Magnetic resonance imaging (MRI) and arthroscopy in the detection of meniscal degenerations: Correlation of arthroscopy and MRI with histology findings

Arthroscopy: The Journal of Arthroscopic and Related Surgery 10(6):634-640 Published by Raven Press, Ltd. © 1994 Arthroscopy Association of North America

Magnetic Resonance Imaging (MRI) and Arthroscopy in the Detection of Meniscal Degenerations: Correlation of Arthroscopy and MRI with Histology Findings J. Raunest, M.D., H. Hftzinger, M.D., and K. F. Biirrig, M.D.

Stmamary: In a prospective double-blind study, the capability of magnetic resonance imaging (MRI) and arthroscopy in the detection and grading of meniscal degenerations is evaluated by correlating MRI findings and arthroscopic diagnoses with a histologic grading model. In 82.8% of our results, grading based on MRI studies corresponded with the histologic grading classification. In 12 instances a meniscal degeneration verified at light microscopy was not detected at MRI, whereas in 15 cases tomography yielded a false-positive result. The overall accuracy was calculated to be 0.93 with a specificity of 0.79 and a sensitivity of 0.96. Concerning the evaluation of meniscal degenerations, MRI provides a positive predictive value of 0.95 and a negative predictive value of 0.82. Compared with the diagnostic specificity of the anterior and posterior zones, that of the intermediate segment of the meniscus is significantly reduced (p < 0.001). At arthroscopy, meniscal degenerations were diagnosed with an overall accuracy of 38.8%, a sensitivity of 27.5%, and a specificity of 75.5%. In 80 cases of grade 3 abnormalities, five false-negative diagnoses were made initially. These results suggest that MRI offers a valuable diagnostic potential providing reliable information about the internal consistency of the meniscus complementary to diagnostic arthroscopy. Key Words: Magnetic resonance imagingmDiagnostic arthroscopy--Meniscus--Degeneration.

Magnetic r e s o n a n c e imaging (MRI) is widely used in the evaluation o f meniscal disorders. In an earlier prospective and double-blinded study we demonstrated that MRI had an accuracy of 78% in the delineation o f menical ruptures (I), a p o o r diagnostic rate c o m p a r e d with the 95% accuracy provided by arthroscopy (2). Comparable results have

been reported by Reicher et al. and M a n c o et al. (3,4). H o w e v e r , little is known so far about the capability o f MRI in the diagnosis of internal foci of degeneration in the knee joint meniscus. True sensitivity and specificity are difficult to access in clinical studies using arthroscopy as a standard method of validation because this procedure is confined to inspection and probing of the meniscal surfaces. The current study was designed to evaluate the capability o f MRI and arthroscopy in the detection o f internal meniscus degeneration and to correlate an MRI grading system of degeneration with histologic findings.

From the Department of General Surgery and Traumatology, Heinrich-Heine-University, IXisseldorf (J.R.); Department of Radiology, Klinikum Niederberg, Velbert (H.H.); and Department of Pathology, St~dtische Krankenanstalten, Hildesheim (K.F.B.), Germany. Address correspondence and reprint requests to Priv.-Doz. Dr. J. Raunest, Abteilung ~ r Allgemein. und Unfallchirurgie, der Heinrich-Heine-Universit~t,Moorenstrasse 5, 13-40225Diisseldorf 1, Germany. No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.

MATERIALS AND METHODS In an experimental trial, MRI was performed on 40 knee specimens obtained within 12 h after au-

634

MR IMAGING A N D A R T H R O S C O P Y I N MENISCAL DEGENERATIONS

topsy. The specimen included a 10-cm portion of the femoral and tibial shaft. Care was taken not to affect the joint cavity before MRI and arthroscopy examination. Knee joints with a previous history of operation or injury were excluded. Subjects ranged in age from 21 to 89 years (mean 71) with equal representation of men and women. As expected, the population represented an increased prevalence of meniscal degenerations (386 pathologic findings in 480 specimens) with a predominance for severe alterations in women. MRI was performed in axial, sagittal, coronal, and oblique planes with the use of a 1.5-tesla superconducting imager (General Electric, Milwaukee, WI). A GRE sequence for localization was made in the axial plane with a repetition time (TR) of 110 msec, and echo time (TE) of 21 msec, and a flip angle of 30°. Sections having a thickness of 4 mm with interslice gaps of zero mm were obtained with a 12.0-cm field of view in a 256 x 256 pixel acquisation matrix. In the coronal and oblique planes, a spin echo of 800/25 msec (TR/TE) was used to emphasize the T1 characteristics of the image with a 20.0-cm field of view format. Contiguous 5-mm interleaved sections were imaged with the knee in 10°-20° of external rotation. Sagittal images were obtained in a T2-weighted mode using a TR of 2,000 msec and a TE of 20-80 msec. The thickness of sections was 4.0 mm with interslice gaps of 1.5 mm.

635

An MRI grading system introduced by Crues et al. (5) was used to classify meniscal degenerations on the basis of objective criteria. The radial image plane best defined meniscal pathology and had to be corroborated in at least one of the other three planes. The following definitions were used in this study (Fig. 1): Grade 0- normal, homogenous hypoechoic structure Grade 1: globular increased MRI signal intensity not adjacent to either the femoral or the tibial articular meniscal surface Grade 2: primarily linear signal intensity within the meniscus not extending to an articular surface Grade 3: increased MRI signal intensity in combination with a separation of the meniscal structure, limited to the internal meniscal structure (grade 3A) or extending at least to one meniscal articular margin (grade 3B) After MRI, arthroscopic examination of the cadaver knees was performed. The joint cavity was irrigated with Ringer's solution. For the examination of the medial meniscus, an anterolateral approach was used for the scope, and the probe was inserted anteromedially. For the lateral meniscus the diagnostic procedure was re~,ersed. A meniscus tear was verified by inspection and probing. The diagnosis of a meniscal degeneration was estab1A, B

FIG. 1. MRI grading system. A: Grade 0: normal hypoechoic structure. B: Grade 1: globular increased MRI signal intensity not adjacent to the articular surface of the meniscus. C: Grade 2: linear signal intensity within the meniscus. D: Grade 3: increased MRI signal intensity combined with a separation of the meniscal structure.

1C,D

Arthroscopy, Vol. 10, No. 6, 1994

636

J. R A U N E S T E T AL.

lished on the basis of a changed color of the meniscus and/or a pathologic consistency at probing. After the diagnostic procedures, the menisci were removed in total. Perpendicular to the horizontal plane of the meniscus, serial sections each having a thickness of 4 wm were obtained from two regions of the anterior, intermediate, and posterior part of the meniscus. With the help of a special threedimensional guiding system we could ascertain that the regions of interest were identical in both MRI and histology examination with a spatial resolution of 20 wm. The menisci were stained with hematoxylin-eosin and prepared for light microscopy examination. A total of 480 different imaging and histology studies were correlated. According to Copenhaver (6), the following definitions were used for a classification on a histologic level (Fig. 2): Stage 0: homogenous eosinophilic staining collagen reinforced ground substance with normal chondrocytes

Stage 1: discrete foci of mucinous, hyaline, or myxoid degeneration and reduction of the chondrocyte concentration Stage 2: bands of mucinous degeneration bordering hypocellular regions of the meniscus without presence of a distinct cleavage plane Stage 3: mucinous degenerations with fibrocartilagenous separation MRI, arthroscopy, and histology stages were evaluated separately by a radiologist, arthroscopist, and pathologist, thus ensuring the conditions of a prospective double-blind protocol and gaining independence between the diagnostic test and the procedure of validation. Interobserver variability was controlled by ensuring that one experienced radiologist interpreted the imaging studies, one arthroscopist performed the endoscopic examinations, and one pathologist classified the degree of degeneration on the basis of histologic examinations. The ×2 test and the Fisher exact probability test were performed to calculate statistical significance,

2A,B

2C,D

FIG. 2. Histologic classification criteria (hematoxylin-eosin stain, original magnfl%ation ×150). A: Stage 0: homogenous staining collagen with interposed chondrocytes. B: Stage 1: discrete foci of mucinous degeneration. C: Stage 2: bands of mucinous degeneration. D: Stage 3: mucinous degeneration combined with fibrocartilagenous separation. Arthroscopy, Vol. t0, No. 6, 1994

MR IMAGING A N D A R T H R O S C O P Y I N MENISCAL DEGENERATIONS

which was defined at an oL level of p > 0.05. The latter test was used when the ceil frequencies expected were very small. RESULTS Table 1 compares the MRI findings with the corresponding histologic changes. An 82.8% correspondence was found between the classification of degeneration at MRI and histology. The imaging studies showed 19 instances in which a meniscal degeneration seen at light microscopy was not detected. In 19 of 480 measurements the imaging studies suggested a meniscal degeneration, whereas the histologic examination confirmed an intact meniscus. Concerning the degree of the degeneration present in a total of 480 observations, MRI resulted in an overestimation of the degeneration in 53 cases, whereas in 38 instances the degeneration was underrated compared with light microscopy diagnosis. With regard to the correct detection of a degenerative lesion, MRI presented with an overall accuracy of 0.93 (Table 2). The imaging studies gave a sensitivity of 0.96 with a corresponding specificity of 0.79. Concerning the diagnosis of a meniscal degeneration, our studies yielded a positive predictive value of 0.95 and a negative predictive value of 0.82, representing an excellent to good level of diagnostic certainty both in patients who had a positive result and in those who had a negative result. In order to consider the influence implied by the prevalence of a meniscal degeneration in the material presented here, pretest and posttest probabilities were calculated on the basis of an 81% prevalence. A distinct difference between the pretest and posttest probabilities should be present in a test considered useful for screening. In our assessment, the pretest probability of a positive result was 0.81 and the posttest probability increased to 0.95. The pretest probability of a negative result was 0.19 compared with a posttest probability of 0.18. These T A B L E 1. Classification o f the degeneration at M R !

and histologic examination Histologic examination Stage Stage Stage Stage

0 1 2 3

MRI Grade 0

Grade 1

Grade 2

Grade 3

75 9 4 6

9 116 9 3

6 15 133 7

4 7 12 65

637

data indicate a high reliability in cases with a positive diagnosis. The imaging studies show considerable differences in the accuracy of the diagnosis related to the site and to the respective segments of the menisci (Fig. 3). Thus, the imaging diagnosis presented with a significantly reduced certainty concerning the lateral meniscus (p < 0.05). The topographical relationship of the different meniscal Segments to the plane of tomography gives rise to partial volume effects, leading to misinterpretations of the MR images. This phenomenon primarily affected the intermediate zone between the anterior and posterior horn of the meniscus. In this segment, diagnostic specificity was 0.71 for the medial and 0.63 for the lateral side, reflecting a significant discrepancy compared with an average of 0.84 in the other segments (p < 0.001). Sensitivity was reduced in the lateral intermediate zone with a mean value of 0.86, differing significantly from the 0.98 value in the anterior and posterior zones (p < 0.01). Table 3 reflects the diagnostic abilities of arthroscopy in the detection of meniscal degenerations. As to be expected, no close correlations were found among grade 1 and 2 abnormalities. However, in a grade 3 lesion, arthroscopy offered an increased diagnostic reliability with a total of 76 true-positive and five false-negative results, indicating an accuracy of 93.8%. Table 4 correlates the arthroscopic gradings with findings in MRI studies and shows a poor correlation in grade 1 and 2 lesions. DISCUSSION In our study we confirmed a close correlation between the intensity of intrameniscal signal at MRI and the prevalence of degenerative foci on the basis of a statistical confirmation. In contradistinction to studies published by other investigators (4,7) who found identical diagnostic qualities in the anterior, posterior, and intermediate segments of the meniscus, we demonstrated considerable differences in the diagnostic ability of MRI in the respective segments of the meniscus. In our experience, even a sophisticated technique of tomography using four imaging planes in combination with a widespread variation of TE and TR parameters could not compensate for this restriction. In a preliminary report Stoller et al. (8) made an attempt to define the pathologic correlations with the pattern of signal intensities produced by the menisci on MRI using a grading system proposed by Arthroscopy, VoL 10, No. 6, 1994

638

J. R A U N E S T E T AL. T A B L E 2. Diagnostic abilities o f M R I in the evaluation o f degenerative changes Predictive value

Test resultsa

Overall Medial meniscus Lateral meniscus

Pretest probability

Posttest probability

Accuracy

Sensitivity

Specificity

TP

FN

FP

TN

+

-

+

-

+

-

0.93 0.95 0.90

0.96 0.99 0.93

0.79 0.81 0.76

314 161 153

19 6 13

19 11 8

75 41 34

0.95 0.95 0.94

0.82 0.94 0.72

0.81 0.80 0.81

0.19 0.20 0.19

0.95 0.95 0.94

0.18 0.06 0.28

TP, true positive; FN, false negative; FP, false positive; TN, true negative.

Crues (5). Based on evaluation of 12 cadaver menisci, the investigators found a one-to-one correlation and concluded that MRI provides a high diagnostic certainty in the delineation of meniscal degenerations. H o w e v e r , the limited number of specimens did not allow for a statistical analysis. The histologic staging underlying our evaluation was determined using well-established criteria of degeneration on light microscopy preparations (6). Although these features comprise the typical morphology of degeneration, they cannot imply the diagnosis of a pathologic state per se in clinical practice. A strict distinction between a pathologic state of meniscal degeneration that implies a specific indication for conservative or operative treatment and morphological alterations, common with old age or extensive mechanical stress, is necessary to use this grading system as a basis for a specified therapy (9). In this respect, morphological changes corresponding to a stage 1 lesion are frequently seen in asymptomatic individuals and represent a response to mechanical stress and loading with increased

Medial Meniscus

production of mucopolysaccharide ground substance (10,1 I). In a stage 2 lesion a distinct fibrocartilagenous separation is observed, usually in the area of the middle perforating collagen bundles (12). As a response to recurrent chronic trauma, these fibers are preferred sites for the accumulation of mucopolysaccharide ground substance, which appears as a linear structure of increased signal intensity at MRI. These patients may be asymptomatic, but a stage 2 lesion must be considered as a precursor to frank tears (10,13). Therefore, these patients are at high risk for developing symptomatic meniscal tears (14-16). A separation of the meniscal structure together with a grade 3 lesion is not evident in arthroscopy provided that the tear does not extend to the articular surface of the meniscus. According to Smillie, such "closed" lesions are characteristic of the early symptomatic horizontal cleavage lesion (13). However, these intrasubstance abnormalities may be detected by probing at arthroscopy or arthrotomy, presenting with a reduction of mechani-

Lateral Meniscus

1 0.8 0.6FIG. 3. Diagnostic accuracy, sensitivity, and specificity of MRI applied to the segments of the medial and lateral meniscus.

0.4. 0.20 anterior intermediate posterior 1

Accuracy

Arthroscopy, Vol. 10, No. 6, 1994

~

anterior intermediate posterior Sensitivity

~

Specificity

MR IMAGING A N D ARTHROSCOPY I N MENISCAL DEGENERATIONS TABLE 3. Diagnostic abilities of arthroscopy in the

degenerative changes Arthroscopy

Histology

Normal meniscus

Pathological consistency at probing; changed color

Stage 0 Stage 1 Stage 2 Stage 3

71 125 96 5

23 22 62 ~

Ruptured meniscus; visible degeneration 0 0 0 76

"Visible degenerations were not rated in presence of ruptures.

cal integrity. With regard to the altered rheological characteristics, loading forces may lead to buckling with increased tension to the meniscocapsular insertion, thus becoming symptomatic (17,18). The normal meniscus substance produces no significant signal on either TI- or T2-weighted images, whereas degenerative changes lead to an increased signal intensity that is most conspicuous on T1weighted, intermediate-weighted, and gradientecho images (12). This intrameniscal signal may in part arise by damaged macromolecules absorbing water and increasing the local spin density. The rotation rates of the absorbed water molecules, which are slowed by interaction with organic macromolecules, cause a bright signal on short TE images (19,20). The decrease of the transitorial motion leads to less motional narrowing and a shorter T2 of the absorbed water so that its signal characteristics are emphasized on short TE images, comparing with water within macromolecules rather than free water of the joint space (12). Therefore, short TE sequences should be used as a standard for screening in clinical use. Some anatomic and pathologic conditions may lead to misinterpretations of the MR image. Thus, TABLE 4. Correlation o f grading at M R ! with findings

at diagnostic arthroscopy MRI grading Arthroscopy

Grade 0

Grade 1

Grade 2

Grade 3

Normal meniscus Pathologic consistency at probing; changed color Ruptured meniscus; visible degeneration

69

111

101

16

21

17

49

20

4

9

11

52

639

small areas of fibrillation or fraying on the free margin of the meniscus might be missed on 5-mm imaging acquisations. Areas of linear degeneration perpendicular to the plane of tomography could be undergraded to a stage 1 lesion. The topographic relationship of the lateral anterior horn to the transverse ligament in combination with a large branch of the lateral inferior genicular artery may be mistaken for a grade 3 signal intensity (21). Finally, the popliteus tendon sheath next to the posterior segment of the lateral meniscus might appear as a grade 3 vertical rupture (22). With regard to MRI, arthroscopy has limitations in the assessment of tears arising from the tibial articular surface in the peripheral rim of the meniscus (23,24). Prearthroscopic evaluation of the menisci with the help of MRI may therefore facilitate investigation in areas of degeneration and potential tears not visible to the endoscopist. Complementary to arthroscopic examination, MRI offers a valuable diagnostic potential providing reliable information about the internal condition of the meniscus, which might be especially useful in assessing the need for meniscal reconstruction and in the follow-up after reconstructive surgery. REFERENCES 1. Raunest J, Oberle K, L6hnert J, H6tzinger H. The clinical value of magnetic resonance imaging in the evaluation of meniscal disorders. J Bone Joint Surg [Am] 1991;73:11-6. 2. L6hnert J, Raunest J. Die Arthroskopie des Kniegelenkes-Eine Analyse aus 3500 Arthroskopien. Orthop Praxis 1986; 23:8-11. 3. Manco LG, Lozman J, Coleman ND, Kavanaugh JH, Bilfield BS, Dougherty G. Noninvasive evaluation of knee meniscal tears: preliminary comparison of MR imaging and CT. Radiology 1987;163:727-30. 4. Reicher MA, Hartzman S, Duckwiler GR, Bassett LW, Anderson LJ, Gold RH. Meniscal injuries: detection using MR imaging. Radiology 1986;159:753-7. 5. Crues JV, Mink J, Levy TL, Lotysch M, Stoller DW. Meniscal tears of the knee: accuracy of magnetic resonance imaging. Radiology 1987;164:445-8. 6. Copenhaver WM, Kelly DR, Wood RL. Bailey's textbook of histology. 17th ed. Baltimore, MD: Williams & Wilkins, 1978:170-8. 7. Reicher MA, Hartzman S, Bassett LW, Mandelbanm B, Duckwiler G, Gold RH. MR imaging of the knee. Part II. Traumatic disorders. Radiology 1987;162:547-51. 8. Stoller DW, Martin C, Crues JV, Kaplan L, Mink JH. Meniscal tears: pathologic correlation with MR imaging. Radiology 1987;163:731-35. 9. Holder J, Haghighi P, Pathria MN, Trudell D, Resnick D. Meniscal changes in elderly: correlation of MR imaging and histologic findings. Radiology 1992;184:221-5. 10. Ferrer-Roca O, Vialta C. Lesions of the meniscus. Part I:

Arthroscopy, Vol. 10, No. 6, 1994

640

11. 12. 13. 14. 15. 16. 17. 18.

J. R A U N E S T

macroscopic and histologic findings. Clin Orthop 1980;146: 301-7. Kornick J, Trefelner E, McCarthy S, Lange R, Lynch K, Jokl P. Meniscal abnormalities in the asymptomatic population at MR imaging. Radiology 1990;177:463-5. Mink JH, Reicher MA, Crues JV, Deutsch AL. MRI o f the knee. 2nd ed. New York: Raven, 1993. Smillie IS. Clinical features ofinternalderangements relative to the menisci. In: Injuries of the knee joint. 4th ed. London: E. & S. Livingstone, 1970:70-97. Dillon EH, Pope CF, Joki P, Lynch K. The clinical significance of stage 2 meniscai abnormalities on magnetic resonance knee images. Magn Res Imaging 1990;8:411-5. Kaplin PA, Nelson NL, Garvin KL, Brown DE. MR of the knee: the significance of high signal in the meniscus that does not clearly extend to the surface. A JR 1991;156:333-6. Reining JW, McDevitt ER, Ove PN. Progression of meniscal degenerative changes in college football players: evaluation with MR imaging. Radiology 1991;181:255-7. Crues JV, Ryu R. MRI of the knee: Part II. Surg Rounds Orthop 1989:27-33. Day B, MacKenzie WG, Shim SS, Leung G. The vascular

Arthroscopy, VoL 10, No. 6, 1994

ET AL.

19.

20.

21.

22.

23.

24.

and nerve supply of the human meniscus. Arthroscopy 1985; 1:58-62 FuUerton GD, Potter JL, Dornbluth NC. NMR relaxation of protons in tissue and other macromolecular water solutions. Magn Reson Imaging 1982;1:209-28. Koenig SH, Brown RD. The importance of motion of water for magnetic resonance imaging. Invest Radiol 1985;20:297305. Mink JH, Stoller DW, Martin C, Crues J. MR of the knee: pitfalls in interpretation (abstr.). In: Book of abstracts: Society of Magnetic Resonance in Medicine 1986. Vol. 4. Berkeley, Calif.: Society of Magnetic Resonance in Medicine, 1986:1195. Lfhnert J, Raunest J. Die topographie am lateralen meniskushinterhorn-eine arthroskopisch-anatomische studie. Unfallchir 1984;10:142-5. Ireland J, Trickey EL, Stoker DJ. Arthroscopy and arthrography of the knee: A critical review. J Bone Joint Surg 1980;62-B:3--6. Jackson RW, DeHaven KE. Arthroscopy of the knee. Clin Orthop 1975;107:87-92.