The role of ultrasound in the diagnosis of meniscal tears and degeneration compared to MRI and arthroscopy

Objectives: To investigate the role of Ultrasound (US) in the diagnosis of meniscus degeneration and tears as compared to Magnetic resonance imaging (MRI) and arthroscopy. Methods: Thirty-five patients were included in the study. Patients were divided into 2 groups according to their ages (≤ 35 years and > 35 years). After MRI and US, 22 of the patients were also evaluated with arthroscopy. Comparisons were made using kappa agreement test and chi-square and Mc-Nemar tests where appropriate. Results: US exam was not as efficient as MRI in detection of degeneration. For tears, the sensitivity and specificity of US were 90.9% and 63.6% and for MR, 93.3% and 100% respectively. Medial meniscus posterior horn was the best visualized portion by US while visualization of the main bodies was limited. In the young group, the sensitivity and specificity of US were 80% and 100% compared to 66.7% and 75% in the elder group. Conclusions: US is not a suitable substitute of MRI in the routine diagnostic evaluation of meniscus lesions. However, in selected cases such as young patients, traumatic cases and cases with a contraindication for MRI, US may find a role as a quick exam to stratify patients for further evaluation.


Introduction
The knee is one of the most active joints in humans and is very prone to sports injuries.The most common meniscal pathologies are tears and degenerative changes.The diagnostic accuracy of magnetic resonance imaging (MRI) is similar to arthroscopy therefore; MRI is accepted as the gold standard technique for evaluation the internal derangements of menisci including tears and degeneration (1,2).However MRI is not always available on demand; does not allow dynamic testing and is a rather lengthy and expensive imaging modality.Ultrasound (US) on the other hand is an inexpensive, widely available and non-invasive technique which also allows dynamic imaging but there are concerns regarding the diagnostic accuracy (2,3).
In this study our aim was to compare US with MRI in detection of meniscal tears and degeneration.The US and MRI exams of patients who have been referred to radiology department because of clinically suspected meniscal pathology were compared with arthroscopic surgery findings to determine the accuracy of both imaging modalities.The study has been tailored so that the medial and the lateral menisci and the anterior, posterior and body portions of the menisci were examined separately by each modality and also the patients were grouped according to age.In this way, we wanted to assess the contribution of the meniscus anatomy and patient age to the power and the shortcomings of the US exam.

Patient selection
All the patients involved in the study were referred to our department after a clinical exam by an orthopedist who presumed presence of meniscal pathologies of either tear or degeneration.All participants were informed about the study and gave their informed consent.Institutional board review was also obtained.A total of 35 patients who had knee pain and locking were enrolled in the study.All of them had MRI and sonographic exams.Twentytwo patients had undergone arthroscopic surgery as well.
All the US and MRI exams were performed by radiologists who were blinded to the results of the other test.MRI and sonographic exams were usually performed on the same day or a few days a part.MRI was the first exam to be performed.It was decided to

Ünlü EN et al.
the arthroscopy according to the radiological and clinical findings by orthopedic surgeons.The medial and lateral menisci as anterior horn, posterior horn and body were evaluated for the presence of tears and degeneration.Later the groups were divided according to age as ≤ 35 years and > 35 years and the results of the studies were compared.

Ultrasound imaging technique
Sonographic exams (Toshiba Aplio SSA-774/80) were performed with 7.5 to 12 MHz probes in prone and supine positions through the anterior, lateral and posterior approaches using static and dynamic techniques.Most of the imaging was done in the longitudinal plane.In the static technique, the anterior horns of the menisci were imaged in supine position with the knee in full extension and 30−45 degrees of flexion.Then the probe was moved to sides to visualize the bodies of the menisci.Afterwards the patients were placed prone and posterior horns were imaged with the knee in extension first then 45 degrees flexion.In dynamic imaging, the knee was subjected to mild internal and external varus stresses to allow better imaging of the menisci using movements.

Ultrasound interpretation
Normal: A triangular, homogeneously echogenic meniscus with no internal heterogeneous echo changes or differences (Figure 1a,b).

Meniscus degeneration:
Loss of homogeneous internal echo structure, linear or nodular hypoechoic/echogenic areas which do not involve an articular surface.
Meniscus rupture: Linear or echogenic clefts extending to the articular margins of the meniscus or through the meniscus, sudden changes in the meniscal contours, blunting of the medial surfaces.

Magnetic resonance imaging technique
The MRI studies were performed using a superconductive magnet of 1.5 Tesla (Signa, GE Medical Systems, Milwaukee, Wisconsin).Dedicated knee coils were used in all studies and the patients were placed in supine position with the knee in extension in all exams.The imaging parameters and sequences used in the MR exam were shown in Table 1.

MRI interpretation
The signal intensity changes of the anterior horns, posterior horns and the bodies of the menisci were assessed using a grading system to detect degeneration and tears.The grading system was as follows: Grade 1: Nodular or punctate signal changes within the menisci which do not reach the articular surfaces.
Grade 2: Linear signal changes which do not reach the articular surface   Arthroscopic studies were performed by orthopedic surgeons with the aim of diagnosis and treatment.The detection of tears was made by visual inspection of the surgeons.

Statistical analysis
MRI was regarded as the gold standard exam for degeneration and only the MRI and US findings were compared statistically for meniscal degeneration because no information regarding the inner structure of the menisci was obtained with arthroscopy.κ coefficient was used to measure the concordance between the studies.
Arthroscopic studies were regarded as the gold standard for meniscal tears and sensitivity and specificity of US and MRI were computed accordingly using McNemar test to prove differences in diagnostic values.A P value < 0.05 was considered significant.The agreement between studies including US vs. MRI, US vs. arthroscopy, MRI vs. arthroscopy was all calculated using the κ coefficient.A rating scale for κ values suggests the following correlations for various κ values: κ values between 1 and 0.75 indicated good agreement while values between 0.5−0.75indicated moderate agreement.κ < 0.05 indicated poor agreement.The agreement between the studies for detection of degeneration using the κ test was less than 0.5 (P = 0.123) and two studies were found statistically in poor agreement.Moderate agreement was present only for the lateral meniscus anterior horn (LA) ( κ: 0.5−0.75,P = 0.0017) while for others (medial meniscus anterior horn (MA), medial meniscus posterior horn (MP) and lateral meniscus posterior horn (LP)) κ value was less than 0.5 suggestive of poor agreement (Table 2).
Among the studied total of 35 patients, a total of 24 meniscal tears in 22 patients were detected by US and a total of 45 tears in 27 patients were detected by MRI.Of those 22 patients who had undergone arthroscopy, a total of 21 tears in 14 patients were detected.The distribution of tears according to the meniscal regions and number of patients as detected by US, MR and arthroscopy were shown in Table 3.
Most of the tears were located in the MP region (Figures 2a,2b).
The sensitivity and specificity of US in detection of meniscal tears were 90.9% and 63.6% respectively.A tear detected by US was not seen on arthroscopy (9.1% false positivity) and 4 tears were missed on US when compared to arthroscopy (36.4% false negativity).
The sensitivity and specificity of MRI to detect a meniscal tear were 93.3% and 100% respectively.One tear reported by MRI was not confirmed by arthroscopy (6.7% false positivity) and no false negative cases were detected (Figures 3a, 3b and Table 4).The sensitivity and the specificity of the US and MRI exams and the agreement between the studies regarding the regions of the menisci were shown in Table 5. US performed best in the posterior horns of medial and lateral menisci.The overall sensitivity of US was best at the LP (75%), and specificity was best at the MP (100%).There was moderate agreement between US -MRI, and US -arthroscopy for these regions (κ = 0.5−0.75,P = 0.001), whereas for medial meniscus anterior horn the correlation was quite poor (κ < 0.5).The performance of MRI was relatively poor in the medial meniscus anterior horn as compared with the posterior horns and there was moderate agreement between MRI and arthroscopy for this region (κ > 0.5, p = 0.001) as compared to a much better correlation (κ > 0.5, P = 0.001) in other parts of the menisci (Table 5).
In one case an intrameniscal cyst was detected in US along with a meniscal tear which was confirmed with MRI study as well.To evaluate the contribution of age, the patients were divided into two groups, > 35 years and ≤ 35 years.
The sensitivity and specificity of US and MRI exams and the level of agreement between studies were all shown in Table 4.While the level of agreement between US vs MRI (κ = 0.5−0.75P = 0.005), and US vs arthroscopy (κ > 0.75, P = 0.016) were high in young patients, the agreement decreased as patients grew older (> 35 years); (κ < 0.5, P = 0.17 for US-MRI and κ = 0.5−0.75,P = 0.164 for US−arthroscopy).Likely higher sensitivity and specificity results were reported for US and MRI in the younger age group as compared with the older age group (Table 6).
The sensitivity and specificity of US in the young age group was 80% and 100% respectively whereas 66.7% and 75% for the older age group.There was only one false negative case in the younger age group with US and no false positives; whereas in the older age group, three false negative cases (33.3%) and one false positive case (14.3%) was detected.The sensitivity and specificity of MRI in the young age group was 100% with no false positive or negative cases.In the elder group the sensitivity was 100%, whereas the specificity was 75%.No false negative cases were present despite one false positive case (Table 6).

Discussion
Meniscal lesions are a major cause of knee pain and have adverse effects on the proper functioning of the knee joint.Tears and degenerations constitute the majority of meniscal lesions and correct diagnosis is important because surgery or arthroscopy is indicated in evidence of a tear.Studies had shown that clinical examination alone does not provide sufficient data to indicate arthroscopic examinations (4).Therefore imaging studies are almost always needed.Arthroscopy is the gold standard exam for meniscal lesions which can be used both for diagnosis and for treatment at the same time, however invasiveness is a disadvantage and also peripheral tears and tears located at the inferior part of MP may be missed if the study was not carefully and throughly conducted (5).
The most widely used exam for detection of meniscal pathology is MRI.The diagnostic accuracy of MRI have been widely studied in the literature and the sensitivity and specificity of MRI in detection of meniscal tears are both reported to be above 90% (6,7).The sensitivity of MRI in our study was 93.3% and the specificity 100%.The multiplanar imaging capability, high soft tissue resolution without bony artifacts and accurate evaluation of other components of the knee joint are the major advantages of MRI however the exam is expensive, not always readily available and has specific contraindications (claustrophobia, pacemakers, neurostimulator devices, implants, clips and foreign bodies) (7).
Recently interest has grown whether US can be used as an alternative exam to MRI in meniscal imaging.US, in contrast to MRI is a widely available and a low cost exam and has an increasing role in the musculoskeletal applications as well (8,9).However one main disadvantage of US is its being widely operator dependent and image degradation related to technical factors or artifacts from bony and soft tissue (3).In this study we wanted to determine whether US can be used as an alternative to MRI for detection of tears and degeneration.MRI is regarded as the gold standard in the evaluation of degeneration and is superior to arthroscopy (10).In our study, the agreement between MRI and US regarding degeneration was statistically poor (κ < 0.5, P = 0.123) and more degenerative lesions were detected by MR than US.MRI showed a definite superiority as compared with US in detection of degeneration.US was unable to demonstrate the echo changes from some parts of the menisci especially the bodies but on MRI, the meniscus was evaluated as a whole.Some authors suggest that it is possible to see all the meniscal parts with US and high sensitivity and specificity values are reported for detection of meniscal lesions including degeneration (4,11,12).Our results suggest that US is not a suitable test for detection of meniscal degeneration and performs poorly in differentiation of tears from degeneration.
Sometimes with MRI it is hard to determine whether a grade 2 signal change represents a degeneration or a tear.Fisher et al. in their series of 1914 patients had detected tears with arthroscopy in 17% of patients with grade 2 degeneration (13).None of our patients with grade 2 signal change had tear in arthroscopy.
Studies related to the use of ultrasound for detection of meniscal tears date back to 1980's (11,12,(14)(15)(16).For a non-invasive test to substitute arthrography and arthroscopy, high accuracy values are needed and the meniscus should be visualized in full without artifacts (11).US was evaluated in this regard however controversial results were obtained (3).In a limited number of in vitro studies performed on cadavers, most cited, the study by Richter et al. and the study by Riedl et al., US was regarded as a successful diagnostic tool (15,16).However both had reported that US was limited in differentiating types of tears, and especially problematic in detection of radial and oblique tears.In a later study by DeMaesseneer et al., US was also noted to be insensitive to depicting meniscocapsular separation (17).Failure to detect bucket handle tears which had dislocated to the intercondylar notch was related to limited beam penetration in US (18).Likewise in our study we were unable to detect three bucket handle tears by US which were delineated by MRI and arthroscopy.
In this study full demonstration of echo signals from certain parts of the menisci especially the bodies were not possible and better sensitivities were achived in the posterior horns as compared with the anterior.Grobbelaar et al., reported that US was very limited in the evaluation of intraarticular structures (19).Lee and Bouffard stated that if correct positioning and appropriate transducers were selected, most of the meniscal pathologies could be recognized by US (20).MP was the best visualized region in our study and the sensitivity was 63.6 and specificity was 100%, for LP, the sensitivity was 75% and specificity 94.4%.In contrast to our study, last studies reported that the US is a very useful method for detecting the lesions of all parts of the menisci In this study, detection of meniscal tears in the population who was young and with history of trauma was relatively easier as compared to patients who are older and with history of osteoarthritis.Consistent with this observation, we have divided the groups into two as less than 35 and more than 35 years old and compared the studies.We found that the sensitivity and specificity of US was 80% and 100% respectively in the young group as compared to 66.7% and 75% in the elder group.Correlation of US with arthroscopy was high in the first group and low in the second.We have decided that age factor strongly influences the success of US exams.Alizadeh at al., in their study of 74 cases suggested that the ultrasonography as an effective initial investigation for tears of medial meniscus in the group of 30 years old or less.The sensitivity, specificity, positive and negative predictive values and accuracy of ultrasonography in detecting medial meniscal tears in this age group were 100, 88.9, 96.5, 100, 97.3 % at this study (21).
The limitation of our study is small patient population.

Conclusion
US is not a suitable substitute of MRI in the routine diagnostic evaluation of meniscus lesions.MRI is more sensitive in detection of tears and degeneration than US.However in selected cases such as younger age group, traumatic cases and cases with a contraindication for MRI, US may find a role as a quick exam to stratify patients for further evaluation.Especially in patients' ≤ 35 years of age, the specificity and sensitivity of US increase considerably.US is relatively more successful in visualization of the posterior horns of the menisci but performs poor to visualize bodies of the menisci properly.

Grade 3 :
Signal changes either linear or nodular which extend to at least one articular surface.Using the grading system, meniscal lesions were classified:Normal: Homogeneous signal Meniscal degeneration: Grade 1 and/or grade 2 signal changes.Meniscal tear: Grade 3 signal changes.

Table 6 .Figure 1 .
Figure 1.Normally homogeneous appearance of posterior horns of lateral (a) and medial (b) menisci on US.

Figure 2 .
Gray scale finding of linear echogen tear touching inferior articular surface of the MP on US (a) is seen.On MRI of the same patient, sagittal T2* GE sequence shows grade 3 signal intensity indicating tear (b).

Figure 3 .
Figure 3. Hipoechoic area at MP considered as degeneration according to US is seen (a).Horizontal tear touching inferior articular surface is revelaed by T2* GE sequence on MRI of the same patient (b).

Table 2 .
The distribution of degeneration and comparative agreement levels as detected by imaging modalities in respective regions of the menisci.LA: lateral meniscus anterior horn, LB: lateral meniscus body, LP: lateral meniscus posterior horn, MA: medial meniscus anterior horn, MB: medial meniscus body, MP: medial meniscus posterior horn.n: number of patients.κ: agreement level.

Table 3 .
The distribution of number of meniscal tears with respect to regions of the menisci.LA: lateral meniscus anterior horn, LB: lateral meniscus body, LP: lateral meniscus posterior horn, MA: medial meniscus anterior horn, MB: medial meniscus body, MP: medial meniscus posterior horn, n: number of patients with tears/total number of examined patients.

Table 4 .
The sensitivity, specificity US and MRI and the κ agreement levels between US, MRI and arthroscopy.
Because the inner structure of the meniscus could not be assessed by arthroscopy, only the US and MRI findings for degeneration were evaluated.Ultrasound detected meniscus degeneration in 15 of the 35 patients (43%) whereas MRI detected meniscus degeneration in 23 of the 35 patients (65.7%).The degeneration related to the meniscal parts as detected by MRI and US were shown in Table2.No degeneration was noted in the body of the menisci by US therefore no comparative statistics was made for this region.Most of the degeneration was detected in the medial meniscus posterior horn (11 patients, 40.7%).

Table 5 .
The sensitivity and specificity of US and MR and the agreement between the studies with respect to regions of the menisci, LP: lateral meniscus posterior horn, MA: medial meniscus anterior horn, MP: medial