探究核磁共振检查及关节镜下测量肩袖损伤大小的一致性。

选取北京大学人民医院创伤骨科2016年12月至2017年11月冈上肌撕裂患者212例，其中核磁共振检查与术中诊断一致者205例，全层撕裂112例、部分撕裂93例。对于全层撕裂，在核磁共振及关节镜下，分别测量撕裂处前后径与内外径，经Bland-Altman检验，进行一致性分析。对于部分撕裂，在核磁共振及关节镜下分别测量撕裂处深度，并根据Ellman标准做出分型，经Kappa检验，比较二者一致性。

对于全层撕裂组，两种方法测量内外径差值的均数为-0.4 mm，标准差为2.2 mm，95%一致性界限为（-4.8，4.0） mm，有96.4%的点位于一致性界限范围内；对于前后径差值的均数为-0.1 mm，标准差为2.3 mm，计算出的95%一致性界限为（-4.6,4.4） mm，有95.5%的点位于一致性界限范围内。部分撕裂组93例患者，按Ellman分级，I度29例、II度47例、III度17例。术中镜下测量值按Ellman分级，I度30例、II度45例、III度18例。经Kappa一致性检验，标准差为0.06，95% CI:0.620~0.856，加权后Kappa值为0.738。

对于全层撕裂，核磁共振检查与关节镜下对于内外径与前后径的测量值具有高度的一致性，可准确定量化预测撕裂程度。而二者对于部分撕裂程度的测量同样具有良好的一致性，但未达到理想界限，因此核磁共振检查结果不能作为部分撕裂诊断及分型的唯一标准，临床中需综合其他检查结果，作出准确的判断。

Rotator cuff injury is one of the most common tendon injuries in adults. Rotator cuff injury has a high incidence of 22% in 65 years old and can increase to over 62% in 80 years old. With the rapid development of rotator cuff surgery technology, the expenditure of medical systems in various countries is increasing year by year, and the disease itself can bring great pain to patients with obvious pain at night, affecting daily work and life. Therefore, early and accurate diagnosis of rotator cuff injury is of great significance. Preoperative imaging examination of the nature and size of rotator cuff injury is of great value to the selection of treatment and prognosis. Magnetic resonance imaging (MRI) has become one of the most commonly used examination methods for preoperative diagnosis of rotator cuff injury due to its non-invasive, good tissue illumination and accurate information of muscle atrophy, degree of fat infiltration and labial injury. It has been widely used in clinical diagnosis and treatment. Some studies have shown that arthroscopic measurement of rotator cuff tear has high accuracy and can be used as the gold standard for detection. Previous studies in our department have shown that MRI has good specificity and sensitivity for rotator cuff injury, and is an effective auxiliary examination method. As far as we know, there is no systematic quantitative analysis of the consistency between rotator cuff tears of different degrees in MRI and specific values measured under arthroscopy. The choice of treatment and prognosis of rotator cuff tear vary with the degree of rotator cuff tear. Therefore, it is of great significance to accurately determine the value of rotator cuff tear before surgery for the selection of correct treatment and improvement of healing rate and postoperative function.

To determine the accuracy of the MRI for detection and measurement of the size of rotator cuff tears, including full-thickness tear and partial-thickness tear, with arthroscopic findings used as the standard.

From December 2016 to November 2017, 212 patients were selected from the department of orthopaedics and traumatology in our hospital, MRI examination and intraoperative diagnosis were performed in 205 cases, including 112 cases with full-thickness tear and 93 cases with partial tear. For full-thickness tears, the anteroposterior diameters and medial-lateral diameters of the tears were measured under MRI and arthroscopy, and Bland-Altman test was used for consistency analysis. For partial tears, the depth of the tear was measured under MRI and arthroscopy, and the newe classification was made according to Ellman classification. The consistency of the two classifications were compared by Kappa test.

For the full-thickness tear, the mean difference of the tear length between the two methods was -0.4 mm, and the standard deviation was 2.2 mm. 95% conformance boundaries are (-4.8, 4.0) mm. 96.4% of the points were within the range of LOA. For the tear width group, the mean difference was -0.1 mm, and the standard deviation was 2.3 mm. The calculated 95% consistency limit was (-4.6, 4.4) mm. 95.5% of the points were located within the LOA range. For 93 patients in the partial tear group, 29 patients were classified as Ellman I, 47 patients as Ellman II, and 17 patients as Ellman III. The tearing depth was measured according to Ellman classification, with 30 patients classified as Ellman I, 45 patients as Ellman II, and 18 patients as Ellman III. After Kappa consistency test, the standard deviation was 0.06 and 95% CI was 0.620-0.856. The weighed Kappa value was 0.738.

For full-layer tears, MRI and arthroscopy have a high degree of consistency in the measured values of medial-lateral diameters and anteroposterior diameters. It can accurately predict the degree of tear quantitatively. The measurement of partial tear degree also has good consistency, but does not reach the ideal limit. The results of MRI examination cannot be used as the only standard for the diagnosis and classification of partial tears. In clinical practice, the results of other examinations should be integrated to make an accurate judgment.