Year : 2004 | Volume
: 14 | Issue : 1 | Page : 33--40
MR Imaging of knee with arthroscopic correlation in twisting injuries
JP Singh, L Garg, R Shrimali, V Setia, V Gupta
Mohan Dai Oswal Cancer Hospital, G.T. Road, Sherpur by Pass, Ludhiana, India
J P Singh
82, New Mehar Singh Colony, Tripari, Patiala, Punjab
The purpose of this study is to study MRI appearances of meniscal and cruciate ligament injuries of knee. Know the role of different sequences in diagnosing these injuries. Compare the MRI findings with arthroscopic findings and appearances reported in literature. The objective is also to study limitations of MRI in detecting meniscal and cruciate ligament injuries of the knee. Materials and Methods: One hundred and seventy three patients with history of twisting injuries to the knee and suspected to have meniscal and/or cruciate ligamentous tears have been studies using Signa Contour (GE) MR machine with a superconducting magnet and field strength of 0.5 tesla using a Quadrature (QD) extremity coil. Sequences and planes used were spin echo T1, fast spin echo T2, Gradient Echo (GRE) and Short Tau Inversion Recovery (STIR) in direct coronal, sagittal and axial planes using 4mm slice thickness. Patient was placed supine with knee externally roated 15-20 degree minues and flexed 5-10 degree minues. Results: Sensitivity, Specificity and Accuracy of MRI in detecting meniscal and cruciate ligamentous injuries as compared to arthroscopy is as follows: Medial Meniscus (MM): 96.5 percent,98.28 percent, 97.69 percent; for Lateral Meniscus (LMP:87 percent, 99.29 percent, 97.11 percent, Anterior Cruciate Ligament (ACL):98.72 percent, 98.94 percent, 98.84 percent; and for Posterior Cruciate Ligament (PCL): 98.72 percent, 98.94 percent, 98.84 percent. Most commonly Medial Meniscus (57 patients:32 percent) and anterior cruciate ligament (78 patients:45 percent) injury was seen involving the posterior horn of Medial Meniscus (45 patients:79 percent), and midsubstance Anterior Cruciate Ligament (53 patients:68 percent) respectively. Conclusion: MRI is a very useful non-invasive diagnostic modality having high sensitivity, specificity and accuracy in the identification of meniscal and cruciate ligament injuuries. MRI should be done in very patient of suspected ligamentous injury, to be posted for arthroscopy thus preventing unwanted arthroscopies.
|How to cite this article:|
Singh J P, Garg L, Shrimali R, Setia V, Gupta V. MR Imaging of knee with arthroscopic correlation in twisting injuries.Indian J Radiol Imaging 2004;14:33-40
|How to cite this URL:|
Singh J P, Garg L, Shrimali R, Setia V, Gupta V. MR Imaging of knee with arthroscopic correlation in twisting injuries. Indian J Radiol Imaging [serial online] 2004 [cited 2020 Jul 12 ];14:33-40
Available from: http://www.ijri.org/text.asp?2004/14/1/33/28551
The most significant advances in knee imaging have been made in the realm of MR imaging, which has clearly emerged as a primary tool, to guide the management of knee pain. With the development of new sequences, improved SNR (signal to noise ratio), higher resolution, reduced artifacts, shorter imaging times and improved accuracy, MRI has changed the traditional algorithm for workup of knee joint pathology, particularly when meniscal and cruciate ligamentous tears in cases of twisting injuries to the knee is suspected. MRI has made it possible to look into the injured knee noninvasively thereby avoiding invasive procedures and further morbidity.
Materials and Methods
This was a prospective double blinded study. The senior radiologist was blind to clinical findings and the orthopedic surgeon was blinded to MRI diagnosis.
Inclusion criteria: All patients of twisting injuries to either or both knees who underwent both MR imaging and arthroscopy of the knee during a 5 yr period from January1998 to December 2002.
Exclusion criteria: Patients with neoplasms, inflammatory or infectious disorders or patients suffering significant injury to the knee between time of MRI and arthroscopy.
173 patients with knee complaints and suspected to have meniscal or ligament tears were studied using Signa Contour (GE) MR machine with a superconducting magnet and field strength of 0.5 T using a QD extremity coil. The patient population consisted of 113 males and 60 females,majority ranging in the third decade. We used SE, fast sequences such as GRE T2*, FSE T2 and STIR sequences. The three standard imaging planes used are the direct coronal, sagittal and axial views. We examined the knee in these three planes using a FOV of 16X16 cm, 256 X 256 matrix, and 3 mm slice thickness. An axial acquisition through patellofemoral joint is used as initial localizer for Subsequent sagittal and coronal plane images. Patient was placed in supine position with the knee in a closely coupled QD extremity coil. The knee is externally rotated 15-20°, in order to facilitate the visualization of ACL completely on sagittal images. The knee was flexed 5-10°.
MRI was performed within a time period of 6 days to 30 days from the date of injury. The time lag between MRI and arthroscopy was 1 day to 30 days with an average of 7 days.
Out of 173 patients, 60 were women (34.68%) and 113 men (65.31%). 93 (53.76%) had derangements of the right knee while 80 (46.24%) had left sided derangements. The patients age range from 17 yrs- 51 yrs (Mean age-30 yrs). 78 patients (45.08%) showed ACL tears, 57 patients (31.8%) showed MM tears, 28 patients (16.18%) showed M tears and 10 patients (5.78%) showed PCL tears [Tables 1 and 2].
Cruciate ligament tears:Out of 173 patients, 78 cases (45.08%) showed ACL tears. 10 patients (50.78%) showed PCL tears. Out of 78 patients of ACL tears, 53 patients (67.94%) had mid substance tears, 15 patients (19.23%) had tears at the femoral attachment and 10(12.8%) at the tibial attachment were detected, on MRI. In our study, the most common tear location was at mid-substance. On Arthroscopy, mid substance tears were seen in 52 patients (66.67%), femoral attachment tears in 15 cases (19.23%) and tibial attachment tears in 11 patients (14.1%) were detected. [Table 3], [Figure 1],[Figure 2],[Figure 3],[Figure 4]
In this study we found hyper intensity in the ligament as the most common sign which was seen in 52 patients (66.67%), 16(20.51%) patients showed discontinuity and 10(12.82%) patients showed non visualization of ACL. [Table 4]
Anterior cruciate ligament tears
Arthroscopically 78 tears were detected.
MRI findings showed:
True positive- 77 Sensitivity-98.72%
True negative-94 Specificity-98.94%
False positive-1 Positive predictive value-98.71%
False negative-1 Negative predictive value-98.94%
Meniscal tears : Out of 173 patients, 57 (38.23%) showed medial meniscal and 28 patients (29.41%) showed lateral meniscal tears on MRI. Only grade 3 tears were compared as grade1 and 2 tears do not reach upto the articular surface and hence are invisible to arthroscopic surface evaluation. Out of 57 patients of medial meniscal tears, most common tear location was at posterior horn. In our study, we found posterior horn tear in 45 patients (78.95%) and anterior horn tears in 12 patients (21.05%) on MRI.
On arthroscopy posterior horn tears were seen in 42 patients, anterior horn tears in 11 patients and middle one third tears in four patients. [Table 5] and [Table 6], [Figure 7],[Figure 8],[Figure 9],[Figure 10],[Figure11],[Figure 12],[Figure 13]
Out of 28 patients of lateral meniscus tears posterior horn tears was seen most commonly, in 25 patients (89.29%) and anterior horn tears were seen in three patients (10.71%). On Arthroscopy posterior horn tears were seen in 24 patients and anterior horn tears in 4 patients while middle one third tears were seen in three patients. [Table 5] and [Table 7],[Figure 14],[Figure 15],[Figure 16],[Figure 17]
Medial meniscal tears -
Arthroscopically 57 tears were identified.
MRI findings showed
True positive-55 Sensitivity-96.5%
True negative-114 Specificity-98.28%
False positive-2 Positive predictive value-96.49%
False negative-2 Negative predictive value-98.27%
Lateral meniscal tears -
Arthroscopically 31 tears were identified.
MRI findings showed
True positive-27 Sensitivity-87%
True negative-141 Specificity-99.29%
False positive-1 Positive predictive value-96.42%
False negative-4 Negative predictive value-97.24%
In this study of meniscal and cruciate ligament injuries of knee by MRI, we studied 173 patients, this included 113 men and 60 women, 93 right and 80 left knee joints. Majority of the patients were in their third decade.
Out of 173 patients 78 patients (45.09%) showed ACL tears, 57 patients (32.95%) showed MM tears, 28 patients (16.18%) showed LM tears and 10 patients (5.78%) showed PCL tears.
Cruciate ligament tears:
Out of 173 patients, 78 patients (45.08%) showed ACL tears. 10 patients (50.78%) showed PCL tears. Out of 78 patients of ACL tears, 53 patients (67.94%) had mid substance tears, 15 patients (19.23%) had tears at the femoral attachment and 10(12.8%) at the tibial attachment were detected, on MRI. In our study, the most common tear location was at mid-substance. On Arthroscopy, mid substance tears were seen in 52 patients (66.67%), femoral attachment tears in 15 patients (19.23%) and tibial attachment tears in 11 patients (14.1%) were detected. Berquist et al  in their study reported midsubstance tears as the most common type (90%).
In this study we found hyper intensity in the ligament as the most common sign which was seen in 52 patients (66.67%), 16(20.51%) cases showed discontinuity and 10 (12.82%) cases showed non visualization of ACL which are corresponding with Gentili et al .
In tears of the anterior cruciate ligament, the sensitivity specificity and accuracy was found to be 98%, 98% and 98% which were corresponding to Fischer et al study .
In our study the positive predictive value and negative predictive value was 98.7% and 98.9% respectively.
In an attempt to identify and explain the errors in interpretation, we undertook a critical review of our false positive and false negative findings at MRI examination.
There was one false positive case. He had both primary and secondary findings. The cause for the primary finding to be present was nonvisualization of the ACL and for the secondary findings is unclear; they may have reflected the physiologic ligamentous laxity.
There was 1 false negative case that had partial tear as seen on arthroscopy, which were misinterpreted as normal. MR imaging in this patient demonstrated a linear band of intact fibres normally oriented in expected location of the ACL. The results of two large studies [13,14] showed that MR imaging has relatively low sensitivity (40%-75%) but moderate to high specificity (62%-94%) in diagnosis of partial tears.
T2 weighted images showed clearly the signal intensity changes seen with these tears as excellent contrast is provided by normal low signal intensity of ligaments Mink et al .
10 patients of PCL tears were detected accurately both by MRI and Arthroscopy. The use of MRI to identify PCL tears has proven to be extremely accurate. This might be expected in light of the fact that the PCL is usually very easily visualized as a homogenous, continuous low-signal structure. Several studies have reported sensitivity, specificity, accuracy, positive predictive value and negative predictive value to be 99-100%. In our study too the sensitivity, specificity, accuracy, positive predictive value and negative predictive value was 100%.
Meniscal tears: Out of 173 patients, 57 (38.23%) showed medial meniscal and 28 cases (29.41%) showed lateral meniscal tears on MRI which is corresponding with La Prade and colleagues who reported that medial meniscal tears are twice as common as lateral meniscal tears.
Out of 57 patients of medial meniscal tears, most common tear location was at posterior horn. In our study, we found posterior horn tear in 45 patients (78.95%), anterior horn tears in 12 patients (21.05%) on MRI. On arthroscopy posterior horn tears were seen in 42 patients, anterior horn tears in 11 patients and middle one third tears in four patients.
Out of 28 patients of lateral meniscus tears posterior horn tears was seen most commonly, in 25 patients (89.29%) and anterior horn tears was seen in three patients (10.71%). On Arthroscopy posterior horn tears were seen in 24 patients and anterior horn tears in four patients while middle one third tears were seen in three patients.
Out of 173 patients, Gr III tear of medial meniscus was seen in 57 patients (32.95%), Gr II in 16 patients (9.25%), Gr I in 20 (11.56%) and normal in 80 patients (46.24%).
Lateral meniscus, Gr III tears were seen in 28 patients (16.18%), Gr II 12 patients (6.94%), Gr I in 14(8.1%) patients while 119 (68.78%) patients had normal lateral menisci. It shows that all the tears documented on arthroscopy were Gr III tears in accordance with Crues et al .
Silva and Silver  have studied the probability of a tear being identified at arthroscopy for each grade MRI signal. The probability of a tear with grade I signal is about 5%, grade II 17-20% and grade III 80%-95%.
In our study the sensitivity, specificity and accuracy for detecting medial meniscal tears was 96.5%, 98.28% and 97.69% respectively which were corresponding to the Fischer et al .
In the detection of Lateral Meniscal tears the sensitivity, specificity and accuracy was 87%, 99.29% and 97.11% respectively in accordance with the Fischer et al  study. The accuracy of MRI in identifying medial meniscal tear has been reported to range from 72%-98%, with sensitivity of 77%-97.5% and a specificity of 37%-100%.
The occurrence of the false positive meniscal tears at MRI imaging has been noted earlier. There are explanations for this apparent discrepancy between findings at MR Imaging and arthroscopy .
Misinterpretation of normal anatomy like Meniscofemoral ligaments etc.The presence of intrasubstance tears which are not seen on arthroscopy.The operator dependence of ArthroscopyThe presence of loose bodies
Out of the 57 Medial Meniscal tears, there were two false positive (3.51%) and two false negative (3.51%) observations.
In our study, the 2 false positive MRI involved the posterior horn of the meniscus. The exact cause of the false positives in detection of tears of the posterior horn of the medial meniscus in our study was not apparent. It may be attributed to the presence of intra-meniscal tears not communicating with the articular surface of the meniscus or misinterpretation of normal anatomy or operator/ procedure dependant drawback of arthroscopy; .
The degeneration and tears of menisci demonstrated as high signal intensity were due to imbibed synovial fluid. These tears are better demonstrated in short TE images like T1 PD and GRET2* images. This has been explained by Stoller et al in their study as- the interaction of synovial fluid with large macromolecules in menisci slows rotational rate of protons and shortens T1 and T2 values.
In our study, we found that the T2* weighted GRE images clearly depicted the tears than FSE images supported by Rubin et al .
Medial collateral ligament was found to be torn in two patients, which was also seen on MRI in both the patients.
Of the 28 patients with lateral meniscal tears who have been followed up by arthroscopy, one patient was falsely reported as positive on MRI which was probably due to pseudotear appearance of lateral meniscus caused by Meniscofemoral ligament. No new tears were found on arthroscopy. Gr I and II tears which are considered as precursors for formation of meniscal tears, are overlooked on arthroscopy. Raunest et al  reported that arthroscopic and arthrographic surface evaluations are insensitive to Gr I and II intrasubstance degenerative changes. We took a critical review of all the false positive meniscal tears reported on arthroscopy. However even on review study arthroscopic findings could not be correlated. Levinson et al  from their study found that the accuracy of arthroscopy in identification of inferior surface tears of posterior horn of Medial Meniscus is as low as 45-65%.
Bucket handle tears: In our study 4 patients showed bucket handle tears in medial meniscus which has been confirmed by subsequent arthroscopy. This is corresponding with study by Singson et al  who reported that medial meniscus bucket handle tears are more common than lateral meniscus. Of the four patients with bucket handle tears, in three patients displaced fragment is seen in intercondylar notch [Figure 9], one patient showed double PCL sign [Figure 9]  and one patient showed absent bowtie sign ). Watts et al  described that double PCL sign is 98% specific but 32 % sensitive and absent bowtie sign 62% specific and 71% sensitive. All four patients showed more mobility of medial meniscus in association with ACL tears.
Summary and Conclusion
The reports of this and several earlier series, that MRI is superior to arthroscopy in the diagnosis of meniscal and cruciate ligament tears has been substantiated by this study. MRI is a useful non-invasive modality having high sensitivity, specificity and accuracy in the diagnosis of meniscal and cruciate ligament injuries.
Majority of patients were in the age group of 30-40yr and this included 113 men and 60 women. Commonest lesion detected in our study was ACL tear and Medial meniscal tears.
In our study as in other studies FSE T2 and GRE T2* sequences were found to be most sensitive to the spectrum of meniscal and cruciate ligament tears while SE T1 sequence was useful to correlate pathology other than ligament tears. STIR sequences were useful to evaluate associated bone contusions.
MR is unique in its ability to evaluate the internal structure as well as the surface of the meniscus.
MR is advantageous in conditions where arthroscopy is not useful like peripheral meniscal tears and inferior surface tears.
MR is more sensitive in detection of multiple meniscal tears that may be overlooked on arthroscopy.
MR is more sensitive than arthroscopy in detection of grade I and II intrasubstance degeneration, precursors to formation of meniscal tears.
MR is less sensitive than arthroscopy in detecting partial ACL tears.
MR being noninvasive does not involve morbidity associated with arthroscopy.
Many anatomic variants can mimic tears on MRI.
MRI should be done in every patient of suspected ligamentous injury, to be posted for arthroscopy thus preventing unwanted arthroscopies.
|1||Berquist TH. Magnetic resonance techniques in musculoskeletal diseases. Rheum Clin North Am 1991; 17: 599-615. |
|2||Gentili A, Seeger LL, Yao L, DoHM. ACL tear:Indirect sign at MRI: Radiology 1994; 193: 835-840. |
|3||Fischer SP, Fox JM, Del Pizzo W et al . Accuracy of diagnosis from magnetic resonance imaging of the knee: a multi centre analysis of one thousand and fourteen patients. J Bone Joint Surg (AM) 1991; 73A: 2-10. |
|4||Mink JH, Levy T, Crues JV. Tears of the ACL and menisci of the knee: MR evaluation. Radiology 1988; 167: 769-774. |
|5||La Prade RF, Burnett QM, Veenstra MA, Hodgman CG. The prevalence of abnormal MRI findings in asymptomatic knees. Am J Sp Med 1994; 171: 761-766. |
|6||Crues JV, Mink JH, Levy T, Lotysch M, Stoller DW. Meniscal tears of the knee. Accuracy of MR imaging. Radiology 1987; 164: 445-448. |
|7||Rubin DA, Kneeland JB, Listerud J, Underberg E, Davis SJ. MR diagnosis of meniscal tears of the knee: value of FSE vs conv SE pulse sequences. AJR 1994; 162: 1131-1138. |
|8||Raunest J, Hotzinger H, Burrig KF. MRI and arthroscopy in detection of meniscal degenerations. Arthroscopy 1994; 10: 624-630. |
|9||Levinson ME, Baker BE. Prearthrotomy diagnostic evaluation of the knee: review of 100 cases, diagnosed by arthrography and arthroscopy. AJR 1980; 134: 107-114. |
|10||Singson RD, Feldman F, Staron R, Kiernan H. MR imaging of the displaced bucket handle tear of the medial meniscus. AJR 1991; 156: 121. |
|11||Helms CA, Laorr A, Cannon WD. The absent bowtie sign in bucket handle tears of the menisci in the knee. AJR 1998; 170(1), 57-61. |
|12||Watt S, Tonya H, Nigel R. The value of the absent Bowtie sign in MRI of Bucket handles tears. Clinical Radiology 2000; 55: 622-626. |
|13||Yao L, Gentili A, Lee K. Partial ACL rupture: an MR diagnosis? Skeletal Radiology 1995; 24: 247-251. |
|14||Umans H, Wimphfeimer O, Haramati N, et al . Diagnosis of partial tear of the anterior cruciate ligament of the knee: Value of MR imaging. AJR 1995; 165: 893-897. |
|15||Silva I, Silver DM. Tears of the meniscus revealed by MRI. J Bone Joint Surg (Am) 1988; 70A:199-202.|