Michael B. Zlatkin

Magnetic resonance imaging of the shoulder. Sensitivity, specificity, and predictive value.

The sensitivity, specificity, and predictive value of magnetic resonance imaging in the diagnosis of lesions of the rotator cuff, glenohumeral capsule, and glenoid labrum were evaluated in ninety-one patients and fifteen asymptomatic volunteers. Magnetic resonance imaging demonstrated 100 per cent sensitivity and 95 per cent specificity in the diagnosis of complete tears, and it consistently predicted the size of the tear of the rotator cuff. There was a definite correlation between atrophy of the supraspinatus muscle and the size of a complete, chronic tear of the rotator cuff. The sensitivity and specificity of magnetic resonance imaging in the differentiation of tendinitis from degeneration of the cuff were 82 and 85 per cent, and in the differentiation of a normal tendon from one affected by tendinitis with signs of impingement the sensitivity and specificity were 93 and 87 per cent. The formation of spurs around the acromion and acromiocalvicular joint correlated highly with increased age of the patient and with chronic disease of the rotator cuff. The sensitivity and specificity of magnetic resonance imaging in the diagnosis of labral tears associated with glenohumeral instability were 88 and 93 per cent. The study showed that high-resolution magnetic-resonance imaging is an excellent non-invasive tool in the diagnosis of lesions of the rotator cuff and glenohumeral instability.

In vivo and ex vivo magnetic resonance imaging evaluation of early disc degeneration with histopathologic correlation

The in vivo and ex vivo microanatomic appearance of early disc degeneration were identified by magnetic resonance imaging and correlated with their respective histopathologic findings. Five cadaver spines (18 discs) and 25 patient studies (122 discs) all imaged at 1.5 Tesla were studied. Two signs of early degenerative disc disease were found: infolding and the central dot. Infolding of the central fibers of the outer annulus coalesced into a central dot of low signal intensity that was seen on both the ex vivo and in vivo images. Infolding was seen 29 of 122 times, and the central dot was observed 15 to 122 times on the in vivo images. A later form of degenerative disc disease was identified as a separation of the nucleus pulposus from the hyaline cartilage end-plate. This separation was seen as a linear area of either low or high signal intensity on the ex vivo images but only as a band of high signal intensity on the in vivo spin-echo 2,500-msec/80-msec images. Only 7 of 122 in vivo discs showed this separation. Internal herniation of nucleus pulposus into the outer annulus was seen only on the ex vivo images. Early degenerative disc disease may exist before there is loss of disc height or signal intensity on the long time-to-repetition (TR)/time-to-echo (TE) magnetic resonance images.

Combined Mr Imaging and Spectroscopy of Bone and Soft Tissue Tumors

Twenty-three patients with bone and soft tissue tumors were studied with combined magnetic resonance (MR) imaging and spectroscopy. The MR examinations were utilized to determine the size, internal characteristics, and relationships of the tumor to the surrounding tissues. They also determined the optimal placement of the surface coil. The surface coil profile was the localization technique utilized. Four patients were also studied with one-dimensional chemical shift localization. Tumors were grouped according to histologic type, degree of muscle contamination, size, and extent of necrosis. Quantitative comparison among the groups was carried out by comparing the mean ratios of the low-energy phosphate portion of the spectra [phosphomonoester (PME), Pi, phosphodiester (PDE)] to beta-nucleotide triphosphate (NTP). Tumor spectra typically showed a relative elevation in PME, Pi, and PDE and a relative decrease in phosphocreatinine. No characteristic spectra were observed for individual tumor types. Contamination of the tumor spectra from surrounding muscle impaired interpretation of the spectral data. Tumor size and extent of necrosis were important determinants of the relative degree of abnormally elevated metabolite peaks (PME, Pi, PDE). A trend toward a higher mean PME/beta-NTP ratio was observed among high-grade lesions. Combined MR imaging and spectroscopy is a useful way to study tumor metabolism. Muscle contamination is a significant problem in analysis of the spectra. Better localization techniques are required.

Integrated magnetic resonance imaging and phosphorus spectroscopy of soft tissue tumors

Eighteen patients with soft tissue masses underwent integrated magnetic resonance imaging (MRI) and phosphorus spectroscopy (31P‐MRS) to evaluate benign and malignant tumor morphology and metabolism. Spectra from soft tissue tumors had a significantly higher proportion of phosphate in the low‐energy portion of the 31P spectrum (P < 0.001) with a concomitant decrease in phosphocreatine (P < 0.01) compared with 31P spectra from normal muscle. Malignant tumors had a mean pH of 7.35 ± 0.13 which was greater than that of muscle tissue with a mean pH of 7.08 ± 0.07 (P < 0.001). All tumors had greater relative levels of phosphomonoesters, inorganic phosphate, and phosphodiesters compared with those in muscle tissue but considerable variability among tumors was noted due to tumor size, extent of tumor necrosis, and muscle contamination. Integrated MRI/MRS studies are necessary to provide exact localization of the tumor and a more correct interpretation of the 31P‐MRS data.

Magnetic resonance imaging and magnetic resonance spectroscopy of bone tumors and bone marrow disease.

The authors have made use of an integrated magnetic resonance imaging/spectroscopy (MRI/MRS) examination to study seven patients with a variety of bone tumors The spatial localization method used in the 31P portion of the examination was surface coil localization and a one-dimensional chemical shift imaging method (3 cases). The authors found that the precision of spatial localization was critical in many of these cases, since most of these bone tumors were surrounded by muscle tissue that contained high concentrations of phosphocreatine (PCr). For this reason, they suggest that the metabolite ratios should be referenced to the adenosine triphosphate (β-NTP) resonance rather than PCr. The phosphate monoester (PME) to β NTP ratio was elevated as compared with normal muscle in all of the bone tumors studied. The authors found that all of these tumors exhibited pHs between 7.0 and 7.2, which are similar to the values found for normal muscle. They also show the feasibility of using a line-selective proton chemical shift imaging sequence with high spatial resolution for investigating changes in the fatty composition of bone marrow. This method is illustrated in an example of a patient with advanced avascular necrosis in the femoral heads.

Magnetic resonance pathology of the rotator cuff.

Summary: This article reviews the current issues in imaging of the rotator cuff with magnetic resonance imaging (MRI). Important technical issues are discussed, including some of the newer pulse sequences, surface coils, and MR arthrography. Debate continues over what constitutes the normal appearance of the rotator cuff. Issues regarding studies of asymptomatic volunteers, pitfalls in imaging, and MR artifacts such as magic angle effects are discussed. The article also reviews the basic concepts of MR interpretation of pathologic lesions of the cuff, including tendinopathy, partial and complete tears, and the relationships to rotator cuff impingement.

Magnetic resonance imaging of the shoulder: Sensitivity, specificity, and predictive value

The sensitivity, specificity, and predictive value of magnetic resonance imaging in the diagnosis of lesions of the rotator cuff, glenohumeral capsule, and glenoid labrum were evaluated in ninety-one patients and fifteen asymptomatic volunteers. Magnetic resonance imaging demonstrated 100 per cent sensitivity and 95 per cent specificity in the diagnosis of complete tears, and it consistently predicted the size of the tear of the rotator cuff. There was a definite correlation between atrophy of the supraspinatus muscle and the size of a complete, chronic tear of the rotator cuff. The sensitivity and specificity of magnetic resonance imaging in the differentiation of tendinitis from degeneration of the cuff were 82 and 85 per cent, and in the differentiation of a normal tendon from one affected by tendinitis with signs of impingement the sensitivity and specificity were 93 and 87 per cent. The formation of spurs around the acromion and acromiocalvicular joint correlated highly with increased age of the patient and with chronic disease of the rotator cuff. The sensitivity and specificity of magnetic resonance imaging in the diagnosis of labral tears associated with glenohumeral instability were 88 and 93 per cent. The study showed that high-resolution magnetic-resonance imaging is an excellent non-invasive tool in the diagnosis of lesions of the rotator cuff and glenohumeral instability.

The glenohumeral joint

Shoulder pain is a common clinical problem. Although its causes are multiple, its clinical symptoms and signs can frequently overlap. To plan the most effective means of therapy, an accurate diagnosis must be made. Radiographic evaluation is an important and necessary adjunct to the clinical examination in patients with shoulder pain. Noninvasive imaging modalities, including plain radiography, radionuclide studies, tomography, and computed tomography (CT), often reveal nonspecific findings. Invasive examinations such as arthrography1,2 and conventional and computed arthrotomography3–6 may yield a more specific diagnosis but are not without morbidity.7 Ultrasound has been reported to be a useful technique in the evaluation of the shoulder for rotator cuff abnormalities.8 but it is limited in the scope of diseases it can assess and is highly dependent on the experience of the examiner. Recent technical advances have resulted in significant improvement in the ability to obtain diagnostic images of the shoulder with magnetic resonance imaging (MRI). These include off-center field of view and oblique imaging,9,10 as well as improvements in surface coils. We have used MRI in a large number of patients with shoulder pain. MRI has been effective in the noninvasive diagnosis of many shoulder disorders, particularly those that result from rotator cuff disease. It may replace most of the other techniques used to evaluate the shoulder. This article will review current experience with this modality and discuss relevant technical, anatomic, and pathologic issues.

Integrated Magnetic Resonance Imaging and 31P-Magnetic Resonance Spectroscopy of Soft Tissue Masses

Magnetic resonance proton imaging (MRI) of tumors provides excellent anatomic detail by defining the size, location and shape of the tumor in relationship to surrounding organs, muscle compartments, fascial planes, bone and neurovascular structures. The signal intensity on MRI depends on the biophysical properties of tissues such as proton density and the intrinsic relaxation times of the protons being imaged (T1 and T2 values). Variations in these parameters provide excellent contrast between normal and pathological tissue even when there is no difference in x-ray absorption. MRI permits the identification of necrosis, edema, hemorrhage, cystic degeneration and fibrosis providing more useful information relating to tumor staging as compared with other imaging techniques. These capabilities may lead to improved surgical management of soft tissue masses (1). Recent studies (2–5) indicate that MRI is superior to computerized tomography in the assessment of soft tissue masses.