Cranial cruciate ligament deficiency is common in dogs, with an estimated 1 million surgeries performed in the United States to treat this condition in 2003.1 As many as 80% of dogs with CCL deficiency have concurrent medial meniscal lesions.2–4 Appropriate treatment of this concurrent damage is considered important in optimizing recovery and functional outcome following surgical treatment of CCL deficiency.5
Likewise, meniscal tears that occur after CCL surgery or that may have gone unrecognized at the time of CCL surgery are among the more common causes of surgical failure and manifest as persistent or recurrent lameness in affected dogs.6,7 Appropriate diagnosis and treatment of these meniscal tears are also important for the resolution of lameness. In addition, evidence exists to suggest that a notable percentage of dogs with CCL deficiency treated without surgery may have a successful outcome.8 However, successful nonsurgical management of dogs with CCL rupture may depend in part on the status of the medial meniscus. Accordingly, an ability to accurately assess the menisci with a noninvasive approach could have clinical relevance in each of these clinical situations.
Physical examination is considered suboptimal in the diagnosis of meniscal lesions because detection of a meniscal click has a low sensitivity (27.6% to 58.3%) for this purpose.7,9–11 Conversely, ultrasonography and MRI are 2 noninvasive diagnostic modalities that have been evaluated for their usefulness in diagnosing meniscal lesions in dogs and for which encouraging data have been reported. The sensitivity of ultrasonography for detecting meniscal (medial or lateral) lesions was 82% and 90% in 2 studies,11,12 and the specificity was 93% in both studies.
Although encouraging, these previously reported data are somewhat limited in fully depicting the usefulness of ultrasonography for assessing meniscal damage because only a total of 23 stifle joints were assessed in those studies.11,12 Further, the reference standard used in the second study12 included craniomedial arthrotomy which reportedly has a lower sensitivity than arthroscopy for detecting meniscal lesions.13 In addition, the ultrasonography assessments in both studies11,12 were performed by radiologists with extensive experience. Ultrasonography is considered a highly user-dependent skill, and it remains unclear whether ultrasonography can provide reliable results when used by someone with less experience.
Multiple studies14–24 have been performed to assess the usefulness of MRI for examination of stifle joints in dogs. Some of these studies15,17,22 have revealed excellent accuracy of MRI for evaluating the medial meniscus, with a reported sensitivity of up to 100%. However, arthrotomy was used as the reference standard for at least some dogs in all studies15,16,22,23 in which the accuracy of high-field MRI was evaluated. As a result, the previously reported and quantified sensitivities and specificities of high-field MRI may have been influenced by the use of arthrotomy as the reference standard, given that arthroscopy is more sensitive than arthrotomy for detecting meniscal lesions and is considered the gold standard for detecting meniscal lesions in dogs.4,13,25–27
To the authors' knowledge, the accuracy of ultrasonography and MRI has not been compared for detecting medial meniscal lesions in dogs. Such comparison would be relevant given that ultrasonography is more widely available and less expensive to perform than MRI. In addition, studies28–31 directly comparing ultrasonography and MRI for detecting meniscal lesions in people have revealed clinically useful, and sometimes superior, results with ultrasonography. For those reasons, the primary goal of the study reported here was to quantify and compare the accuracy of ultrasonography and MRI for diagnosing clinically relevant medial meniscal lesions in dogs, with arthroscopy used as the reference standard. For this study, ultrasonography was performed by a veterinarian with limited ultrasonography experience to investigate the broader applicability of this imaging modality for use in veterinary practice.
Additional study goals included determination of the accuracy of testing for a meniscal click or signs of pain during stifle joint flexion for diagnosing medial meniscal lesions requiring surgical treatment in dogs. We also sought to assess agreement between 2 surgeons as to which meniscal lesions required surgical treatment on the basis of arthroscopic findings. This last objective was relevant because arthroscopy has been cited as the gold standard for diagnosing meniscal lesions in dogs27 and was to be used as the gold standard for diagnostic test comparisons in the study reported here. However, we were unaware of any studies conducted to quantify agreement between surgeons regarding which meniscal lesions do or do not require surgical treatment.
On the basis of findings in previous studies,11,12,22,23 we hypothesized that both ultrasonography and MRI would be clinically useful, with diagnostic CCRs between 80% and 90%. Given findings from previous investigations,7,9,10 we expected that when used alone, testing for a meniscal click or signs of pain during stifle joint flexion would have a low CCR for diagnosing medial meniscal lesions. Finally, we hypothesized that interobserver agreement when assessing medial meniscal lesions via arthroscopic video recordings would be high because menisci can be easily seen arthroscopically.
Materials and Methods
Dogs
Dogs enrolled in this prospective study were recruited through the Orthopedic Surgery Service at the University of Georgia. Enrollment criteria were that the dog needed to have received a tentative diagnosis of CCL deficiency on the basis of its clinical history and results of general physical and orthopedic examinations (including detection of a cranial drawer sign), radiographic assessment, and preanesthetic hematologic evaluation. Dogs that had previously undergone surgery of the affected stifle joint were excluded. All owners were provided with an explanation of the study objectives and methodology and signed enrollment consent forms. They were informed that participation would involve examination of the stifle joint via ultrasonography, MRI, and arthroscopy; dogs were required to receive all 3 types of assessments to be included in the data analysis. The study protocol was approved by the Clinical Research Committee at the University of Georgia.
Orthopedic examination
All dogs were examined by 1 investigator (SPF) prior to enrollment. All were tested for a meniscal click and signs of pain on stifle joint flexion. Meniscal-click testing was performed when dogs were awake and again when they were either sedated or anesthetized. If a consistent meniscal click was identified through several range-of-motion cycles, whether awake, sedated, or anesthetized, the dog was classified as having a positive meniscal-click response.
Sedation and general anesthesia
If general anesthesia was used when MRI or ultrasonographic examinations were performed, dogs were treated with various sedation and anesthetic induction protocols but anesthesia was consistently maintained with isoflurane. If dogs were sedated for ultrasonographic examination, dexmedetomidine hydrochloridea (5 μg/kg [2.3 μg/lb]) and nalbuphine hydrochlorideb (0.5 mg/kg [0.23 mg/lb]) were administered IV. If dogs were chemically immobilized for MRI by IV drug administration only a catheter was placed in a cephalic vein and dogs were given dexmedetomidine (7 μg/kg [3.2 μg/lb]), nalbuphine (0.5 mg/kg), and ketamine hydrochloridec (1 mg/kg [0.45 mg/lb]). Supplemental administration of these medications was performed as necessary to maintain the dog immobile throughout the MRI examination.
MRI examination
A 1.5-T MRI clinical scannerd was used for all MRI examinations. Dogs were positioned in dorsal recumbency with pelvic limbs extended and stifle joints flexed to approximately 135°. All images were acquired by use of a 6-channel flexible receive-only body matrix coil, which was conformed to the limbs. Parallel imaging was used in all sequences. The following settings were used for slice thickness, echo time, repetition time, echo train, matrix size, field of view, and number of acquisitions, respectively, in the various sequences: 2-D TSE proton-density fat-suppressed sequence in the sagittal plane, 3 mm with no gap, 43 milliseconds, 2,000 milliseconds, 11, 320 × 320 pixels, 16 × 16 cm, and 1 radial (BLADE) reconstruction; 2-D dual echo TSE sequence in the transverse plane, 3 mm with no gap, 33 and 90 milliseconds, 4,200 to 4,300 milliseconds, 6, 320 × 256 pixels, 14 × 14 cm, and 1 radial (BLADE) reconstruction; 2-D T1 TSE sequence in the dorsal plane, 2.5 mm with no gap, 11 milliseconds, 406 milliseconds, 5, 370 × 320 pixels, 16 × 12.7 cm, and 3; and single-slab 3-D TSE proton-density sequence with a slab-selective, variable excitation pulse (sampling perfection with application of optimized contrasts by use of different flip-angle evolution; SPACE) in either the sagittal or dorsal plane with reconstructions in other planes, 0.6 mm, 39 milliseconds, 1,100 milliseconds, 51, 282 × 282 pixels, 16 × 16 cm, and 2.
Each set of MRI scans of the medial meniscus was evaluated, and a diagnosis was made by 1 radiologist (SPH) who was blinded to all other examination findings. Reported guidelines32 on corresponding MRI findings and meniscal diagnoses were used for categorization and diagnosis of medial meniscal lesions in the study. The menisci were ultimately classified as having damage for which surgical treatment was recommended (positive result) or as normal (undamaged; Figure 1) or with inconsequential lesions for which surgical treatment was not recommended (negative result).
Sagittal 2-D TSE proton-density fat-suppressed MRI view of the stifle joint of a dog showing the caudal pole of an undamaged medial meniscus. For classification purposes, this meniscus was considered one for which treatment was not necessary.
Citation: Journal of the American Veterinary Medical Association 251, 1; 10.2460/javma.251.1.71
Surgery was recommended for menisci with full-thickness longitudinal tears with (ie, bucket handle) or without displacement (Figure 2), flap tears (ie, bucket-handle tears in which one of the arms was torn), a folded caudal pole of the medial meniscus, and peripheral detachment of the caudal pole. Similarly, surgery was recommended for complex tears including double bucket-handle tears, menisci with multiple tears, and severe destruction or maceration of the caudal pole of the medial meniscus. Interstitial tearing that was diffuse enough that it was considered relevant to function, or that would displace with arthroscopic probing (for arthroscopic assessment); large horizontal tearing that was considered relevant to function; and radial tears > 3 but typically > 5 mm in length were all considered requiring of surgical treatment.
Sagittal 2-D TSE proton-density fat-suppressed MRI view of the stifle joint of a dog showing the caudal pole of a medial meniscus with a full-thickness vertical longitudinal tear (arrow). For classification purposes, this meniscus was considered one for which treatment was necessary.
Citation: Journal of the American Veterinary Medical Association 251, 1; 10.2460/javma.251.1.71
Medial meniscal lesions for which surgical (arthroscopic) treatment was deemed unnecessary included small, partial-thickness, vertical longitudinal tears that would not displace with probing and were inconsequential to mechanical function at the time of discovery; small interstitial tears that were not affecting mechanical function; small horizontal intraparenchymal tears; and degenerate menisci (as seen on MRI) without any tearing of any type noted. Small radial tears < 2 mm in length were not considered clinically relevant or requiring of surgery for purposes of this study and diagnostic test evaluation.
Ultrasonographic examination
The same surgeon (SPF) performed all ultrasonographic examinations and used an ultrasonographic machinee with a 5- to 18-mHz linear array probe to evaluate the cranial, middle, and caudal aspect of the medial meniscus as described elsewhere.12 This surgeon was not blinded to the physical and orthopedic examination findings (including detection of a meniscal click or signs of pain on stifle joint flexion) but was blinded to MRI findings. This surgeon had limited experience with ultrasonography, having performed ultrasonographic examination of approximately 20 to 30 stifle joints before the study began, including validation of ultrasonographic findings with intraoperative findings.
For the examination, dogs were positioned in lateral recumbency, with the affected stifle joint down and clipped of hair. The stifle joint was imaged longitudinally by placing the ultrasound probe on the joint such that the long axis of the linear probe was parallel to the long axis of the limb, thereby providing a cross-sectional image of the medial meniscus. The examination began with the transducer placed against the craniomedial aspect of the stifle joint approximately midway between the patellar tendon and the medial collateral ligament. The probe was moved caudally while maintaining the longitudinal orientation of the probe, until the stifle joint had been imaged from the craniomedial aspect to the caudomedial corner of the joint. The transducer was maintained perpendicular to the portion of meniscus being evaluated at all points. Because the meniscus is C-shaped, this required that the probe be pointed toward the caudolateral aspect of the stifle joint when the probe was on the craniomedial aspect of the joint and that the probe be eventually oriented craniolaterally when the probe had reached the caudomedial aspect of the joint. The stifle joint was positioned at approximately 135° of flexion. Imaging in this manner was repeated as deemed necessary to allow assessment of meniscal integrity.
The shape of the medial meniscus, heterogeneity of the echogenicity of the meniscus, presence of local effusion around the meniscus, and displacement or extrusion of the meniscus were ultrasonographically evaluated. An undamaged meniscus is definitively triangular in cross section and has homogenous echogenicity (Figure 3). If the margins of the triangular meniscus were severely abnormal or the meniscus did not appear clearly triangular in cross section, this was considered abnormal (Figure 4). Similarly, if the meniscus appeared notably heterogenous, this was considered abnormal (Figure 5). Local accumulation of effusion, manifest as a pocket of hypoechogenicity immediately adjacent to the meniscus, was also considered abnormal (Figure 6), as was displacement of the abaxial margin of the meniscus, referred to as the white line (Figure 7).
Ultrasonographic image of the stifle joint of a dog showing an undamaged medial meniscus. Notice the well-defined, homogenous hyperechoic triangular structure of the caudal pole of the meniscus. Also notice the abaxial border of the meniscus, sometimes referred to as the white line (arrows). The femoral condyle is to the left and the tibia to the right.
Citation: Journal of the American Veterinary Medical Association 251, 1; 10.2460/javma.251.1.71
Ultrasonographic images of the stifle joint of a dog showing an abnormal caudal pole with a full-thickness vertical longitudinal tear (A) and undamaged cranial horn (B) of the medial meniscus. A—Notice the distorted shape of the medial meniscus. B—Notice that the undamaged cranial horn retains a well-defined, hyperechoic triangular structure.
Citation: Journal of the American Veterinary Medical Association 251, 1; 10.2460/javma.251.1.71
Ultrasonographic image of the stifle joint of a dog showing the heterogenous echogenicity (arrows) of an abnormal medial meniscus.
Citation: Journal of the American Veterinary Medical Association 251, 1; 10.2460/javma.251.1.71
Ultrasonographic image of the stifle joint of a dog showing effusion (arrow) adjacent to the medial meniscus.
Citation: Journal of the American Veterinary Medical Association 251, 1; 10.2460/javma.251.1.71
Ultrasonographic image of the stifle joint of a dog showing a medial meniscus in which the abaxial margin (white line) is abaxially displaced (arrows).
Citation: Journal of the American Veterinary Medical Association 251, 1; 10.2460/javma.251.1.71
The medial meniscus was categorized as abnormal with surgical treatment recommended if ≥ 2 of the aforementioned criteria (shape, echogenicity, presence of effusion, and displacement) were considered abnormal. Conversely, the meniscus was categorized as normal or surgical treatment not recommended if ≥ 3 of those criteria were considered normal. More specific diagnoses such as longitudinal or radial tearing were not made by use of ultrasonography. All ultrasonographic examinations were performed within 2 days prior to the arthroscopic examination.
Arthroscopic examination
The same surgeon (SPF) who performed the ultrasonographic examinations also performed a complete examination of the affected stifle joint of each dog using a lateral parapatellar arthroscopy portal and a medial parapatellar instrument portal. An intra-articular stifle joint distractorf was placed immediately lateral to the distal aspect of the patella in situations in which additional distraction was deemed beneficial for thoroughly evaluating the meniscus. The medial meniscus was probed in all examined joints, and the surgeon recorded his initial diagnosis and performed treatment on the basis of the intraoperative assessment. The procedure was video recorded. Surgery was then performed to manage the CCL-deficient stifle joint in most dogs, whereas some dogs had meniscal treatment only.
Four to 8 months after arthroscopy, the video recordings were reviewed by the same surgeon who had performed the initial arthroscopy (SPF) and an additional surgeon (JLC) working independently and who was blinded to all results from all aspects of the study. All medial menisci were classified by each surgeon as requiring surgical treatment or not requiring surgical treatment. The specific diagnoses for which surgical treatment were recommended were the same as those described for MRI examinations.
Statistical analysis
Interobserver agreement regarding results of arthroscopic examination of medial meniscal lesions by the 2 surgeons was quantified by calculation of the Cohen κ statistic and use of the intraoperative assessment of the meniscus by the surgeon performing the arthroscopy. Intraobserver agreement (for SPF) in diagnoses made during the initial arthroscopic examination and subsequent review of the video recordings were similarly quantified. The CCR, sensitivity, and specificity of meniscal-click testing, flexion-induced-pain testing, ultrasonography, and MRI for diagnosing medial meniscal lesions requiring surgical treatment were determined, with arthroscopic examination used as the reference standard. The 95% CIs for sensitivity and specificity estimates were calculated as described elsewhere.33 Two sets of CCRs, sensitivities, and specificities were calculated: one for which the arthroscopic examination was performed by the surgeon (SPF) who was unblinded to the findings of physical, orthopedic, and ultrasonographic examinations, and the other for which the arthroscopic examination was performed by the surgeon (JLC) who used video recordings and who was blinded to the results of all diagnostic tests.
In addition, because the surgeon performing the ultrasonographic examination was not blinded to the results of the physical examination and may have been influenced by such knowledge, an additional set of calculations was performed to provide an unbiased assessment of the accuracy of ultrasonography for detecting medial meniscal lesions requiring surgical treatment. Previous studies7,9–11 have shown that a negative result of meniscal-click testing on orthopedic examination is diagnostically unhelpful. Therefore, the CCR, sensitivity, and specificity of ultrasonography (vs arthroscopy) were also calculated exclusively for those stifle joints for which no meniscal click was detected.
Results
Dogs and stifle joints
Thirty-one stifle joints of 26 dogs were included in the study. Orthopedic examination revealed a meniscal click in 6 (19%) stifle joints and signs of pain on flexion for 4 (13%) stifle joints. Ultrasonographic and MRI examinations revealed that the medial meniscus had damage for which surgical treatment was recommended in 21 (68%) and 15 (48%) stifle joints, respectively.
For 26 (84%) stifle joints (23 [88%] dogs), MRI examination had been performed with the dogs chemically immobilized via IV drug administration. General anesthesia was used for MRI examination of 5 (16%) stifle joints in 3 (12%) dogs, including bilateral stifle joint examination in 2 dogs and MRI examination and arthroscopy (plus tibial plateau leveling osteotomy) in the same anesthetic episode for another dog. The MRI examinations were performed within 2 days prior to arthroscopic assessment for 29 (94%) stifle joints; MRI preceded arthroscopy by 3 days for 1 (3%) joint and by 6 days for 1 (3%) other joint. Ultrasonographic examination of 5 stifle joints was performed with dogs chemically immobilized; 26 stifle joints were ultrasonographically examined while dogs were anesthetized and immediately preceding arthroscopy.
Intra- and interobserver arthroscopic agreement
At the time of initial arthroscopic examination, 20 of 31 (65%) stifle joints were judged to contain medial meniscal lesions necessitating surgical treatment by the surgeon who had not been blinded to findings of physical and orthopedic examinations or ultrasonography. When the video recordings were reviewed by the same surgeon 4 to 8 months later, grading of medial meniscal lesions remained the same except for 1 joint for which the surgeon had initially deemed the meniscus as requiring surgical treatment and performed a meniscal release at that time. However, on reexamination by the same surgeon, the medial meniscus was deemed abnormal but not sufficiently damaged to require surgical treatment. Hence, intraobserver agreement (repeatability) for arthroscopic examination was high (30/31 [97%] joints; κ = 0.93; 95% CI, 0.80 to 1.00).
When the initial arthroscopic assessments by the unblinded surgeon were compared with those of the blinded surgeon, agreement was again high (29/31 [94%]; κ = 0.85; 95% CI, 0.66 to 1.00). For the stifle joints of 2 dogs, the blinded surgeon believed that the medial meniscal lesions warranted surgical treatment, but the unblinded surgeon did not believe surgical treatment was warranted. In 1 dog, release of the medial meniscus was performed because, although treatment was considered unnecessary, the owner was disinclined to accept a risk of future meniscal injury and surgery. For the second dog, no surgical treatment of the meniscus was performed. This patient had a successful outcome as indicated through a conversation with the owner 10 months after arthroscopic evaluation and tibial plateau leveling osteotomy, inconsistent with a requirement for surgical treatment of the medial meniscus.
Diagnostic test comparisons
With each of the 2 surgeons' arthroscopic evaluations as the reference standard, CCRs, sensitivities, and specificities of meniscal-click testing, flexion-induced-pain testing, MRI, and ultrasonography for diagnosing medial meniscal lesions requiring surgical treatment were summarized (Table 1). The CCR and sensitivity were relatively low for meniscal-click testing and flexion-induced-pain testing, although the specificity of both of these tests was relatively high. Ultrasonography had a numerically higher CCR and sensitivity than MRI but a lower specificity. Magnetic resonance imaging had 100% specificity regardless of the surgeon performing the arthroscopic assessment.
Correct classification rate (%), sensitivity (%), and specificity (%) of various diagnostic tests, with arthroscopic examination of the affected stifle joint by each of 2 surgeons* as reference standards, for diagnosing medial meniscal lesions in the stifle joints (n = 31) of 26 dogs with CCL deficiency.
| Surgeon 1 | Surgeon 2 | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Diagnostic test | No. correctly classified as positive | No. correctly classified as negative | CCR | Sensitivity (95% CI) | Specificity (95% CI) | No. correctly classified as positive | No. correctly classified as negative | CCR | Sensitivity (95% CI) | Specificity (95% CI) |
| Meniscal click | 5 | 10 | 48 | 25 (10–49) | 91 (57–100) | 5 | 8 | 42 | 22 (9–46) | 89 (51–99) |
| Signs of pain on stifle-joint flexion | 3 | 10 | 42 | 15 (4–39) | 91 (57–100) | 3 | 8 | 35 | 14 (4–36) | 89 (51–99) |
| MRI of the stifle joint | 15 | II | 84 | 75 (51–90) | 100 (68–100) | 15 | 9 | 77 | 68 (45–82) | 100 (63–100) |
| Ultrasonography of the stifle joint | 19 | 9 | 90 | 95 (73–100) | 82 (48–97) | 19 | 7 | 84 | 86 (64–96) | 78 (40–96) |
Surgeon 1 (SPF) was not blinded to the physical and orthopedic examination findings (including detection of a meniscal click or signs of pain on stifle joint flexion), but was blinded to MRI findings when diagnosing medial meniscal lesions via arthroscopy. This surgeon also had limited experience with ultrasonography, having performed ultrasonographic examination of approximately 20 to 30 stifle joints before the study began, including validation of ultrasonographic findings with intraoperative findings. Surgeon 2 (JLC) used video recordings of the arthroscopic examinations performed by surgeon 1 for diagnosing medial meniscal lesions and was blinded to the results of all other diagnostic tests.
The CCR, sensitivity, and specificity of ultrasonography for diagnosing medial meniscal lesions were also calculated from data pertaining to only the 25 stifle joints with negative results of meniscal-click testing. With arthroscopic examination by the unblinded surgeon used as the reference standard, the CCR of ultrasonography was 88%, sensitivity was 93% (95% CI, 66% to 100%), and specificity was 80% (95% CI, 44% to 96%). With arthroscopic examination by the blinded surgeon used as the reference standard, these values were 80%, 82% (56% to 95%), and 75% (36% to 96%), respectively.
Discussion
The primary objective of the study reported here was to quantify and compare the CCR, sensitivity, and specificity of ultrasonography and MRI for diagnosing medial meniscal lesions requiring surgical treatment in dogs with CCL deficiency. In addition, we sought to assess the usefulness of ultrasonography for this purpose when the operator lacked extensive experience with this diagnostic technique. Findings indicated that ultrasonography was associated with a CCR of 80% to 90% and a sensitivity of 82% to 95% when including all manners of assessing the data, such as the use of both surgeons' arthroscopic assessments as the reference standard and when including or omitting dogs with a palpable meniscal click. These values, compared with the greater specificity (91%) but poorer CCR and sensitivity of meniscal-click testing, suggest that a veterinarian without extensive ultrasonography experience could use an orthopedic examination coupled with ultrasonography to non-invasively assess the medial meniscus in dogs with CCL deficiency. However, if veterinarians elect to use ultrasonography for this purpose, we recommend that they practice the technique and evaluate their accuracy by comparing their ultrasonographic assessments to intraoperative findings rather than presume that the study data could be generalized to their own clinical scenarios.
Values for sensitivity (68% or 75%, depending on the reference standard) and specificity (100%) of MRI for diagnosing medial meniscus lesions in the present study were identical to those reported for a previous study23 involving 14 dogs and a 1.5-T MRI unit. However, the accuracy of MRI in the present and previous study23 was lower than that reported for other MRI studies.15,22 Exact causes for the discrepancy among studies remain unclear, but the differences suggest possible subjectivity in high-field MRI assessments of meniscal integrity, as has been demonstrated for low-field MRI assessments of meniscal lesions.19 In addition, arthroscopy is reportedly more sensitive than arthrotomy for detecting meniscal lesions.4,13,25,26 The present study represents the first in which arthroscopy was used exclusively as the reference standard for quantifying the sensitivity and specificity of high-field MRI for detecting medial meniscus lesions. Therefore, the lower sensitivity of MRI obtained in the present versus previous studies may have been in part attributable to the use of a more sensitive reference standard to which MRI was compared.
Additional objectives of the present study included assessment of intra- and interobserver agreement in the arthroscopic assessment of medial meniscal lesions. This objective was pertinent because arthroscopy was used as the reference (gold) standard to which ultrasonography and MRI were compared. However, despite arthroscopy being considered the gold standard for diagnosing medial meniscal lesions in dogs, agreement among surgeons as to which meniscal lesions warrant surgical treatment has not previously been established.27 In the study reported here, intra- and interobserver agreement were high when the same arthroscopic video recordings were used, providing support for the use of arthroscopy as the reference standard.
An additional goal of the present study was to quantify the sensitivity and specificity of orthopedic examination for detecting meniscal lesions in dogs with CCL deficiency. Sensitivity of meniscal-click testing was low (22% or 25%, depending on the reference standard) and consistent with other reported values for the sensitivity of this test (27% to 58%).7,9–11 On the other hand, specificity was high (89% or 91%, depending on the reference standard), which is also consistent with other reported values for the specificity of this test (94% and 96%)9,10 Sensitivity of flexion-induced-pain testing was low (14% or 15%, depending on the reference standard), but specificity was high (89% or 91%). This is in contrast to a sensitivity of 77% and specificity of only 57% achieved in another study.9 The discrepancy in findings likely reflects that the surgeon in our study was more conservative in assessing a positive result of flexion-induced-pain testing, resulting in a lower sensitivity but higher specificity. Overall, we conclude that detection of signs of pain on stifle joint flexion cannot be used alone to accurately identify dogs with medial meniscal lesions requiring surgical treatment and that negative results of meniscal-click testing alone should not be used to rule out medial meniscal lesions.
Findings of the present study should be evaluated in light of the study limitations. One surgeon performed the orthopedic, ultrasonographic, and arthroscopic examinations; therefore, these 3 assessments were not performed in a blinded manner. However, this represents the realistic situation in which a veterinarian would not be blinded to the physical and orthopedic examination findings when performing ultrasonography to evaluate and treat medial meniscal lesions in dogs with lameness isolated to the stifle joint. Moreover, to overcome this limitation, we also calculated the accuracy of ultrasonography for detecting medial meniscal lesions using only stifle joints in which no meniscal click was detected on orthopedic examination (25/31). Likewise, the bias introduced by having the unblinded surgeon perform the arthroscopy was addressed by having a blinded surgeon review the arthroscopic video recordings. We believe that the different analyses performed provided both conservative and optimistic quantification of the diagnostic accuracy of ultrasonography. The CCRs, sensitivities, and specificities of ultrasonography were generally consistent regardless of whether a meniscal click was detected and regardless of which surgeon's arthroscopic assessment was used as the reference standard. The results reported here are also consistent with those of previous studies.11,12 Collectively, findings suggested that ultrasonography and MRI are clinically useful diagnostic techniques for assessment of the medial meniscus in dogs with CCL deficiency
Acknowledgments
Supported by Hitachi-Aloka Medical America and SCiL Animal Care Company.
ABBREVIATIONS
| CCL | Cranial cruciate ligament |
| CCR | Correct classification rate |
| CI | Confidence interval |
| TSE | Turbo spin echo |
Footnotes
Dexdomitor, Zoetis Animal Health, Parsippany NJ.
Nubain, Hospira, Lake Forest, Ill.
Ketaset (100 mg/mL), Henry Schein, Plainview, NY.
Magnetom Symphony ATS, Siemens, Malvern, Pa.
Noblus, Hitachi-Aloka Medical America Inc, Wallingford, Conn.
Ventura stifle thrust level, Imex Veterinary Inc, Longview, Tex.
References
1. Wilke VL, Robinson DA, Evans RB, et al. Estimate of the annual economic impact of treatment of cranial cruciate ligament injury in dogs in the United States. J Am Vet Med Assoc 2005; 227:1604–1607.
2. Gambardella PC, Wallace LJ, Cassidy F. Lateral suture technique for management of anterior cruciate ligament rupture in dogs: a retrospective study. J Am Anim Hosp Assoc 1981; 17:33–38.
3. Ralphs SC, Whitney WO. Arthroscopic evaluation of menisci in dogs with cranial cruciate ligament injuries: 100 cases (1999–2000). J Am Vet Med Assoc 2002; 221:1601–1604.
4. Ritzo ME, Ritzo BA, Siddens AD, et al. Incidence and type of meniscal injury and associated long-term clinical outcomes in dogs treated surgically for cranial cruciate ligament disease. Vet Surg 2014; 43:952–958.
5. Franklin SP, Gilley RS, Palmer RH. Meniscal injury in dogs with cranial cruciate ligament rupture. Compend Contin Educ Pract Vet 2010; 32:E1–E10.
6. Metelman LA, Schwarz PD, Salman M, et al. An evaluation of 3 different cranial cruciate ligament surgical stabilization procedures as they relate to postoperative meniscal injuries: a retrospective study of 665 stifles. Vet Comp Orthop Traumatol 1995; 8:118–123.
7. Case JB, Hulse D, Kerwin SC, et al. Meniscal injury following initial cranial cruciate ligament stabilization surgery in 26 dogs (29 stifles). Vet Comp Orthop Traumatol 2008; 21:365–367
8. Wucherer KL, Conzemius MG, Evans R, et al. Short-term and long-term outcomes for overweight dogs with cranial cruciate ligament rupture treated surgically or nonsurgically J Am Vet Med Assoc 2013; 242:1364–1372.
9. Dillon DE, Gordon-Evans WJ, Griffon DJ, et al. Risk factors and diagnostic accuracy of clinical findings for meniscal disease in dogs with cranial cruciate ligament disease. Vet Surg 2014; 43:446–450.
10. Neal BA, Ting D, Bonczynski JJ, et al. Evaluation of meniscal click for detecting meniscal tears in stifles with cranial cruciate ligament disease. Vet Surg 2015; 44:191–194.
11. Arnault F, Cauvin E, Viguier E, et al. Diagnostic value of ultrasonography to assess stifle lesions in dogs after cranial cruciate ligament rupture: 13 cases. Vet Comp Orthop Traumatol 2009; 22:479–485.
12. Mahn MM, Cook JL, Cook CR, et al. Arthroscopic verification of ultrasonographic diagnosis of meniscal pathology in dogs. Vet Surg 2005; 34:318–323.
13. Pozzi A, Hildreth BE III, Rajala-Schultz PJ. Comparison of arthroscopy and arthrotomy for diagnosis of medial meniscal pathology: an ex vivo study. Vet Surg 2008; 37:749–755.
14. Marino DJ, Loughin CA. Diagnostic imaging of the canine stifle: a review. Vet Surg 2010; 39:284–295.
15. Barrett E, Barr F, Owen M, et al. A retrospective study of the MRI findings in 18 dogs with stifle injuries. J Small Anim Pract 2009; 50:448–455.
16. Galindo-Zamora V, Dziallas P, Ludwig DC, et al. Diagnostic accuracy of a short-duration 3 Tesla magnetic resonance protocol for diagnosing stifle joint lesions in dogs with nontraumatic cranial cruciate ligament rupture. BMC Vet Res 2013; 9:40.
17. Gonzalo-Orden JM, Altonaga JR, Gonzalo Cordero JM, et al. Magnetic resonance imaging in 50 dogs with stifle lameness. Eur J Companion Anim Pract 2001; 11:115–118.
18. Martig S, Konar M, Schmökel HG, et al. Low-field MRI and arthroscopy of meniscal lesions in ten dogs with experimentally induced cranial cruciate ligament insufficiency. Vet Radiol Ultrasound 2006; 47:515–522.
19. Böttcher P, Armbrust L, Blond L, et al. Effects of observer on the diagnostic accuracy of low-field MRI for detecting canine meniscal tears. Vet Radiol Ultrasound 2012; 53:628–635.
20. McCartney W, McGovern F. Optimising the MRI protocol for imaging of the canine stifle meniscus using a low field system. Int J Appl Res Vet Med 2011; 9:392–395.
21. McCartney WT, McGovern F. Use of low-field MRA to presurgically screen for medial meniscus lesions in 30 dogs with cranial cruciate deficient stifles. Vet Rec 2012; 171:47
22. Blond L, Thrall DE, Roe SC, et al. Diagnostic accuracy of magnetic resonance imaging for meniscal tears in dogs affected with naturally occuring cranial cruciate ligament rupture. Vet Radiol Ultrasound 2008; 49:425–431.
23. Olive J, d'Anjou MA, Cabassu J, et al. Fast presurgical magnetic resonance imaging of meniscal tears and concurrent subchondral bone marrow lesions. Study of dogs with naturally occurring cranial cruciate ligament rupture. Vet Comp Orthop Traumatol 2014; 27:1–7
24. Böttcher P, Brühschwein A, Winkels P, et al. Value of low-field magnetic resonance imaging in diagnosing meniscal tears in the canine stifle: a prospective study evaluating sensitivity and specificity in naturally occurring cranial cruciate ligament deficiency with arthroscopy as the gold standard. Vet Surg 2010; 39:296–305.
25. Plesman R, Gilbert P, Campbell J. Detection of meniscal tears by arthroscopy and arthrotomy in dogs with cranial cruciate ligament rupture: a retrospective, cohort study. Vet Comp Orthop Traumatol 2013; 26:42–46.
26. Thieman KM, Tomlinson JL, Fox DB, et al. Effect of meniscal release on rate of subsequent meniscal tears and owner-assessed outcome in dogs with cruciate disease treated with tibial plateau leveling osteotomy. Vet Surg 2006; 35:705–710.
27. Wustefeld-Janssens BG, Pettitt RA, Cowderoy EC, et al. Peak vertical force and vertical impulse in dogs with cranial cruciate ligament rupture and meniscal injury. Vet Surg 2016; 45:60–65.
28. Alizadeh A, Babaei Jandaghi A, Keshavarz Zirak A, et al. Knee sonography as a diagnostic test for medial meniscal tears in young patients. Eur J Orthop Surg Traumatol 2013; 23:927–931.
29. Cook JL, Cook CR, Stannard JP, et al. MRI versus ultrasonography to assess meniscal abnormalities in acute knees. J Knee Surg 2014; 27:319–324.
30. Shetty AA, Tindall AJ, James KD, et al. Accuracy of handheld ultrasound scanning in detecting meniscal tears. J Bone Joint Surg Br 2008; 90:1045–1048.
31. Timotijevic S, Vukasinovic Z, Bascarevic Z. Correlation of clinical examination, ultrasound sonography, and magnetic resonance imaging findings with arthroscopic findings in relation to acute and chronic lateral meniscus injuries (Erratum published in J Orthop Sci 2014; 19:375). J Orthop Sci 2014; 19:71–76.
32. Nguyen JC, De Smet AA, Graf BK, et al. MR imaging-based diagnosis and classification of meniscal tears. Radiographics 2014; 34:981–999.
33. Newcombe RG. Two-sided confidence intervals for the single proportion: comparison of seven methods. Stat Med 1998; 17:857–872.
