- Open Access
- Open Peer Review
MRI efficacy in diagnosing internal lesions of the knee: a retrospective analysis
© Nikolaou et al; licensee BioMed Central Ltd. 2008
- Received: 23 February 2008
- Accepted: 02 June 2008
- Published: 02 June 2008
Many surgeons tend to believe that MRI is an accurate, non invasive diagnostic method, enough to lead to decisions for conservative treatment and save a patient from unnecessary arthroscopy. We conducted a retrospective study to investigate the accuracy of the MRI of the knee for the detection of injuries of the meniscus, cruciate ligaments and articular cartilage, in comparison with the preoperative clinical examination and intraoperative findings. Between May 2005 and February 2006 102 patients after clinical examination were diagnosed with meniscal or cruciate injury and underwent definitive treatment with arthroscopy. 46 of these patients fulfilled the inclusion criteria. The accuracy, sensitivity, specificity, negative and positive predictive values of the MRI findings were correlated with the lesions identified during arthroscopy. The diagnostic performance of the initial clinical examination was also calculated for the meniscal and cruciate ligament injuries.
The accuracy for tears of the medial, lateral meniscus, anterior and posterior cruciate ligaments and articular cartilage was 81%, 77%, 86%, 98% and 60% respectively. The specificity was 69%, 88%, 89%, 98% and 73% respectively. The positive predictive value was 83%, 81%, 90%, 75% and 53% respectively. Finally, the clinical examination had significant lower reliability in the detection of these injuries.
MRI is very helpful in diagnosing meniscal and cruciate ligament injuries. But in a countable percentage reports with false results and in chondral defects its importance is still vague. The arthroscopy still remains the gold standard for definitive diagnosis.
- Positive Predictive Value
- Cruciate Ligament
- Posterior Cruciate Ligament
- Lateral Meniscus
- Medial Meniscus
Arthroscopy is considered as "the gold standard" for diagnosis of traumatic intraarticular knee lesions . However, arthroscopy is an invasive procedure that requires hospitalization and anaesthesia, thus presenting all the potential complications of a surgical procedure . Since it's introduction in the 1980's Magnetic Resonance Imaging (MRI) has gained in popularity as a diagnostic tool of the musculoskeletal disorders . Especially the knee is the most frequent examined joint with MRI. Many surgeons tend to believe that MRI is an accurate, non invasive diagnostic method of the knee injuries, enough to lead to decisions for conservative treatment and save a patient from unnecessary arthroscopy. Nevertheless, even nowadays, remains very expensive. Taking in account that health-economics play important role in patients management, many questions arise regarding when and how often one must ask for an MRI when clinical examination has already confirm the diagnosis of meniscal tear or cruciate ligament rupture . The opposite question might be more important; is negative MRI enough to prevent unnecessary arthroscopy, when clinical examination suggests a meniscal or cruciate ligament injury?
With the purpose of investigating the accuracy of magnetic resonance imaging in patients with clinical signs of traumatic intraarticular knee lesions, we compared its findings with those obtained from the subsequent arthroscopies.
After obtaining the approval of the hospital ethics committee, we retrospectively reviewed the case notes of patients who had been clinically diagnosed with meniscal or cruciate injury, between May 2005 and February 2006 in our institution. Patients who had subsequently undergone further examination with MRI and were definitively treated with arthroscopy were then identified. We adhered to the Standards for Reporting of Diagnostic Accuracy (STARD) criteria for design and presentation of diagnostic studies 
Patients that in plain X-rays had fractures, loose bodies or signs of severe osteoarthritis were excluded from the study. Additionally, patients that after the MRI examination have had new injury to the same knee, before the arthroscopy or delayed to undergo arthroscopy for more than 3 months, were also excluded.
All patients had thorough clinical examination from two experienced knee surgeons prior to the MRI. Clinical examination focused on meniscal injury and cruciate ligament injury. The tests used in the clinical diagnosis were: the anterior-posterior drawer test, the Lachman test, the pivot shift test for the diagnosis of cruciate ligament injuries and the Apley's and McMurray's test for the meniscal injuries .
MRI examinations were performed in 2 different diagnostic centres. The MRI scanners were two 1.5 tesla units (Philips Medical Systems). T1 and T2 weighted images in coronal, axial and sagittal planes were obtained. Slice thickness ranged from 3 to 5 mm. The films were interpreted from 2 experienced knee radiologists who were aware of the result of the clinical examination as this was written at the initial referral letter. Any abnormalities of the cruciate ligaments, menisci or hyaline cartilage were described on a standard form. Preoperatively each MRI was also assessed by the surgeon performing the arthroscopy. In the case of different opinions between the two, the radiologists' diagnosis was considered more reliable.
All arthroscopies were performed by 2 experienced knee surgeons in a hospital environment with complete preoperative and postoperative care. A 4 mm Karl-Storz arthroscope with a 30-degree angle was used. Standard arthroscopic portals were used. ; the inferolateral portal for the arthroscope, and the inferomedial portal for the probe. Before any intervention, all knee interior structures were examined with the probe. Chondral defects were classified as positive if were more than 2nd grade according to the Outerbridge classification  and measured more than 1 cm in diameter. A cruciate ligament was considered to be torn if it was completely disrupted at one of its attachments to bone or in its substance, or if laxity (partial tear) could be demonstrated with a probe. All arthroscopic findings were photographed and registered. For further evaluation all arthroscopic findings were considered accurate and served as reference base.
MRI diagnoses and clinical findings were placed into one of four categories after arthroscopic evaluation. A result was considered a true-positive if the clinical or MRI diagnosis was confirmed by arthroscopic evaluation. A result was considered a true-negative if the diagnosis of no tears was confirmed by arthroscopy. A result was considered a false-positive if the arthroscopy was negative but the results were positive at the clinical examination or on the MRI. If the arthroscopy was positive but the clinical examination and MRI were negative, this was considered a false-negative result.
Statistical analysis was used to calculate sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV), in order to assess the realibility of the clinical and MRI results. 95% confidence intervals for sensitivity and specificity, as well as positive (LR+) and negative likelihood ratios (LR-) and areas under the ROC curve (AUC) were calculated.
Demographic baseline data for the patients that fulfilled the inclusion criteria (N = 46)
Demographic baseline data
Mean Age (Range)
32 (18 – 45)
Mechanism of injury (N)
Non sport injury
No history of injury
Mean delay from injury to MRI (weeks) (N = 34)
Mean delay from MRI to arthroscopy (weeks) (N = 46)
2.7 (0 – 12)
Total number of arthroscopic findings in patients that had previously examined with MRI
Medial meniscus tears
Lateral meniscus tears
Results of the data analysis: Accuracy, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), positive likelihood ratio (LR+), negative likelihood ratio (LR-) and area under the ROC curve (AUC) of magnetic resonance imaging to evaluate lesions of the medial meniscus, lateral, meniscus, anterior cruciate ligament, posterior cruciate ligament, and articular cartilage.
Medial meniscus tears
Lateral meniscus tears
Sensitivity (95% CI)
83% (63 – 93)
62% (30 – 81)
83% (60 – 94)
100% (31 – 100)
42% (21 – 66)
Specificity (95% CI)
69% (41 – 88)
88% (68 – 97)
89% (70 – 97)
98% (80 – 97)
73% (52 – 88)
Results of the data analysis for the clinical examination. Results were significantly inferior to MRI.
Medial meniscus tears
Lateral meniscus tears
Sensitivity (95% CI)
65% (44 – 82)
30% (13 – 54)
68% (46 – 84)
Specificity (95% CI)
50% (26 – 73)
75% (53 – 89)
77% (54 – 91)
The purpose of this study was to demonstrate the diagnostic value of MRI in diagnosing the presence or absence of the most common injuries of the knee; the meniscus tears, the cruciate ligament ruptures and the chondral defects.
There are studies that support the view that the diagnostic accuracy of the MRI could affect in a critical way the treatment pathway of knee injuries. McKenzie et al  have studied 332 patients' diagnosis before and after MRI. The diagnosis was initially based upon the clinical examination and the therapeutic procedure was decided before MRI. 57 from 113 clinically positive before MRI meniscal tears were not confirmed with MRI. This result leaded to revaluation and differentiation of treatment in 62% of the patients. From those patients programmed for surgery only 38% finally underwent arthroscopy. In another study, Weinstabl et al  randomly distributed patients with positive meniscus rupture tests in two groups. All the patients of the first group had MRI examination before arthroscopy. In this group only 2% of patients didn't have positive findings during arthroscopy. Second group patients underwent arthroscopy, based only to the findings of the clinical examination. In this group, only in 30% of patients arthroscopy confirmed the findings of clinical examination.
However, in our study, MRI showed false results in significant proportion. For example as far as medial meniscus concerns there were 5 false positive and 5 false negative diagnosis whereas for lateral meniscus there were 8 false positive and 3 false negative diagnosis (PPV 83% and 81%, NPV 69% and 74% for medial and lateral meniscus tears respectively). As far as the chondral lesions concerns the MRI results were even more inferior with PPV and NPV reaching 53% and 63% respectively.
The posterior cruciate ligament can be examined very well with MRI. Bibliography refers accuracy in ruptures higher than 90% [19, 20, 1]. In our study we evaluated only 3 PCL ruptures and all were identified by MRI. At the same time, one false positive result occurred (accuracy 98%, sensitivity 100%, specificity 98%). Even though our results agree with the bibliography data, the number of cases is too small for statistical significant conclusions. However, surgeons must always bear in mind that PCL is difficult to investigate during arthroscopy because of its anatomic position, and many times there are arthroscopic false negative results.
Other authors however, like Heron et al , have shown that MRI can satisfactory reveal the 2nd and 3rd grade chondral defects as well as damages at the patellar articular cartilage, but is not accurate for smaller injuries like fibrilization or small fissuring in articular hyaline cartilage. Similar results were reported from Ochi et al  who showed that the sensitivity of MRI increased (from 40% to 71%) when MRI reading was done retrospectively, after the arthroscopic findings were registered. Especially, in chondral lesions with full thickness loss of cartilage and large-deep erosions the retrospectively calculated MRI sensitivity was 100% and 75% respectively. On the other hand site surface injuries, fibrillization or shallow small cuts were not well described, not even post-arthroscopically. Furthermore, according to Mori et al , usage of modern, improved techniques, can not only reveal the size of chondral lesions but to distinguish partial from full depth chondral damages as well.
There is no doubt that the radiologist's experience and training are very important factors in interpretation of MRI. At the same time reliable statistical data of the diagnostic value of MRI are also related with the independent base of reference. Regarding knee MRI, in most of the studies and in our study as well, the base of reference is arthroscopy. This presupposes that arthroscopy is 100% accurate and allows for the diagnosis of every possible knee pathology. This is not always the case [25, 16] ; arthroscopy is a technical demanding procedure and the results are varying according to surgeons' experience, especially in difficult cases. From the 13 false positive results of our study, the majority referred to posterior meniscus tear. Nevertheless the belief is that, even in these cases, the meniscus pathology existed but failed to be discovered during arthroscopy [12, 26]. Especially the inferior surface of posterior aspect of the medial meniscus is difficult to be reached with a probe and often rupture at that point can be missed. Nowadays, the overall accuracy of arthroscopy varies between 70–100%, depending on the surgeons' experience [16, 27–29]. This fluctuation inevitably raises questions, regarding the reliability of the MRI results classification on true or false .
In the everyday practice, based on clinical examination that comes first, surgeons decide whether must proceed to further laboratory tests, MRI, conservative or surgical treatment. But how precise can clinical examination be? There seems to be disagreement regarding the answer to this question. Investigations support that the accuracy of clinical examination compared with arthroscopic findings ranges between 64–85% [31, 32]. Rose et al  found that clinical examination is as accurate as MRI in diagnosing meniscal tears and ACL ruptures, so they concluded that MRI because of its high cost is not necessary in patients with clinical suspicion of meniscus and cruciate ligament tears. Similar conclusion was reported by Boden et al  who supported that when clinical examination sets the diagnosis of meniscus damage, MRI will not change treatment decisions.
On the other hand, Ruwe et at  reported that preoperative MRI can prevent unnecessary arthroscopy in 50% of the patients, so is of great value and must be done preoperatively. Boeree et al  believe that clinical examination is of minor significance with sensitivity in diagnosing medial meniscus, lateral meniscus and ACL tear of 67%, 48% and 55% respectively. Similar conclusions were reported by Jackson et al  who concluded that negative MRI for meniscus or cruciate ligament tears can discourage diagnostic arthroscopy even if clinical examination is positive for injury. The results of our study come in agreement with these studies, confirming a quite low diagnostic performance of the clinical examination (Table 4).
In summary, from our results, the accuracy of MRI in medial and lateral meniscus tear was 81% and 77% respectively, whilst for ACL and PCL rupture was 86% and 98% respectively. In the existing bibliography the accuracy of MRI reaches 90% in medial meniscus and ACL injuries, is lesser in lateral meniscus injury and slightly higher in PCL injuries [19, 26, 20, 37, 1, 38]. Most of the studies agree that MRI has low accuracy and sensitivity as far as chondral defects concerns [12, 23, 39]. The same has been shown in the current study, with the accuracy to be only 60% and the sensitivity and specificity 42% and 73% respectively.
In conclusion, the present study supports that MRI is very helpful in diagnosing meniscal and cruciate ligament injuries. But in a countable percentage reports with false results and in chondral defects its importance is still vague. Nowadays patients' expectations are maximal and taking in account that MRI false or misleading results can be as high as 20% to 30% in specific knee pathologies it is concluded that arthroscopy still remains the gold standard in diagnosing the internal knee lesions. Undoubtedly new techniques and more powerful tomographers will improve MRI's accuracy leading to better diagnostic equipment in knee injuries.
In any case, what one must always have in mind is that diagnosis alone is no the end point of the treatment and does not solve the problem. It is the beginning of new thoughts and actions one must follow to achieve accurate prognosis and correct treatment. In order to plan and apply the correct treatment pathways, the most important is not statistics or cost effectiveness data. Clinical experience and adequacy of the surgeon always have the greatest value, when it comes to the assurance of the patient optimal treatment.
Funding was neither sought nor obtained.
- Fischer SP, Fox JM, Del Pizzo W, Friedman MJ, Snyder SJ, Ferkel RD: Accuracy of diagnoses from magnetic resonance imaging of the knee. A multi-center analysis of one thousand and fourteen patients. J Bone Joint Surg Am. 1991, 73 (1): 2-10.PubMedGoogle Scholar
- Sherman OH, Fox JM, Snyder SJ, Del Pizzo W, Friedman MJ, Ferkel RD, Lawley MJ: Arthroscopy--"no-problem surgery". An analysis of complications in two thousand six hundred and forty cases. J Bone Joint Surg Am. 1986, 68 (2): 256-265.PubMedGoogle Scholar
- Lee JK, Yao L, Phelps CT, Wirth CR, Czajka J, Lozman J: Anterior cruciate ligament tears: MR imaging compared with arthroscopy and clinical tests. Radiology. 1988, 166 (3): 861-864.View ArticlePubMedGoogle Scholar
- Gelb HJ, Glasgow SG, Sapega AA, Torg JS: Magnetic resonance imaging of knee disorders. Clinical value and cost-effectiveness in a sports medicine practice. Am J Sports Med. 1996, 24 (1): 99-103. 10.1177/036354659602400118.View ArticlePubMedGoogle Scholar
- Bossuyt PM, Reitsma JB, Bruns DE, Gatsonis CA, Glasziou PP, Irwig LM, Lijmer JG, Moher D, Rennie D, de Vet HC: Towards complete and accurate reporting of studies of diagnostic accuracy: the STARD initiative. BMJ. 2003, 326 (7379): 41-44. 10.1136/bmj.326.7379.41. 2003/01/04View ArticlePubMedPubMed CentralGoogle Scholar
- Benjaminse A, Gokeler A, van der Schans CP: Clinical diagnosis of an anterior cruciate ligament rupture: a meta-analysis. J Orthop Sports Phys Ther. 2006, 36 (5): 267-288.View ArticlePubMedGoogle Scholar
- Outerbridge RE: The etiology of chondromalacia patellae. J Bone Joint Surg Br. 1961, 43-B: 752-757.PubMedGoogle Scholar
- Mackenzie R, Dixon AK, Keene GS, Hollingworth W, Lomas DJ, Villar RN: Magnetic resonance imaging of the knee: assessment of effectiveness. Clin Radiol. 1996, 51 (4): 245-250. 10.1016/S0009-9260(96)80340-0.View ArticlePubMedGoogle Scholar
- Weinstabl R, Muellner T, Vecsei V, Kainberger F, Kramer M: Economic considerations for the diagnosis and therapy of meniscal lesions: can magnetic resonance imaging help reduce the expense?. World J Surg. 1997, 21 (4): 363-368. 10.1007/PL00012254.View ArticlePubMedGoogle Scholar
- Dixon AK: Magnetic resonance imaging of meniscal tears of the knee. J Bone Joint Surg Br. 1996, 78 (2): 174-176.PubMedGoogle Scholar
- Tyrrell RL, Gluckert K, Pathria M, Modic MT: Fast three-dimensional MR imaging of the knee: comparison with arthroscopy. Radiology. 1988, 166 (3): 865-872.View ArticlePubMedGoogle Scholar
- Spiers AS, Meagher T, Ostlere SJ, Wilson DJ, Dodd CA: Can MRI of the knee affect arthroscopic practice? A prospective study of 58 patients. J Bone Joint Surg Br. 1993, 75 (1): 49-52.PubMedGoogle Scholar
- Herman LJ, Beltran J: Pitfalls in MR imaging of the knee. Radiology. 1988, 167 (3): 775-781.View ArticlePubMedGoogle Scholar
- Watanabe AT, Carter BC, Teitelbaum GP, Bradley WG: Common pitfalls in magnetic resonance imaging of the knee. J Bone Joint Surg Am. 1989, 71 (6): 857-862.PubMedGoogle Scholar
- Mackenzie R, Palmer CR, Lomas DJ, Dixon AK: Magnetic resonance imaging of the knee: diagnostic performance studies. Clin Radiol. 1996, 51 (4): 251-257. 10.1016/S0009-9260(96)80341-2.View ArticlePubMedGoogle Scholar
- Ireland J, Trickey EL, Stoker DJ: Arthroscopy and arthrography of the knee: a critical review. J Bone Joint Surg Br. 1980, 62-B (1): 3-6.PubMedGoogle Scholar
- Hodler J, Haghighi P, Trudell D, Resnick D: The cruciate ligaments of the knee: correlation between MR appearance and gross and histologic findings in cadaveric specimens. AJR Am J Roentgenol. 1992, 159 (2): 357-360.View ArticlePubMedGoogle Scholar
- Rose NE, Gold SM: A comparison of accuracy between clinical examination and magnetic resonance imaging in the diagnosis of meniscal and anterior cruciate ligament tears. Arthroscopy. 1996, 12 (4): 398-405.View ArticlePubMedGoogle Scholar
- Bassett LW, Grover JS, Seeger LL: Magnetic resonance imaging of knee trauma. Skeletal Radiol. 1990, 19 (6): 401-405. 10.1007/BF00241788.View ArticlePubMedGoogle Scholar
- Trieshmann HW, Mosure JC: The impact of magnetic resonance imaging of the knee on surgical decision making. Arthroscopy. 1996, 12 (5): 550-555.View ArticlePubMedGoogle Scholar
- Speer KP, Spritzer CE, Goldner JL, Garrett WE: Magnetic resonance imaging of traumatic knee articular cartilage injuries. Am J Sports Med. 1991, 19 (4): 396-402. 10.1177/036354659101900414.View ArticlePubMedGoogle Scholar
- Mori R, Ochi M, Sakai Y, Adachi N, Uchio Y: Clinical significance of magnetic resonance imaging (MRI) for focal chondral lesions. Magn Reson Imaging. 1999, 17 (8): 1135-1140. 10.1016/S0730-725X(99)00033-8.View ArticlePubMedGoogle Scholar
- Ochi M, Sumen Y, Kanda T, Ikuta Y, Itoh K: The diagnostic value and limitation of magnetic resonance imaging on chondral lesions in the knee joint. Arthroscopy. 1994, 10 (2): 176-183.View ArticlePubMedGoogle Scholar
- Heron CW, Calvert PT: Three-dimensional gradient-echo MR imaging of the knee: comparison with arthroscopy in 100 patients. Radiology. 1992, 183 (3): 839-844.View ArticlePubMedGoogle Scholar
- Quinn SF, Brown TF: Meniscal tears diagnosed with MR imaging versus arthroscopy: how reliable a standard is arthroscopy?. Radiology. 1991, 181 (3): 843-847.View ArticlePubMedGoogle Scholar
- Crues JV, Mink J, Levy TL, Lotysch M, Stoller DW: Meniscal tears of the knee: accuracy of MR imaging. Radiology. 1987, 164 (2): 445-448.View ArticlePubMedGoogle Scholar
- Dandy DJ, Jackson RW: The diagnosis of problems after meniscectomy. J Bone Joint Surg Br. 1975, 57 (3): 349-352.PubMedGoogle Scholar
- Jackson RW, Abe I: The role of arthroscopy in the management of disorders of the knee. An analysis of 200 consecutive examinations. J Bone Joint Surg Br. 1972, 54 (2): 310-322.PubMedGoogle Scholar
- Boeve BF, Davidson RA, Staab EV: Magnetic resonance imaging in the evaluation of knee injuries. South Med J. 1991, 84 (9): 1123-1127. 10.1097/00007611-199109000-00016.View ArticlePubMedGoogle Scholar
- Mackenzie R, Keene GS, Lomas DJ, Dixon AK: Errors at knee magnetic resonance imaging: true or false?. Br J Radiol. 1995, 68 (814): 1045-1051.View ArticlePubMedGoogle Scholar
- Gillies H, Seligson D: Precision in the diagnosis of meniscal lesions: a comparison of clinical evaluation, arthrography, and arthroscopy. J Bone Joint Surg Am. 1979, 61 (3): 343-346.PubMedGoogle Scholar
- Simonsen O, Jensen J, Mouritsen P, Lauritzen J: The accuracy of clinical examination of injury of the knee joint. Injury. 1984, 16 (2): 96-101. 10.1016/S0020-1383(84)80007-8.View ArticlePubMedGoogle Scholar
- Boden SD, Labropoulos PA, Vailas JC: MR scanning of the acutely injured knee: sensitive, but is it cost effective?. Arthroscopy. 1990, 6 (4): 306-310.View ArticlePubMedGoogle Scholar
- Ruwe PA, Wright J, Randall RL, Lynch JK, Jokl P, McCarthy S: Can MR imaging effectively replace diagnostic arthroscopy?. Radiology. 1992, 183 (2): 335-339.View ArticlePubMedGoogle Scholar
- Boeree NR, Watkinson AF, Ackroyd CE, Johnson C: Magnetic resonance imaging of meniscal and cruciate injuries of the knee. J Bone Joint Surg Br. 1991, 73 (3): 452-457.PubMedGoogle Scholar
- Jackson DW, Jennings LD, Maywood RM, Berger PE: Magnetic resonance imaging of the knee. Am J Sports Med. 1988, 16 (1): 29-38. 10.1177/036354658801600105.View ArticlePubMedGoogle Scholar
- Barry KP, Mesgarzadeh M, Triolo J, Moyer R, Tehranzadeh J, Bonakdarpour A: Accuracy of MRI patterns in evaluating anterior cruciate ligament tears. Skeletal Radiol. 1996, 25 (4): 365-370. 10.1007/s002560050096.View ArticlePubMedGoogle Scholar
- Justice WW, Quinn SF: Error patterns in the MR imaging evaluation of menisci of the knee. Radiology. 1995, 196 (3): 617-621.View ArticlePubMedGoogle Scholar
- Munk B, Madsen F, Lundorf E, Staunstrup H, Schmidt SA, Bolvig L, Hellfritzsch MB, Jensen J: Clinical magnetic resonance imaging and arthroscopic findings in knees: a comparative prospective study of meniscus anterior cruciate ligament and cartilage lesions. Arthroscopy. 1998, 14 (2): 171-175.View ArticlePubMedGoogle Scholar
- Oei EH, Nikken JJ, Verstijnen AC, Ginai AZ, Myriam Hunink MG: MR imaging of the menisci and cruciate ligaments: a systematic review. Radiology. 2003, 226 (3): 837-848. 10.1148/radiol.2263011892. 2003/02/26View ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.