Interobserver Reliability When Classifying MR Imaging of the Lumbar Spine: Written Instructions Alone Do Not Suffice

Detection and Characterization of Endplate Structural Defects on CT: A Diagnostic Accuracy Study

STUDY DESIGN

Diagnostic test study.

OBJECTIVE

To determine the reliability and validity or diagnostic accuracy of two previously described endplate structural defect (EPSD) assessment methods.

SUMMARY OF BACKGROUND DATA

Studies of EPSD may further the understanding of pathoanatomic mechanisms underlying back pain. However, clinical imaging methods used to document EPSD have not been validated, leaving uncertainty about what the observations represent.

MATERIALS AND METHODS

Using an evaluation manual, 418 endplates on CT sagittal slices obtained from 19 embalmed cadavers (9 men and 10 women, aged 62-91 yr) were independently assessed by two experienced radiologists and a novice for EPSD using the two methods. The corresponding micro-CT (µCT) from the harvested T7-S1 spines were assessed by another independent rater with excellent intra-rater reliability (k=0.96).

RESULTS

Inter-rater reliability was good for the presence (k=0.60-0.69) and fair for specific phenotypes (k=0.43-0.58) of EPSD. Erosion, for which the Brayda-Bruno classification lacked a category, was mainly (82.8%) classified as wavy/irregular, while many notched defects (n=15, 46.9%) and Schmorl's nodes (n=45, 79%) were recorded as focal defects using Feng's classification. When compared to µCT, endplate fractures (n=53) and corner defects (n=28) were routinely missed on CT. Endplates classified as wavy/irregular on CT corresponded to erosion (n=29, 21.2%), jagged defects (n=21, 15.3%), calcification (n=19, 13.9%), and other phenotypes on µCT. Some focal defects on CT represented endplate fractures (n=21, 27.6%) on µCT. Overall, with respect to the presence of an EPSD, there was a sensitivity of 70.9% and a specificity of 79.1% using Feng's method, and 79.5% and 57.5% using Brayda-Bruno's method. Poor to fair inter-rater reliability (k=0.26-0.47) was observed for defect dimensions.

CONCLUSION

There was good inter-rater reliability and evidence of criterion validity supporting assessments of EPSD presence using both methods. However, neither method contained all needed EPSD phenotypes for optimal sensitivity, and specific phenotypes were often misclassified.

Grading lumbar foraminal stenosis - Interrater agreement of radiologists and radiology trainees before and after education of a standardised grading scale

INTRODUCTION

Lumbar foraminal stenosis is a key contributor to low back pain. Imaging, particularly MRI, is commonly used in the assessment of foraminal stenosis, contributing to treatment planning. The adoption of a standardised grading system to try and improve inter-rater agreement is thought to be of importance. Our study aims to assess the variability of grading lumbar foraminal stenosis amongst reporting doctors, determine whether education about a validated grading scale increases agreement, and determine if these changes persist over time.

METHODS

A single-site study involving MRI reporting registrars/radiologists was performed. Participants were shown select MRI images and asked to grade the degree of stenosis in each on a 4-point scale. Subsequently, they were educated about Lee et al's grading system and asked to re-grade the cases 1 and 6 weeks later. The level of agreement was calculated using Gwet's AC1 coefficient and Krippendorff's Alpha.

RESULTS

The baseline level of agreement was substantial (AC1 = 0.71). This decreased to a moderate level of agreement post-intervention (AC1 = 0.575 at 1-week, P-value 0.033 and AC1 = 0.598 at 6 weeks, P-value 0.012). A grading of severe stenosis was 21% more likely 6 weeks post-education.

CONCLUSION

The baseline agreement at our institution was substantial, thought to be due to the single-centre nature of the study. Moderate agreement was achieved after education regarding the Lee et al.'s scale, in-line with other studies, with changes maintained at 6 weeks, showing retention of the scale parameters. Grading of severe stenosis was more common post intervention.

Reliability and Accuracy of the Outerbridge Classification in Staging of Cartilage Defects

OBJECTIVE

The decision on whether or not and how to treat a local cartilage defect is still made intraoperatively based on the visual presentation of the cartilage and findings from indentations with an arthroscopic probe. The treatment decision is then usually based on grading according to established classifications systems, which, therefore, need to have high reliability and accuracy. The aim of the present study was to evaluate the reliability and accuracy of the Outerbridge classification in staging cartilage defects.

METHODS

We performed an observer arthroscopic study using the Outerbridge classification on seven fresh-frozen human cadaveric knees, which collectively exhibited nine cartilage defects. To evaluate accuracy, defect severity was verified through histological examination. Interrater and intrarater reliabilites were calculated using Cohen's kappa and the intra-class correlation coefficient (ICC 3.1).

RESULTS

The interrater and intrarater reliability for the Outerbridge classification ranged from poor to substantial, with 0.24 ≤ κ ≤ 0.70 and κ = 0.55 to κ = 0.66, respectively. The accuracy evaluated by comparison with the histological examination was 63% overall. The erroneous evaluations were, however, still often at the discrimination of grade 2 and 3. We did not find any relationship between higher experience and accuracy or intraobserver reliability. Taken together, these results encourage surgeons to further use diagnostic arthroscopy for evaluating cartilage lesions. Nevertheless, especially in grade 2 and 3, deviations from the histology were observed. This is, however, the point where a decision is made on whether to surgically address the defect or not.

CONCLUSION

Diagnostic arthroscopy is the standard for cartilage lesion assessment, yet interobserver reliability is fair to substantial. Caution is warranted in interpreting varied observer results. The accuracy of the "simpler" Outerbridge classification is insufficient compared to histological examinations, highlighting the need for improved techniques in guideline-based intraoperative decision-making.

A systematic review of validated classification systems for cervical and lumbar spinal foraminal stenosis based on magnetic resonance imaging

PURPOSE

Foraminal stenosis is commonly investigated with radiological methods in patients with radiating pain in extremities. However, there is a lack of consensus regarding the methodology to assess compression of the nerve roots. This systematic review was performed to identify validated classification systems for foraminal stenosis in the lumbar and cervical spine based on magnetic resonance imaging (MRI).

METHODS

A systematic search was conducted according to the PRISMA guidelines. The search included Cochrane, Embase, Medline and PubMed databases going back 30 years and up to September 2021. Three categories of words were used in different variations; foraminal stenosis, MRI and scoring. For inclusion, at least one word from each category had to be present. Articles suggesting classification systems or reporting on their validation were selected for inclusion.

RESULTS

A total of 823 articles were identified and all abstracts were reviewed. Subsequently, a full-text review of 64 articles was performed and finally 14 articles were included. A total of three validated classification systems were found for the cervical and lumbar spine. The remaining 11 articles reported on validation or suggested modifications of the classification systems.

CONCLUSION

The three classification systems demonstrated moderate to good reliability and have all been shown feasible in the clinical setting. There is however a need for further studies testing the validity of these classifications in relation to both clinical findings and to surgical outcome data.

Spine Medical Image Segmentation Based on Deep Learning

The aim was to further explore the clinical value of deep learning algorithm in the field of spinal medical image segmentation, and this study designed an improved U-shaped network (BN-U-Net) algorithm and applied it to the spinal MRI medical image segmentation of 22 research objects. The application value of this algorithm in MRI image processing was comprehensively evaluated by accuracy (Acc), sensitivity (Sen), specificity (Spe), and area under curve (AUC). The results show that the image processing time of fully convolutional network (FCN) algorithm and U-Net algorithm is greater than 6 min, while the processing time of BN-U-Net algorithm is only 5-10 s, and the processing time is significantly shortened (P < 0.05). The Acc, Sen, and Spe results of BN-U-Net segmentation algorithm were 94.54 ± 3.56%, 88.76 ± 2.67%, and 86.27 ± 6.23%, respectively, which were significantly improved compared with FCN algorithm and U-Net algorithm (P < 0.05). In summary, the improved U-Net network algorithm used in this study significantly improves the quality of spinal MRI images by automatic segmentation of MRI images, which is worthy of further promotion in the field of spinal medical image segmentation.

Introduction and reproducibility of an updated practical grading system for lumbar foraminal stenosis based on high-resolution MR imaging

In this paper we sought to develop and assess the reproducibility of an updated 6-point grading system for lumbar foraminal stenosis based on the widely used Lee classification that more accurately describes lumbar foraminal stenosis as seen on high-resolution MRI. Grade A indicates absence of foraminal stenosis. Grades B, C, D and E indicate presence of foraminal stenosis with contact of the nerve root with surrounding anatomical structures (on one, two, three or four sides for B, C, D and E respectively) yet without morphological change of the nerve root. To each grade, a number code indicating the location of contact between the nerve root and surrounding anatomical structure(s) is appended. 1, 2, 3 and 4 indicate contact of the nerve root at superior, posterior, inferior and anterior position of the borders of the lumbar foramen. Grade F indicates presence of foraminal stenosis with morphological change of the nerve root. Three readers graded the lumbar foramina of 101 consecutive patients using high-resolution T2w (and T1w) MR images with a spatial resolution of beyond 0.5 mm³. Interreader agreement was excellent (Cohen’s Kappa = 0.866–1). Importantly, 30.6%/31.6%/32.2% (reader 1/reader 2/ reader 3) of foramina were assigned grades that did not appear in the original Lee grading system (grades B and D). The readers found no foramen that could not be described accurately with the updated grading system. Thus, an updated 6-point grading system for lumbar foraminal stenosis is reproducible and comprehensively describes lumbar foraminal stenosis as seen on high-resolution MRI.

Relationship between Machine-Learning Image Classification of T2-Weighted Intramedullary Hypointensity on 3 Tesla Magnetic Resonance Imaging and Clinical Outcome in Dogs with Severe Spinal Cord Injury

Early prognostic information in cases of severe spinal cord injury can aid treatment planning and stratification for clinical trials. Analysis of intraparenchymal signal change on magnetic resonance imaging has been suggested to inform outcome prediction in traumatic spinal cord injury. We hypothesized that intraparenchymal T₂-weighted hypointensity would be associated with a lower potential for functional recovery and a higher risk of progressive neurological deterioration in dogs with acute, severe, naturally occurring spinal cord injury. Our objectives were to: 1) demonstrate capacity for machine-learning criteria to identify clinically relevant regions of hypointensity and 2) compare clinical outcomes for cases with and without such regions. A total of 95 dogs with complete spinal cord injury were evaluated. An image classification system, based on Speeded-Up Robust Features (SURF), was trained to recognize individual axial T₂-weighted slices that contained hypointensity. The presence of such slices in a given transverse series was correlated with a lower chance of functional recovery (odds ratio [OR], 0.08; confidence interval [CI], 0.02-0.38; p < 10^-3) and with a higher risk of neurological deterioration (OR, 0.14; 95% CI, 0.05-0.42; p < 10^-3). Identification of intraparenchymal T₂-weighted hypointensity in severe, naturally occurring spinal cord injury may be assisted by an image classification tool and is correlated with functional recovery.