The medical records and radiographs of dogs admitted to the Veterinary Medical Teaching Hospital of Konkuk University for CrCL rupture from March 2007 to June 2013 were evaluated retrospectively. Thirty-three stifle joints from 26 dogs showing clinical recovery after extracapsular surgery (ES) were included in this study. Initial inclusion criteria were complete CrCL rupture confirmed during surgery and that the affected stifle joint was stabilized by ES. Dogs with medial patellar luxation (MPL) were included because the affected animals were mostly small breeds. Dogs that had undergone tibial tuberosity transposition as well as ones with meniscal damage, collateral ligament damage, or a caudal cruciate ligament rupture were excluded from the investigation.

Radiographs were evaluated to select dogs with preoperative plain and stress views along with postoperative plain views of the stifle joint within 1 month after surgery. The postoperative stress view was not examined because it was not taken into account during surgery. Radiographs of stifle joints with ≤ 2 mm superimposition of the femoral condyles and little to no degenerative joint disease were assessed.

Medical records of dogs presented to the Veterinary Medical Teaching Hospital of Konkuk University from March 2007 to June 2013 were evaluated retrospectively. A total of 105 normal stifle joints from 79 dogs were examined. The animals were confirmed to be free of stifle joint disorders based on medical history, orthopedic examination, and radiographs. Radiographs were evaluated to select dogs with mediolateral views of stifle joints that had a ≤2 mm superimposition of femoral condyles.

Mediolateral radiographs of stifle joints were digitized for analysis using S View software (ver. 1.5.5.3; Comed Medical Systems, Korea). Preoperative plain and stress mediolateral views of the affected stifle joint along with postoperative plain mediolateral views were obtained. All radiographic views focused on the tibiofemoral joint space and extended at least to the level of the mid-diaphysis of the tibia and femur. Landmarks for measurements were based on those described in a previous study by Dennler et al. [2].

The flexion angle of the stifle joint (β) was measured from two long axes of the femur and tibia that were defined as the lines between the two shafts, midpoints at one-third and two-thirds of the femoral and tibial lengths (Fig. 1). The angle between the proximal border of the patellar ligament and tibial plateau (α) was measured on each radiograph. The tibial plateau line extended from the point on the cranial margin of the tibial plateau at which the CrCL inserted to the point on the caudal margin of the tibia at which the caudal cruciate ligament was inserted. Angles α and β in all patients were evaluated by the same observer. In our preliminary study (unpublished data) evaluating the reliability of measuring the patellar ligament angle, 12 normal stifle joints of six beagles were included. Mediolateral radiographic plain and stress views of the stifle joints were obtained, and plain radiography was performed three times for three groups of flexion angles (angle β 40~60°, 60~80°, and 80~100°). The PLA was evaluated according to stifle joint flexion by three independent observers. Each observer measured three times and calculated the average angle α. Significant differences in the PLA were observed among the flexion angle groups (p < 0.05). No significant difference in PLA was observed between the plain and stress views. The inter-class correlation coefficient of the PLA obtained from each observer was 0.971 (p < 0.05). The intra-class correlation of measurement for the PLA was 0.940 (p < 0.05).

All dogs underwent lateral or medial parapatellar arthrotomy to explore the affected stifle joint. Remnants of the torn CrCL were resected, and joint lavage was performed with sterile saline. Stabilization procedures using extracapsular methods included lateral fabellotibial suturing along with modified medial and lateral retinacular imbrication techniques. Fisher, polyester, or polyethylene line was used as the implant material. Sutures were manually tied or connected with a crimp. Closure was performed using a routine technique. All patients had resistance to the cranial drawer test and showed complete clinical recovery after surgery.

The data analysis was performed using SPSS for Windows (ver. 19.0; SPSS, USA). The mean and range for angles α and β were calculated. Relationships between angles α and β were analyzed by simple linear regression with a 95% confidence interval (CI). PLA preoperative and postoperative views were compared using an independent t-test. The capacity of PLA measurements to identify a CrCL rupture was evaluated using receiver operating characteristic (ROC) curves. Differences were considered significant if p < 0.05.

For the 33 stifle joints, 20 surgeries were performed on the right stifle joint and 13 were conducted for the left stifle joint. Seven dogs underwent bilateral surgery. Twelve stifle joints with CrCL rupture also had MPL and were corrected by trochleoplasty and lateral imbrication with ES. The 26 dogs included four sexually intact females, six neutered females, one sexually intact male, and 15 neutered males. Breeds included Yorkshire terrier (n = 7), Maltese (n = 6), poodle (n = 3), and one each of Shih Tzu, Cocker spaniel, Siberian husky, mixed, Bichon Frise, Coton de Tulear, beagle, chowchow, American Eskimo, and Cane Corso. The mean age of the dogs at the time of surgery was 7.66 ± 2.54 years (range, 3.0~12.0 years), and mean body weight was 9.87 ± 12.49 kg (range, 1.7~52.0 kg).

Preoperative angle α of the 33 stifle joints ranged from 68.10° to 109.77° (mean, 89.00 ± 11.00°) and flexion angle β ranged from 39.01° to 98.76° (mean, 62.61 ± 18.26°). Postoperative angle α ranged from 79.41° to 116.71° (mean, 95.86 ± 8.26°) while flexion angle β ranged from 35.01° to 97.78° (mean, 55.48 ± 15.19°). The linear regression equation (Fig. 2) for α versus β before surgery was: α = -0.78β + 118 (r² = 0.602; p < 0.001). The linear regression equation for α versus β after surgery was: α = -0.56β + 113 (r² = 0.314; p < 0.001).

Mean age of the normal dogs was 6.78 ± 4.10 years (range, 0.5~15.0 years). Mean body weight was 9.37 ± 10.44 kg (range, 1.1~49.0 kg). The breeds included Maltese (n = 25), Yorkshire terrier (n = 10), poodle (n = 7), miniature Schnauzer (n = 1), miniature pinscher (n = 3), Pomeranian (n = 4), beagle (n = 2), white terrier (n = 4), Dalmatian (n = 2), Jindo (n = 3), Siberian husky (n = 1), Pekinese (n = 2), French bulldog (n = 3), Cocker spaniel (n = 8), golden retriever (n = 1), Labrador retriever (n = 5), Chihuahua (n = 3), border collie (n = 1), silky terrier (n = 1), Shih Tzu (n = 4), Bichon Frise (n = 1), American Eskimo (n = 1), and mixed breed (n = 11). The mean angle α was 97.35 ± 7.72°° (range, 75.32~119.44°) and the mean angle β was 58.60 ± 15.45° (range, 27.21~94.04°) for all dogs. The linear regression equation for angles α versus β (Fig. 2) was: α = -0.32β + 116 (r² = 0.416; p < 0.001).

Comparisons of the PLA were performed on radiographs taken with a normal standing angle (40~60° flexion angle) and general positioned angle during the radiographic examination (60~80° flexion angle). Extremely extended or flexed stifles were excluded. Significant differences were observed in the PLA between the preoperative plain view and preoperative stress view (p = 0.003), preoperative stress view and normal stifles (p = 0.000), and preoperative stress view and postoperative plain view (p = 0.001) with a 40~60° flexion angle (panel A in Fig. 3). Significant differences in the PLA were also found between preoperative plain view and normal stifles (p = 0.008), preoperative stress view and normal stifles (p = 0.000), preoperative plain view and postoperative plain view (p = 0.039), and preoperative stress view and postoperative plain view (p = 0.002) with a 60~80° flexion angle (panel B in Fig. 3).

The ROC curve for the stress view (panel A in Fig. 4) with a 40~60° flexion angle indicated that the area under the curve (AUC) was 0.949 (95% CI, 0.897~1.000; p = 0.000). If the cutoff value was 92.40°, the PLA measurement had a sensitivity of 90.0% and specificity of 100%. The ROC curve for the plain view (panel B in Fig. 4) demonstrated the AUC was 0.686 (95% CI, 0.500~0.873; p = 0.040). With a 94.51° of cutoff value, the PLA measurement had a sensitivity of 86.0% and specificity of 61.5%.

The ROC curve for the stress view (panel C in Fig. 4) with a 60~80° flexion angle showed that the AUC was 0.945 (95% CI, 0.881~1.000; p = 0.000). With a 90.55° cutoff value, the PLA measurement had a sensitivity of 83.9% and specificity of 100%. The ROC curve for the plain view (panel D in Fig. 4) indicated that the AUC was 0.785 (95% CI, 0.605~0.965; p = 0.004). With a 90.76° cutoff value, the PLA measurement had a sensitivity of 83.9% and specificity of 75.0%.