(Radiology. 1999;210:183-188.)
© RSNA, 1999
Repeat Tear of Postoperative Meniscus: Potential MR Imaging Signs
Philip S. Lim, MD1,
Mark E. Schweitzer, MD1,
Manoj Bhatia, MD1,
Vincenzo Giuliano, MD1,
Perry P. Kaneriya, MD1,
Russ M. Senyk, BS1,
Marcelino Oliveri, DO2,
William Johnson, DO3,
Bernard Amster, DO3 and
Laurence Parker, PhD1
1 Department of Radiology, Thomas Jefferson University Hospital, 132 S 10th St, 1096 Main Bldg, Philadelphia, PA 19107 (P.S.L., M.E.S., M.B., V.G., P.P.K., R.M.S., L.P.)
2 Orthopedic Consultants, Lancaster, Pa (M.O.)
3 Orthopedic Associates, Langhorne, Pa (W.J., B.A.).
 |
Abstract
|
|---|
PURPOSE: To determine the usefulness of five magnetic resonance (MR) signs of repeat tears of the postoperative meniscus.
MATERIALS AND METHODS: Fat-saturated proton-density–weighted and T2-weighted MR images in 20 patients who had undergone at least two knee surgeries were reviewed. Second-look arthroscopic confirmation of meniscal status was available in all patients. MR images of 39 menisci (10 repeat tears, 11 postoperative menisci without repeat tears, three tears in previously normal menisci, and 15 normal menisci) were evaluated for (a) a linear area of abnormal signal intensity extending to an articular surface on intermediate-weighted images, (b) abnormal meniscal morphology, (c) likelihood of a typical postoperative appearance, (d) fluid extending into the linear area on T2-weighted images, and (e) cartilage defects. Two radiologists rated their confidence on a five-point scale for the presence of these signs and were compared for level of agreement.
RESULTS: The intraclass correlation coefficient (range, 0.75–0.91) indicated a high level of agreement. Areas under the receiver operating characteristic curves were large (range, 0.70–0.87) for all five signs, with a line (grade 3 signal intensity) and fluid within a line having the highest areas. These two signs also had the highest statistical scores (
2 = 14.12 and 13.30; P < .002; odds ratio, 1.588 and 1.599 for the line and the fluid within the line, respectively).
CONCLUSION: The two best MR signs of repeat tear of the postoperative meniscus were the presence of a line and fluid within the line extending to an articular surface.
Index terms: Knee, arthrography, 4524.122, 4525.122 • Knee, ligaments, menisci, and cartilage, 4524.4852, 4525.4852 • Knee, MR, 4524.121415, 4524.121416, 4525.121415, 4525.121416 • Knee, surgery, 4524.45, 4525.45
|
Introduction
|
|---|
Magnetic resonance (MR) imaging is accurate in diagnosing meniscal tears in menisci that have not been operated on or that are "virgin" menisci (1–3). However, linear areas of abnormal signal intensity are often seen in untorn, postoperative menisci at MR imaging (Fig 1), making it difficult to diagnose a repeat meniscal tear (4,5). Previous studies in which the presence of joint fluid within this linear area of abnormal signal intensity and intraarticular injection of gadolinium-based contrast material for MR arthrography have been used have improved MR accuracy in the diagnosis of repeat meniscal tear (6,7). Since the role of MR imaging in postoperative patients with recurrent knee symptoms is to differentiate repeat meniscal tears from other causes of knee pain, the purpose of our study was to determine the accuracy of five MR signs in diagnosing repeat tears of the postoperative meniscus.
View larger version (10K):
[in this window]
[in a new window]
|
Figure 1. Intrameniscal signal intensity conversion. Diagrams illustrate the potential of creating an intrasubstance grade 3 signal intensity abnormality after meniscectomy. A, torn meniscus. The dotted lines show the planned incision. The surgeon resects the torn portion of the virgin meniscus after probing it. This resected fragment usually does not contain the entire linear area of abnormal signal intensity (straight line) seen on the MR images. B, Untorn postoperative meniscus. A grade 3 signal intensity abnormality now extends to the articular surface and mimics a tear. (Modified, with permission, from reference 4.)
|
|
|
MATERIAL AND METHODS
|
|---|
A total of 39 menisci in 20 patients were retrospectively studied. One patient had undergone total meniscectomy. Inclusion criteria were at least two knee surgeries, with an interval knee MR imaging examination. Findings of this interval MR imaging examination were then evaluated.
Ten menisci in 10 patients had a repeat tear at the second surgery (nine men, one woman; age range, 25–59 years; mean age, 43 years; time between MR imaging and the second surgery, 0.3–4.0 months; mean time, 1.3 months; time between the two surgeries, 3–108 months; mean time, 45 months). Ten patients composed the control population (five men, five women; age range, 22–45 years; mean age, 32 years; time between MR imaging and the second surgery, 0.3–6.0 months; mean time, 3.2 months; time between the two surgeries, 2–216 months; mean time, 62 months). Meniscal surgery had previously been performed in 11 menisci but these menisci had no repeat tear. Three menisci had a tear in virgin menisci, and 15 menisci were normal at both surgeries. Surgical reports in all patients were used as the standard of reference in determining the presence of a repeat meniscal tear.
The surgical report for the initial surgery was available in only three patients. Therefore, we evaluated the degree of meniscectomy on the basis of the sagittal MR images in the remaining seven patients. Combining these data, we found that five had undergone 0%–25% meniscal resection and five had undergone 25%–75% meniscal resection.
Imaging was performed with a 1.5-T unit (Signa; GE Medical Systems, Milwaukee, Wis) by using a knee coil. Images from only a proton-density–weighted (1,000–1,300/20 [repetition time msec/echo time msec]) sequence with fat saturation in the sagittal and coronal planes and a T2-weighted fast spin-echo (SE) sequence (4,200–4,500/100–105) with fat saturation in the sagittal plane were evaluated for the purpose of this study. In addition, meniscal windowing of the sagittal and coronal proton-density–weighted and sagittal T2-weighted images were available. The usual field of view was 12–16 cm, the section thickness was 4 mm, and the intersection gap was 1 mm. The matrix was 256 x 256.
Two radiologists (M.B., V.G.) who were blinded to patient identification, clinical and surgical history, frequency of retorn menisci, and initial MR imaging readings rated their degree of confidence of five signs: (a) a high-signal-intensity linear area extending to an articular surface of the meniscus on proton-density–weighted images (traditional grade 3 signal intensity in meniscus [4,5]), (b) abnormal shape of the meniscus (evaluation of the ratio of anterior to posterior horns, presence of sharp triangular margins, and absence of meniscal flounce [8]), (c) meniscal shape likely to be postoperative (Fig 2), (d) fluid within the line that extended to the articular surface on T2-weighted images (linear area of fluid in the meniscus), and (e) cartilage defects (altered thickness and modified signal intensity). As concerns the likelihood that the MR imaging appearance was of a typical postoperative shape, the readers kept in mind that the surgeon tries to maintain as much meniscus as possible, leaves no sharp margins, tries not to violate the meniscocapsular junction, and keeps a meniscoid shape (9,10). These described signs were chosen because they have been used to determine the presence of a meniscal tear in either virgin or postoperative menisci (1,4,6). Also, the presence of cartilage defects may help explain knee pain in a postoperative patient (11–13). The following confidence scale in grading the signs was used: "1" indicated definitely absent; "2," probably absent; "3," unsure; "4," probably present; and "5," definitely present. When determining the number of true-positive (TP), true-negative (TN), false-positive (FP), and false-negative (FN) findings and the sensitivity, specificity, and accuracy for each sign, a total cutoff score of 8 was used (the sum of confidence scores given by both readers; a cutoff score of 8 or more indicated a total confidence score showing that the readers believed a meniscal retear was present; a score of 7 or less indicated absence of meniscal retear).
View larger version (10K):
[in this window]
[in a new window]
|
Figure 2. Possible postoperative meniscal shapes of the posterior horn, as seen on sagittal images. A, Diagram shows meniscus after 0%–25% meniscal resection, with reshaping of the inner margin. The inner margin appears blunted. B, Diagram shows meniscus after approximately 50% meniscal resection. C, Diagram shows meniscus after approximately 75% meniscal resection. D, Diagram shows meniscal regrowth after partial resection. The meniscus has an irregular, wavy contour with a slightly higher signal intensity (stippling) than that of the virgin meniscus on images obtained with a proton-density–weighted MR sequence.
|
|
By using all the scores for each meniscus by both readers, interrater reliability was measured with the intraclass correlation coefficient (14), and the areas under the receiver operating characteristic (ROC) curves (SAS/STAT Software: Changes and Enhancements through Release 6.12. Cary, NC: SAS Institute, 1997; pages 423–470) for each sign were calculated. A logistic regression analysis was then performed to determine the best MR sign of repeat meniscal tear (SAS/STAT Software: Changes and Enhancements through Release 6.12. Cary, NC: SAS Institute, 1997; pp 423–470). Further statistical evaluation was also performed to determine whether there was a significant difference in the readers' scores among the repeat tears in patients who had undergone 0%–25% or 25%–75% initial meniscal resection. The mean total scores for each of these two subgroups were calculated to see if there was a statistical difference in the usefulness of the signs.
|
RESULTS
|
|---|
Meniscal repeat tears in the experimental population were distributed at second surgery as follows: posterior horn of medial meniscus, six menisci; posterior horn of lateral meniscus, one meniscus; anterior horn of lateral meniscus, one meniscus; all portions of the medial meniscus, one meniscus; all portions of the lateral meniscus, one meniscus. Table 1 illustrates the degree of interrater correlation and areas under the ROC curves. Table 2 illustrates the number of TP, TN, FP, and FN findings and gives the sensitivity, specificity, and accuracy for both the experimental and control populations when a total confidence score of eight was used as a cutoff. When the mean rater scores were calculated in the experimental subgroups of patients who had undergone 0%–25% meniscectomy and 25%–75% meniscectomy, no statistical difference was observed.
The presence of a line that extended to an articular surface on proton-density–weighted images and the observation of fluid within the line on T2-weighted images had the same number of TP (n = 9), TN (n = 23), FP (n = 3), and FN (n = 4) findings and the same sensitivity (69%), specificity (88%), and accuracy (82%) (Figs 3–5). Observations of these two signs by both readers had the highest areas under the ROC curve (0.85 and 0.87). By using logistic regression analysis to determine the sign that best correlated with a meniscal repeat tear, the presence of a line within the meniscus (traditional grade 3 signal intensity) was found to have the highest value (2 score of 14.12; P < .002; odds ratio, 1.588). Very similar values were calculated for the presence of fluid within the line (2 score of 13.30; P < .002; odds ratio, 1.599). An abnormally shaped meniscus and the postoperative appearance of a meniscus had very similar numbers of TP, TN, FP, and FN findings and sensitivity, specificity, and accuracy. Both of these morphologic signs overall had lower sensitivity (54% and 64%), specificity (73% and 77%), and accuracy (both 67%) in helping the reader detect repeat meniscal tears.
View larger version (143K):
[in this window]
[in a new window]
|
Figure 3a. Proved repeat tear of the medial meniscus in a 46-year-old man. The tear is documented as an intrameniscal linear area of abnormal signal intensity extending to the articular surface and as fluid within the line (TP repeat meniscal tear). (a) Sagittal proton-density–weighted fat-suppressed image (1,133/20) shows an oblique linear area of abnormal signal intensity (arrow) extending to the undersurface of the meniscus. (b) Sagittal T2-weighted fast SE fat-suppressed image (4,500/80) shows fluid within the line (straight arrow) and a moderate joint effusion (curved arrow).
|
|
View larger version (179K):
[in this window]
[in a new window]
|
Figure 3b. Proved repeat tear of the medial meniscus in a 46-year-old man. The tear is documented as an intrameniscal linear area of abnormal signal intensity extending to the articular surface and as fluid within the line (TP repeat meniscal tear). (a) Sagittal proton-density–weighted fat-suppressed image (1,133/20) shows an oblique linear area of abnormal signal intensity (arrow) extending to the undersurface of the meniscus. (b) Sagittal T2-weighted fast SE fat-suppressed image (4,500/80) shows fluid within the line (straight arrow) and a moderate joint effusion (curved arrow).
|
|
View larger version (110K):
[in this window]
[in a new window]
|
Figure 5a. Proved repeat meniscal tear at the site of initial surgery in a 52-year-old man (TP repeat meniscal tear). (a) Initial sagittal proton-density–weighted fat-suppressed image (1,033/20) shows a mild irregularity (arrow) at the undersurface of the meniscus that was a tear at surgery. (b) Initial sagittal T2-weighted fast SE fat-suppressed image (4,233/105) shows mild irregularity at the undersurface but no fluid within the line. (c) Sagittal proton-density–weighted fat-suppressed image (1,183/20) and (d) corresponding sagittal T2-weighted fast SE fat-suppressed image (4,233/100) obtained 24 months after a and b. c shows a linear area of abnormal signal intensity (arrow) extending to the undersurface of the meniscus. d shows fluid within the line (arrow), indicating a tear that has extended from the initial meniscal surgical site.
|
|
View larger version (112K):
[in this window]
[in a new window]
|
Figure 5b. Proved repeat meniscal tear at the site of initial surgery in a 52-year-old man (TP repeat meniscal tear). (a) Initial sagittal proton-density–weighted fat-suppressed image (1,033/20) shows a mild irregularity (arrow) at the undersurface of the meniscus that was a tear at surgery. (b) Initial sagittal T2-weighted fast SE fat-suppressed image (4,233/105) shows mild irregularity at the undersurface but no fluid within the line. (c) Sagittal proton-density–weighted fat-suppressed image (1,183/20) and (d) corresponding sagittal T2-weighted fast SE fat-suppressed image (4,233/100) obtained 24 months after a and b. c shows a linear area of abnormal signal intensity (arrow) extending to the undersurface of the meniscus. d shows fluid within the line (arrow), indicating a tear that has extended from the initial meniscal surgical site.
|
|
View larger version (87K):
[in this window]
[in a new window]
|
Figure 5c. Proved repeat meniscal tear at the site of initial surgery in a 52-year-old man (TP repeat meniscal tear). (a) Initial sagittal proton-density–weighted fat-suppressed image (1,033/20) shows a mild irregularity (arrow) at the undersurface of the meniscus that was a tear at surgery. (b) Initial sagittal T2-weighted fast SE fat-suppressed image (4,233/105) shows mild irregularity at the undersurface but no fluid within the line. (c) Sagittal proton-density–weighted fat-suppressed image (1,183/20) and (d) corresponding sagittal T2-weighted fast SE fat-suppressed image (4,233/100) obtained 24 months after a and b. c shows a linear area of abnormal signal intensity (arrow) extending to the undersurface of the meniscus. d shows fluid within the line (arrow), indicating a tear that has extended from the initial meniscal surgical site.
|
|
View larger version (94K):
[in this window]
[in a new window]
|
Figure 5d. Proved repeat meniscal tear at the site of initial surgery in a 52-year-old man (TP repeat meniscal tear). (a) Initial sagittal proton-density–weighted fat-suppressed image (1,033/20) shows a mild irregularity (arrow) at the undersurface of the meniscus that was a tear at surgery. (b) Initial sagittal T2-weighted fast SE fat-suppressed image (4,233/105) shows mild irregularity at the undersurface but no fluid within the line. (c) Sagittal proton-density–weighted fat-suppressed image (1,183/20) and (d) corresponding sagittal T2-weighted fast SE fat-suppressed image (4,233/100) obtained 24 months after a and b. c shows a linear area of abnormal signal intensity (arrow) extending to the undersurface of the meniscus. d shows fluid within the line (arrow), indicating a tear that has extended from the initial meniscal surgical site.
|
|
False interpretations were present in both our control and experimental populations. In three patients who had severe degenerative cartilage loss and virgin menisci, both readers interpreted an abnormal shape as a postoperative appearance (Fig 6). Degenerative fraying was seen at arthroscopy in one virgin meniscus; the MR images were interpreted as normal. Also, a 2.5-mm tear of the inner rim at the posterior horn of the medial meniscus was seen at surgery; the MR images were interpreted as probably normal. One meniscus had no repeat tear at surgery, but both readers interpreted it as the presence of a line extending to the articular surface and fluid within the line. Another patient had a repeat tear of the anterior horn of the lateral meniscus, but both readers interpreted it as only an abnormal shape and a postoperative appearance without a linear area of abnormal signal intensity (Fig 7). Two patients had previously undergone meniscectomy but had no repeat tear, which both readers interpreted as a linear area of abnormal signal intensity without fluid within the line. One patient had previously undergone meniscectomy and had only fraying, which both readers interpreted as the presence of all five signs (Fig 8).
View larger version (80K):
[in this window]
[in a new window]
|
Figure 6. Abnormal meniscal morphology in a 34-year-old woman with intractable knee pain due to severe tricompartmental osteoarthritis; the patient had not undergone prior meniscal resection (FP finding). Sagittal proton-density–weighted fat-suppressed image (1,000/20) shows that the anterior horn of the medial meniscus (arrow) is diminutive.
|
|
View larger version (63K):
[in this window]
[in a new window]
|
Figure 7a. Proved meniscal repeat tear at the anterior horn of the lateral meniscus in a 30-year-old man (FN finding). (a) Sagittal proton-density–weighted fat-suppressed image (1,183/20) shows an abnormal shape of the anterior horn and a likely postoperative appearance but no linear area of abnormal signal intensity. The linear area of abnormal signal intensity extending to the articular surface of the posterior horn (arrow) was due to prior meniscal resection, but no repeat tear was seen at second surgery. (b) Coronal proton-density–weighted fat-suppressed image (1,000/20) shows an abnormal shape of the anterior horn of the lateral meniscus (arrow).
|
|
View larger version (55K):
[in this window]
[in a new window]
|
Figure 7b. Proved meniscal repeat tear at the anterior horn of the lateral meniscus in a 30-year-old man (FN finding). (a) Sagittal proton-density–weighted fat-suppressed image (1,183/20) shows an abnormal shape of the anterior horn and a likely postoperative appearance but no linear area of abnormal signal intensity. The linear area of abnormal signal intensity extending to the articular surface of the posterior horn (arrow) was due to prior meniscal resection, but no repeat tear was seen at second surgery. (b) Coronal proton-density–weighted fat-suppressed image (1,000/20) shows an abnormal shape of the anterior horn of the lateral meniscus (arrow).
|
|
View larger version (125K):
[in this window]
[in a new window]
|
Figure 8a. Proved meniscal fraying rather than repeat tear in a 22-year-old woman (FP meniscal tear at site of original surgery). (a) Initial sagittal proton-density–weighted fat-suppressed image (1,066/20) shows an oblique linear area of abnormal signal intensity (arrows) extending to the undersurface of the meniscus. (b) Initial sagittal T2-weighted fast SE fat-suppressed image (4,233/100) shows no fluid within the line. (c) Sagittal proton-density–weighted fat-suppressed image (1,133/20) obtained 7 months after a shows a more extensive linear area of abnormal signal intensity (arrows) extending to the superior and inferior articular surfaces of the meniscus. (d) Corresponding sagittal T2-weighted fast SE fat-suppressed image (4,233/100) shows faint fluid within the line (arrow) extending to the superior articular surface of the meniscus that was not present on the initial MR study. At surgery, only meniscal fraying was seen.
|
|
View larger version (138K):
[in this window]
[in a new window]
|
Figure 8b. Proved meniscal fraying rather than repeat tear in a 22-year-old woman (FP meniscal tear at site of original surgery). (a) Initial sagittal proton-density–weighted fat-suppressed image (1,066/20) shows an oblique linear area of abnormal signal intensity (arrows) extending to the undersurface of the meniscus. (b) Initial sagittal T2-weighted fast SE fat-suppressed image (4,233/100) shows no fluid within the line. (c) Sagittal proton-density–weighted fat-suppressed image (1,133/20) obtained 7 months after a shows a more extensive linear area of abnormal signal intensity (arrows) extending to the superior and inferior articular surfaces of the meniscus. (d) Corresponding sagittal T2-weighted fast SE fat-suppressed image (4,233/100) shows faint fluid within the line (arrow) extending to the superior articular surface of the meniscus that was not present on the initial MR study. At surgery, only meniscal fraying was seen.
|
|
View larger version (131K):
[in this window]
[in a new window]
|
Figure 8c. Proved meniscal fraying rather than repeat tear in a 22-year-old woman (FP meniscal tear at site of original surgery). (a) Initial sagittal proton-density–weighted fat-suppressed image (1,066/20) shows an oblique linear area of abnormal signal intensity (arrows) extending to the undersurface of the meniscus. (b) Initial sagittal T2-weighted fast SE fat-suppressed image (4,233/100) shows no fluid within the line. (c) Sagittal proton-density–weighted fat-suppressed image (1,133/20) obtained 7 months after a shows a more extensive linear area of abnormal signal intensity (arrows) extending to the superior and inferior articular surfaces of the meniscus. (d) Corresponding sagittal T2-weighted fast SE fat-suppressed image (4,233/100) shows faint fluid within the line (arrow) extending to the superior articular surface of the meniscus that was not present on the initial MR study. At surgery, only meniscal fraying was seen.
|
|
View larger version (124K):
[in this window]
[in a new window]
|
Figure 8d. Proved meniscal fraying rather than repeat tear in a 22-year-old woman (FP meniscal tear at site of original surgery). (a) Initial sagittal proton-density–weighted fat-suppressed image (1,066/20) shows an oblique linear area of abnormal signal intensity (arrows) extending to the undersurface of the meniscus. (b) Initial sagittal T2-weighted fast SE fat-suppressed image (4,233/100) shows no fluid within the line. (c) Sagittal proton-density–weighted fat-suppressed image (1,133/20) obtained 7 months after a shows a more extensive linear area of abnormal signal intensity (arrows) extending to the superior and inferior articular surfaces of the meniscus. (d) Corresponding sagittal T2-weighted fast SE fat-suppressed image (4,233/100) shows faint fluid within the line (arrow) extending to the superior articular surface of the meniscus that was not present on the initial MR study. At surgery, only meniscal fraying was seen.
|
|
|
DISCUSSION
|
|---|
Several authors have described the frequent presence of a line extending to an articular surface (grade 3 signal intensity) in postoperative menisci that makes the MR distinction of a repeat tear from an untorn postoperative meniscus difficult (4,5). After probing the torn virgin meniscus, the orthopedic surgeon resects the torn portion. This resected fragment, however, is usually not the entire linear area of abnormal signal intensity interpreted on the MR images to be the tear, and grade 3 signal intensity remains in the postoperative meniscus and is referred to as "intrameniscal signal conversion" (4,5).
Farley et al (6) provided a useful algorithm in trying to determine the presence of a repeat tear in a meniscus that had previously been arthroscopically repaired: a full-thickness grade 3 signal intensity area with an increase in signal intensity on T2-weighted images, a tear at a location other than the old repair site, or a displaced meniscal fragment. Applegate et al (7) reported improved accuracy in diagnosing arthroscopically proved repeat tears with MR arthrography when patients had resection of more than 25% of the meniscus, as intraarticular gadolinium was seen within a line that extended to an articular surface.
In our study, a sensitivity of 69% for the findings of fluid in the line extending to the articular surface confirms the 60% sensitivity reported by Farley et al (6). We also had an accuracy of 82% in detecting repeat tears by using the presence of a grade 3 signal intensity abnormality or fluid within the line. When one considers our small sample size, this value also confirms the 65% accuracy reported by Applegate et al (7) when they used conventional MR imaging alone in diagnosing repeat tears. Our study differed from the study by Applegate et al in several other ways. A population-selection bias was present in their study, where all 39 patients had previously undergone meniscal surgery. Our study, although including only patients who had previously undergone arthroscopy, included some patients who did not previously undergo meniscal surgery. Another difference may be that more of our experimental subjects may have undergone resection of less than 25% of the meniscus during the initial surgery. This is possible if we erred in estimating the degree of meniscal resection on the sagittal views of the meniscus in seven patients.
It has previously been reported that MR signs for detecting repeat tears in patients who have undergone prior 0%–25% meniscal resection are similar to the signs in patients with virgin menisci (4,5). Also, early preliminary studies of a small number of menisci have shown 100% agreement between arthroscopy and conventional MR imaging in detecting repeat tears in the 0%–25% resection group, approximately 64% agreement in the 25%–75% group, and approximately 85% concordance in the greater than 75% group (4). However, we found no statistical difference in the readers' scores between the two subgroups with repeat meniscal tears who had either 0%–25% or 25%–75% initial meniscectomy. If we are correct in our assessment of the degree of initial meniscectomy, then perhaps the presence of a grade 3 signal intensity abnormality in a postoperative meniscus in both subgroups with repeat meniscal tears is a better MR sign than initially reported.
The presence of fluid within the line that extended to an articular surface had the highest area under the ROC curve. The absence of a joint effusion that can imbibe the torn meniscus may limit the usefulness of this sign. By using MR arthrography, Applegate et al (7) reported 87% and 100% accuracies in detecting meniscal repeat tears in patients who had undergone 25%–75% and greater than 75% initial meniscal resections, respectively. Similarly, Katz et al (15) reported improved accuracy in differentiating postoperative meniscal changes from repeat tear by using MR arthrography. Whether this is a contrast or distention effect is unknown, although Applegate et al (16) and Hall (17) believe the value of MR arthrography lies in the distention effect. We found a high specificity (88%) but a moderate sensitivity (69%) for detecting repeat tears by using linear areas of fluid extending to the articular surface of the meniscus. These findings suggest that the presence of fluid in the line should dissuade one from using intraarticular contrast material. However, intraarticular contrast material may help find a tear in the absence of a linear area of abnormal signal intensity at conventional MR imaging.
An abnormal shape and a typical postoperative appearance had much lower sensitivities, specificities, and accuracies with higher numbers of FN and FP findings in menisci with repeat tears. This supports the finding reported by Smith and Totty (11) that the diagnosis of repeat tears of meniscal segments with marked contour irregularity should be made cautiously, since this irregularity can mimic a tear.
The presence of cartilage abnormalities was seen in both our control and experimental groups, and there were large numbers of FP and FN findings with use of this sign alone. This finding was expected in our population, since cartilage defects are a common cause of pain in patients who have previously undergone knee arthroscopy. In addition, conventional MR sequences, as were used in this study, are of limited usefulness to evaluate articular cartilage.
Another limitation may be that our standard of reference relied on several different arthroscopists' interpretations of what constituted a tear. Difficulty in establishing an arthroscopic diagnosis of meniscal tear has been studied by Quinn and Brown (18). First, a surgeon's interpretation can be based on subjective criteria; what one arthroscopist considers a tear of the free margin may be considered a frayed edge by another. Second, all segments of the posterior horn of the meniscus cannot be depicted with arthroscopy. Instead, the method of probing and compressing the articular surfaces of the meniscus is accepted by arthroscopists as adequate examination of this region. This limitation may have affected our results.
Other limitations of our study were the small number of patients, especially the small number of patients with virgin menisci in which a tear was confirmed at the second surgery. Another limitation that may have hindered detection of meniscal repeat tears was the absence of T2-weighted coronal imaging in our imaging protocol. Also, patients in the experimental group had undergone variable degrees of initial meniscal surgeries.
In conclusion, the two best MR imaging signs of repeat tear of the postoperative meniscus were the presence of a grade 3 signal intensity abnormality and the presence of fluid within a line that extended to an articular surface. Surprisingly, we found that a line extending to the articular surface remains an accurate sign of repeat meniscal tear.
View larger version (180K):
[in this window]
[in a new window]
|
Figure 4a. Proved repeat tear of the medial meniscus in a 49-year-old man. The tear is documented as an intrameniscal linear area of abnormal signal intensity extending to the articular surface and as fluid within the line (TP repeat meniscal tear). (a) Sagittal proton-density–weighted fat-suppressed image (1,133/20) shows a linear area of abnormal signal intensity extending from the superior (double arrows) to the inferior (single arrow) articular surfaces of the meniscus. (b) Sagittal T2-weighted fast SE fat-suppressed image (4,488/100) shows fluid within the line (straight arrows) and a small joint effusion. Subtle femoral cartilage loss (curved arrows) also is seen.
|
|
View larger version (173K):
[in this window]
[in a new window]
|
Figure 4b. Proved repeat tear of the medial meniscus in a 49-year-old man. The tear is documented as an intrameniscal linear area of abnormal signal intensity extending to the articular surface and as fluid within the line (TP repeat meniscal tear). (a) Sagittal proton-density–weighted fat-suppressed image (1,133/20) shows a linear area of abnormal signal intensity extending from the superior (double arrows) to the inferior (single arrow) articular surfaces of the meniscus. (b) Sagittal T2-weighted fast SE fat-suppressed image (4,488/100) shows fluid within the line (straight arrows) and a small joint effusion. Subtle femoral cartilage loss (curved arrows) also is seen.
|
|
|
Footnotes
|
|---|
Address reprint requests to M.E.S.
From the 1997 RSNA scientific assembly.
Abbreviations: FN = false-negative
FP = false-positive
ROC = receiver operating characteristic
SE = spin echo
TN = true-negative
TP = true-positive
Author contributions: Guarantors of integrity of entire study, P.S.L., M.E.S.; study concepts and design, M.E.S.; literature research, P.S.L., M.E.S.; data acquisition, P.S.L., M.B., V.G., P.P.K., R.M.S., M.O., W.J., B.A.; data analysis, P.S.L., M.E.S., L.P.; statistical analysis, P.S.L., M.E.S.; manuscript preparation, editing, and review, P.S.L., M.E.S.
Received January 28, 1998;
revision requested April 7, 1998; revision received May 29, 1998;
accepted September 10, 1998.
|
References
|
|---|
-
Reicher MA, Steven H, Duckwiler GR, Bassett LW, Anderson LJ, Gold RH. Meniscal injuries: detection using MR imaging. Radiology 1986; 159:753-757.[Abstract/Free Full Text]
-
Crues JV, Mink J, Levy TL, Lotysch M, Stoller DW. Meniscal tears of the knee: accuracy of MR imaging. Radiology 1987; 164:445-448.[Abstract/Free Full Text]
-
Fischer SP, Fox JM, Del Pizzo W, Friedman MJ, Snyder SJ, Ferkel RD. Accuracy of diagnoses from magnetic resonance imaging of the knee. J Bone Joint Surg [Am] 1991; 73:2-10.[Abstract/Free Full Text]
-
Deutsch Al, Mink JH. The post-operative knee. In: Mink JH, Reicher M, Crues JV, Deutsch AL, eds. MRI of the knee. 2nd ed. New York, NY: Raven, 1993; 237-290.
-
Deutsch AL, Mink JH, Fox JM, et al. Peripheral meniscal tears: MR findings after conservative treatment or arthroscopic repair. Radiology 1990; 176:485-488.[Abstract/Free Full Text]
-
Farley TE, Howell SM, Love KF, Wolfe RD, Neumann CH. Meniscal tears: MR and arthrographic findings after arthroscopic repair. Radiology 1991; 180:517-522.[Abstract/Free Full Text]
-
Applegate GR, Flannigan BD, Tolin BS, Fox JM, Pizzo WD. MR diagnosis of recurrent tears in the knee: value of intraarticular contrast material. AJR 1993; 161:821-825.[Abstract/Free Full Text]
-
Chew FS. Medial meniscal flounce: demonstration on MR imaging of the knee (letter). AJR 1990; 155:199.[Medline]
-
Metcalf RW. The torn medial meniscus. In: Parisien JS, eds. Arthroscopic surgery. New York, NY: McGraw-Hill, 1988; 93-110.
-
Newman AP, Daniels AU, Burks RT. Principles and decision making in meniscal surgery. Arthroscopy 1993; 9:33-51.[Medline]
-
Smith DK, Totty WG. The knee after partial meniscectomy: MR imaging features. Radiology 1990; 176:141-144.[Abstract/Free Full Text]
-
Brahme SK, Fox Jm, Ferkel RD, et al. Osteonecrosis of the knee after arthroscopic surgery: diagnosis with MR imaging. Radiology 1991; 178:851-853.[Abstract/Free Full Text]
-
Deutsch AL, Mink JH, Fox JM, Friedman MJ, Howell SM. The postoperative knee. Magn Reson Q 1992; 8:23-54.[Medline]
-
Portney LG, Watkins MP. Statistical measures of reliability. In: Portney LG, Watkins MP, eds. Foundations of clinical research: applications to practice. Norwalk, CT: Appleton & Lange, 1993; 505-528.
-
Katz LD, Lynch K, Ruwe P, Jokl P, Kelley J. MR arthrography of the knee in the postoperative meniscus in patients with knee pain (abstr). Radiology 1996; 201(P):169.
-
Applegate GR, Flannigan BD, Tolin BS, Fox JM, Pizzo WD. MR diagnosis of recurrent tears in the knee: value of intraarticular contrast material (reply). AJR 1994; 162:1502-1503.[Medline]
-
Hall FM. MR diagnosis of recurrent meniscal tears: how soon we forget (letter). AJR 1994; 162:1502-1503.
-
Quinn SF, Brown TF. Meniscal tears diagnosed with MR imaging versus arthroscopy: how reliable a standard is arthroscopy?. Radiology 1991; 181:843-847.[Abstract/Free Full Text]
This article has been cited by other articles:
|
|
|
|
 
T. G. Sanders and M. D. Miller
A Systematic Approach to Magnetic Resonance Imaging Interpretation of Sports Medicine Injuries of the Knee
Am. J. Sports Med.,
January 1, 2005;
33(1):
131 - 148.
[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]
|
|
|
|
|
|
|
M. J. Vives, D. Homesley, M. G. Ciccotti, and M. E. Schweitzer
Evaluation of Recurring Meniscal Tears with Gadolinium-Enhanced Magnetic Resonance Imaging: A Randomized, Prospective Study
Am. J. Sports Med.,
November 1, 2003;
31(6):
868 - 873.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
C. Mutschler, B. C. Vande Berg, F. E. Lecouvet, P. Poilvache, J.-E. Dubuc, B. Maldague, and J. Malghem
Postoperative Meniscus: Assessment at Dual-Detector Row Spiral CT Arthrography of the Knee
Radiology,
September 1, 2003;
228(3):
635 - 641.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
L. M. White, M. E. Schweitzer, D. Weishaupt, J. Kramer, A. Davis, and P. H. Marks
Diagnosis of Recurrent Meniscal Tears: Prospective Evaluation of Conventional MR Imaging, Indirect MR Arthrography, and Direct MR Arthrography
Radiology,
February 1, 2002;
222(2):
421 - 429.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
Abstract
Introduction
