Quantitative evaluation of canine cranial cruciate ligament disease by stress radiography

BACKGROUND To establish a quantitative parameter reflecting displacement of the tibia relative to the femoral condyle in dogs with damage to the cranial cruciate ligament (CrCL). ANIMALS One hundred and forty-eight client-owned dogs with intact (n=34), partial rupture (n=43), or complete rupture (n=93) of the CrCL confirmed by arthroscopy of the stifle joint. PROCEDURES The Cranial Tibial Displacement Index (CTDI) was measured on mediolateral radiographs obtained in lateral recumbency with the tarsal and stifle joints at 90° of flexion. The sensitivity and specificity of the CTDI for assessment of CrCL damage was compared with that of the cranial drawer test, tibial compression test, fat pad sign, and radiographic OA score. mean CTDI were 0.8±0.2, and 1.2±0.3 in the groups an the differences sensitivity, specificity, CTDI for distinguishing respectively.

CONCLUSIONS AND CLINICAL RELEVANCE The CTDI, which quantifies cranial displacement of the tibia, could be an objective indicator reflecting the grade of CrCL damage. CrCL injury could be evaluated indirectly from the CTDI score, which may allow a minimally invasive early diagnosis of a ruptured CrCL.

Background
Rupture of the cranial cruciate ligament (CrCL) is a common injury in dogs, and the main cause of osteoarthritis (OA) and degenerative disease in the stifle joint 1 . Although rupture of the anterior cruciate ligament can occur acutely with trauma in humans, most cases of CrCL rupture (CrCLR) occur secondary to chronic degenerative changes combined with chondrometaplasia of the CrCL 2 . Although the precise mechanism of degeneration of the CrCL has not been clarified, some causative factors have been reported. These include trauma, immune-mediated mechanisms, age-related degeneration, obesity, an excessive tibial plateau angle (TPA), and conformational abnormalities, such as patellar luxation and a narrowed intercondylar notch 3 − 6 . These are collectively referred to as CrCL disease 2,7 . However, this disease has been recognized as a degenerative process that ultimately leads to partial or complete ligament rupture, instability, and secondary degenerative joint disease 8, 9 . Instability of the stifle joint following CrCLR is an important pathophysiologic mechanism leading to development of progressive OA and stifle joint "organ" failure. 8,10− 13 Although an early partial tear is likely to progress to a complete tear if left untreated because of the degenerative cascade of effects that occurs, the benefits of surgery may not outweigh the risks. However, early diagnosis of a CrCL injury and treatment by tibial-plateau-leveling osteotomy may protect against further disruption of the CrCL, lending stability to the joint and decreasing the risk of meniscal injury and damage to the articular cartilage. 14 Therefore, for maintenance of normal stifle joint function, it is important to diagnose and treat a partial rupture early and prevent transition to osteoarthritis and complete rupture.
A number of techniques and medical imaging systems have been used to diagnose canine CrCLR, including palpation, radiography, arthroscopy, computed tomography (CT), magnetic resonance imaging (MRI), and ultrasonography. In clinical practice, CrCLR is usually diagnosed by orthopedic and radiographic examination. A cranial drawer test 4 (CDT) or tibial compression test (TCT) is used to evaluate craniocaudal instability of the stifle joint 15 . It was determined that the CDT alone or combined with the TCT could not be used to differentiate the cause of stifle instability associated with rupture of the CrCL, caudal cruciate ligament, or total cruciate ligament 16 . It has been pointed out that many dogs with chronic CrCLR do not have palpable instability of the stifle joint because of chronic periarticular fibrosis, and the sensitivity of the CDT and TCT was surprisingly low when performed on conscious patients. 17 Therefore, more objective tools are needed to accurately detect disorders of the stifle joint. Radiographs showing joint effusion, osteophytosis, or cranial tibial displacement support a diagnosis of CrCLR. Radiologically, there might be evidence of osteophyte formation and/or changes to the infrapatellar fat pad shadow 4,18 . A stress view radiograph of the stifle might be helpful for observing cranial displacement of the tibia in dogs with CrCLR 19 . Arthroscopy is a diagnostic and therapeutic procedure that provides a direct view of the articular surface for the purpose of integral evaluation of the joint. The disadvantage is that it is an invasive procedure that requires major technologic infrastructure and many hours of specialized training 20 .
Ultrasonography, which involves non-invasive, non-ionizing radiation and is portable and less expensive, is a useful technique for evaluation of intraarticular proliferation of reactive fibrotic tissue in unstable stifle joints with CrCLR as a result of chronic synovitis 21 . However, expert operators are mandatory for ultrasonography to be useful. Furthermore, ultrasonography is not an accurate test for CrCLR, but is specific for the pathologic changes in the soft tissue that occur as a consequence of joint instability 21,22 .
Other useful noninvasive techniques are scintigraphy 23 , CT, and MRI 24 ; however, these modalities are not routinely available in veterinary medicine. Arthroscopy and MRI may be 5 useful for confirming the diagnosis; however, arthroscopy is invasive and MRI is expensive 23,25 . Meanwhile, CT emits a high level of radiation and has limited ability to detect a fracture. Again, CT and MRI require expert operators to be useful. Undeniably, all these medical imaging systems have some disadvantages.
We have developed a quantitative method to evaluate instability of the stifle joint related to insufficiency of the CrCL. We evaluated and compared the extent of CrCL damage using an index measured on radiographs taken in the stress position in dogs with intact stifle joints and in those with partial or complete CrCLR confirmed by arthroscopy. and results of clinical examination available. Information on duration of lameness of the affected limb was obtained from the owner. All the owners provided informed consent, and participation in the study did not impact decisions regarding clinical management.

Study groups
The study population included a control group, a group with partial CrCLR, and a group 6 with complete CrCLR. The control group consisted of 10 dogs which have CrCLR on the opposite side and 12 no abnomal dogs without any orthopedic disease that were presented to our hospital. The 34 hind limbs in this group were confirmed to be healthy on the basis of a normal physical examination (with no lameness) and normal orthopedic, radiographic, and arthroscopic examination results. Our hospital is secondary hospital and specializes in orthopedic surgery especially stifle arthroscopy, so some aggressive owners come to our hospital to inspect their dogs with stifle arthroscopy and we always CrCL was graded as intact (no damage, with continuity), partially ruptured (damage present, continuity preserved), or completely ruptured (damage present, continuity disrupted). The majority of the lesions were close to the femoral insertion site and there were no distal lesions ( Fig. 1). At last, all inspection including arthroscopy was conducted under orner's consent. In particularly, the dogs classified as normal were inspected by 7 orner's strong request.

Stress radiography
Radiography was performed without sedation or anesthesia. During this examination, we referred to the TCT, which is an orthopedic procedure performed at the time of clinical diagnosis of CrCLR. Mediolateral radiographs were obtained with the dogs in lateral recumbency and the tarsal and stifle joints at 90° of flexion with the tibia parallel to the digital image capture device. The X-ray beam was centered over the stifle joint and collimated to the tarsus, entire tibia, and distal half of the femur. If the intercondylar eminences were not exactly superimposed, the midpoint of each eminence was identified.
The femoral condyles and talar trochlea were superimposed to achieve the correct rotational alignment. The radiographs were obtained using a digital system and assessed using a DICOM viewer (Sedecal, Entrad). Radiographs of the stifle joints with superimposition of the femoral condyles ≤ 2 mm and flexion angles of 60°-120° were assessed. 30 If the femoral condyles were not exactly superimposed, a best-fit circle was drawn around each femoral condyle and the center of each condyle was identified.
The radiographs were digitized for analysis and analyzed using the DICOM viewer software. The cranial and caudal aspects of the medial tibial condyle were used as landmarks for the proximal tibial joint orientation line ( The mechanical axis defined as the straight line extending from the midpoint between the apices of the two tibial intercondylar eminences of the tibial plateau to the center of the 8 circle created by the tarsus was used, as described elsewhere 34,36,37 . The tibial anatomic axis was defined as the line connecting the midpoint between the cranial and caudal cortex of the tibia at 50% and 75% of the length of the tibial shaft (Fig. 4). 38− 41 All measurements were made by the same investigator (TK).
The angle of the stifle joint was measured as the angle formed between the long axis of the distal femur and the tibia in all cases 42 . This angle (''O'') was determined by drawing a segment (FW) between the femoral cortices at a distance equal to the length of the femoral condyle from the proximal extent of the trochlea. A segment (''D'') parallel to the FW was drawn 20 mm proximal to the FW. A line was then drawn joining the centers of the FW and D segments to define the long axis of the distal femur. The anatomic axis of the tibia was drawn as previously defined. The stifle joint angle was measured at the intersection of these two axes (Fig. 5). In all stifles, the actual angle of the tarsal joint were also measured from the two long axes of the previous tibial axis and the tarsal axis that was defined as the lines between the two shafts, midpoints at the intertarsal joint, and the tarsal over metacarpal joint (Fig. 6) 30 . For all stifle joints, the TPA was measured on a mediolateral radiograph acquired in the stress position according to the method devised by Warzee et al 43 .

Measurement of Cranial Tibial Displacement Index
A circle was drawn on the mediolateral radiograph to fit the circumference of the femoral condyles (the average of the circle both lateral and medial femoral condyles, more fit in femoral condyles). The center of the femoral condyle was then determined. Next, the radius of the femoral condyle (C) was measured in centimeters. A line was drawn from the center of the femoral condyle parallel to the line of the mechanical axis of the tibia. The distance (α, in centimeters) between them was then measured. The ratio of α to C, i.e., the tibial cranial displacement index (CTDI), was calculated to correct for differences in body size (Fig. 7).
We determined the cut-off CTDI values for intact and partially ruptured CrCLs. In the present study, to determine the normal range of the CTDI, we initially excluded measurements obtained at the extremes on either side and then calculated the mean of the values obtained in the normal group.

Palpation
The CDT and TCT were performed in lateral recumbency without sedation or anesthesia in all cases using the techniques and procedures described, respectively, by Jerram and Radiographic OA score The severity of OA in the stifle joints of dogs in the CrCLR group was graded on radiographs using an established scoring system 48 . Scores obtained using this system range from 0 (no OA sign, i.e., osteophytes absent) to 4 (severe, osteophytes present on the patella, femoral trochlear groove, medial and lateral femoral condyles, and tibial plateau, i.e., subchondral sclerosis of the femoral condyles).

Sensitivity, specificity, and positive and negative predictive values
The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of the CDT, TCT, fat pad sign, ROS, and CTDI in the diagnosis of CrCL were compared. For the CDT and TCT, the result was recorded as "positive" (over 0) or "negative". CTDI cut-off values of 0.62 and 1.01 were identified as the thresholds for an intact CrCL and a partial CrCLR, respectively. Sensitivity is defined as the proportion of true-positive results identified correctly and specificity is the proportion of true-negative results identified correctly. The PPV is the proportion of positive tests that are truly positive and the NPV is the proportion of negative tests that are truly negative.

Statistical analysis
Tukey's honestly significant difference (HSD) test and the Games-Howell test were used to identify statistically significant between-group differences in patient age and body weight, joint angle, and TPA between the three study groups. The relationships between arthroscopic CrCL grade and the study variables (α, C, α/C ratio, stifle angle, tarsal angle, TPA, age, body weight, and duration of lameness) were determined using the Spearman's rank correlation. The data are presented as the mean and standard deviation or as the number or percentage. A p-value < 0.05 was considered statistically significant for all comparisons.

Arthroscopy
Thirty-four stifle joints were classified as intact (no damage or vascularization of the CrCL, complete continuity), 43 as partially ruptured (various degrees of damage and vascularization, most being close to the femoral insertion site, but preserved continuity of the CrCL), and 93 as completely ruptured. The caudal cruciate ligament was confirmed to be intact by arthroscopic observation in all three-study groups (Table 1).

Demographic and clinical data
The 148 dogs were divided into an intact group, partial rupture group, and complete rupture group based on arthroscopic and intraoperative findings. Detailed information on sex, age, and body weight are provided in Table 1. There was no significant between-group difference in age (intact vs complete, P = 0.029; intact vs partial, P = 0.682; partial vs complete, P = 0.184). The dogs in the complete and partial rupture groups were significantly heavier than those in the intact group (both P = 0.001).

Tibial plateau angle
The mean TPA was 25.4 ± 3.9° (range, 14.6°-32.2°) in the intact group, 27.0°±4.3°( range, 13.5°-37.0°) in the partial rupture group, and 28.6°±4.2° (range, 20.0°-37.0°) in the complete rupture group (Table 1). There was a significant difference in TPA between the intact group and the complete rupture group (P = 0.001) but not between the intact and partial rupture groups or between the partial and complete rupture groups (P = 0.21 and P = 0.1, respectively).

Cranial Tibial Displacement Index
In the intact group, the mean α value was 3.0 ± 1.1 mm, the mean C value was 6.0 ± 13 0.3 mm, and the mean CTDI score was 0.5 ± 0.1; the respective values were 5.8 ± 2.2 mm, 7.9 ± 0.4 mm, and 0.8 ± 0.2 in the partial rupture group and 8.3 ± 3.0 mm, 7.3 ± 0.2 mm, and 1.2 ± 0.3 in the complete rupture group. The mean CTDI cut-off values for distinguishing between the partial and complete rupture groups and between the intact and partial rupture groups were 1.01 and 0.62, respectively. The radius of the femoral condyle was significantly larger in the complete and partial rupture groups than in the intact group (P = 0.002 and P = 0.011, respectively), with no significant difference between the partial rupture group and the complete rupture group (P = 0.38). There were significant differences in the α value and CTDI score between the three study groups (P < 0.001; Table 2) The mean actual stifle angle on stress radiography was 88.  (Table 1).
There were no significant between-group differences in the stifle angle (intact vs complete, P = 0.458; intact vs partial, P = 0.05; partial vs complete, P = 0.232). The tarsal joint angle was significantly smaller in the complete rupture group than in the intact group (P = 0.002); however, there was no significant difference between the intact and partial rupture groups (P = 0.433) or between the partial and complete rupture groups (P = 0.076).

Palpation
The CDT and TCT results were negative in all dogs in the intact group. In the partial rupture group, the CDT result was positive in 67.6% of dogs, the mean CTDI scores in the joints that were CDT-positive and CDT-negative were 0.8 ± 0.3 and 0.8 ± 0.2, respectively, the TCT result was positive in 52.9% of cases, and the mean CTDI scores in the joints that were TCT-positive and TCT-negative were 0.8 ± 0.3 and 0.8 ± 0.2, respectively (Fig. 9). In the complete rupture group, the CDT result was positive in 94.4% of dogs, the mean CTDI scores in the joints that were CDT-positive and CDT-negative were 1.2 ± 0.3 and 0.8 ± 0.2, respectively, the TCT results was positive in 93.1% of cases, and the mean CTDI scores in the joints that were TCT-positive and TCT-negative were 1.2 ± 0.3 and 0.9 ± 0.2, respectively.
Correlations between CTDI and the variables measured were determined, and the results are summarized in Table 3. No association of CTDI with age, sex, angle, TPA, weight, or duration of lameness was detected in any of the study groups (Table 2).

Fat pad sign
None of the dogs in the intact group had the fat pad sign. In the partial rupture group, a grade 0 fat pad sign was present in 25.6% of cases, a grade 1 sign in 39.5%, and a grade 2 sign in 34.9%. The mean CTDI score was 0.8 ± 0.3 for the joints with a grade 0 fat pad sign, 0.7 ± 0.2 for those with a grade 1 sign, and 0.8 ± 0.2 for those with a grade 2 sign. In the complete rupture group, a grade 0 fat pad sign was present in 16.1% of cases, a grade 1 sign in 41.9%, and a grade 2 sign in 41.9%. The mean CTDI score was 1.2 ± 0.3 for the joints with a grade 0 fat pad sign, 1.2 ± 0.4 for those with a grade 1 sign, and 1.2 ± 0.3 for those with a grade 2 sign.

Sensitivity, specificity, PPV, and NPV
The sensitivity, specificity, PPVs and NPVs values for the CDT, TCT, fat pad sign, ROS, and CTDI score are summarized in Table 4.

Discussion
There was a significant difference in the CTDI score between the intact group, partial rupture group, and complete rupture group in this study, suggesting that the CTDI, which quantifies the cranial displacement of the tibia relative to the femoral condyle, could be an objective indicator of the degree of CrCL damage. Our present findings suggest that we may be able to indirectly evaluate damage to the CrCL from the CTDI score, potentially enabling minimally invasive early diagnosis of CrCLR (Fig. 10).
The significant difference in the femoral radius between the partial rupture group and the intact group may reflect differences in body weight and physique. The tarsal joint angle was significantly lower in the complete rupture group than in intact group. The target joint angle on radiography was set to 90°. We adopted this stored position taking into account the reproducibility at the time of acquiring the images. However, in CrCLR cases, the tibia is displaced forward and the anatomic axis is also tilted forward, so the tarsal joint tends to flex easily 24 . Although the influence of the difference in the tarsal joint angle is unknown, it is considered that the 90° imaging method by visual observation is useful because a significant difference in the CTDI score was found in this study. However, in a recent report, it was shown that cranial displacement of the tibia is maximized when the tarsal joint is overflexed. 24 Measurement at the time of overflexion should be included in a further study, because it is possible that the difference in CTDI due to the degree of damage will become more awkward.
It is considered that various factors, including TPA, OA scores, the hamstring muscles, medial buttress, periarticular fibrosis, and intra-articular structure, would have an influence on the CTDI 49 − 54 . There are several studies to demonstrate that these factors affect CrTT and lead to degenerative joint disease and CrCLR 55 − 62 . Especially as for muscle tension, our study implements under no anesthesia. In this circumstance, some dogs are not corporate inspection because of such as fear or tension, may not indicate true CTDI. It is considered to be an issue. When the CrCLR is chronic and the circumference of the joint has become fibrotic and fluctuation is reduced, 90° of stress during palpation or radiography may reduce the thrust forward propulsion force. No correlation was found between the days since onset and the CTDI score in our study.
However, the CTDI score may be decreased in cases where the stifle joint is stabilized by periarticular fibrosis because of a chronic course. Furthermore, in our study, the complete rupture group had a higher TPA than the intact group. CrTT is generated by loading the body weight to the stifle joint surface and TPA. Increases in posterior inclination of the tibial plateau have been associated with increases in cranial translation of the tibia during monopodial stance in humans 63 . This finding suggests a similar situation in canine stifle joints. When a stifle joint in a dog with CrCLR show greater posterior inclination of the tibial plateau, cranial translation of the tibia might increase. In previous experiments using canine cadavers, the tibial plateau leveling procedure was consistently successful in eliminating cranial translation of the tibia during axial loading 43,55 . This finding means that the CrTT could change with the degree of TPA, and a greater CrTT means greater CTDI. However, it is difficult to quantify the force generated by the hamstring muscles, the antagonistic effects of the medial buttress, and periarticular fibrosis against CrTT.
The fat pad sign, CDT, and TCT are diagnostic tests for CrCLR in clinical practice, and have been extensively studied 65 . Based on that earlier study, we compared the positive rates for each examination performed in this study. In our study, retention of joint fluid was confirmed at about 80% of joints in the complete rupture group, 93% of which showed no abnormality in CDT or TCT. However, in the partial rupture group, about 30% did not show retention of joint fluid and no abnormality was detected in palpation. Therefore, we compared the CTDI with the fat pad sign with or without palpation in the complete rupture group and the partial rupture group. In the complete rupture group, only four cases that were negative by palpation were mildly low, but the other values were similar in the other 95% of cases. In the partial rupture group, the CTDI score was almost the same regardless of the level of joint fluid retention or findings on palpation. Therefore, it is possible that the CTDI can evaluate damage that cannot be detected by palpation or the fat pad sign.
According to a study of humans, the partial rupture is categorized as 25-50% or 50-75% ligament rupture 64 . On the other hand, in veterinary orthopedics, partial rupture of the canine CrCL is classified clinically as functioning or non-functioning depending on the function of the remaining ligament as qualitative. So study of the relationship between the CTDI score and the degree of damage as quantitative is needed in the future.
The sensitivity of the CTDI was higher than that of the CDT, TCT, fat pad sign, and ROS.
We evaluated the fat pad sign as an indicator of OA of the stifle joint in the present study.
The criteria used were selected based on high reproducibility 37,38 . The fat pad sign is one of the earliest and most consistent findings in dogs with CrCLR and is consistent with joint effusion, edema of the fat pad, or periarticular fibrosis 50 − 52 . However, the fat pad sign and ROS only evaluate evidence of secondary arthritis of the stifle joint such as tumor or IMPA, so are not specific indicators of CrCLR.
Moreover, both the CDT and TCT have the potential to yield false-negative results. I suggested that both these tests gave similar results for a recent CrCLR, but are unreliable in the diagnosis of chronic failure of the CrCL 4,44,66 . There is no single test or sign has been shown to be universally applicable or reliable. In summary, measurement of the CTDI may yield more objective results and a more reliable diagnosis.
In one study, the sensitivity of the CDT when performed on conscious patients was surprisingly low. The CDT in conscious dogs is perhaps the most commonly used diagnostic test for CrCLR, and the findings of our study indicate that reliance on this test in conscious patients will result in significant numbers of CrCL cases being misdiagnosed as cruciate-normal, and 40% of dogs with proven CrCL failure do not have cranial instability detectable by the CDT. 17 It is normal that craniocaudal movement beyond the 0-2 mm of mobility found in a normal stifle joint. On the other hand, if a partial tear is present, the cranial drawer sign may reveal only 2-3 mm of instability when the test is performed with the stifle flexed and no instability with the stifle in extension 5 . In addition, one study found that 12 of 25 dogs with partial rupture of the CrCL had no detectable cranial drawer sign in response to manipulation of the involved stifle 67 . Hence, it is not surprising that veterinarians encounter difficulties in diagnosing partial ruptures of the CrCL with palpation. Subjective evaluation on palpation frequently varied between observers and agreed poorly. Therefore, a diagnosis of CrCLR may be qualitative and 19 subjective. Our results show that CTDI may be an objective index for diagnosis of CrCLR under no anesthesia and sedation with minimally invasive.

Conclusion
In our study, radiograph showed that the CTDI may be the most beneficial, objective, and minimally invasive diagnostic technique for indirect evaluation of the CrCL.
However, our study has some limitations and was performed in a small population, so further research is needed to identify differences according to canine breed, sex, age, muscle strength in the hind leg as well as the influence of joint fibrosis, comparison of the diagnosis rate using the fat pad sign or palpation, and classification of a partial tear.

20
Written informed consent was obtained from all the study participants, including consent to participate and to publish the findings.

Consent for publication
Not applicable

Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.
During the student at Tottori University, I worked on TRPA1 channel under the guidance of Pharmacology professor Ohta T.
Last year I presented a poster at ACVS in Pheonix and oral presentation at AiSVS. Now I have submitted the paper about CrCLR to BMC with Dr. Brian Beale who is a small animal surgical specialist in the US.

Figure 1
Evaluation of the degree of damage to the cranial cruciate ligament by arthroscopy, i.e., intact (no damage, with continuity, partial tear (with damage and continuity, no ligament is completely ruptured, but none is intact), or complete tear (damage, no continuity). The majority of lesions were close to the femoral insertion site with no distal lesion.   FW, femoral width, which is measured at a distance equal to FCL from X; D, a segment parallel to FW and drawn 20 mm proximal to FW. Measurement of the tibial cranial displacement index. A circle to fit the circumference of the femoral condyles (average of the circle for both the lateral and medial femoral condyles, more fit in the femoral condyles) was drawn, and then the center of the femoral condyle was determined and the radius of the femoral condyle was measured (C). A parallel line was drawn from the center of the femoral condyle to the tibial mechanical axis line. The distance (α) between these two points was measured. The ratio of α to C , known as the Tibial cranial Displacement Index (CTDI), was measured.
35 Figure 8 Outcome of measurement of the Tibial Cranial Displacement Index in the intact, partial rupture, and complete rupture groups.
36 Figure 9 Comparison of the Tibial Cranial Displacement Index for each test.
37 Figure 10 Comparison of Tibial Cranial Displacement Index score, radiographic results, and arthroscopic findings at the stifle joint between each of the three study groups.