| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Musculoskeletal Imaging |
1 From the Departments of Radiology (C.M., B.C.V.B., F.E.L., B.M., J.M.) and Orthopedic Surgery (P.P., J.E.D.), Cliniques universitaires St Luc, Université Catholique de Louvain, 10 av Hippocrate, 1200 Brussels, Belgium. Received May 17, 2002; revision requested July 12; final revision received November 27; accepted January 14, 2003. Address correspondence to B.C.V.B. (e-mail: vandeberg@rdgn.ucl.ac.be).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Spiral CT arthrography was performed in 20 patients who presented with pain after partial meniscectomy. Findings at the initial reading of the images and at two retrospective independent readings were compared with those at second-look arthroscopy performed in all patients. At initial interpretation, conventional criteria for meniscal tear were used, including partial- or full-thickness tear of any size and meniscal separation. At retrospective interpretation, criteria for meniscal tear included large partial- and full-thickness tear and meniscal separation but not small partial-thickness tear. Sensitivity and specificity for the detection of tear of the postoperative menisci were calculated for initial and retrospective readings.
RESULTS: At initial interpretation, the sensitivity and specificity for the detection of tear of the postoperative menisci were 100% and 78%, respectively. At retrospective interpretation, the sensitivity and specificity for the detection of tear of the postoperative menisci were 79% and 89% at reading 1 and 93% and 89% at reading 2, respectively.
CONCLUSION: Spiral CT arthrography is valuable for the assessment of postoperative menisci, but the application of conventional definitions of meniscal tear to arthrographic findings in postoperative menisci can lead to overestimation of the clinical importance of meniscal lesions.
© RSNA, 2003
Index terms: Knee, CT, 452.12115 • Knee, ligaments, menisci, and cartilage, 452.4852
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Dual–detector row spiral computed tomography (CT) performed after single–contrast arthrography in the knee has been shown accurate in the assessment of the native meniscus (15,16). The purpose of our study was to assess the effectiveness of this technique, hereafter referred to as spiral CT arthrography, in the evaluation of the postoperative meniscus for recurrent or residual meniscal tear. We used second-look arthroscopy as the reference standard.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Between February 1, 1998, and December 30, 2000, a total of 1,193 spiral CT arthrographic procedures were performed at our institution in the assessment of internal derangements of the knee. We retrospectively reviewed the questionnaires administered to the patients before spiral CT arthrography, which included inquiries regarding a history of meniscal surgery.
Our study population consisted of 20 consecutive patients in whom spiral CT arthrography and second-look arthroscopy were performed after partial meniscectomy at our institution. No patient was excluded in whom spiral CT arthrography and second-look arthroscopy were performed subsequent to partial meniscectomy. The 18 men and two women in the study group had a mean age ± SD of 41.1 years ± 13 (range, 18–68 years; median, 43 years) at the time of imaging. Orthopedic surgeons from our institution referred all patients for spiral CT arthrography of the right knee (n = 11) or left knee (n = 9) because of clinical suspicion of recurrent or residual meniscal tear. This patient group had undergone a total of 23 partial meniscectomy procedures (17 medial and six lateral); three patients had undergone both medial and lateral partial meniscectomy. No meniscal suture had been used in any of the patients during the initial arthroscopic surgery. Four patients had an associated complete tear of the anterior cruciate ligament that was left untreated at initial arthroscopy. The mean time between initial arthroscopic surgery and subsequent spiral CT arthrography was 40.1 months (range, 4–156 months; median, 24 months). Informed consent was obtained from all patients prior to spiral CT arthrography.
The approval of the ethics committee at our institution was not solicited for our retrospective study, because spiral CT arthrography was performed routinely at our institution to assess internal derangement of the knee (15). Our institutional review board likewise did not require that we obtain either its approval or patients’ informed consent for our review of medical records and imaging findings.
Arthrography
A volume of 10 mL of ionic contrast material consisting of ioxaglate meglumine and ioxaglate sodium with 320 mg of iodine per milliliter (Hexabrix 320; Guerbet, Aulnay-sous-bois, France), mixed with 1 mL of 0.1% adrenaline hydrochloride solution containing 1 mg of epinephrine per milliliter (Parke-Davis, Zaventem, Belgium), was injected under fluoroscopic observation through a 20-gauge needle placed in the suprapatellar pouch of the knee joint after sterile skin preparation (17,18). This quantity of contrast material is routinely used at our institution for this procedure, and the potential advantage of a larger amount was not assessed. Epinephrine also is routinely used at conventional arthrography to slow the process of contrast material resorption, and we did not assess the potential for complications secondary to intraarticular injection of epinephrine. After intraarticular injection, patients were asked to walk about the room and to perform full-range knee flexions (19,20). The radiologist performing the procedure used fluoroscopy to verify homogeneous coating of the menisci and articular surfaces by contrast material. If coating of the cartilage and menisci appeared incomplete at fluoroscopy, the patient was asked to perform additional knee flexions. No complications were encountered. Patients walked to the nearby CT scanner.
CT Scanning
Spiral CT arthrography was performed with a dual–detector row scanner (Twin RTS, Real Time Scanning; Marconi Medical Systems, Cleveland, Ohio) in spiral scanning mode. All patients were imaged while supine with 15° knee flexion, because mild knee flexion enables clearer depiction of the anterior cruciate ligament (16,21). After a lateral projection scout image was obtained, a 65–85-second spiral CT arthrographic scan was obtained of the area between the apex of the patella and the tibial plateaus. Spiral CT arthrography was performed at 140 kVp and 135 mAs, with a focus of 0.75 mm. A dynamic oscillating focus was used (22). The field of view was 430 mm. The table speed was 0.75 mm/sec (effective pitch, 0.75), and the effective section thickness was 0.55 mm.
For reconstruction, a 360° linear interpolation algorithm, a high-frequency kernel with 14 line pairs per centimeter, an increment of 0.3 mm (60% of the nominal section thickness), and a zoom factor of 1.94 were used. Images were reconstructed within a 512 x 512 matrix, and in-plane resolution was 0.43 mm. With a pitch of 0.75 and a 40% overlap between sections, longitudinal resolution was 0.3 mm—equivalent to the reconstruction increment (23,24). All reconstructed images were prospectively stored on erasable optical disks.
Lesion Criteria and Image Analysis
Initial interpretation.—Images obtained with spiral CT arthrography in routine clinical practice were interpreted initially by one of three staff musculoskeletal radiologists (B.C.V.B., F.E.L., J.M.), with 10, 6, and 25 years of experience, respectively. Difficult cases were discussed and a consensus interpretation was reached. The radiologists were aware of each patient’s clinical history but did not know whether the patient would undergo arthroscopy. Initial interpretation was based on film images showing coronal, sagittal oblique, and transverse views. The coronal imaging plane was parallel to the posterior bicondylar line. The sagittal oblique imaging plane was perpendicular to the bicondylar line, with a 10° internal rotation. The transverse imaging plane was parallel to the meniscus on sagittal oblique and coronal images of each femorotibial compartment. The gap between sections was 1.6 mm for sagittal and coronal reformations and 0.1 mm for transverse reformations.
At initial analysis, a normal postoperative meniscus was defined by the presence of contour irregularity, including truncation, blunting, flattening, and/or rounding of the inner borders, and on the absence of meniscal separation, intrameniscal contrast material, and displaced meniscal fragments. A postoperative meniscus was considered torn if it demonstrated peripheral meniscal separation, meniscal tear, and/or a displaced meniscal fragment (15). Peripheral meniscal separation was defined by the presence of contrast material between the meniscus and the articular capsule, after the exclusion of normal anatomic variations (25). Tear was defined by the presence of intrameniscal contrast material. Displaced meniscal tears consisted of tears in which meniscal fragments had become detached from their normal anatomic location and lodged in the intercondylar notch, the superior meniscal recess, the inferior meniscal recess, or the femorotibial articular space (15). Loose bodies were not considered necessarily to indicate a torn postoperative meniscus, because a free meniscal fragment cannot be differentiated from a free hyaline cartilage fragment at spiral CT arthrography.
Retrospective interpretation.—Two retrospective readings of images from the 20 spiral CT arthrographic examinations were performed independently by three radiologists as follows: Reading 1 was performed by one of the three musculoskeletal staff radiologists (B.C.V.B.) with a fellow in musculoskeletal radiology (C.M.) by using a workstation (Omnipro; Silicon Graphics, Mountain View, Calif); film images were not used. Reading 2 was performed by another musculoskeletal staff radiologist (F.E.L.) by using the same film images that had been used by the surgeons who performed the meniscectomies. Sagittal oblique, transverse, and coronal reformations were printed by using bone settings (window width, 1,900 HU; window level, 450 HU) and a zoom factor of 2. The two musculoskeletal staff radiologists involved in the retrospective readings (B.C.V.B., F.E.L.) also had been involved in the initial interpretations. At retrospective interpretation, all readers were aware that the patients had undergone initial arthroscopic surgery and second-look arthroscopy. They were blinded to the clinical history of the patients, as well as to the findings and treatments performed at initial arthroscopic surgery and second-look arthroscopy. They were blinded also to the frequency of recurrent meniscal tears found at second-look arthroscopy in patients in our study group.
The readers evaluated each case to determine the status of the menisci at spiral CT arthrography. A postoperative meniscus was considered normal if it demonstrated altered shape with or without small partial-thickness tear as defined by the presence of intrameniscal contrast material involving less than one-third of the meniscus length and height. A postoperative meniscus was considered torn if it showed peripheral meniscal separation, full-thickness tear, large partial-thickness tear, and/or a displaced meniscal fragment. Meniscal separation and displaced meniscal fragments were defined as they had been defined at initial interpretation. Full-thickness tear was defined by the presence of intrameniscal contrast material that extended through the entire height (from the upper to the lower surface) or depth (up to the articular capsule) of the meniscus. Large partial-thickness tear was defined by the presence of intrameniscal contrast material involving more than one-third of the meniscus length or height but not the entire height or depth.
Interobserver and interreading comparison.—Fourteen months after the retrospective interpretations were completed, the radiologic reports from the initial interpretation were given to the staff radiologist and the fellow who had performed retrospective reading 1 to determine independently the status of the menisci on the basis of the initial interpretations. All administrative and clinical data had been deleted from the reports to prevent patient recognition. Statistical analysis was performed to assess the level of agreement between these two readers in determining whether the postoperative menisci were torn on the basis of the initial reports.
Arthroscopy
Two experienced orthopedic surgeons from our institution performed all knee arthroscopic procedures. The mean time between spiral CT arthrography and second-look arthroscopy was 10 weeks (range, 1–28 weeks; median, 8 weeks). Hard-copy sagittal oblique, coronal, and transverse reformatted spiral CT arthrographic images were available to the surgeons. Original reports included meniscal lesion description and notation of the presence or absence of displaced meniscal fragments.
Results of second-look arthroscopy were used as the standard of reference for determining the presence or absence of recurrent or residual meniscal tear. At the time of arthroscopy, meniscal lesion patterns, including tear, peripheral separation, and contour irregularities, were noted on charts. Drawings that showed the locations and configurations of all important meniscal abnormalities, including displaced fragments, were made. Arthroscopic procedures were not videotaped. The arthroscopic criteria for a normal postoperative meniscus were the presence a smooth meniscal remnant without mobile or displaced meniscal fragment and without meniscal separation. Superficial change with smooth contour irregularities was not considered to indicate abnormality (26–28). The arthroscopic criteria for a recurrent or residual meniscal tear included the identification of an unstable meniscal fragment, flap tear, or extensive meniscal cleavage.
Statistical Analysis
Statistical analysis of the data included determination of sensitivity and specificity and of positive and negative predictive values with 95% confidence intervals for the detection of recurrent or residual tear of postoperative menisci with spiral CT arthrography.
Agreement between the initial reading and the two retrospective readings, and between retrospective readings 1 and 2, was evaluated by using 
statistics. values of 0.20 or less were considered to indicate poor agreement; of 0.21–0.40, fair agreement; of 0.41–0.60, moderate agreement; of 0.61–0.80, substantial agreement; and of 0.81–1.00, almost perfect agreement. Data on menisci that had not been surgically repaired at initial arthroscopy were not taken into account. There was no separate analysis of the results observed in medial and lateral postoperative menisci.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
|
|
|
|
|
|
Agreement between the two retrospective readings for detection of recurrent or residual meniscal tear was substantial ( = 0.649). For detection of intrameniscal contrast material, agreement between the initial reading and retrospective reading 1 also was substantial ( = 0.777) and agreement between the initial reading and retrospective reading 2 was almost perfect ( = 0.893). Interobserver agreement was perfect ( = 1.00) between the staff radiologist and the fellow in determining whether the postoperative menisci were torn on the basis of the initial reports.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Findings at retrospective analysis demonstrated that spiral CT arthrography of the knee enabled detection of recurrent or residual tears of postoperative menisci, with sensitivities of 79% and 93% and specificities of 89% and 89% for readings 1 and 2, respectively. At retrospective analysis, criteria for normal or almost normal postoperative menisci included abnormal meniscal shape without intrameniscal contrast material, and abnormal meniscal shape with intrameniscal contrast material that involved less than one-third of the meniscal length and height. The first criterion applies to a postoperative meniscus with a continuous surface and without meniscal separation or displaced meniscal fragment, and the second criterion applies to a small partial-thickness tear. The second criterion was derived from the literature (27,29,30) and was used successfully in a previous study to enable recognition of stable meniscal lesions in native menisci (15). The categorization of small partial tear as a "normal" finding at spiral CT arthrography in postoperative menisci is based on the assumption that small irregularities are clinically unimportant (27,29,30); the main goal of partial meniscectomy is to excise mobile fragments and leave the residual meniscus rim intact, stable, and reasonably well contoured (27,29,30).
At retrospective analysis of spiral CT arthrographic images, the criteria for a finding of torn postoperative meniscus included the presence of full-thickness tear, displaced meniscal fragment, meniscal separation, or large partial-thickness tear involving at least one-third of the meniscus length or height. The rationale for the use of these criteria was that they were found to enable recognition of unstable meniscal tears with a sensitivity of 97% and specificity of 90% in a prior study of native menisci (15). Most likely, the value of spiral CT arthrography in the assessment both of postoperative and of native menisci derives from its spatial resolution and multiplanar capacity. In the postoperative meniscus, however, the high spatial and contrast resolution achieved with spiral CT arthrography enable detection of minute details of the meniscal surface, and the observer must decide the clinical importance of these surface changes. Despite this difficulty, in our study there was substantial interobserver agreement in the recognition of clinically important meniscal lesions at retrospective interpretation.
The indication for spiral CT arthrography in the assessment of symptomatic knee after partial meniscectomy has yet to be determined. Cartilage damage may develop progressively after meniscal surgery (31), and spiral CT arthrography enables delineation of cartilage lesions, which typically are filled with contrast material (32). Its value in the detection of bone marrow changes, which frequently develop after meniscectomy (33), is likely to be limited (34). A study in which each patient underwent spiral CT arthrography and MR imaging before second-look arthroscopy would have provided data enabling a comparison of both imaging modalities. Other constraints inherent in the use of spiral CT arthrography but not specifically in the postoperative knee include the limited ability to evaluate for ligamentous injury, the invasiveness of ionizing radiation, and the possibility of complications ensuing from intraarticular injection of iodinated contrast material.
The results of the current study are limited. First, the use of arthroscopy as the standard of reference for evaluation of meniscal tear has intrinsic limitations, including difficulty in probing the posterior horns of the menisci (35). Estimation of the residual postoperative meniscus at inspection and palpation may depend on the skill of the surgeon and on the presence of associated knee lesion (26,30,31,36).
Second, for ethical reasons, the surgeons performing second-look arthroscopy were not blinded to the findings at spiral CT arthrography. Imaging findings influenced the decision to proceed or not to proceed to second-look arthroscopy, with a subsequent selection bias favoring meniscal tear in the evaluation of the menisci at arthroscopy. It is possible that patients with unclear clinical findings and with recurrent or residual meniscal lesions that were occult at spiral CT arthrography did not undergo arthroscopic surgery. The results of our study, therefore, apply to a subset of patients who had clinical symptoms or imaging findings severe enough to warrant arthroscopy.
Third, a limited number of postoperative menisci were studied, and consequently, the 95% CIs are large. Further studies should be performed in a larger group of patients. The limited number of cases did not enable assessment of the statistical significance of the difference between the initial and retrospective readings.
Fourth, the size of the postoperative meniscus was not taken into account. Size appears to be a determinant of the value of conventional MR imaging (2,3,11) but has not been considered in studies of MR arthrography (9,10).
Fifth, the fact that two of the three musculoskeletal staff radiologists who performed the retrospective readings were involved in the initial interpretation of the spiral CT arthrographic images could have introduced recall bias. This possibility is remote, however, because of the long period of time during which cases were collected for this study.
Finally, the use of statistics to indicate agreement between initial and retrospective readings and between retrospective readings must be interpreted carefully. Several radiologists, not one, performed the initial readings. In addition, the two retrospective readings were performed differently: reading 1 was performed by using a CT workstation, and reading 2 was performed with film images. Finally, reading 1 was performed by a staff musculoskeletal radiologist in consensus with a fellow in musculoskeletal radiology. Inter- and intraobserver agreement therefore should be reassessed in a further study.
In conclusion, spiral CT arthrography of the knee enabled assessment of postoperative menisci with sensitivities and specificities depending on the criteria considered to indicate residual or recurrent tear. The application of conventional definitions of meniscal tear to arthrographic findings in postoperative menisci can lead to overestimation of the clinical importance of meniscal lesions.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, B.C.V.B.; study concepts, B.C.V.B., F.E.L., P.P.; study design, B.C.V.B., F.E.L., J.M.; literature research, C.M.; clinical studies, B.C.V.B., F.E.L., J.M., P.P., J.E.D.; data acquisition, F.E.L., B.C.V.B., J.M., P.P., J.E.D.; data analysis/interpretation, B.C.V.B., C.M.; statistical analysis, B.C.V.B.; manuscript preparation, B.C.V.B., C.M.; manuscript definition of intellectual content, B.C.V.B., J.M., F.E.L., C.M., B.M.; manuscript editing, B.C.V.B., C.M.; manuscript revision/review, all authors; manuscript final version approval, B.C.V.B.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  N. Pujol, L. Panarella, T. A. S. Selmi, P. Neyret, D. Fithian, and P. Beaufils Meniscal Healing After Meniscal Repair: A CT Arthrography Assessment Am. J. Sports Med., August 1, 2008; 36(8): 1489 - 1495. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Moser, A. Moussaoui, M. Dupuis, V. Douzal, and J.-C. Dosch Anterior Approach for Knee Arthrography: Tolerance Evaluation and Comparison of Two Routes Radiology, December 1, 2007; 246(1): 193 - 197. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. R. McCauley MR Imaging Evaluation of the Postoperative Knee Radiology, January 1, 2005; 234(1): 53 - 61. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
