Keywords
ultrasonography - magnetic resonance imaging - meniscus - knee - arthroscopy
Nearly 1 million meniscal injuries occur in the United States annually.[1] Currently, meniscal pathology is most often diagnosed based on history, clinical
examination, magnetic resonance imaging (MRI), and/or arthroscopic visualization.[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22] Early and accurate diagnosis of meniscal pathology is vital for determining type
and timing of treatment, as well as prognosis for return to function in the short
term and degree of morbidity in the long term. In the United States, preoperative
diagnosis of meniscal pathology is largely based on MRI of the affected knee. Diagnostic
MRI may not be performed until weeks or even months after injury, and while MRI is
often considered the “gold standard” diagnostic imaging modality for detection of
meniscal abnormalities, it is associated with misdiagnosis in 14 to 47% of cases.[7]
[10]
[13]
[14]
[21]
[22] MRI is also costly and is not readily available to a large number of patients for
either financial or logistical reasons, or both. Because early and accurate diagnosis
of meniscal pathology is imperative for treatment planning and prognostication, this
imaging modality may not be efficient and effective for optimal management of patients
with meniscal pathology.
Ultrasonographic examination of the knee shows promise for being an effective diagnostic
tool for assessing meniscal pathology with the potential to overcome many of the shortcomings
of MRI. Ultrasonography has been used for diagnosis of meniscal pathology in veterinary
medicine for more than a decade with reported sensitivities and specificities as high
as 90 and 92.9%, respectively.[23] Studies in human patients report sensitivities ranging from 30 to 100% and specificities
ranging from 71.4 to 98% for sonographic diagnosis of meniscal pathology.[2]
[3]
[4]
[5]
[6]
[8]
[9]
[12]
[15]
[16]
[17]
[18]
[19]
[20] The most recent studies report the highest sensitivities, specificities, and correct
classification rates (CCRs), likely as a result of improvements in technology, training,
and experience. Based on the recent results, the costs and availability, the portability
of the equipment, and the safety associated with ultrasonography, the authors of these
studies concluded that ultrasonography is a clinically useful diagnostic technique
for meniscal pathology. However, the capabilities of ultrasonography for clinical
diagnosis of meniscal abnormalities in patients have not been fully evaluated. Therefore,
the objective of this study was to determine the usefulness of ultrasonography for
diagnosis of meniscal pathology in patients with acute knee pain and compare its diagnostic
accuracy to MRI in a clinical setting.
Methods
With Institutional Review Board approval, patients (n = 71) with acute knee pain were prospectively enrolled through informed consent.
Patients were included in the study if they were 14 years of age and older, they presented
with acute knee pain, and they had a high probability for requiring surgical intervention
for diagnosis and/or treatment of the affected knee(s). Patients were excluded when
these criteria were not met or informed consent was not granted and documented.
Epidemiologic data, sonographic imaging findings, MRI reports, and surgery reports
and images (when available) were recorded in the patients' medical records and the
dedicated database for the study. Data were only retrieved for analysis.
Preoperative MRI was performed on each affected knee using the hospital's standard
equipment and protocols and read by faculty radiologists trained in musculoskeletal
MRI. All MRIs were performed using 1.5 T units with dedicated knee coils. Sequences
included were at the discretion of the attending radiologists based on their preferences
and standard-of-care clinical practice. The attending radiologist reviewed the MRI
and subjectively classified each meniscus as normal, torn (with description of location
and type when possible), or showing abnormal signal, but not torn. MRI reports were
entered into the medical record and data specific to meniscal findings were entered
into a separate database by another investigator blinded to all other diagnostic findings.
At a separate appointment, each patient underwent ultrasonographic examination of
the affected knee conducted by one ultrasonographer blinded to patient history, physical
examination findings, and MRI findings. Ultrasonographic examination was performed
by one of two investigators trained in musculoskeletal ultrasonography using a portable
ultrasound machine (Logiq i, GE Health Care, Milwaukee, WI) with a 10 to 14 MHz linear
transducer. Each patient was initially placed supine with the knee extended, so that
the anterior horns of the menisci could be examined. The knee was then flexed to 90
degrees and the probe rotated laterally to examine the anterior cruciate ligament
(ACL) (this is a dynamic examination with the knee being serially extended). The patient
was then placed prone so the posterior horns of each meniscus could be examined. Each
meniscus was evaluated for displacement, echogenicity, shape, and associated effusion
([Fig. 1]). Scores were used for determining strength of correlations between ultrasonographic
findings and basic science data for another portion of this research. The presence
or absence of meniscal pathology was determined based on the ultrasonographer's overall
assessment using the four criteria in the standardized assessment protocol to subjectively
determine deviations from normal. Location of the tear was determined, however, characterization
of type of tear was not attempted. Rather, the goal was to determine whether or not
meniscal pathology warranting surgical intervention was present. Findings consistent
with joint effusion, synovial thickening, and osteophytosis were also recorded. Ultrasonography
reports were entered into the medical record and data specific to meniscal findings
were entered into a separate database by another investigator blinded to all other
diagnostic findings.
Fig. 1 Ultrasound assessment form developed for evaluations.
Based on indication and with informed patient consent, complete arthroscopic examination
of affected knees was performed by one of three faculty orthopedic surgeons to assess
and record all joint pathology, which served as the reference standard for determining
presence, type, and severity of meniscal pathology present. Both menisci were assessed
by visualization and palpation with an arthroscopic probe and all findings were subjectively
described and recorded. Arthroscopy reports were entered into the medical record and
data specific to meniscal findings were entered into a separate database by another
investigator blinded to all other diagnostic findings.
Arthroscopic data were used as the reference standard for statistical analyses. Data
were compiled and analyzed by a separate investigator to determine sensitivity, specificity,
positive predictive value (PPV), negative predictive value (NPV), correct classification
rate (CCR), likelihood ratios (LR[+], LR[−]), and odds ratios.
Results
Patients included in the study ranged in age from 15 to 73 years with a mean age of
37.2 years. Forty patients (56%) were male and 31 (44%) were female.
Based on arthroscopic evaluation, 59 patients had identifiable meniscal pathology
(49 medial, 18 lateral, 4 both) and 12 patients had no definitive meniscal pathology
identified.
[Table 1] provides sensitivity, specificity, PPV, NPV, CCR, and likelihood ratios for MRI
and ultrasonographic assessments of meniscal pathology. Based on odds ratio analysis,
ultrasonography was two times more likely than MRI to correctly determine presence
or absence of meniscal pathology diagnosed based on arthroscopic assessments ([Fig. 2]).
Fig. 2 Representative images from cases included in this study showing normal (A, B) and
pathologic (C–F) menisci based on arthroscopic assessment, which were all correctly
diagnosed by ultrasonography while one (F) was incorrectly diagnosed by magnetic resonance
imaging.
Table 1
Sensitivity, specificity, PPV, NPV, CCR, LR+, and LR− for MRI and ultrasonographic
assessments of meniscal pathology
|
Modality
|
Sensitivity
|
Specificity
|
PPV
|
NPV
|
CCR
|
LR+
|
LR−
|
|
Ultrasound
|
91.2%
|
84.2%
|
94.5%
|
76.2%
|
89.5%
|
5.78
|
0.10
|
|
MRI
|
91.7%
|
66.7%
|
84.6%
|
80.0%
|
81.1%
|
2.75
|
0.13
|
Abbreviations: CCR, correct classification rate; LR+, positive likelihood ratio; LR−,
negative likelihood ratio; NPV, negative predictive value; PPV, positive predictive
value.
For MRI, incorrect diagnoses involved missing lateral meniscal degeneration, missing
medial radial tears, missing medial longitudinal tears, missing lateral tears when
both menisci were torn, incorrectly reporting lateral tears as medial tears, and reporting
tears not identified arthroscopically.
For ultrasonography, incorrect diagnoses involved missing medial and lateral radial
tears, missing lateral tears when both menisci were torn, and reporting tears not
identified arthroscopically.
Discussion
This study produced data that are in agreement with previous work suggesting that
ultrasonography is a useful tool for diagnosis of meniscal pathology with potential
advantages over MRI. By using a standardized assessment system for meniscal ultrasonography,
consistent and accurate information was obtained. This assessment system provides
ultrasonographers with a methodology for progressing efficiently along the learning
curve of this technique to obtain results similar to ours in clinical practice. In
this prospective study, ultrasonographic assessment was two times more likely to correctly
determine the presence or absence of meniscal pathology in patients with acute knee
pain with a CCR of 89.5% compared with 81.1% for MRI. This means that MRI is “wrong”
almost twice as often as ultrasound for determining the presence or absence of meniscal
pathology as assessed in this study.
Importantly, the greatest statistical differences between the two techniques were
noted for specificity and positive likelihood ratios. These differences provide further
evidence for the benefits of ultrasonography for being associated with a higher probability
to be correct in determining the absence of meniscal pathology and differentiating
patients with meniscal pathology from those without meniscal pathology. Based on the
extremely large volume of meniscal injuries treated each year—approximately 1 million
per year in the United States alone[1]—and the fact that the majority of these injuries require surgical intervention,
the authors suggest these findings are of critical clinical importance.
The advantages of ultrasound as a diagnostic test would translate into the avoidance
of unnecessary surgeries and related morbidity, as well as profound savings in related
health care costs. Similarly, financial benefits related to costs for diagnostic imaging
could also be realized. At the time of study initiation, our institution's charges
were $384 to perform and read an ultrasonographic study of the knee and $1,211 for
MRI of the knee as described. Coupled with the high safety of ultrasound and its relative
comfort for patients with claustrophobia, phonophobia, or magnet-sensitive implants,
these features make ultrasound very attractive for clinical use in diagnostic imaging
for meniscal pathology.
Another potential advantage of ultrasonography for diagnosis of meniscal pathology
is related to the portability of current equipment. High-quality ultrasound machines
are available in laptop and hand-held versions. A portable, laptop version was used
for imaging in the present study. This factor provides the potential for point-of-injury
diagnostic imaging, which for athletic activities includes the potential for “on-the-field”
diagnostics and decision making. As an example, American football has one of the highest
rates of injury associated with an athletic activity with more than one-third of all
football injuries involving the knee.[24]
[25]
[26] Football players are nearly six times more likely to suffer knee injuries requiring
surgery than the general population.[27] A study that evaluated athletes at the National Football League Combine for incoming
rookies described a meniscal tear/injury incidence of 12.4 per 100 players, and indicated
meniscectomy as the most common procedure performed on these athletes with an incidence
of 10.3 meniscectomies per 100 players.[28] Based on current protocols, there is typically at least a 24- to 72-hour delay before
diagnostic imaging in the form of MRI is performed, which renders the patient and
medical staff unable to make an informed decision regarding treatment and prognosis,
and potentially allows for further damage to occur. This same scenario could be multiplied
by the thousands by application to male and female soccer players around the world—as
well as any other athletic activity for people of all ages—many of whom have no access
(logistical and/or financial) to MRI. Ultrasonography provides a portable, cost effective,
and accurate tool for earlier diagnosis for these patients so optimal management can
be efficiently pursued.
It should be pointed out that these findings apply only to diagnosis of meniscal pathology.
While ultrasonographic assessment of the knee can include evaluation of medial collateral
ligament, lateral collateral ligament, posterior lateral corner, joint capsule and
limited regions of the ACL, PCL, and articular cartilage, assessment of these structures
was not included in the present study and there are aspects of pathology of the knee
that cannot be determined using ultrasonography. Therefore, it should be emphasized
that MRI has advantages over ultrasonography for comprehensive diagnostic imaging
of the knee and ultrasonographic assessment cannot be used alone to completely assess
the injured knee. Knee ultrasonography as described in this study can only be applied
to targeted evaluation of the menisci.
The limitations of the present study primarily involve the number and type of patients
included and the relative inequity among imagers performing the two different diagnostic
imaging techniques. Certainly, a larger number and spectrum of patients need to be
included in a multicenter trial that validates the findings of this study before conclusive
arguments regarding preference for use of ultrasonography over MRI for diagnosis of
meniscal pathology can be made. Similarly, additional imagers—perhaps using a standardized
and optimized MRI protocol—would need to be included for both modalities to ensure
that these data are broadly applicable. It could also be argued that higher strength
(3 T) magnets with advanced software would likely provide improved diagnostic capabilities
and compare more favorably with ultrasonography for assessment of meniscal pathology.
However, that level of MRI technology is not widely available for routine clinical
use at this time and therefore would not provide a “real life” standard-of-care comparison
for broad application.
At this point in development, the ultrasonographic assessment methodology we employ
is not based on quantitative or even semiquantitative measures. While scores were
assigned in each of the four categories included in the standardized system, those
scores were used for determining strength of correlations in other aspects of this
research and not used as thresholds for presence or severity of meniscal pathology.
However, the use of the standardized criteria and technique described resulted in
consistent and accurate results from two different ultrasonographers in terms of correctly
classifying presence or absence and location of meniscal pathology in patients presenting
for acute knee pain.
In summary, this study provides evidence for the use of standardized ultrasonographic
assessment of the knee as a useful tool for diagnosis of meniscal pathology with potential
advantages over MRI. Based on these data and available portable equipment, ultrasonography
could be considered for use as a point-of-injury diagnostic modality for meniscal
injuries.
