SymTC: A symbiotic Transformer-CNN net for instance segmentation of lumbar spine MRI.

Intervertebral disc disease, a prevalent ailment, frequently leads to intermittent or persistent low back pain, and diagnosing and assessing of this disease rely on accurate measurement of vertebral bone and intervertebral disc geometries from lumbar MR images. Deep neural network (DNN) models may assist clinicians with more efficient image segmentation of individual instances (discs and vertebrae) of the lumbar spine in an automated way, which is termed as instance image segmentation. In this work, we proposed SymTC, an innovative lumbar spine MR image segmentation model that combines the strengths of Transformer and Convolutional Neural Network (CNN). Specifically, we designed a parallel dual-path architecture to merge CNN layers and Transformer layers, and we integrated a novel position embedding into the self-attention module of Transformer, enhancing the utilization of positional information for more accurate segmentation. To further improve model performance, we introduced a new data synthesis technique to create synthetic yet realistic MR image dataset, named SSMSpine, which is made publicly available. We evaluated our SymTC and the other 16 representative image segmentation models on our private in-house dataset and public SSMSpine dataset, using two metrics, Dice Similarity Coefficient and the 95th percentile Hausdorff Distance. The results indicate that SymTC surpasses the other 16 methods, achieving the highest dice score of 96.169 % for segmenting vertebral bones and intervertebral discs on the SSMSpine dataset. The SymTC code and SSMSpine dataset are publicly available at https://github.com/jiasongchen/SymTC.

Comparison of image quality and diagnostic efficacy of routine clinical lumbar spine imaging at 0.55T and 1.5/3T.

PURPOSE: To compare image quality, assess inter-reader variability, and evaluate the diagnostic efficacy of routine clinical lumbar spine sequences at 0.55T compared with those collected at 1.5/3T to assess common spine pathology. METHODS: 665 image series across 70 studies, collected at 0.55T and 1.5/3T, were assessed by two neuroradiology fellows for overall imaging quality (OIQ), artifacts, and accurate visualization of anatomical features (intervertebral discs, neural foramina, spinal cord, bone marrow, and conus / cauda equina nerve roots) using a 4-point Likert scale (1 = non-diagnostic to 4 = excellent). For the 0.55T scans, the most appropriate diagnosis(es) from a picklist of common spine pathologies was selected. The mean ± SD of all scores for all features for each sequence and reader at 0.55T and 1.5/3T were calculated. Paired t-tests (p ≤ 0.05) were used to compare ratings between field strengths. The inter-reader agreement was calculated using linear-weighted Cohen's Kappa coefficient (p ≤ 0.05). Unpaired VCG analysis for OIQ was additionally employed to represent differences between 0.55T and 1.5/3T (95 % CI). RESULTS: All sequences at 0.55T were rated as acceptable (≥2) for diagnostic use by both readers despite significantly lower scores for some compared to those at 1.5/3T. While there was low inter-reader agreement on individual scores, the agreement on the diagnosis was high, demonstrating the potential of this system for detecting routine spine pathology. CONCLUSIONS: Clinical lumbar spine imaging at 0.55T produces diagnostic-quality images demonstrating the feasibility of its use in diagnosing spinal pathology, including osteomyelitis/discitis, post-surgical changes with complications, and metastatic disease.

Automated Detection and Measurement of Dural Sack Cross-Sectional Area in Lumbar Spine MRI Using Deep Learning.

Lumbar spine magnetic resonance imaging (MRI) is a critical diagnostic tool for the assessment of various spinal pathologies, including degenerative disc disease, spinal stenosis, and spondylolisthesis. The accurate identification and quantification of the dural sack cross-sectional area are essential for the evaluation of these conditions. Current manual measurement methods are time-consuming and prone to inter-observer variability. Our study developed and validated deep learning models, specifically U-Net, Attention U-Net, and MultiResUNet, for the automated detection and measurement of the dural sack area in lumbar spine MRI, using a dataset of 515 patients with symptomatic back pain and externally validating the results based on 50 patient scans. The U-Net model achieved an accuracy of 0.9990 and 0.9987 on the initial and external validation datasets, respectively. The Attention U-Net model reported an accuracy of 0.9992 and 0.9989, while the MultiResUNet model displayed a remarkable accuracy of 0.9996 and 0.9995, respectively. All models showed promising precision, recall, and F1-score metrics, along with reduced mean absolute errors compared to the ground truth manual method. In conclusion, our study demonstrates the potential of these deep learning models for the automated detection and measurement of the dural sack cross-sectional area in lumbar spine MRI. The proposed models achieve high-performance metrics in both the initial and external validation datasets, indicating their potential utility as valuable clinical tools for the evaluation of lumbar spine pathologies. Future studies with larger sample sizes and multicenter data are warranted to validate the generalizability of the model further and to explore the potential integration of this approach into routine clinical practice.

Technology and Tool Development for BACPAC: Qualitative and Quantitative Analysis of Accelerated Lumbar Spine MRI with Deep-Learning Based Image Reconstruction at 3T.

OBJECTIVES: To evaluate whether combining fast acquisitions with deep-learning reconstruction can provide diagnostically useful images and quantitative assessment comparable to standard-of-care acquisitions for lumbar spine magnetic resonance imaging (MRI). METHODS: Eighteen patients were imaged with both standard protocol and fast protocol using reduced signal averages, each protocol including sagittal fat-suppressed T2-weighted, sagittal T1-weighted, and axial T2-weighted 2D fast spin-echo sequences. Fast-acquisition data was additionally reconstructed using vendor-supplied deep-learning reconstruction with three different noise reduction factors. For qualitative analysis, standard images as well as fast images with and without deep-learning reconstruction were graded by three radiologists on five different categories. For quantitative analysis, convolutional neural networks were applied to sagittal T1-weighted images to segment intervertebral discs and vertebral bodies, and disc heights and vertebral body volumes were derived. RESULTS: Based on noninferiority testing on qualitative scores, fast images without deep-learning reconstruction were inferior to standard images for most categories. However, deep-learning reconstruction improved the average scores, and noninferiority was observed over 24 out of 45 comparisons (all with sagittal T2-weighted images while 4/5 comparisons with sagittal T1-weighted and axial T2-weighted images). Interobserver variability increased with 50 and 75% noise reduction factors. Deep-learning reconstructed fast images with 50% and 75% noise reduction factors had comparable disc heights and vertebral body volumes to standard images (r2≥ 0.86 for disc heights and r2≥ 0.98 for vertebral body volumes). CONCLUSIONS: This study demonstrated that deep-learning-reconstructed fast-acquisition images have the potential to provide noninferior image quality and comparable quantitative assessment to standard clinical images.

Effect of Clinical Decision Support-Generated Report Cards Versus Real-Time Alerts on Primary Care Provider Guideline Adherence for Low Back Pain Outpatient Lumbar Spine MRI Orders.

OBJECTIVE: The purpose of this study is to determine whether the type of feedback on evidence-based guideline adherence influences adult primary care provider (PCP) lumbar spine (LS) MRI orders for low back pain (LBP). MATERIALS AND METHODS: Four types of guideline adherence feedback were tested on eight tertiary health care system outpatient PCP practices: no feedback during baseline (March 1, 2012-October 4, 2012), randomization by practice to either clinical decision support (CDS)-generated report cards comparing providers to peers only or real-time CDS alerts at order entry during intervention 1 (February 6, 2013-December 31, 2013), and both feedback types for all practices during intervention 2 (January 14, 2014-June 20, 2014, and September 4, 2014-January 21, 2015). International Classification of Disease codes identified LBP visits (excluding Medicare fee-for-service). The primary outcome of the likelihood of LS MRI order being made on the day of or 1-30 days after the outpatient LBP visit was adjusted by feedback type (none, report cards only, real-time alerts only, or both); patient age, sex, race, and insurance status; and provider sex and experience. RESULTS: Half of PCPs (54/108) remained for all three periods, conducting 9394 of 107,938 (8.7%) outpatient LBP visits. The proportion of LBP visits increased over the course of the study (p = 0.0001). In multilevel hierarchic regression, report cards resulted in a lower likelihood of LS MRI orders made the day of and 1-30 days after the visit versus baseline: 38% (p = 0.009) and 37% (p = 0.006) for report cards alone, and 27% (p = 0.020) and 27% (p = 0.016) with alerts, respectively. Real-time alerts alone did not affect MRI orders made the day of (p = 0.585) or 1-30 days after (p = 0.650) the visit. No patient or provider variables were associated with LS MRI orders being generated on the day of or 1-30 days after the LBP visit. CONCLUSION: CDS-generated evidence-based report cards can substantially reduce outpatient PCP LS MRI orders on the day of and 1-30 days after the LBP visit. Real-time CDS alerts do not.

Incidental Renal Lesions on Lumbar Spine MRI: Who Needs Follow-Up?

OBJECTIVE: Incidentally discovered renal lesions on lumbar spine MRI are a common occurrence. Many follow-up recommendations are generated by radiologists encountering renal lesions to help characterize the finding as a benign cyst or a more complex, potentially malignant lesion. We hypothesized that analysis of T2-weighted imaging features of incidentally discovered renal lesions could reliably distinguish complex renal lesions from simple cysts. MATERIALS AND METHODS: Two independent readers retrospectively evaluated 149 renal lesions identified on lumbar spine MRI examinations. Presence or absence of a complex renal lesion was determined using T2-weighted imaging only. Using dedicated renal cross-sectional imaging examinations as the reference standard, statistical analysis was performed to determine the accuracy of lumbar spine MRI in predicting a complex and potentially neoplastic renal lesion. RESULTS: Of 149 renal lesions, 115 were simple cysts, and 34 were complex renal lesions (20 Bosniak II cysts, nine renal cell carcinomas, three Bosniak IIF cysts, and two angiomyolipomas). Lumbar spine MRI readers identified 72 lesions as simple cysts and 77 lesions as complex renal lesions. Reader sensitivity for detection of a complex renal lesion on lumbar spine MRI was 94% (95% CI, 80-99%); specificity, 63% (95% CI, 53-72%); positive predictive value, 43% (95% CI, 37-49%); and negative predictive value, 97% (95% CI, 90-99%). Readers correctly identified all neoplastic and potentially neoplastic lesions (≥ Bosniak IIF). Interreader agreement was excellent (κ = 0.84). CONCLUSION: Follow-up imaging may not be required in all cases of incidentally discovered renal lesions on lumbar spine MRI. Analysis of T2-weighted imaging alone appears to reliably rule out neoplastic and potentially neoplastic complex renal lesions.

Lumbar spine MRI versus non-lumbar imaging modalities in the diagnosis of sacral insufficiency fracture: a retrospective observational study.

BACKGROUND: Sacral insufficiency fractures (SIFs) are a common cause of lower back pain in the elderly. However, because clinical symptoms are frequently vague and nonspecific and can mimic lumbar spine pathologies, initial imaging in SIF patients is frequently targeted at the lumbar spine rather than the sacrum, resulting in delayed diagnosis. The purpose of this study is to show the proportions of modalities used in diagnosing SIF in practice and to compare the clinical and imaging features of SIF diagnosed by lumbar spine MRI (L-spine MRI) with those diagnosed by non-lumbar imaging modalities (bone scan, pelvic bone CT, pelvis MRI). METHODS: Forty-two patients with SIF were enrolled in this study. SIFs diagnosed by L-spine were assigned to group 1 and SIFs diagnosed by non-lumbar imaging modalities (bone scan, pelvic bone CT, pelvis MRI) were assigned to group 2. The clinical and imaging features of SIFs were assessed and compared between two groups. RESULTS: SIF were more commonly diagnosed by L-spine MRI (group 1: n = 27, 64.3%) than non-lumbar imaging modalities (group 2: n = 15, 35.7%), which was comprised of pelvic bone CT (n = 6, 14.3%), bone scan (n = 5, 11.9%), and pelvis MRI (n = 4, 9.5%). Lower back pain, radiating pain and comorbid other causes of pain were more frequently identified in group 1. Fracture involving bilateral sacral ala with horizontal component was the most common shape and S2 being the most commonly involved horizontal component, without significant difference between two groups. CONCLUSION: SIFs are more commonly diagnosed by L-spine MRI than non-lumbar imaging modalities, because of symptoms that mimic lumbar spine pathology and variable comorbid causes of pain. To know that L-spine MRI commonly reveal SIF and to be familiar with SIF features on L-spine MRI would help increase sensitivity in detecting this commonly underrecognized entity and achieve earlier and more appropriate management.

Reducing Inappropriate Lumbar Spine MRI for Low Back Pain: Radiology Support, Communication and Alignment Network.

PURPOSE: The aim of this study is to evaluate the impact of educational sessions on reducing lumbar spine MRI inappropriateness for uncomplicated low back pain and to present our institutional experience on the use of ACR's Radiology Support, Communication and Alignment Network (R-SCAN) program toward achieving appropriateness. METHODS: The R-SCAN web portal was accessed to register a project. Using order entry data, the number of lumbar spine MRI orders placed per month at three family medicine clinics was assessed over a 10-month period. After educational presentations were given at those three clinics highlighting the American College of Physicians and Choosing Wisely campaign imaging guidelines, the number of MRI orders placed was reassessed over an additional 10 months. For a subset of these exams, the ACR Appropriateness Criteria rating of the lumbar spine MRIs were compared between the pre- and posteducation periods. A P value < .05 was considered statistically significant. RESULTS: The average number of monthly MRIs ordered from all three clinics combined was 6.3 during the posteducation period, which was significantly less than during the pre-education period of 10.0 (P = .009). The combined average ACR Appropriateness Criteria rating made at all three clinics was 5.8 after educational sessions, which was significantly higher than the rating of 4.7 before educational sessions (P = .014). CONCLUSION: Clinician education, facilitated by R-SCAN, resulted in a reduced number of MRI lumbar spine studies performed for uncomplicated low back pain and improved appropriateness of those studies as measured by the ACR Appropriateness Criteria rating.

Simultaneous Multislice-Based 5-Minute Lumbar Spine MRI Protocol: Initial Experience in a Clinical Setting.

BACKGROUND AND PURPOSE: Previous studies have used parallel imaging (PI) techniques to decrease spine magnetic resonance imaging (MRI) protocol acquisition times. Recently developed MRI sequences allow even faster acquisitions. Our purpose was to develop a lumbar spine MRI protocol using PI with GRAPPA (generalized autocalibrating partially parallel acquisition) and a simultaneous multislice (SMS)-based sequence and to evaluate its diagnostic performance compared to a standard lumbar spine MRI protocol. METHODS: Ten patients were scanned in a 3-Tesla scanner (MAGNETOM Skyra, Siemens Healthcare). Each patient was imaged using a standard protocol and an optimized fast protocol acquiring the same contrasts and planes. The fast protocol included sagittal T1, T2, and fat suppressed T2 sequences accelerated with GRAPPA and an SMS-based axial T2-weighted sequence using a high-density spine coil (Siemens MR, 30 channel spine). Two blinded neuroradiologists independently assessed image quality and diagnostic accuracy for clinically relevant imaging findings. RESULTS: The fast protocol acquisition time was 5:28 minutes, compared with 16:30 minutes for the standard protocol. Both protocols had a similar performance for definition of anatomical structures, diagnostic quality, and identification of clinically relevant imaging findings. There were more artifacts in the SMS Turbo Spin Echo (P = .014) sequence without compromising diagnostic performance. Artifacts in the remaining non-SMS sequences were similar in both protocols (P > .180). The sensitivity, specificity, and accuracy of the 5-minute protocol were 92.3%, 100.0%, and 99.6%, respectively, for the clinically relevant findings (P = 1.0, interrater agreement .57). CONCLUSIONS: A 5-minute SMS-based MRI protocol for lumbar spine imaging is feasible and can be achieved without significant impact in the overall diagnostic quality.

Limited magnetic resonance imaging of the lumbar spine has high sensitivity for detection of acute fractures, infection, and malignancy.

OBJECTIVE: The objective of this study is to determine how a limited protocol MR examination compares to a full conventional MR examination for the detection of non-degenerative pathology such as acute fracture, infection, and malignancy. MATERIALS AND METHODS: A sample of 349 non-contrast MR exams was selected retrospectively containing a 3:1:1:1 distribution of negative/degenerative change only, acute fracture, infection, and malignancy. This resulted in an even distribution of pathology and non-pathology. A limited protocol MR exam was simulated by extracting T1-weighted sagittal and T2-weighted fat-saturated (or STIR) sagittal sequences from each exam and submitting them for blinded review by two experienced musculoskeletal radiologists. The exams were evaluated for the presence or absence of non-degenerative pathology. Interpretation of the limited exam was compared to the original report of the full examination. If either reader disagreed with the original report, the case was submitted for an unblinded adjudication process with the participation of a third musculoskeletal radiologist to establish a consensus diagnosis. RESULTS: There were five false negatives for a sensitivity of 96.9 % for the limited protocol MR exam. Infection in the psoas, paraspinal muscles, and sacroiliac joint, as well as acute fractures in transverse processes and sacrum were missed by one or more readers. No cases of malignancy were missed. Overall diagnostic accuracy was 96.0 % (335/349). CONCLUSIONS: MR imaging of the lumbar spine limited to sagittal T1-weighted and sagittal T2 fat-saturated (or STIR) sequences has high sensitivity for the detection of acute fracture, infection, or malignancy compared to a conventional MR examination.

Should We Recommend Ultrasonography for an Incidental Thyroid Nodule on Additional Cervicothoracic Sagittal T2-Weighted Image of Lumbar Spine MRI?

PURPOSE: To determine whether we should recommend ultrasonography (US) for an incidental thyroid nodule identified by additional cervicothoracic sagittal T2-weighted image (C-T sag T2WI) of lumbar spine magnetic resonance imaging (MRI). MATERIALS AND METHODS: A retrospective study of 61 patients who underwent both lumbar spine MRI and thyroid US between December 2011 and April 2015 was conducted. For all US-found thyroid nodules > 1 cm, investigators evaluated whether there was any correlation between thyroid nodule detectability by C-T sag T2WI and US features such as echogenicity, composition, or suspicion of malignancy. RESULTS: Solid hypoechoic (2/4; 50%) or mixed echoic nodules (4/8; 50%) appeared to be found relatively more easily by C-T sag T2WI than more benign-looking solid isoechoic (1/4; 25%) or spongiform nodules (0/6; 0%). Among six nodules with ultrasonographic suspicion for malignancy, only one nodule was detected by C-T sag T2WI. CONCLUSION: If an incidental thyroid nodule is seen by C-T sag T2WI, it would be better to recommend thyroid US for identifying malignancy.

The Value of Additional Cervicothoracic Spine Sagittal T2-weighted Images Included in Routine Lumbar Spine MR Imaging

PURPOSE: To evaluate the usefulness of cervicothoracic spine sagittal T2-weighted images (CT SAG T2WIs) included in routine lumbar spine MRI. MATERIALS AND METHODS: Institutional review board approval was obtained and informed consents were waived for this retrospective study. The study group comprised 2,113 patients who underwent lumbar spine MRI from January 2005 to December 2005. CT SAG T2WIs were added in the routine lumbar spine MRIs. Radiologic reports were reviewed retrospectively for pathologic lesions on CT SAG T2WIs by one radiologist. Information of additional cervical or thoracic spine MRI and/or CT for further evaluation of positive findings on CT SAG T2WIs and their treatment were collected by retrospectively reviewing medical records. RESULTS: The CT SAG T2WIs revealed 142 pathologic lesions in 139 (6.58%) of the 2,113 patients. They were easily obtained without positional change in a scan time of less than 2 minutes. Additional cervical or thoracic spine MRI and/or CT for positive findings on CT SAG T2WIs were performed in 13 patients. Seven patients underwent surgical treatment. CONCLUSION: CT SAG T2WIs included in routine lumbar spine MRI were useful in finding the pathologic lesions in cervicothoracic spine for the patients who assumed to have lesions in lumbar spine.