| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 All authors: Department of Radiology, University of Michigan Medical Center, 1500 E. Medical Center Dr., TC-2910G, Ann Arbor, MI 48109-0326.
Received February 17, 2000;
accepted after revision March 28, 2000.
Presented at the annual meeting of the American Roentgen Ray Society, New
Orleans, May 1999.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Sixteen consecutive MR imaging examinations were retrospectively and independently evaluated by two musculoskeletal radiologists for primary signs (graft signal, orientation, fiber continuity, complete discontinuity, and thickness) and secondary signs (anterior tibial translation, uncovered posterior horn lateral meniscus, posterior cruciate ligament hyperbuckling, and abnormal posterior cruciate ligament line) of anterior cruciate ligament reconstruction graft tear in 15 patients with follow-up arthroscopy. Results were compared with arthroscopy, and both receiver operating characteristic curves and kappa values for interobserver variability were calculated.
RESULTS. Arthroscopy revealed four full-thickness graft tears, seven partial-thickness tears, and five intact grafts. Of the primary signs, graft fiber continuity in the coronal plane and 100% graft thickness in the sagittal or coronal plane were most valuable in excluding full-thickness tear. Complete discontinuous graft in the coronal plane also was valuable in diagnosis of full-thickness tear. Of the secondary signs, anterior tibial translation and uncovered posterior horn lateral meniscus assisted in differentiating graft tear (partial or full thickness) from intact graft. The other primary and secondary signs were less valuable. Kappa values were highest for graft fiber continuity and graft discontinuity in the coronal plane.
CONCLUSION. Full-thickness anterior cruciate ligament graft tear can be differentiated from partial-thickness tear or intact graft by evaluating for graft fiber continuity (coronal plane), complete graft discontinuity (coronal plane), and graft thickness (coronal or sagittal plane).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The goal of this study was to determine whether any MR imaging features allow distinction between intact ACL reconstruction graft, partial-thickness graft tear, and full-thickness graft tear, using arthroscopy as the gold standard. The significance of primary or intrinsic ACL graft features and secondary associated MR imaging findings is determined.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
All patients were examined with a 1.5-T MR imaging system (Signa Advantage; General Electric Medical Systems, Milwaukee, WI). Because of the retrospective nature of this study, the MR imaging protocols varied but included the following: proton density—weighted fast spin-echo images with fat saturation in the coronal and axial planes in 16 examinations and in the sagittal plane in 11, proton density—weighted spin-echo sagittal images in 13 examinations, proton density-weighted fast spin-echo coronal images in 11 examinations, proton density—weighted gradient-echo sagittal images in five examinations, T1-weighted spin-echo sagittal images in three examinations (two of which received IV gadopentetate dimeglumine [Magnevist; Berlex Imaging, Wayne, NJ]), proton density—weighted spin-echo sagittal images with fat saturation in one examination, T1-weighted spin-echo sagittal images with fat saturation in one examination after IV gadopentetate administration, and T1-weighted spin-echo coronal images with fat saturation after IV gadopentetate administration in one examination. The TR range/TE range for T1-weighted images was 300-900/10-17 and for proton density—weighted images, 1000-4500/12-17. Gradient-echo images were obtained with a flip angle of 30° and a TR/TE of 33/15. The echo train length for fast spin-echo images was eight. The number of excitations was one to two. The slice thickness and slice gap for each imaging plane were 3- or 4-mm thick and 1-mm gap for the sagittal plane (except for gradient echo, 1.5-mm thick and 0-mm gap), 4-mm thick and 1- or 0.5-mm gap for the coronal plane, and 10-mm thick and 2-mm gap for the axial plane. Sagittal MR imaging was performed with the knee in 0-10° of external rotation to obtain images sagittal to the plane of the ACL.
Arthroscopy was completed by one of three orthopedic surgeons who specializes in sports medicine. Each had knowledge of the MR imaging report before arthroscopy. Arthroscopic results were reviewed from surgery reports. Information regarding patient history and physical examination findings (Lachman test) was reviewed from medical records. The Lachman test is a noninvasive clinical test of ACL integrity classified as grade I (proprioceptive appreciation of a positive test), grade II (visible anterior tibial translation), grade III (passive subluxation of the tibia with the patient supine), or grade IV (patient can actively subluxate the proximal tibia) [12].
The prospective MR imaging reports of the 16 examinations were reviewed. The initial MR imaging interpretation was made by one of six fellowshiptrained musculoskeletal radiologists. Each MR imaging report was categorized by one fellowship-trained musculoskeletal radiologist as intact ACL graft, partial-thickness tear, or full-thickness graft tear. These reports were then compared with the arthroscopic results to determine prospective sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of MR imaging in determining ACL graft integrity.
A retrospective review of the MR images of the 16 examinations was then completed by two fellowship-trained musculoskeletal radiologists who were unaware of the arthroscopic results. Each observer independently reviewed the 16 examinations in random order for various primary and secondary findings on the MR images relative to the ACL graft. All MR imaging sequences and planes for each case were available for review. The primary signs concentrated on ACL graft fiber signal, orientation, and continuity. More specifically, for each MR image plane, images were assessed for the following primary signs: diffuse increased ACL graft signal intensity (Fig. 1A,1B), location of focal increased graft signal if present (proximal, middle, or distal), graft orientation on sagittal images (either taut between femur and tibia or horizontal or lax), complete ACL graft discontinuity (Fig. 1A,1B), the presence of any ACL graft fiber continuity (Fig. 2A,2B), and focal graft thinning (100%, 50-99%, or <50% thickness) (Fig. 3A,3B). Evaluation for secondary signs of ACL graft tear included anterior tibial translation (posterior cortex of mid lateral tibia translated >5 mm anterior to the posterior cortex of the femur on sagittal images) (Fig. 4), uncovered posterior horn of lateral meniscus (line drawn superior from posterior cortex of lateral tibia intersects the posterior horn of lateral meniscus on sagittal images) (Fig. 4), posterior cruciate ligament (PCL) hyperbuckling (posterior concavity of PCL on sagittal images), and abnormal posterior PCL line (line tangential to posterior margin of distal PCL does not intersect femur in distal 5 cm on sagittal images). Additionally, each observer determined placement of the ACL graft in the tibial tunnel relative to the tibial plateau on sagittal images as first, second, third, or fourth quartiles from anterior to posterior. Lastly, each observer was asked to separately grade their confidence in diagnosing full-thickness graft tear and partial-thickness graft tear by assigning a grade ranging from 1 to 5 (1 = definitely no, 2 = probably no, 3 = unsure, 4 = probably yes, and 5 = definitely yes).
|
|
|
|
|
|
|
Analysis of the retrospective MR imaging results included number and percentage of occurrence of the primary and secondary signs for each arthroscopic category (intact, partial tear, and complete tear). Retrospective MR imaging results were then compared with the arthroscopic results to determine sensitivity, specificity, positive predictive value, negative predictive value, and accuracy. Because three graft conditions existed in the gold standard (intact graft, partial-thickness graft tear, and full-thickness graft tear), categories were combined to create two-by-two contingency tables for this analysis resulting in full-thickness tear versus other conditions (intact graft or partial-thickness tear), partial-thickness tear versus other conditions (full-thickness tear or intact graft), and any type of graft tear (partial or full) versus intact graft. Kappa values for interobserver variability were calculated for each of the primary and secondary MR imaging results (0.21-0.40 = fair agreement, 0.41-0.60 = moderate agreement, 0.61-0.80 = substantial agreement, 0.81-1.0 = almost perfect agreement). Receiver operating characteristic curves were constructed for each observer, and the areas under each curve were compared.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Indications for reassessment of the knee was available in 10 patients and included reevaluation of the ACL graft and meniscus (n = 6) and evaluation for meniscal tear (n = 4). Repeat arthroscopy revealed four full-thickness tears of the ACL graft, seven partial-thickness tears, and five intact grafts. Available medical records pertaining to the physical examination findings revealed Lachman test results as follows: negative results in one case, grade I in seven cases, grade II in three cases, and grade III in three cases. Of the four full-thickness graft tears, three cases had a grade III Lachman test and one case had a grade II. Of the six partial-thickness graft tears, two cases had a grade II Lachman test and four cases had grade I. Of the four intact ACL grafts, three cases had a grade I Lachman test and one case had a negative Lachman test.
Prospective MR Imaging Results
The initial prospective MR imaging reports were categorized as
full-thickness graft tear in two cases, partial-thickness tear in three, and
intact graft in 11. Of the four full-thickness graft tears at arthroscopy, two
were interpreted as full-thickness tears and two as partial-thickness tears on
MR imaging. Of the seven partial-thickness tears at arthroscopy, all seven
cases were interpreted as intact graft on MR imaging. Of the five intact ACL
grafts at arthroscopy, four were interpreted as intact and one as a
partial-thickness tear on MR imaging.
The diagnosis of full-thickness graft tear versus other condition (partial-thickness or intact graft) on MR imaging using arthroscopy as the gold standard resulted in 50% sensitivity, 100% specificity, 100% positive predictive value, 86% negative predictive value, and 87.5% accuracy. The diagnosis of partial-thickness graft tear versus other conditions (full-thickness tear or intact graft) resulted in 0% sensitivity, 67% specificity, 0% positive predictive value, 46% negative predictive value, and 37.5% accuracy. The diagnosis of any graft tear versus intact graft resulted in 36% sensitivity, 80% specificity, 80% positive predictive value, 36% negative predictive value, and 50% accuracy.
Retrospective MR Imaging: Primary Signs
ACL graft signal.—Diffuse increased signal in the region of
the ACL graft on proton density—weighted images was present in 50% of
the full-thickness graft tears, 57% of partial-thickness tears, and 10% of
intact grafts (Figs.
1A,1B
and 5). The diagnosis of
full-thickness graft tear versus other conditions (partial-thickness tear or
intact graft) resulted in 50% sensitivity, 70% specificity, 22% positive
predictive value, 85% negative predictive value, and 65% accuracy. The
diagnosis of partial-thickness ACL graft versus other conditions
(full-thickness tear or intact graft) resulted in a 43% sensitivity, 72%
specificity, 72% positive predictive value, 60% negative predictive value, and
59% accuracy. The diagnosis of ACL graft tear (partial- or full-thickness)
versus intact graft resulted in a 45% sensitivity, 90% specificity, 94%
positive predictive value, 46% negative predictive value, and 59% accuracy.
The kappa value for interobserver variability was 0.41 (moderate agreement)
for diffuse increased graft signal.
|
Focal increased signal in the ACL graft on proton density—weighted images was identified with 75% of full-thickness graft tears (distal in 25%, middle in 12%, middle and distal in 12%, and proximal in 25%), 28% of partial-thickness tears, and 40% of intact grafts (Fig. 5). The diagnosis of full-thickness graft tear versus other conditions (partial-thickness tear or intact graft) resulted in 75% sensitivity, 67% specificity, 41% positive predictive value, 90% negative predictive value, and 68% accuracy. The diagnosis of partial-thickness ACL graft versus other conditions (full-thickness tear or intact graft) resulted in 28% sensitivity, 44% specificity, 29% positive predictive value, 43% negative predictive value, and 39% accuracy. The diagnosis of ACL graft tear (partial or full thickness) versus intact graft resulted in 45% sensitivity, 60% specificity, 71% positive predictive value, 33% negative predictive value, and 50% accuracy. The kappa value for interobserver variability was 0.75 (substantial agreement) for focal increased graft signal.
Because of the potential association between increased graft signal from graft impingement [11] and abnormal placement of tibial tunnel used in ACL reconstruction [13], location of the tibial tunnel on MR imaging was assessed. All the tibial tunnels were classified as residing within the second or third quartiles or the central 50% of the tibial plateau on sagittal images. The kappa value for interobserver variability was 0.24 (fair agreement) for tibial tunnel placement.
ACL graft orientation.—The orientation of the ACL graft on sagittal images was found to be lax or horizontal in 12.5% of full-thickness graft tears, 43% of partial-thickness tears, and 10% of intact grafts (Fig. 5). The diagnosis of full-thickness graft tear versus other conditions (partial-thickness tear or intact graft) resulted in a 13% sensitivity, 71% specificity, 13% positive predictive value, 71% negative predictive value, and 56% accuracy. The diagnosis of partial-thickness ACL graft versus other conditions (full thickness or intact graft) resulted in a 43% sensitivity, 89% specificity, 75% positive predictive value, 67% negative predictive value, and 69% accuracy. The diagnosis of ACL graft tear (partial or full thickness) versus intact graft resulted in a 32% sensitivity, 90% specificity, 88% positive predictive value, 38% negative predictive value, and 50% accuracy. The kappa value for interobserver variability was 0.33 (fair agreement) for lax or horizontal ACL graft orientation.
ACL graft discontinuity.—Complete ACL graft discontinuity was evaluated in the sagittal and coronal planes (Fig. 1A,1B). In the sagittal plane, complete graft discontinuity was seen in 50% of full-thickness tears, 14% of partial-thickness tears, and 10% of intact grafts. In the coronal plane (Fig. 5), complete graft discontinuity was seen in 75% of full-thickness tears, 14% of partial-thickness tears, and 0% of intact grafts. Using both sagittal and coronal planes, complete graft discontinuity was seen in 50% of full-thickness graft tears and 0% of partial-thickness tears and intact grafts. Using the coronal plane, the diagnosis of full-thickness graft tear versus other conditions (partial-thickness tear or intact graft) resulted in a 75% sensitivity, 92% specificity, 75% positive predictive value, 92% negative predictive value, and 87% accuracy. Using the coronal and sagittal planes, the sensitivity was 50%; specificity, 100%; positive predictive value, 100%; negative predictive value, 86%; and accuracy, 87%. Using the coronal plane in the diagnosis of partial-thickness ACL graft versus other conditions (full-thickness tear or intact graft), the sensitivity was 14%; specificity, 67%; positive predictive value, 25%; negative predictive value, 50%; and accuracy, 44%. Using the coronal and sagittal planes, the sensitivity was 0%; specificity, 78%; positive predictive value, 0%; negative predictive value, 50%; and accuracy, 44%. Using the coronal plane in the diagnosis of ACL graft tear (partial or full thickness) versus intact graft, the sensitivity was 36%; specificity, 100%; positive predictive value, 100%; negative predictive value, 42%; and accuracy, 50%. Using the coronal and sagittal planes, the sensitivity was 18%; specificity, 100%; positive predictive value, 100%; negative predictive value, 36%; and accuracy, 44%. The kappa value for interobserver variability was 0.13 (poor agreement) for complete graft fiber discontinuity in the sagittal plane and 1.00 (almost perfect agreement) for the coronal plane.
ACL graft fiber continuity.—MR images were also evaluated for ACL graft fiber continuity (identification of any intact fibers) in the sagittal and coronal planes (Fig. 2A,2B). The results show that the coronal plane outperformed the sagittal plane; the accuracy in diagnosing intact graft or partial-thickness tear was 78% for the sagittal plane and 88% for the coronal plane. Therefore, usefulness of graft fiber continuity on coronal MR images was assessed. The diagnosis of intact graft or partial-thickness tear versus full-thickness tear resulted in a 92% sensitivity, 75% specificity, 92% positive predictive value, 75% negative predictive value, and 88% accuracy. The diagnosis of intact ACL graft versus other conditions (full-thickness or partial-thickness tear) resulted in a 100% sensitivity, 36% specificity, 42% positive predictive value, 100% negative predictive value, and 56% accuracy. The kappa value for interobserver variability was 1.00 (almost perfect agreement) for graft fiber continuity in the coronal plane.
Graft thickness.—MR images were assessed for ACL graft thickness to determine presence of focal thinning (Fig. 3A,3B). A 100% graft thickness (no thinning) on coronal MR images in the diagnosis of intact graft versus partial- or full-thickness tear resulted in a 60% sensitivity, 73% specificity, 48% positive predictive value, 81% negative predictive value, and 68% accuracy. A 100% graft thickness (no thinning) on MR images in the diagnosis of intact graft or partial-thickness tear versus full-thickness tear was assessed as follows. Using the sagittal plane, sensitivity was 41%; specificity, 100%; positive predictive value, 100%; negative predictive value, 35%; and accuracy, 56%. Using the coronal plane, the sensitivity was 50%; specificity, 100%; positive predictive value, 100%; negative predictive value, 40%; and accuracy, 62%. The kappa value for interobserver variability was 0.58 (moderate agreement) for graft thinning in the sagittal plane and 0.74 (substantial agreement) for graft thinning in the coronal plane.
Retrospective MR Imaging Results: Secondary Signs
Anterior tibial translation.—Abnormal anterior tibial
translation on sagittal images was present in 25% of the full-thickness graft
tears, 57% of partial-thickness tears, and 10% of intact grafts (Figs.
4 and
5). The diagnosis of
full-thickness graft tear versus other conditions (partial-thickness tear or
intact graft) resulted in a 25% sensitivity, 63% specificity, 20% positive
predictive value, 71% negative predictive value, and 53% accuracy. The
diagnosis of partial-thickness ACL graft versus other conditions
(full-thickness or intact graft) resulted in a 57% sensitivity, 84%
specificity, 73% positive predictive value, 73% negative predictive value, and
72% accuracy. The diagnosis of ACL graft tear (partial or full thickness)
versus intact graft resulted in a 46% sensitivity, 90% specificity, 93%
positive predictive value, 43% negative predictive value, and 60% accuracy.
The kappa value for interobserver variability was 0.60 (moderate agreement)
for anterior tibial translation.
Uncovered posterior horn of lateral meniscus.—Uncovered posterior horn of the lateral meniscus on sagittal images was present in 12.5% of the full-thickness graft tears, 29% of partia-thickness tears, and 0% of intact grafts (Figs. 4 and 5). In the diagnosis of full-thickness graft tear versus other conditions (partial-thickness tear or intact graft), the sensitivity was 13%; specificity, 83%; positive predictive value, 17%; negative predictive value, 74%; and accuracy, 66%. The diagnosis of partial-thickness ACL graft tear versus other conditions (full-thickness tear or intact graft) resulted in a 29% sensitivity, 95% specificity, 84% positive predictive value, 63% negative predictive value, and 66% accuracy. The diagnosis of ACL graft tear (partial or full thickness) versus intact graft resulted in a 23% sensitivity, 100% specificity, 100% positive predictive value, 38% negative predictive value, and 52% accuracy. The kappa value for interobserver variability was 0.29 (fair agreement) for uncovered posterior horn of lateral meniscus.
PCL hyperbuckling.—Hyperbuckling of the PCL on mid sagittal images was present in 37.5% of the full-thickness graft tears, 43% of partial-thickness tears, and 30% of intact grafts (Fig. 5). The diagnosis of full-thickness graft tear versus other conditions (partial-thickness tear or intact graft) resulted in a 38% sensitivity, 63% specificity, 27% positive predictive value, 75% negative predictive value, and 57% accuracy. The diagnosis of partial-thickness ACL graft versus other conditions (full thickness or intact graft) resulted in a 43% sensitivity, 67% specificity, 49% positive predictive value, 61% negative predictive value, and 56% accuracy. The diagnosis of ACL graft tear (partial or full-thickness) versus intact graft resulted in a 41% sensitivity, 70% specificity, 76% positive predictive value, 35% negative predictive value, and 50% accuracy. The kappa value for interobserver variability was 0.48 (moderate agreement) for PCL hyperbuckling.
Abnormal PCL line.—Abnormal PCL line on mid sagittal images was present in 25% of the full-thickness graft tears, 50% of partial-thickness tears, and 30% of intact grafts (Fig. 5). The diagnosis of full-thickness graft tear versus other conditions (partial-thickness tear or intact graft) resulted in a 25% sensitivity, 59% specificity, 19% positive predictive value, 69% negative predictive value, and 50% accuracy. The diagnosis of partial-thickness ACL graft versus other conditions (full-thickness tear or intact graft) resulted in a 50% sensitivity, 73% specificity, 56% positive predictive value, 67% negative predictive value, and 63% accuracy. The diagnosis of ACL graft tear (partial- or full-thickness tear) versus intact graft resulted in a 41% sensitivity, 70% specificity, 75% positive predictive value, 36% negative predictive value, and 50% accuracy. The kappa value for interobserver variability was 0.25 (fair agreement) for an abnormal PCL line.
Receiver Operating Characteristic Curve Analysis
In the diagnosis of partial-thickness ACL graft tear versus other
conditions (full-thickness tear or intact graft) on MR imaging, receiver
operating characteristic curves were constructed for each observer
(Fig. 6). The area under the
curve equaled 0.62 for the first observer and 0.40 for the second observer.
There was no significant difference between the areas under each curve
(p > 0.05). In the diagnosis of full-thickness ACL graft tear
versus other conditions (partial-thickness tear or intact graft) on MR
imaging, receiver operating characteristic curves were constructed for each
observer (Fig. 7). The area
under the curve equaled 0.97 for the first observer and 0.95 for the second
observer with no significant difference between the two areas (p >
0.05).
|
|
The grading results indicating the confidence of each retrospective observer in the diagnosis of partial-thickness or full-thickness graft tear was compared with the prospective MR imaging interpretations. Of the three MR imaging examinations originally interpreted as showing a partial-thickness tear, the first examination (with an intact graft at arthroscopy) had average grades of 2 and 5 for diagnosis of full-thickness and partial-thickness graft tear, respectively. The other two examinations (with full-thickness tears at arthroscopy) had average grades of 4.5 and 2.5, respectively, for the second examination, and 2 and 5, respectively, for the third. Of the two MR imaging examinations originally interpreted as showing a full-thickness tear, each had average grades of 4 and 3, respectively (both patients had a full-thickness tear at arthroscopy). Of the 11 MR imaging examinations originally interpreted as showing an intact graft, seven had a partial-thickness tear at arthroscopy, with average grades of 1 and 1, 1 and 3.5, 1.5 and 3, 1 and 1.5, 1 and 2, 2.5 and 4.5, and 2 and 4, respectively. The remaining four examinations with normal findings at arthroscopy had average grades of 1 and 1, 1 and 1.5, 1 and 2.5, and 1.5 and 3 for diagnosis of a full-thickness tear and a partial-thickness tear, respectively.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Identification of graft fiber continuity and complete graft discontinuity
was best achieved on the coronal images (almost perfect interobserver
agreement, 
= 1.00). Noting complete discontinuous graft on both
sagittal and coronal planes increased specificity and negative predictive
value to 100% in the diagnosis of a full-thickness graft tear. However,
agreement between observers in evaluating graft discontinuity in the sagittal
plane was poor ( = 0.13). Graft thickness was also determined to be an
important primary sign of ACL graft tear, relating most closely to the
continuity of graft fibers. If 100% graft thickness was identified in the
coronal or sagittal plane, the diagnosis of a full-thickness tear could be
excluded (specificity and positive predictive value of 100%, substantial
interobserver agreement) (Fig.
2A,2B).
Other primary signs of ACL graft tear evaluated in this study were less valuable. Increased graft signal is insensitive in diagnosing graft tear. However, diffuse increased signal has a 94% positive predictive value for the diagnosis of full-thickness or partial-thickness graft tear (noting low sensitivity of 45% and only moderate interobserver agreement) (Fig. 1A,1B). Lax graft fiber orientation is also not useful in diagnosing a tear (sensitivity of 32%, fair interobserver agreement). However, taut reconstruction graft between the distal femur and tibial tunnel is present in most intact grafts (90% specificity). It is unclear why only 12.5% of the full-thickness graft tears had graft laxity. This finding may be related to scar tissue from the prior surgery or the intrinsic properties of tendon graft, which may behave differently from native ligamentous tissue.
Regarding the secondary signs of ACL graft tear evaluated with MR imaging in this study, we found none particularly useful. For instance, the sensitivities in diagnosing graft tear ranged from 13% to 50%, and interobserver agreement for the secondary signs only ranged from fair to moderate. However, it should be noted that both anterior tibial translation and uncovering of the posterior horn of the lateral meniscus showed high specificities (90% and 100%, respectively) and high positive predictive values (93% and 100%, respectively) in differentiating any graft tear from an intact graft (Fig. 4). This suggests that the presence of these findings is helpful in predicting graft tear (keeping in mind that the low sensitivities indicate that most graft tears did not have these associated signs). It is unclear why many of these secondary signs associated with ACL tear were present with intact grafts. Because the secondary signs generally evaluate for anterior knee instability, perhaps these findings indicate that the intact grafts in this study did not provide the same degree of physiologic stability as seen with a native ACL. This theory is supported by the fact that three of the four intact grafts had mild instability or a grade I Lachman test. Another explanation would relate to inaccuracy in identification of the secondary MR imaging signs (fair to moderate interobserver agreement).
The topic of ACL graft signal intensity deserves further comment. It has been reported that increased signal intensity of clinically stable ACL grafts increases up to 12 months after surgery and then decreases over the subsequent 12 months [7, 10]. This increase has been attributed to revascularization and cellular infiltration [5] and has been considered an indeterminate finding in the assessment of graft integrity [4]. Of the five intact ACL grafts confirmed at arthroscopy in this study, one had diffuse increased signal and two had focal increased signal. The patient who had diffuse increased signal had a time interval of 56 months from ACL graft placement to MR imaging, and the remaining four patients with intact ACL grafts had an average time interval of 27 months. Likewise, the two patients with focal increased signal also had a longer time interval from ACL graft surgery to MR imaging (46-month average compared with 25-month average for the remaining intact ACL grafts). These findings suggest that both diffuse and focal increased graft signal may persist well beyond 12 months in an intact ACL graft. However, we also found that diffuse increased signal was seen in 50% of full-thickness tears, so considerable overlap exists. It should be noted that focal increased signal in the distal aspect of ACL grafts has been reported with graft impingement [11]. Of the intact grafts in this study with focal increased signal, none of the patients had abnormal anterior tibial tunnel placement. Therefore, graft impingement is likely not the cause of increased signal in these patients [13].
Our results show relative insensitivity in the detection of a partial-thickness ACL graft tear on MR imaging (Fig. 3A,3B), which is highlighted by the receiver operating characteristic curve data (Fig. 6). In addition, all seven of the partial-thickness graft tears were prospectively interpreted as normal on MR imaging. One possible reason for this result is that the arthroscopic diagnosis of partial-thickness graft tear by our surgeons takes into consideration the graft tautness in addition to anatomic appearance. Four of the seven arthroscopically diagnosed partial-thickness tears were described as lax at arthroscopy without gross evidence of fiber disruption. It is possible that although morphologically intact, these grafts were functional failures leading to the false-negative MR imaging interpretations. Fortunately, the differentiation between partial-thickness graft tear and intact ACL graft was not clinically significant in our cases. Of the seven partial-thickness graft tears, three received no treatment and the remaining four grafts were débrided without graft reconstruction. Differentiation of full-thickness graft tear from other conditions (intact graft or partial-thickness tear) remains clinically important; the receiver operating characteristic curve data (Fig. 7) show that the two observers were effective in this distinction.
Our study has limitations. The study sample was relatively small. Some selection bias was introduced by considering cases only arthroscopically proven. However, four of the patients were examined for reasons other than ACL graft evaluation thereby reducing this potential selection bias. We believe the benefit of arthroscopy as the optimal gold standard outweighs the limitations of the resulting selection bias. It is possible that intrinsic abnormalities of the ACL identified on MR imaging may not be recognized with arthroscopy. Additionally, bias was introduced because the orthopedic surgeons had knowledge of the original MR imaging results before arthroscopy. Several types of ACL grafts were used in this study. Because of the retrospective nature of this study, we were unable to standardize the MR imaging protocols and sequences. No specialized ACL specific sequences were used. With regard to the primary and secondary MR imaging findings, an attempt was made to limit their subjectivity by defining strict criteria. Any remaining variability in interpreting these signs that may have been a limitation is addressed with calculation of kappa values.
In summary, we found MR imaging to be accurate for discriminating complete ACL graft tear from partial-thickness tear and intact graft. The most useful MR imaging signs were graft fiber continuity (in the coronal plane), complete graft discontinuity (in the coronal plane), and graft thickness (in either the coronal or sagittal plane). The ability to distinguish partial-thickness tears from full-thickness tears and intact graft is limited.
Acknowledgments
We thank Mohammed Kabeto of Consortium for Health Outcomes, Innovation, and
Cost Effectiveness Studies for statistical analysis.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  T. R. McCauley MR Imaging Evaluation of the Postoperative Knee Radiology, January 1, 2005; 234(1): 53 - 61. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. R. McCauley, A. Elfar, A. Moore, A. H. Haims, P. Jokl, J. K. Lynch, P. A. Ruwe, and L. D. Katz MR Arthrography of Anterior Cruciate Ligament Reconstruction Grafts Am. J. Roentgenol., November 1, 2003; 181(5): 1217 - 1223. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
= observer 1; 
= observer 2.