Magnetic Resonance Imaging (MRI) of the Extremities

Number: 0171

Policy

1. Aetna considers magnetic resonance imaging (MRI) studies of the knee medically necessary when any of the following criteria is met:

   1. Detection, staging, and post-treatment evaluation of tumor of the knee; or
   2. Fitting of implants for total knee arthroplasty; or
   3. Persistent knee pain/swelling and/or instability (giving way) not associated with an injury and not responding to at least 3 weeks of conservative therapy; or 
   4. Persistent knee pain/swelling and/or instability (giving way) secondary to an injury and not responding to conservative therapy when multi-view x-rays have ruled out a fracture or loose body in the knee and the clinical picture remains uncertain.  Conservative therapy consists of a combination of rest, ice, compression, elevation, non-steroidal anti-inflammatory drugs (NSAIDs), crutches, and range of motion (ROM) exercises; or
   5. Persistent true locking of the knee indicative of a torn meniscus or loose body.  (True locking is defined as more than a momentary locking of the joint with the knee in a flexed position, as compared to the sensation of momentary “catching” that many individuals experience in extension.); or
   6. Suspected bone infection (i.e., osteomyelitis); or
   7. Suspected osteochondritis dissecans or suspected osteonecrosis, if the clinical picture, including x-rays, is not confirmatory.

2. Aetna considers knee MRI experimental and investigational for all other indications, including any of the following circumstances because its effectiveness for indications other than the ones listed above has not been established:

   1. If arthroscopy or ligament reconstruction is definitely planned and the MRI findings are unlikely to change the planned treatment; or
   2. If the clinical picture (i.e., history, physical examination, x-rays, etc.) is diagnostic with high degree of certainty of a torn meniscus, loose body, or osteochondritis dessicans; or
   3. To diagnose or evaluate rheumatoid arthritis or degenerative joint disease.

3. Aetna considers MRI of the extremities (e.g., hands, knees, feet, etc.) experimental and investigational for diagnosing or monitoring arthritis because its effectiveness for these indications has not been established.

4. Aetna considers MRI of the extremities experimental and investigational for the following (not an all-inclusive list) because its effectiveness for these indications has not been established.

   1. Assessment of perfusion in diabetic foot ulcer
   2. Diagnosis of chronic exertional compartment syndrome
   3. Diagnosis suspected upper extremity deep vein thrombosis
   4. Evaluation and/or monitoring of disease progression in facioscapulohumeral muscular dystrophy

Background

Magnetic resonance imaging (MRI) has become the premier orthopedic diagnostic tool used in detecting meniscal and anterior cruciate ligament (ACL) tears and has virtually replaced both arthrography and arthroscopy as the diagnostic test of choice.

According to established guidelines from the American College of Rheumatology (2002), disease progression in rheumatoid arthritis (RA) should be followed using standard X-rays of the extremities.  There is no adequate evidence from prospective clinical studies that clinical outcomes are improved by using MRI over standard X-rays for this indication.  Although several studies have shown that MRI can detect early osseous changes, prospective clinical studies are needed to determine how well these early changes can predict development of clinically significant disease, and to determine whether clinical outcomes are improved by initiating therapy in persons with normal X-rays based on MRI findings.  McQueen et al (2001) found that only 25 % of changes detected by MRI progressed to X-ray erosions.  These results raise questions about the nature and pathophysiologic basis of the osseous changes detected by MRI, whether one can predict which of these osseous changes will progress to X-ray erosions, and about the nature of the changes detected by MRI that do not progress to X-ray erosions.

In addition, prospective clinical studies are necessary to determine whether clinical outcomes are improved by using MRI over standard X-rays to monitor disease progression in persons with RA.  Goldbach-Mansky et al (2003) of the National Institute of Arthritis and Musculoskeletal Diseases concluded that “[c]areful validation of MRI findings and the evaluation of MRI as a tool to follow the effect of therapy remain to be performed before MRI may be used as a clinical tool to follow therapy or as a surrogate for evaluating osseous changes over time.”  Boutry et al (2005) evaluated prospectively the use of MRI for differentiating true RA from systemic lupus erythematosus (SLE) or primary Sjogren syndrome in patients who have inflammatory polyarthralgia of the hands but no radiographic evidence of RA.  They concluded that it may be impossible to distinguish between patients with early RA and those without RA (namely those with SLE or primary Sjogren syndrome) by means of MRI.  Graham et al (2005) reported that determining synovial volume in the hand and wrist in patients (n = 10) with juvenile rheumatoid arthritis by MRI is feasible and correlates with total hand swelling assessed on physical examination.  However, these investigators stated that inconsistent or poor correlation with other clinical variables and the clinical definition of improvement requires further study.

Extremity MRI is not considered medically necessary to monitor the progression of arthritis.  The American College of Rheumatology (ACR) Guidelines for the Management of Rheumatoid Arthritis (2002) lists X-rays as the appropriate method of monitoring disease progression.  MRI of the arms and legs may be appropriate for the evaluation of masses, localized infections, non-healing fractures of long bones, and in certain cases, preoperative planning.  Furthermore, the report on extremity MRI in RA by the ACR Extremity MRI Task Force (2006) stated that most of the literature assessing the utility of peripheral joint MRI has used high-field, not low-field extremity MRI; therefore, actual sensitivity, specificity, and predictive value of the low-field scanners available for the practicing rheumatologists are not known.  The report also noted that the marginal benefit of low-field extremity MRI above and beyond standard measures of disease activity and severity (e.g., medical history, physical examination, selective laboratory testing, and radiography of the hands and wrists) has not been rigorously examined in studies published to date.  In summary, the benefits of low-field strength extremity MRI for the diagnosis and management of RA are still being elucidated.

More recently, guidance on the management of rheumatoid arthritis from the National Institute for Health and Clinical Excellence (NICE, 2009) recommended X-ray the hands and feet early in the course of the disease in people with persistent synovitis in these joints.  The Scottish Intercollegiate Guidelines Network (SIGN, 2011) concluded that "[t]he evidence for additional imaging at diagnosis to assess disease activity in early RA [rheumatoid arthritis] is limited and methodologically poor.  The evidence suggests that power Doppler ultrasound may be useful in assessing disease activity and may have predictive value on radiological outcome."  SIGN (2010) guidelines on psoriatic arthritis make no recommendation for magnetic resonance imaging of extremities.

This is consistent with the comments in a recent UpToDate review of the management of rheumatoid arthritis (Venables and Maini, 2011) which states that, although "magnetic resonance imaging (MRI) is a more sensitive technique than plain radiography for identifying bone erosions ... the clinical significance of erosions only detected by MRI awaits elucidation."  Another recent UpToDate review of management of rheumatoid arthritis by Schur et al (2011) stated that although "magnetic resonance imaging (MRI) [is] more sensitive for the detection of cartilage and bone abnormalities ... [it's] role in the process of making therapeutic decisions is presently under investigation."

There has been concern regarding the overuse of MRI.  MRI has come to be perceived by many doctors and patients as the initial and the sine qua non diagnostic tool prior to surgical treatment.  However, MRI should not be used as a routine screening tool in all knee injuries.   Its use should be reserved for clinical situations in which the diagnosis remains in doubt.   MRI does not replace a thorough history and physical examination and traditional multi-view x-rays as primary diagnostic tools.  In a randomized controlled study (n = 87), Nikken and colleagues (2005) concluded that short MRI examination with a low-field-strength MRI system after radiography in initial evaluation of patients with acute wrist trauma has additional value in prediction of treatment need; however, it does not have value in identification of patients who can be discharged without further follow-up.  In a randomized controlled study (n = 109), Oei and associates (2005) concluded that implementation of a dedicated extremity MRI examination, in addition to or instead of radiography, performed in patients with traumatic knee injury improves prediction of the need for additional treatment but does not significantly aid in identification of patients who can be discharged without further follow-up.  They stated that value of a short MRI examination in the initial stage after knee trauma is limited.

Andrish (1996) stated that isolated meniscal injuries are rare in children under the age of 14, but the frequency increases thereafter.  Meniscal tears in children are frequently associated with congenital meniscal abnormalities, while those in adolescents are often associated with ligamentous injuries of the knee.  The combination of recurrent and often dramatic popping and intermittent episodes of locking has been termed the "snapping knee syndrome", and is almost invariably associated with a discoid meniscus.  Although double-contrast arthrography has proved to be a reliable diagnostic technique, MRI is now the modality of choice.  In this regard, Connolly et al (1996) had described the MRI appearance and associated abnormalities of discoid menisci in children.  They noted that discoid meniscus commonly occurs bilaterally.  High intra-meniscal signal is found, especially in symptomatic patients.  The size criteria for diagnosing this condition in children are similar to those for adults.

In a prospective study, McNally et al (2002) examined if MRI of the acutely locked knee can alter surgical decision-making.  The study group comprised patients with a clinical diagnosis of knee locking requiring arthroscopy.  The decision to perform arthroscopy was made by an experienced consultant orthopedic surgeon specializing in trauma and recorded in the patient's notes prior to MRI.  Pre-operative MRI was carried out using a 1.5 T system.  The management was altered from surgical to conservative treatment in 20 (48 %) patients on the basis of the MR findings.  Arthroscopy was limited to patients with an MR diagnosis of a mechanical block, usually a displaced meniscal tear or loose body.  Both patient groups were followed clinically until symptoms resolved.  A total of 42 patients were entered into the study.  MRI identified a mechanical cause for locking in 22 patients (21 avulsion meniscal tears and 1 loose body).  All were confirmed at arthroscopy.  Twenty patients were changed from operative to non-operative treatment on the basis of the MRI findings.  One patient in this group required a delayed arthroscopy for an impinging anterior cruciate ligament stump.  The sensitivity, specificity, and accuracy of MRI in identifying patients who require arthroscopy was 96, 100, and 98 %, respectively.  These investigators concluded that MRI can successfully segregate patients with a clinical diagnosis of mechanical locking into those who have a true mechanical block and those who can be treated conservatively.  They stated that MRI should precede arthroscopy in this clinical setting.

Schurmann et al (2007) stated that complex regional pain syndrome type I (CRPS I) is difficult to diagnose in post-traumatic patients.  As CRPS I is a clinical diagnosis the characteristic symptoms have to be differentiated from normal post-traumatic states.  These researchers compared several diagnostic procedures for diagnosing post-traumatic CRPS I.  A total of 158 patients with distal radial fracture were included in this study.  A detailed clinical examination was carried out 2, 8, and 16 weeks after trauma in conjunction with bilateral thermography, plain radiographs of the hand skeleton, three phase bone scans (TPBSs), and contrast-enhanced magnetic resonance imaging (MRI).  All imaging procedures were assessed blinded.  At the end of the observation period, 18 patients (11 %) were clinically identified as having CRPS I and 13 patients (8 %) revealed an incomplete clinical picture which were defined as CRPS borderline cases.  The sensitivity of all diagnostic procedures used was poor and decreased between the first and the last examinations (thermography: 45 % to 29 %; TPBS: 19 % to 14 %; MRI: 43 % to 13 %; bilateral radiographs: 36 %).  In contrast, a high specificity was observed in the TPBS and MRI at the 8th and 16th-week examinations (TPBS: 96 %, 100 %; MRI: 78 %, 98 %) and for bilateral radiographs 8 weeks after trauma (94 %).  Thermography presented a fair specificity that improved from the 2nd to the 16th week (50 % to 89 %).  The authors concluded that the poor sensitivity of all tested procedures combined with a reasonable specificity produced a low positive predictive value (17 % to 60 %) and a moderate negative predictive value (79 % to 86 %).  These results suggested that those procedures can not be used as screening tests.  Imaging methods are not able to reliably differentiate between normal post-traumatic changes and changes due to CRPS I.  Clinical findings remain the gold standard for the diagnosis of CRPS I and the procedures described above may serve as additional tools to establish the diagnosis in doubtful cases.

Tsai and Beredjiklian (2008) noted that arthritis of the thumb joints is a common problem and remains a significant cause of morbidity in the adult population.  Careful physical examination is critical in the evaluation of these individuals, given the large differential diagnosis of conditions affecting the thumb and the radial side of the wrist.  Because treatment should be specifically directed at the area of pathology, adequate diagnosis is vital.  Plain radiograph examination remains the diagnostic modality of choice in the evaluation of patients with degenerative conditions regarding the hand and wrist.

The American College of Occupational and Environmental Medicine's clinical guideline on "Hand, wrist, and forearm disorders not including carpal tunnel syndrome" (2011) does not recommend MRI for diagnosing tuft fractures as well as phalangeal and metacarpal fractures.

1. Is there a standard diagnostic method available?
2. What ICP threshold criteria should be used for diagnosing CECS? and
3. What are the criteria and options for surgical management? 

Finally, these investigators made statements on the strength of each diagnostic criterion of ICP based on a rigorous standardized process.  The authors searched for studies that investigated ICP measurements in diagnosing CECS in the leg of human subjects, using PubMed, Score, PEDRO, Cochrane, Scopus, SportDiscus, Web of Knowledge, and Google Scholar.  Initial searches were performed using the phrase, "chronic exertional compartment syndrome".  The phrase "compartment syndrome" was then combined, using Boolean connectors ("OR" and "AND") with the words "diagnosis", "parameters", "levels", "localisation," or "measurement".  Data extracted from each study included study design, number of subjects, number of controls, ICP instrument used, compartments measured, limb position during measurements, catheter position, exercise protocol, timing of measurements, mean resting compartment pressures, mean maximal compartment pressures, mean post-exercise compartment pressures, diagnostic criteria used, and whether a reference diagnostic standard was used.  The quality of studies was assessed based on the approach used by the American Academy of Orthopaedic Surgeons in judging the quality of diagnostic studies, and recommendations were made regarding each ICP diagnostic criteria in the literature by taking into account the quality and quantity of the available studies proposing each criterion.  A total of 32 studies were included in this review.  The studies varied in the ICP measurement techniques used; the most commonly measured compartment was the anterior muscle compartment, and the exercise protocol varied between running, walking, and ankle plantarflexion and dorsiflexion exercises.  Pre-exercise, mean values ranged from 7.4 to 50.8 mm Hg for CECS patients, and 5.7 to 12 mm Hg in controls; measurements during exercise showed mean pressure readings ranging from 42 to 150 mm Hg in patients and 28 to 141 mm Hg in controls.  No overlap between subjects and controls in mean ICP measurements was found at the 1-min post-exercise timing interval only showing values ranging from 34 to 55.4 mm Hg and 9 to 19 mm Hg in CECS patients and controls, respectively.  The quality of the studies was generally not high, and the researchers found the evidence for commonly used ICP criteria in diagnosing CECS to be weak.  The authors concluded that studies in which an independent, blinded comparison is made with a valid reference standard among consecutive patients are yet to be undertaken.  There should also be an agreed ICP test protocol for diagnosing CECS because the variability here contributes to the large differences in ICP measurements and hence diagnostic thresholds between studies.  Current ICP pressure criteria for CECS diagnosis are therefore unreliable, and emphasis should remain on good history.  However, clinicians may consider measurements taken at 1 min after exercise because mean levels at this timing interval only did not overlap between subjects and controls in the studies that were analyzed.  Levels above the highest reported value for controls here (27.5 mm Hg) along with a good history, should be regarded as highly suggestive of CECS.  The authors stated that it is evident that to achieve an objective recommendation for ICP threshold, there is a need to set up a multi-center study group to reach an agreed testing protocol and modify the preliminary recommendations they have made.

Krabben et al (2013) stated that MRI is increasingly used to measure inflammation in rheumatoid arthritis (RA) research, but the correlation to clinical assessment is unexplored.  This study determined the association and concordance between inflammation of small joints measured with MRI and physical examination.  A total of 179 patients with early arthritis underwent a 68 tender joint count and 66 swollen joint count and 1.5T MRI of MCP (2-5), wrist and MTP (1-5) joints at the most painful side.  Two readers scored synovitis and bone marrow edema (BME) according to the OMERACT RA MRI scoring method and assessed tenosynovitis.  The MRI data were first analyzed continuously and then dichotomized to analyze the concordance with inflammation at joint examination.  A total of 1,790 joints of 179 patients were studied.  Synovitis and tenosynovitis on MRI were independently associated with clinical swelling, in contrast to BME.  In 86 % of the swollen MCP joints and in 92 % of the swollen wrist joints any inflammation on MRI was present.  In 27 % of the non-swollen MCP joints and in 66 % of the non-swollen wrist joints any MRI inflammation was present.  Vice versa, of all MCP, wrist and MTP joints with inflammation on MRI 64 %, 61 % and 77 %, respectively, were not swollen.  Bone marrow edema, also in case of severe lesions, occurred frequently in clinically non-swollen joints.  Similar results were observed for joint tenderness.  The authors concluded that inflammation on MRI is not only present in clinically swollen but also in non-swollen joints.  In particular BME occurred in clinically non-inflamed joints.  The relevance of subclinical inflammation for the disease course is a subject for further studies.

Also, an UpToDate review on “Diagnosis and differential diagnosis of rheumatoid arthritis” (Venables and Maini, 2014) states that “MRI and ultrasound -- Magnetic resonance imaging (MRI) studies and ultrasonography do not have an established role in the routine evaluation of patients with polyarthritis.  However, MRI and ultrasound are more sensitive than radiography at detecting changes resulting from synovitis and may be helpful in establishing the presence of synovitis in patients with normal radiographs and uncertainty regarding either the diagnosis or the presence of inflammatory changes, such as patients with obesity or subtle findings on examination”.

Furthermore, Axelsen and colleagues (2014) examined the ability of whole-body MRI (WBMRI) to visualize inflammation [synovitis, BME and enthesitis] and structural damage in patients with RA.  The 3T WBMR images were acquired in a head-to-toe scan in 20 patients with RA and at least 1 swollen or tender joint.  Short Tau Inversion Recovery and pre- and post-contrast T1-weighted images were evaluated for readability and the presence/absence of inflammation (synovitis, BME and enthesitis) and structural damage (erosions and fat infiltrations) in 76 peripheral joints, 30 entheseal sites and in the spine.  The readability was greater than 70 % for all individual joints, except for the most peripheral joints of the hands and feet.  Synovitis was most frequent in the wrist, first tarsometatarsal, first CMC joints and glenohumeral joints (67 to 61 %); BME in the wrist, CMC, acromioclavicular and glenohumeral joints (45 to 35 %) and erosions in the wrist, MTP and CMC joints (19 to 16 %).  Enthesitis at greater than or equal to 1 site was registered in 16 patients.  Bone marrow edema was frequently seen in the cervical (20 %) but not the thoracic and lumbar spine, while fat infiltrations and erosions were rare.  The intra-reader agreement was high (85 to 100 %) for all pathologies.  The agreement between WBMRI and clinical findings was low.   The authors concluded that peripheral and axial inflammation and structural damage at joints and entheses was frequently identified by WBMRI, and more frequently than by clinical examination.  They stated that WBMRI is a promising tool for evaluation of the total inflammatory load of inflammation (an MRI joint count) and structural damage in RA patients.

Management of Diabetic Foot Ulceration:

Forsythe and Hinchliffe (2016) noted that evaluation of foot perfusion is a vital step in the management of patients with diabetic foot ulceration, in order to understand the risk of amputation and likelihood of wound healing.  Underlying peripheral artery disease (PAD) is a common finding in patients with foot ulceration and is associated with poor outcomes.  Evaluation of foot perfusion should therefore focus on identifying the presence of PAD and to subsequently estimate the effect this may have on wound healing.  Assessment of perfusion can be difficult because of the often complex, diffuse and distal nature of PAD in patients with diabetes, as well as poor collateralization and heavy vascular calcification.  Conventional methods of evaluating tissue perfusion in the peripheral circulation may be unreliable in patients with diabetes, thus, it may therefore be difficult to determine the extent to which poor perfusion contributes to foot ulceration.  Anatomical data obtained on cross-sectional imaging is important but must be combined with measurements of tissue perfusion (such as transcutaneous oxygen tension) in order to understand the global and regional perfusion deficit present in a patient with diabetic foot ulceration.  Ankle-brachial pressure index is routinely used to screen for PAD, but its use in patients with diabetes is limited in the presence of neuropathy and medial arterial calcification.  Toe pressure index may be more useful because of the relative sparing of pedal arteries from medial calcification but may not always be possible in patients with ulceration.  Fluorescence angiography is a non-invasive technique that can provide rapid quantitative information about regional tissue perfusion; capillaroscopy, iontophoresis and hyper-spectral imaging may also be useful in assessing physiological perfusion but are not widely available.  The authors concluded that there may be a future role for specialized perfusion imaging of these patients, including MRI techniques, single-photon emission computed tomography (SPECT) and positron emission tomography (PET)-based molecular imaging; however, these novel techniques require further validation and are unlikely to become standard practice in the near future.

Evaluation and/or Monitoring of Disease Progression in Facioscapulohumeral Muscular Dystrophy:

Andersen and associates (2017) noted that there is no effective treatment available for facioscapulohumeral muscular dystrophy type 1 (FSHD1), but emerging therapies are underway that call for a better understanding of natural history in this condition.  In a prospective, longitudinal study, these researchers used quantitative MRI to evaluate yearly disease progression in patients with FSHD1.  Ambulatory patients with confirmed diagnosis of FSHD1 (25/20 men/women, age range of 20 to 75 years, FSHD score: 0 to 12) were tested with 359-560-day interval between tests.  Using the MRI Dixon technique, muscle fat replacement was evaluated in para-spinal, thigh, and calf muscles.  Changes were compared with those in FSHD score, muscle strength (hand-held dynamometry), 6-minute-walk-distance (6MWD), 14-step-stair-test, and 5-time-sit-to-stand-test.  Composite absolute fat fraction of all assessed muscles increased by 0.036 (confidence interval [CI]: 0.026 to 0.046, p < 0.001), with increases in all measured muscle groups.  The clinical severity FSHD score worsened (10 %, p < 0.05), muscle strength decreased over the hip (8 %), neck (8 %), and back (17 %) (p < 0.05), but other strength measures, 6MWD, 5-times-sit-to-stand-test, and 14-step-stair-test were unchanged.  Changes in muscle strength, FSHD score, and fat fraction did not correlate.  The authors concluded that this first study to systemically monitor quantitative fat replacement longitudinally in FSHD1 showed that MRI provided an objective measure of disease progression, often before changes can be appreciated in strength and functional tests.  The study indicated that quantitative MRI can be a helpful end-point in follow-up and therapeutic trials of patients with FSHD1.  These preliminary findings need to be validated by well-designed studies.

In a single-center study, Mui and colleagues (2017) examined if quantitative muscle MRI should be added to the clinical trial toolbox for(FSHD by correlating it to clinical outcome measures in a large cohort of genetically and clinically well-characterized patients with FSHD comprising the entire clinical spectrum.  Quantitative MRI scans of leg muscles of 140 patients with FSHD1 and FSHD2 were assessed for fatty infiltration and TIRM hyper-intensities and were correlated to multiple clinical outcome measures.  The mean fat fraction of the total leg musculature correlated highly with the motor function measure, FSHD clinical score, Ricci score, and 6MWD (correlation coefficients -0.845, 0.835, 0.791, -0.701, respectively).  Fat fraction per muscle group correlated well with corresponding muscle strength (correlation coefficients up to -0.82).  The hamstring muscles, adductor muscles, rectus femoris, and gastrocnemius medialis were affected most frequently, also in early stage disease and in patients without leg muscle weakness.  Muscle involvement was asymmetric in 20 % of all muscle pairs and fatty infiltration within muscles showed a decrease from distal to proximal of 3.9 %.  TIRM hyper-intense areas, suggesting inflammation, were found in 3.5 % of all muscles, with and without fatty infiltration.  The authors concluded that this study showed a strong correlation of quantitative muscle MRI with different clinical outcome measures, especially with the motor function measure.  Since quantitative muscle MRI also had good clinimetric properties, it is a promising biomarker representative of disease severity.  For clinical trials, these researchers proposed to include quantitative muscle MRI and the motor function measure in the FSHD trial toolbox.  Moreover, they stated that while the single-center, single-evaluator design of this study avoided interrater variability and increased the reliability of correlations between fat fractions and clinical measures; however, a limitation of this design was the decrease in the external validity.  Furthermore, results regarding disease duration should be interpreted with caution since age at disease onset was collected retrospectively

The above policy is based on the following references: