 1998; 39: 518–523. Therapy XIII. Philadelphia: W.B.Saunders 2000. velopmental elbow diseases in the dog. 2. J Am 24. Dennis R KR, Wrigley RH, Barr F. Handbook of 41. Salg KG, Temwitchitr J, Imholz S et al. Assess- Anim Hosp Assoc; 22: 725–730. small animal radiological differential diagnosis. ment of collagen genes involved in fragmented 8. Fluckiger M. [Elbow dysplasia in the dog]. Saunders; 2001. medial coronoid process development in Labra- Schweiz Arch Tierheilkd 1992; 134: 261–271. 25. Burton NJ, Comerford EJ, Bailey M et al. Digital dor Retrievers as determined by affected sibling- 9. Van Ryssen B, van Bree H. Arthroscopic findings analysis of ulnar trochlear notch sclerosis in La- pair analysis. Am J Vet Res 2006; 67: 1713–1718. in 100 dogs with elbow lameness. Vet Rec 1997; brador Retrievers. J Small Anim Pract 2007; 48: 42. Lewis DD, Parker RB, Hager DA. Fragmented 140: 360–362. 220–224. medial coronoid process of the canine elbow. 10. Grondalen J. Arthrosis with special reference to 26. De Rycke LM, Gielen IM, van Bree H et al. Com- Compend Contin Educ Pract Vet 1989; 11: the elbow joint of young rapidly growing dogs. II. puted tomography of the elbow joint in clinically 703–706, 708, 710, 712–715. Occurrence, clinical and radiographical findings. normal dogs. Am J Vet Res 2002; 63: 1400–1407. 43. Huibregste BA, Johnson AL, Muhlbauer M et al. Nord Vet Med 1979; 31: 69–75. 27. Rovesti GL, Biasibetti M, Schumacher A et al. The effect of treatment of fragmented coronoid 11. Grondalen J. Arthrosis with special reference to The use of the computed tomography in the diag- process on the development of osteoarthritis of the the elbow joint of young rapidly growing dogs. I A nostic protocol of the elbow in the dog: 24 joints. elbow. J Am Anim Hosp Assoc 1994; 190–195. review of the literature. Nord Vet Med 1979; 31: Vet Comp Orthop Traumatol 2002; 15: 35–43. 44. Bennett D, Duff SRI, Kene RO et al. Osteochon- 62–68. 28. Lang J, Busato A, Baumgartner D et al. Compari- dritis dissecans and fragmentation of the coronoid 12. Grondalen J. Arthrosis in the elbow joint of young son of two classification protocols in the evalu- process in the elbow joint of the dog. Vet Rec rapidly growing dogs. III. Ununited medical cor- ation of elbow dysplasia in the dog. J Small Anim 1981; 109: 329–336. onoid process of the ulna and osteochondritis dis- Pract 1998; 39: 169–174. 45. Henry WB, Jr. Radiographic diagnosis and surgi- secans of the humeral condyle. Surgical procedure 29. Haudiquet PR, Marcellin-Little DJ, Stebbins ME. cal management of fragmented medial coronoid for correction and postoperative investigation. Use of the distomedial-proximolateral oblique process in dogs. J Am Vet Med Assoc 1984; 184: Nord Vet Med 1979; 31: 520–527. radiographic view of the elbow joint for examin- 799–805. 13. Grondalen J. Arthrosis in the elbow joint of young ation of the medial coronoid process in dogs. Am 46. Samii VF, Les Clifford M, Schulz KS et al. Com- rapidly growing dogs. VI. Interrelation between J Vet Res 2002; 63: 1000–1005. puted tomographic osteoabsorptiometry of the clinical, radiographical and pathoanatomical 30. Wosar MA, Lewis DD, Neuwirth L et al. Radio- elbow joint in clinically normal dogs. Am J Vet findings. Nord Vet Med 1982; 34: 65–75. graphic evaluation of elbow joints before and after Res 2002; 63: 1159–1166. 14. Grondalen J. Arthrosis in the elbow joint of young surgery in dogs with possible fragmented medial 47. Korbel J MU, Maierl J, Muller-Gerbl M. Ct-Os- rapidly growing dogs. VII. Occurrence in the coronoid process. J Am Vet Med Assoc 1999; 214: teoabsorptiometry in dogs with elbow diseases. Rottweiler breed. Nord Vet Med 1982; 34: 76–82. 52–58. 10th Congress European Society of Veterinary Or- 15. Grondalen J, Grondalen T. Arthrosis in the elbow 31. Berry CR. Evaluation of the canine elbow for thopaedics and Traumatology; 23–26th March joint of young rapidly growing dogs. V. A pathoa- fragmented medial coronoid process. Vet Radiol 2000; Munich; 2000. p. 91–92. natomical investigation. Nord Vet Med 1981; 33: Ultrasound 1992; 33: 273–276. 48. Korbel J, Wilcken R, Huskamp B. Computed to- 1–16. 32. Gemmill TJ, Hammond G, Mellor D et al. Use of mographic examination of the canine elbow joint. 16. Grondalen J, Lingaas F. Arthrosis of the elbow reconstructed computed tomography for the as- Kleintierpraxis 2001; 46: 325–331. joint among Rottweiler dogs. Results from inves-

Vet Comp Orthop Traumatol 1/2009

Downloaded from www.vcot-online.com on 2011-11-23 | ID: 1000542141 | IP: 137.166.68.75 For personal or educational use only. No other uses without permission. All rights reserved. 15 Radiographic analysis of trochlear notch sclerosis

49. Clarke S, Wakeley C, Duddy J et al. Dual-energy 51. Eckstein F, Lohe F, Muller-Gerbl M et al. Stress Correspondence to: X-ray absorptiometry applied to the assessment of distribution in the trochlear notch. A model of Davinia A. Draffan, BVMS CertSAS MRCVS tibial subchondral bone mineral density in os- bicentric load transmission through joints. J Bone Glasgow University Veterinary School teoarthritis of the knee. Skeletal Radiol 2004; 33: Joint Surg Br 1994; 76: 647–653. Small Animal Hospital 588–595. 52. Preston CA, Schulz KS, Kass PH. In vitro deter- 464 Bearsden Road 50. Christensen P, Kjaer J, Melsen F et al. The Sub- mination of contact areas in the normal elbow Glasgow, Scotland chondral Bone of the Proximal Tibial Epiphysis in joint of dogs. Am J Vet Res 2000; 61: 1315–1321. G61 1QH, UK Osteoarthritis of the Knee. Acta Orthopaedica Phone: +44 7799764666, Fax: +44 1413303663 Scandinavica 1982; 53: 889–895. E-mail: [email protected]

Vet Comp Orthop Traumatol 1/2009