SSR white paper: guidelines for utilization and performance of direct MR arthrography

OBJECTIVE

Direct magnetic resonance arthrography (dMRA) is often considered the most accurate imaging modality for the evaluation of intra-articular structures, but utilization and performance vary widely without consensus. The purpose of this white paper is to develop consensus recommendations on behalf of the Society of Skeletal Radiology (SSR) based on published literature and expert opinion.

MATERIALS AND METHODS

The Standards and Guidelines Committee of the SSR identified guidelines for utilization and performance of dMRA as an important topic for study and invited all SSR members with expertise and interest to volunteer for the white paper panel. This panel was tasked with determining an outline, reviewing the relevant literature, preparing a written document summarizing the issues and controversies, and providing recommendations.

RESULTS

Twelve SSR members with expertise in dMRA formed the ad hoc white paper authorship committee. The published literature on dMRA was reviewed and summarized, focusing on clinical indications, technical considerations, safety, imaging protocols, complications, controversies, and gaps in knowledge. Recommendations for the utilization and performance of dMRA in the shoulder, elbow, wrist, hip, knee, and ankle/foot regions were developed in group consensus.

CONCLUSION

Although direct MR arthrography has been previously used for a wide variety of clinical indications, the authorship panel recommends more selective application of this minimally invasive procedure. At present, direct MR arthrography remains an important procedure in the armamentarium of the musculoskeletal radiologist and is especially valuable when conventional MRI is indeterminant or results are discrepant with clinical evaluation.

Quantifying Uncertainty in Deep Learning of Radiologic Images

In recent years, deep learning (DL) has shown impressive performance in radiologic image analysis. However, for a DL model to be useful in a real-world setting, its confidence in a prediction must also be known. Each DL model's output has an estimated probability, and these estimated probabilities are not always reliable. Uncertainty represents the trustworthiness (validity) of estimated probabilities. The higher the uncertainty, the lower the validity. Uncertainty quantification (UQ) methods determine the uncertainty level of each prediction. Predictions made without UQ methods are generally not trustworthy. By implementing UQ in medical DL models, users can be alerted when a model does not have enough information to make a confident decision. Consequently, a medical expert could reevaluate the uncertain cases, which would eventually lead to gaining more trust when using a model. This review focuses on recent trends using UQ methods in DL radiologic image analysis within a conceptual framework. Also discussed in this review are potential applications, challenges, and future directions of UQ in DL radiologic image analysis.

Ischiofemoral impingement syndrome provoked by labor: An unusual case of complete sciatic mononeuropathy

Peripheral neuropathies are a recognized complication of labor in the post-partum period. Herein, we describe an uncommon presentation of sciatic mononeuropathy due to ischiofemoral impingement during labor. A 29-year-old, gravida 4 para 2, female presented post-partum with acute left lower limb paresthesia and left foot drop, following spontaneous vaginal delivery of twins. Neurological examination demonstrated no activation of the left sciatic-innervated muscles and sensory loss in the same distribution. Electromyography (EMG) demonstrated an acute complete left sciatic mononeuropathy. MRI of the lumbosacral plexus and sciatic nerve showed a narrowed quadratus femoris space with mild edema of the muscle, consistent with ischiofemoral impingement syndrome. In addition, there was flattening of the sciatic nerve as it passed through the ischiofemoral space. She was treated conservatively, and at 7-month follow-up, there was marked improvement in muscle strength with ongoing sensory impairment. Repeat EMG demonstrated reinnervation in all sciatic-innervated muscles. This case highlights the risk of a sciatic mononeuropathy secondary to ischiofemoral impingement in the peripartum setting. Future studies are needed to determine if women with a narrow ischiofemoral space at baseline are at increased risk for peripheral nerve injury during labor.

Accuracy of MR arthrography in the detection of posterior glenoid labral injuries of the shoulder

OBJECTIVE

The purpose of this study is to evaluate the accuracy of MR arthrography in detecting isolated posterior glenoid labral injuries using arthroscopy as the reference standard.

METHODS

MR arthrograms of 97 patients with isolated posterior glenoid labral tears by arthroscopy and those of 96 age and gender-matched controls with intact posterior labra were reviewed by two blinded radiologists for the presence and location of posterior labral abnormalities. The sensitivity and specificity of detection of posterior labral tears were calculated as well as the prevalence of associated pathologies. Medical records were reviewed for demographics, history and direction of shoulder instability, and prior surgery.

RESULTS

Posterior labral pathology was detected by MR arthrography with sensitivities of 76% and 84% for readers 1 and 2, and a specificity of 88% for both readers. Kappa value for interreader agreement was 0.91. Twenty-two of twenty-three (96%) tears isolated to the posteroinferior quadrant on arthroscopy were correctly identified on MRI. Commonly associated pathologies included paralabral cyst (38%), humeral fracture (7%), and glenoid fracture (2%). Fifteen of ninety-seven (16%) patients with posterior tears on both arthroscopy and MRI had glenoid rim deficiency on imaging versus no patients with intact posterior labra (p < 0.001). Forty of ninety-seven (41%) patients with posterior tears on arthroscopy had a history of posterior instability versus none without posterior tears. There was no significant difference in tear length on MRI between those with a history of instability and those without (p = 0.56).

CONCLUSION

MR arthrography is accurate in detecting posterior glenoid labroligamentous injuries.

A deep learning algorithm for detecting lytic bone lesions of multiple myeloma on CT

BACKGROUND

Whole-body low-dose CT is the recommended initial imaging modality to evaluate bone destruction as a result of multiple myeloma. Accurate interpretation of these scans to detect small lytic bone lesions is time intensive. A functional deep learning) algorithm to detect lytic lesions on CTs could improve the value of these CTs for myeloma imaging. Our objectives were to develop a DL algorithm and determine its performance at detecting lytic lesions of multiple myeloma.

METHODS

Axial slices (2-mm section thickness) from whole-body low-dose CT scans of subjects with biochemically confirmed plasma cell dyscrasias were included in the study. Data were split into train and test sets at the patient level targeting a 90%/10% split. Two musculoskeletal radiologists annotated lytic lesions on the images with bounding boxes. Subsequently, we developed a two-step deep learning model comprising bone segmentation followed by lesion detection. Unet and "You Look Only Once" (YOLO) models were used as bone segmentation and lesion detection algorithms, respectively. Diagnostic performance was determined using the area under the receiver operating characteristic curve (AUROC).

RESULTS

Forty whole-body low-dose CTs from 40 subjects yielded 2193 image slices. A total of 5640 lytic lesions were annotated. The two-step model achieved a sensitivity of 91.6% and a specificity of 84.6%. Lesion detection AUROC was 90.4%.

CONCLUSION

We developed a deep learning model that detects lytic bone lesions of multiple myeloma on whole-body low-dose CTs with high performance. External validation is required prior to widespread adoption in clinical practice.

Does gadolinium improve magnetic resonance arthrography of the hip beyond fluid distension alone?

AIM

To determine if there is a difference in the detection of labral tears and cartilage lesions using the T2-weighted sequences of magnetic resonance arthrography (MRA) only, which simulate MRA with saline alone, compared with the full examination including T1-weighted sequences.

MATERIALS AND METHODS

One hundred hip MRA examinations performed at 3 T were identified retrospectively. Each study was reviewed by a musculoskeletal radiologist using either the T2-weighted sequences only (without gadolinium-based contrast agent [GBCA] effect) or the entire examination, including T1-weighted sequences (with GBCA effect). Receiver operating characteristic (ROC) curves were calculated for both groups and compared within a non-inferiority framework, using - 0.10 as the limit of non-inferiority.

RESULTS

For labral tears, the difference area under the curve (AUC) was -0.004 (95% confidence interval [CI]: -0.064-0.056, p=0.90). For acetabular cartilage lesions, the difference AUC was 0.011 (95% CI: -0.073-0.096, p=0.79). For femoral cartilage lesions, the difference AUC was -0.059 (95% CI: -0.206-0.087, p=0.43).

CONCLUSION

T2-weighted sequences alone were not inferior in diagnostic accuracy compared with the full examination (including T1-weighted sequences with intra-articular GBCA) in detecting acetabular cartilage lesions and labral tears. Further study with prospective comparison of saline injectate to GBCA-containing injectate may help clarify the necessity of continued GBCA use in MRA of the hip.

Intraosseous "Lipoma" of the Calcaneus Developing in an Intraosseous Ganglion Cyst

Intraosseous lipomas are rare primary bone tumors, most common to the calcaneus. There are several proposed theories of the natural etiology of intraosseous lipomas; however, all lack definitive support. In this report, an 18-year-old man presented with radiologic evidence of a simple bone cyst of the calcaneus. Over a 4-year period, the patient was followed with interval magnetic resonance imaging. The cyst demonstrated progressive development of peripheral intralesional fat with final magnetic resonance imaging features characteristic of an intraosseous lipoma. To our knowledge, this is the first longitudinal study that shows gradual peripheral fat deposition within an intraosseous ganglion cyst, illustrating a potential cause of intraosseous lipomas of the calcaneus.

MRI of pathology-proven peripheral nerve amyloidosis

OBJECTIVE

To highlight the MRI characteristics of pathologically proven amyloidosis involving the peripheral nervous system (PNS) and determine the utility of MRI in directing targeted biopsy for aiding diagnosis.

MATERIALS AND METHODS

A retrospective study was performed for patients with pathologically proven PNS amyloidosis who also underwent MRI of the biopsied or excised nerve. MRI signal characteristics, nerve morphology, associated muscular denervation changes, and the presence of multifocal involvement were detailed. Pathology reports were reviewed to determine subtypes of amyloid. Charts were reviewed to gather patient demographics, neurological symptoms and radiologist interpretation.

RESULTS

Four men and three women with a mean age of 62 ± 11 years (range 46-76) were identified. All patients had abnormal findings on EMG with mixed sensorimotor neuropathy. All lesions demonstrated diffuse multifocal neural involvement with T1 hypointensity, T2 hyperintensity, and variable enhancement on MRI. One lesion exhibited superimposed T2 hypointensity. Six of seven patients demonstrated associated muscular denervation changes.

CONCLUSION

Peripheral nerve amyloidosis is rare, and the diagnosis is difficult because of insidious symptom onset, mixed sensorimotor neurologic deficits, and the potential for a wide variety of nerves affected. On MRI, peripheral nerve involvement is most commonly characterized by T1 hypointensity, T2 hyperintensity, variable enhancement, maintenance of the fascicular architecture with fusiform enlargement, multifocal involvement and muscular denervation changes. While this appearance mimics other inflammatory neuropathies, MRI can readily detect neural changes and direct-targeted biopsy, thus facilitating early diagnosis and appropriate management.

Syntactic and semantic errors in radiology reports associated with speech recognition software

Speech recognition software can increase the frequency of errors in radiology reports, which may affect patient care. We retrieved 213,977 speech recognition software-generated reports from 147 different radiologists and proofread them for errors. Errors were classified as "material" if they were believed to alter interpretation of the report. "Immaterial" errors were subclassified as intrusion/omission or spelling errors. The proportion of errors and error type were compared among individual radiologists, imaging subspecialty, and time periods. In all, 20,759 reports (9.7%) contained errors, of which 3992 (1.9%) were material errors. Among immaterial errors, spelling errors were more common than intrusion/omission errors ( p < .001). Proportion of errors and fraction of material errors varied significantly among radiologists and between imaging subspecialties ( p < .001). Errors were more common in cross-sectional reports, reports reinterpreting results of outside examinations, and procedural studies (all p < .001). Error rate decreased over time ( p < .001), which suggests that a quality control program with regular feedback may reduce errors.

New vertebral fractures after vertebroplasty: 2-year results from a randomised controlled trial

A randomised controlled trial of vertebroplasty (VP) versus placebo assessed the effect of VP on the risk of further vertebral fractures. While no statistically significant between-group differences for new or progressed fracture risk at 12 and 24 months were observed, we observed a consistent trend towards higher risk of any type of fracture in the group undergoing VP. Our analysis was underpowered, and further adequately powered studies are needed to be able to draw firm conclusions about further vertebral risk with vertebroplasty.

PURPOSE

This study seeks to assess the effect of VP on the risk of further radiologically apparent vertebral fracture within two years of the procedure.

METHODS

We conducted a randomised placebo-controlled trial of VP in people with acute osteoporotic vertebral fracture. Eligible participants were randomly assigned to VP (n = 38) or placebo (n = 40). Cement volume and leakage were recorded for the VP group. Plain thoracolumbar radiographs were taken at baseline, 12 and 24 months. Two independent radiologists assessed these for new and progressed fractures at the same, adjacent and non-adjacent levels.

RESULTS

At 12 and 24 months, radiographs were available for 45 (58 %) and 47 (60 %) participants, respectively. There were no between-group differences for new or progressed fractures: 32 and 40 in the VP group after 12 and 24 months compared with 21 and 33 in the placebo group (hazard ratio (HR) 1.80, 95 % confidence interval (CI) 0.82 to 3.94). Similar results were seen when considering only adjacent (HR (95 % CI) 2.30 (0.57 to 9.29)) and non-adjacent (HR (95 % CI) 1.45 (0.55 to 3.81) levels. In all comparisons, there was a consistent trend towards higher risk of any type of fracture in the group undergoing VP. Within the VP group, fracture risk was unrelated to total (HR (95 % CI) 0.91 (0.71 to 1.17)) or relative (HR (95 % CI) 1.31 (0.15 to 11.48)) cement volume or cement leakage (HR (95 % CI) 1.20 (0.63 to 2.31)).

CONCLUSION

For patients undergoing VP, our study did not demonstrate significant increases in subsequent fracture risk beyond that experienced by those with vertebral fractures who did not undergo the procedure. However, because of the non-significant numerical increases observed, studies with adequate power are needed to draw definite conclusions about fracture risk.

Syntactic and Semantic Errors in Radiology Reports Associated With Speech Recognition Software

Speech recognition software (SRS) has many benefits, but also increases the frequency of errors in radiology reports, which could impact patient care. As part of a quality control project, 13 trained medical transcriptionists proofread 213,977 SRS-generated signed reports from 147 different radiologists over a 40 month time interval. Errors were classified as "material" if they were believed to alter interpretation of the report. "Immaterial" errors were subclassified as intrusion/omission or spelling errors. The proportion of errors and error type were compared among individual radiologists, imaging subspecialty, and time periods using .2 analysis and multiple logistic regression, as appropriate. 20,759 (9.7%) reports contained errors; 3,992 (1.9%) contained material errors. Among immaterial errors, spelling errors were more common than intrusion/omission errors (P&lt;.001). Error proportion varied significantly among radiologists and between imaging subspecialties (P&lt;.001). Errors were more common in cross-sectional reports (vs. plain radiography) (OR, 3.72), reports reinterpreting results of outside examinations (vs. in-house) (OR, 1.55), and procedural studies (vs. diagnostic) (OR, 1.91) (all P&lt;.001). Dictation microphone upgrade did not affect error rate (P=.06). Error rate decreased over time (P&lt;.001).

Imaging findings in 22 cases of Schnitzler syndrome: characteristic para-articular osteosclerosis, and the "hot knees" sign differential diagnosis

BACKGROUND AND PURPOSE

Schnitzler syndrome is characterized by chronic urticaria, monoclonal gammopathy, and a risk of developing lymphoproliferative disorders. Patients frequently present with bone pain, fever, arthralgia, and lymphadenopathy. The purpose of this study is to retrospectively review and evaluate potentially attributable imaging abnormalities in a series of patients with clinically diagnosed Schnitzler syndrome.

MATERIALS AND METHODS

Clinical and pathological chart review identified 22 patients meeting Strasbourg criteria for Schnitzler syndrome. Imaging of these patients was retrospectively reviewed by a musculoskeletal radiologist and a radiology fellow to determine and characterize significant "positive" findings that could potentially be attributed to their primary disease process.

RESULTS

Evidence of abnormal bone remodeling was present in 64% (14/22) of patients with 40 bone abnormalities. Lesions were predominantly or entirely sclerotic and most commonly involved the distal femora (n = 11), proximal tibiae (nine), and innominate bones (six). Patterns of involvement were variable, but typically eccentrically intramedullary, cortical, and juxta-articular. Skeletal surveys and targeted radiographs identified only 37% (10/27) of abnormal bones subsequently "positive" by additional modalities. PET and bone scans were all positive in patients with bone sclerosis (13/13 and 17/17 bone sites, respectively).

CONCLUSIONS

Schnitzler syndrome often presents with characteristic osteosclerosis, most commonly around the knees and in the pelvis. In patients with a suggestive clinical history, the radiologist could direct an appropriate imaging strategy or might to be the first to suggest the diagnosis. Bone scan may be the most appropriate initial screening tool in suspected cases.

Case-control study to estimate the performance of dual-energy computed tomography for anterior cruciate ligament tears in patients with history of knee trauma

OBJECTIVE

Computed tomography (CT) is used to assess for fracture after knee trauma, but identification of ligamentous injuries may also be beneficial. Our purpose is to assess the potential of dual-energy computed tomography (DECT) for the detection of complete anterior cruciate ligament (ACL) disruption.

METHODS

Sixteen patients with unilateral traumatic ACL disruption (average of 58 days following trauma) confirmed by MRI, and 11 control patients without trauma, underwent DECT of both knees. For each knee, axial, sagittal, and oblique sagittal images (with DECT bone removal, single-energy (SE) bone removal, and DECT tendon-specific color mapping) were reconstructed. Four musculoskeletal radiologists randomly evaluated the 324 DECT reconstructed series (54 knees with 6 displays) separately, to assess for ACL disruption using a five-point scale (1 = definitely not torn, to 5 = definitely torn). ROC analysis was used to compare performance across readers and displays.

RESULTS

Sagittal oblique displays (mixed kV soft tissue, SE bone removal, and DECT bone removal) demonstrated higher areas under the curve for ACL disruption (AUC = 0.95, 0.93 and 0.95 respectively) without significant differences in performance between readers (p > 0.23). Inter-reader agreement was also better for these display methods (ICC range 0.62-0.69) compared with other techniques (ICC range 0.41-0.57). Mean sensitivity for ACL disruption was worst for DECT tendon-specific color map and axial images (24 % and 63 % respectively).

CONCLUSION

DECT knee images with oblique sagittal reconstructions using either mixed  kV or bone removal displays (either DECT or SE) depict ACL disruption in the subacute or chronic setting with reliable identification by musculoskeletal radiologists.

MRI of wrist ligaments

The optimal imaging protocols for magnetic resonance imaging (MRI) of the wrist ligaments are discussed, including the use of magnetic resonance arthrography, and 3 Tesla (T) versus 1.5 T magnetic field strength. The normal MRI appearance of the triangular fibrocartilage complex, capsular, and interosseous wrist ligaments is briefly covered to point out potential diagnostic pitfalls. Numerous examples of common ligamentous pathology discernible on MRI are provided, along with the latest estimates of diagnostic sensitivity and specificity provided by the literature.