| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Pediatric Imaging |
1 From the Department of Radiology, Hallym University School of Medicine, Seoul, Korea (K.L.); the Mallinckrodt Institute of Radiology, 510 S Kingshighway Blvd, St Louis, MO 63110 (M.J.S., C.F.H.); the Department of Radiology, Stanford University School of Medicine, Palo Alto, Calif (D.M.L.); and the Department of Orthopaedic Surgery, Washington University School of Medicine, St Louis (M.J.M.). Received September 10, 1998; revision requested November 3; final revision received March 17, 1999; accepted April 26. Address reprint requests to M.J.S. (e-mail: siegelm@mir.wustl.edu).
 |
Abstract |
|---|
 TOP Abstract IntroductionMATERIALS and METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: MR images obtained in 43 patients aged 5–16 years who underwent arthroscopy were retrospectively reviewed. Two reviewers evaluated primary findings (abnormal signal intensity, abnormal course as defined by Blumensaat angle, and discontinuity), secondary findings (bone bruise in lateral compartment, anterior tibial displacement, uncovering of posterior horn of lateral meniscus, posterior cruciate ligament line, and posterior cruciate angle), and meniscal and other ligamentous injuries.
RESULTS: There were 19 ACL tears and 24 intact ACLs. Overall sensitivity and specificity of MR imaging in detecting ACL tears were 95% and 88%, respectively. Sensitivities of the primary findings were 94% for abnormal Blumensaat angle; 79%, abnormal signal intensity; and 21%, discontinuity. The specificity of all primary findings was 88% or greater. The sensitivity and specificity of the secondary findings, respectively, were 68% and 88% for bone bruise; 63% and 92%, anterior tibial displacement; 42% and 96%, uncovered posterior horn of lateral meniscus; 68% and 92%, positive posterior cruciate line; and 74% and 71%, abnormal posterior cruciate angle. Fifteen (79%) patients had meniscal tears, and five (26%) had collateral ligament injuries.
CONCLUSION: Primary and secondary findings of ACL tears in young patients have high specificity and are useful for diagnosis.
Index terms: Knee, injuries, 452.4191, 452.42, 452.4852, 452.4857 • Knee, ligaments, menisci, and cartilage, 452.4191, 452.42, 452.4852, 452.4857 • Knee, MR, 452.121411, 452.121412, 452.121416
|
Introduction |
|---|
|
TOPAbstractIntroductionMATERIALS and METHODSRESULTSDISCUSSIONReferences |
|---|
|
MATERIALS and METHODS |
|---|
|
TOPAbstractIntroductionMATERIALS and METHODSRESULTSDISCUSSIONReferences |
|---|
All knee MR examinations were performed with a 1.5-T unit (Magnetom; Siemens, Erlangen, Germany) and a phased-array extremity coil. The knee was placed in an extended position with approximately 15° of external rotation. In all patients, sagittal multiecho (repetition time msec/echo time msec, 2,500–3,600/20–120), coronal T1-weighted (600/12), coronal multiecho (2,500–3,000/17–119), and transverse gradient-echo or turbo T2-weighted sequences were performed. The matrix size with all of the sequences was 256 x 256 with a section thickness of 4 mm; there was no intersection gap.
The MR imaging studies were reviewed retrospectively by two pediatric radiologists (D.M.L., M.J.S.) who were blinded to the arthroscopic findings, clinical histories, and initial MR imaging interpretations. Because this was a retrospective study, all observations were made on the film hard-copy images. Primary signs and secondary criteria were evaluated. The images were also reviewed for the presence or absence of meniscal tears and collateral ligament injuries. The final diagnosis was reached by consensus. The cases of intact and torn ACLs were randomly mixed in this study.
Arthroscopy was performed with the patient under general anesthesia by five orthopedic surgeons. One orthopedic surgeon, who specialized in sports medicine, performed surgery in 31 (72%) of the 43 patients. At arthroscopy, 19 patients had torn ACLs and 24 had normal ACLs. Two of the 24 patients with intact ACLs had tibial spine avulsion fractures, which were defined as an avulsion fracture of the intercondylar eminence of the tibia at the tibial insertion site of the ACL. At the time of surgery, an ACL tear was classified as either complete or partial on the basis of the degree of the pivotal shift, tension response to probing, and number of disrupted fascicles. The tear was considered to be complete if there was a marked pivotal shift, minimal resistance to probing, and disruption of 90% or more of the fascicles. The tear was classified as partial if there was a mild pivotal shift, substantial resistance to probing, and disruption of less than 90% of the fascicles.
The medical charts and questionnaires administered to the patients at the time of clinical examination by the orthopedic surgeon were reviewed to determine the age of injury, date of surgery, and surgical findings. An ACL tear was classified as acute if the interval between the injury and MR imaging was less than 2 weeks, subacute if the interval was 2–8 weeks, and chronic if the interval was longer than 8 weeks (7). On the basis of these criteria, ACL tears were complete in 17 patients and partial in two. The ACL tears were acute in eight, subacute in six, and chronic in five patients.
Evaluation of Primary Findings
The ACL was considered to be normal when it was seen as a continuous linear band of low signal intensity on T1-weighted, intermediate-weighted, and T2-weighted images (Fig 1). The primary signs of tear were an abnormal course, abnormal signal intensity, and discontinuity. The course of the ACL was based on the Blumensaat angle measurement (Fig 1). The Blumensaat line is parallel to the roof of the intercondylar notch. The Blumensaat angle, measured by using a hand-held goniometer, is the angle between the Blumensaat line and a line along the distal portion of the ACL (8). A Blumensaat angle measurement greater than the mean value in all cases was used as the threshold for the diagnosis of ACL tear. Abnormal signal intensity was defined as focally or diffusely increased signal intensity on intermediate- or T2-weighted images with no depiction of the ACL. As described by Falchook et al (9), discontinuity was defined as a focal gap in the ligament or depiction of more than one ligament piece.
|
The posterior cruciate line, as described by Schweitzer et al (11), refers to the position of the posterior cruciate line in relation to the distal femur. This criterion was assessed on the image that best demonstrated the distal portion of the posterior cruciate line. A line was drawn tangent to the posterior margin of the distal posterior cruciate ligament and extended proximally. The posterior cruciate line was considered to be positive for ACL tear if the proximal extension of this line did not intersect the medullary cavity of the femur within 5 cm of its distal aspect (7). This sign was considered to be negative if the proximal extension of the posterior cruciate ligament line intersected the medullary cavity within 5 cm of its distal aspect (Fig 2). The posterior cruciate angle was defined as the point of intersection between lines drawn through the proximal and distal portions of the posterior cruciate ligament (8,12) (Fig 2). A posterior cruciate angle measurement less than the mean value for all cases was used as the threshold for the diagnosis of ACL tear.
|
|
|
|
RESULTS |
|---|
|
TOPAbstractIntroductionMATERIALS and METHODSRESULTSDISCUSSIONReferences |
|---|
The sex of the patient was found to be significantly associated with ACL tears (P < .001). Fourteen (70%) of 20 female patients had tears, whereas five (22%) of 23 male patients had tears. In this study population, a female patient with a knee injury was eight times more likely to have a torn ACL than a male patient (P = .001) (95% CI: 2, 37). The Blumensaat angle was significantly greater in the female patients (P = .007). The mean angle measurement (± SD) in the female patients was 18.7° ± 20.98. In the male patients the mean angle measurement was 1.8° ± 15.80. The number of female patients in whom images showed an abnormal signal intensity, 12 (60%) of 20, was significantly higher (P = .010) than the number of male patients with this finding, five (22%) of 23.
Primary Findings
The primary, secondary, and associated imaging findings of ACL tears are noted in the Table. Primary findings were present in all the patients with ACL tears. Thirteen (68%) of 19 patients had more than one finding; two findings were present in 10 patients, and three findings were present in three.
|
|
|
.008).
|
|
|
|
|
Of the 19 patients with ACL tears, 18 (95%) had secondary findings. Only one patient with an ACL tear had primary findings without secondary findings. Five (21%) of 24 patients with intact ACLs had eight secondary findings. Two of these patients had tibial spine fractures, one had osteochondritis dissecans, and two had no anatomic abnormality at arthroscopy. In these five false-positive cases, MR images showed bone bruises in three, anterior tibial displacement in two, an abnormal posterior cruciate line in two, and an uncovered posterior horn of the lateral meniscus in one. One patient with a tibial spine fracture had four secondary findings—bone bruise, tibial displacement, uncovered meniscus, and abnormal posterior cruciate line; the other patients had one secondary finding each.
On the basis of all the primary and secondary criteria, the reviewers' diagnoses of ACL tears were correct in 18 of 19 cases (sensitivity, 95%). The diagnosis of normal ACL was correct in 21 of 24 cases (specificity, 88%).
Associated Injuries
Associated injuries, all of which were surgically confirmed, included 17 meniscal tears in 15 (79%) of 19 patients with ACL tears; two patients each had two tears. Thirteen of the 17 tears involved the medial menisci, and four involved the lateral menisci. Five (26%) of 19 patients had collateral ligament injuries; four were medial and one was lateral. None of the associated injuries was significantly related to ACL tears (P 
.081) (Table).
All statistical models to predict ACL tears, which were created by means of stepwise logistic regression analysis, including those with forced variable entry, resulted in unstable parameter estimates with entry of the second variable. Therefore, no acceptable models were produced.
|
DISCUSSION |
|---|
|
TOPAbstractIntroductionMATERIALS and METHODSRESULTSDISCUSSIONReferences |
|---|
In our series, on the basis of both primary and secondary imaging findings, the sensitivity of MR imaging for the diagnosis of an ACL tear was 95%, and the specificity was 88%. In a previous series of children (5), we found a sensitivity of 64% and specificity of 94%. The discrepancy in the results between these two studies may reflect differences in technique and the MR imaging criteria used to define tears or differences in the age of the injuries. Delays between the injury and MR imaging are a potential source of false-negative diagnoses. Most of the tears in our study were acute or subacute and complete, which may have increased our ability to detect an abnormality. In our previous study (5), we did not define the ages of the ACL injuries, so we cannot determine whether these were factors in the false-negative diagnoses. Finally, the inclusion of more secondary findings in this study may have contributed to the increased sensitivity of our results.
Patient age was not significantly related to ACL tears. Patient sex, however, was related to ACL tears: Female patients were eight times as likely to have a tear as male patients. This difference may be accounted for by earlier epiphyseal fusion in female patients compared with that in male patients. Alternatively, this difference may have been due to a sampling error in our patient population.
Injury of the ACL is most reliably diagnosed by observing primary findings, which are abnormalities of the ligament itself. In our patient population, the Blumensaat angle and abnormal signal intensity were the most sensitive primary findings for the diagnosis of a tear (sensitivities of 94% and 79%, respectively). As demonstrated in the cumulative probability graph (Fig 5), the probability of an ACL tear increased markedly with increasing Blumensaat angle. Discontinuity of the ACL was seen in only four (21%) patients with ACL tears. These results are similar to those reported by Falchook et al (9) in a series of 92 adult patients, in which the highest diagnostic accuracy was obtained when there was an abnormal course of the ligament (sensitivity, 96%) and high signal intensity (sensitivity, 89%). Discontinuity was similarly found to be an insensitive finding, occurring in only 23% of that patient population.
Numerous secondary or indirect MR imaging findings of ACL tears in adults have been reported (4,8,11,12,14,17–19). In our study population, 95% of patients with ACL tears had more than one secondary finding; one patient had only primary findings. The sensitivity of secondary findings ranged from 42% to 74%, with specificities of 71%–96%. The low to moderate sensitivity and high specificity of secondary findings for the detection of ACL tears in our study population were similar to those reported in adults (4,12,14,17–19).
Although there is little doubt that secondary findings can increase observer confidence in the diagnosis of an ACL tear (17,19), there remains controversy about which findings are the best predictors of tear. The results of our study showed that the most specific secondary findings are an uncovered posterior horn of the lateral meniscus, an abnormal posterior cruciate line, and anterior tibial displacement. Depending on the threshold used for ACL tear, the sensitivity and specificity of the posterior cruciate angle varied considerably. When a value (ie, 115°, which is an appropriate threshold for this study population) slightly greater than the mean for all cases (114.8°) was used as the threshold, the sensitivity and specificity were modest (74% and 71%, respectively). To better demonstrate the relationship between ACL tear and posterior cruciate angle, a graph of the cumulative probability was created (Fig 10). As the graph indicates, the probability of ACL tear increased as the posterior cruciate angle decreased. As the steepness of the curve in this graph indicates, however, this relationship was not as pronounced as that between ACL tear and Blumensaat angle (Fig 5), in which the correlation grew stronger as the steepness of the curve increased.
Several authors (4,10,17,20) have observed a relationship between bone contusion and ACL tears. Among adult patients with ACL tears, 37%–100% will reportedly have this secondary finding (4,8,10,12,20). Bone contusions represent trabecular microfractures caused by impaction of the tibia on the femur at the time of injury. The most frequent mechanism of ACL injury is valgus stress in which the tibia rotates internally with respect to the femur and thus causes the lateral plateau to strike the lateral femoral condyle. Contusions of the posterolateral tibial plateau have been reported to be a highly sensitive (97%–100%) finding of ACL tear in adults (4,8,12). In our patient population, the sensitivity of this finding was 68% (13 of 19 patients). This result is similar to that of Snearly et al (21), who pointed out that bone contusions in locations that are highly suggestive of ACL tears in adults are not sensitive for the detection of ACL tears in adolescent patients. In their series, 13 (72%) of 18 adolescents with typical bone bruises had ACL tears. Snearly et al (21) suggested that these differences in children and adolescents may be explained by the inherent ligamentous laxity in the younger age group. This idea is supported by the findings of other studies (22,23).
Baxter (22) examined 232 children aged 7–14 years with normal knees and found a progressive decrease in ligament laxity with increasing age. Grana and Moretz (23), in another study involving 672 high school students, found that ligamentous laxity was a normal finding, especially in the female students. On the basis of these study results, Snearly et al (21) suggested that the increased ligamentous laxity in younger patients allows an abnormal internal rotation of the tibia on the femur and thus results in the bone contusions seen on MR images obtained in patients with preserved ACLs.
There was one false-negative case in our patient population. Two (11%) patients had partial tears, one of which was misinterpreted as normal. MR imaging in this patient demonstrated a linear band of intact fibers normally oriented in the expected location of the ACL. The results of two large studies (24,25) showed that MR imaging has relatively poor sensitivity (40%–75%) but moderate to high specificity (62%–94%) in the diagnosis of partial tears. Differentiating partial from complete tears is relevant, because a partial tear often can be treated conservatively, whereas a complete tear requires surgical intervention and is therefore associated with potential growth plate violation, which may lead to limb length discrepancies. Our error rate was 50% (one of two patients with partial tears), but our study size was too small to imply that this had any statistical significance.
False-positive findings occurred in six patients. One patient had only a false-positive primary finding; two (both with tibial spine fractures) had both primary and secondary findings; and three had false-positive secondary findings alone. Partial volume averaging with surrounding hematoma may explain the false-positive primary findings, especially in the patients with tibial spine injuries. Tibial spine fractures tend to be caused primarily by either an indirect twisting injury with the foot planted and the leg internally rotated on the femur or a hyperextension of the knee. With this type of injury, there is substantial risk for meniscal and collateral ligament injuries. Theoretically, depending on the severity and mechanism of injury, other soft-tissue structures of the knee could be injured, which would explain the secondary findings in patients with tibial spine injuries. The cause of the false-positive secondary findings in the three other patients is unclear, but they may have reflected the physiologic ligamentous laxity in younger patients.
Although we were unable to build a statistically significant model for the classification of ACL tears, we point out that the combined primary and secondary findings may improve the sensitivity of MR imaging–based diagnoses of ACL tears and be of clinical importance. This is because the specificities of primary and secondary findings are generally high; thus, if a case were positive for a primary or secondary finding, then there would be a strong likelihood that it was truly positive.
In this study population, sensitivity increased when a case was considered positive because either of any two of the primary or secondary findings was positive. This occurred at the cost of a slight decrease in diagnostic specificity. For example, in this study, if the finding of abnormal signal intensity had been combined separately with each of the secondary findings, the new sensitivity and specificity would have been 93% and 88%, respectively. The means of the original sensitivities and specificities of signal intensity and the four secondary findings were 64% and 92%, respectively (Table). Thus, there would have been a considerable gain in sensitivity at a slight cost in specificity. This may account for the 95% sensitivity of our clinical diagnoses, which was higher than any of the sensitivities listed in the Table and the 88% specificity of our clinical diagnoses, which was equal to or lower than all but one of the specificities for primary and secondary findings listed in the Table. However, one should bear in mind that with the sensitivities and specificities in the Table and the sensitivities and specificities that resulted from combining primary and secondary findings, there was a considerable overlap in the CIs for all but the most extreme values. Thus, it is not possible to say with certainty which primary or secondary finding or combination of them would result in optimal diagnostic performance. Such a determination needs to be based on the results of a study with a larger sample size. Despite this, the odds ratios (Table) indicate that both the primary and secondary MR imaging findings of this study were of high diagnostic value for detecting torn ACLs in young patients.
Meniscal tears were frequently associated with ACL tears in our patient population. Tears of the menisci were seen in 79% of the patients with ACL tears. These results were higher than those previously reported in adult series, in which there were tears in approximately 65%–70% of patients (2). The meniscal injuries in our study patients were more frequent in the medial meniscus than in the lateral meniscus. This pattern of involvement is similar to that seen in adults with ACL tears, in which the frequency of medial versus lateral meniscal injury is reportedly 43% versus 32%, respectively (2).
One limitation of our study was the interval between the injury and the MR imaging examination. We studied the images obtained in patients with predominantly acute and subacute injuries. If more patients with chronic tears had been included, then the value of the primary findings and of some of the secondary signs, such as bone bruising, might have been diminished. Conversely, the value of other signs, such as anterior dislocation and meniscal uncoverage, might have increased owing to the ligamentous laxity seen in chronic tears. In addition, the MR imaging examinations in our study were performed with 1.5-T magnets, and it is possible that the results might have been less impressive with low- or mid-field-strength imaging units. Furthermore, our analyses were retrospective rather than prospective; however, we were blinded to the clinical data and arthroscopic results during our assessment. Finally, no inter- or intraobserver reliability tests were performed, but rather the images were interpreted by consensus between two radiologists.
In conclusion, MR imaging of the ACL in young patients with immature skeletal systems is a highly reliable study. The results of our study demonstrated the relative usefulness of primary and secondary findings in the diagnosis of ACL tears in children and adolescents. Primary and secondary findings are highly specific and useful predictors of these tears.
|
Footnotes |
|---|
Author contributions: Guarantor of integrity of entire study, M.J.S.; study concepts and design, M.J.S., K.L.; definition of intellectual content, M.J.S., M.J.M.; literature research, M.J.S., K.L.; clinical studies, M.J.M.; data acquisition, D.M.L., M.J.S.; data analysis, M.J.S., K.L., C.F.H.; statistical analysis, C.F.H.; manuscript preparation, M.J.S., C.F.H.; manuscript editing, M.J.S., M.J.M.; manuscript review, all authors.
|
References |
|---|
|
TOPAbstractIntroductionMATERIALS and METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  C. Shanmugam and N. Maffulli Sports injuries in children Br. Med. Bull., June 1, 2008; 86(1): 33 - 57. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S Ostlere Imaging the knee Imaging, September 1, 2007; 19(3): 249 - 268. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. S.-h. Chiu The anterior tibial translocation sign. Radiology, June 1, 2006; 239(3): 914 - 915. [Full Text] [PDF]
|
|
|
|
 |
|
 J. S. Prince, T. Laor, and J. A. Bean MRI of Anterior Cruciate Ligament Injuries and Associated Findings in the Pediatric Knee: Changes with Skeletal Maturation Am. J. Roentgenol., September 1, 2005; 185(3): 756 - 762. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. S. Oeppen, S. A. Connolly, J. T. Bencardino, and D. Jaramillo Acute Injury of the Articular Cartilage and Subchondral Bone: A Common but Unrecognized Lesion in the Immature Knee Am. J. Roentgenol., January 1, 2004; 182(1): 111 - 117. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S Ostlere Imaging the knee Imaging, December 1, 2003; 15(4): 217 - 241. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P Sharma, K. Luscombe, and N Maffulli Sports injuries in children Trauma, October 1, 2003; 5(4): 245 - 259. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
