Whole body magnetic resonance imaging in healthy children and adolescents. Bone marrow appearances of the axial skeleton

Update on MRI in Rheumatic Diseases

Over the past decade, MRI has significantly advanced the diagnosis of rheumatic disease in both adults and juveniles. In this article, the authors present an update on MRI applications in rheumatology, based on a review of the most recent publications. New developments in adults related to, among others, axial spondyloarthritis, peripheral arthritis, and the whole body-MRI (WB-MRI) are presented. In juveniles, this update addresses the latest advancements in diagnostic MRI of peripheral joints, followed by MRI of the axial skeleton and implementation of the WB-MRI for the screening of inflammation. The authors also discuss topics of interest concerning contrast-enhanced MRI examinations in children.

Preferential involvement of the pelvis and hips along with active sacroiliitis in chronic nonbacterial osteomyelitis: MRI of 97 patients from a single tertiary referral center

OBJECTIVE

To present MRI distribution of active osteitis in a single tertiary referral center cohort of patients with chronic nonbacterial osteomyelitis (CNO).

METHODS

Two musculoskeletal radiologists retrospectively reviewed MRI examinations of all patients with a final clinical diagnosis of CNO over 15 years. Sites of active osteitis at any time during the course of disease were divided into seven groups: (A) mandible, sternum, clavicles, or scapulas; (B) upper extremities; (C) subchondral sacrum and ilium immediately subjacent to sacroiliac joints (active osteitis denoting "active sacroiliitis" here); (D) pelvis and proximal 1/3 of femurs (excluding group C); (E) bones surrounding knees including distal 2/3 of femurs and 1/2 of proximal tibias and fibulas; (F) distal legs (including distal 1/2 of tibias and fibulas), ankles, or feet; (G) spine (excluding group C). Temporal changes of lesions in response to treatment (or other treatment-related changes such as pamidronate lines) were not within the scope of the study.

RESULTS

Among 97 CNO patients (53 males [55%], 44 females; age at onset, mean ± SD, 8.5 ± 3.2 years; age at diagnosis, 10.3 ± 3.3 years), whole-body (WB) MRI was performed in 92%, mostly following an initial targeted MRI (94%). A total of 557 (346 targeted and 211 WB) MRIs were analyzed. Biopsy was obtained in 39 patients (40%), all consistent with CNO or featuring supporting findings. The most common locations for active osteitis were groups D (78%; 95% CI 69‒85%) and C (72%; 95% CI 62‒80%).

CONCLUSION

Pelvis and hips were preferentially involved in this cohort of CNO patients along with a marked presence of active sacroiliitis.

CLINICAL RELEVANCE STATEMENT

When suggestive findings of CNO are identified elsewhere in the body, the next targeted site of MRI should be the pelvis (entirely including sacroiliac joints) and hips, if whole-body MRI is not available or feasible.

KEY POINTS

• Heavy reliance on MRI for diagnosis of CNO underscores the importance of suggestive distribution patterns. • Pelvis and hips are the most common (78%) sites of CNO involvement along with active sacroiliitis (72%). • Pelvis including sacroiliac joints and hips should be targeted on MRI when CNO is suspected.

Whole-Body MRI at Initial Presentation of Chronic Recurrent Multifocal Osteomyelitis, Juvenile Idiopathic Arthritis, Their Overlapping Syndrome, and Non-Specific Arthropathy

(1) Background: Whole-body magnetic resonance imaging (WB-MRI) is central to defining total inflammatory burden in juveniles with arthritis. Our aim was to determine and compare the initial distribution of lesions in the WB-MRI in patients with chronic recurrent multifocal osteomyelitis (CRMO), juvenile idiopathic arthritis (JIA), their overlapping syndrome (OS), and with Non-specific Arthropathy (NA). (2) Methods: This retrospective single center study was performed on an Avanto 1.5-T MRI scanner with a dedicated multichannel surface coil system. A total of 173 pediatric patients were included with the following final diagnoses: CRMO (15.0%), JIA (29.5%), OS (4.6%), and NA (50.9%). (3) Results: Bone marrow edema (BME) was the most common abnormality, being seen in 100% patients with CRMO, 88% with OS, 55% with JIA, and 11% with NA. The bones of the lower extremities were the most affected in all compared entities. Effusion was seen in 62.5% children with OS, and in 52.9% with JIA, and in CRMO and NA, the exudate was sporadic. Enthesitis was found in 7.8% of patients with JIA and 3.8% with CRMO, and myositis was seen in 12.5% of patients with OS and in 3.9% with JIA. (4) Conclusions: The most frequent indication for WB-MRI in our center was JIA. The most common pathology in all rheumatic entities was BME, followed by effusion mainly seen in in OS and JIA. Enthesitis and myositis were less common; no case was observed in NA.

Machine Learning Algorithm: Texture Analysis in CNO and Application in Distinguishing CNO and Bone Marrow Growth-Related Changes on Whole-Body MRI

OBJECTIVE

The purpose of this study is to analyze the texture characteristics of chronic non-bacterial osteomyelitis (CNO) bone lesions, identified as areas of altered signal intensity on short tau inversion recovery (STIR) sequences, and to distinguish them from bone marrow growth-related changes through Machine Learning (ML) and Deep Learning (DL) analysis.

MATERIALS AND METHODS

We included a group of 66 patients with confirmed diagnosis of CNO and a group of 28 patients with suspected extra-skeletal systemic disease. All examinations were performed on a 1.5 T MRI scanner. Using the opensource 3D Slicer software version 4.10.2, the ROIs on CNO lesions and on the red bone marrow were sampled. Texture analysis (TA) was carried out using Pyradiomics. We applied an optimization search grid algorithm on nine classic ML classifiers and a Deep Learning (DL) Neural Network (NN). The model's performance was evaluated using Accuracy (ACC), AUC-ROC curves, F1-score, Positive Predictive Value (PPV), Mean Absolute Error (MAE) and Root-Mean-Square Error (RMSE). Furthermore, we used Shapley additive explanations to gain insight into the behavior of the prediction model.

RESULTS

Most predictive characteristics were selected by Boruta algorithm for each combination of ROI sequences for the characterization and classification of the two types of signal hyperintensity. The overall best classification result was obtained by the NN with ACC = 0.91, AUC = 0.93 with 95% CI 0.91-0.94, F1-score = 0.94 and PPV = 93.8%. Between classic ML methods, ensemble learners showed high model performance; specifically, the best-performing classifier was the Stack (ST) with ACC = 0.85, AUC = 0.81 with 95% CI 0.8-0.84, F1-score = 0.9, PPV = 90%.

CONCLUSIONS

Our results show the potential of ML methods in discerning edema-like lesions, in particular by distinguishing CNO lesions from hematopoietic bone marrow changes in a pediatric population. The Neural Network showed the overall best results, while a Stacking classifier, based on Gradient Boosting and Random Forest as principal estimators and Logistic Regressor as final estimator, achieved the best results between the other ML methods.

Whole-body MRI in children and adolescents: Can T2-weighted Dixon fat-only images replace standard T1-weighted images in the assessment of bone marrow?

OBJECTIVE

When performing whole-body MRI for bone marrow assessment in children, optimizing scan time is crucial. The aim was to compare T2 Dixon fat-only and TSE T1-weighted sequences in the assessment of bone marrow high signal areas seen on T2 Dixon water-only in healthy children and adolescents.

MATERIALS AND METHODS

Whole-body MRIs from 196 healthy children and adolescents aged 6 to 19 years (mean 12.0) were obtained including T2 TSE Dixon and T1 TSE-weighted images. Areas with increased signal on T2 Dixon water-only images were scored using a novel, validated scoring system and classified into "minor" or "major" findings according to size and intensity, where "major" referred to changes easily being misdiagnosed as pathology in a clinical setting. Areas were assessed for low signal on T2 Dixon fat-only images and, after at least three weeks to avoid recall bias, on the T1-weighted sequence by two experienced pediatric radiologists.

RESULTS

1250 high signal areas were evaluated on T2 Dixon water-only images. In 1159/1250 (92.7%) low signal was seen on both T2 Dixon fat-only and T1-weighted sequences while in 24 (1.9%) it was not present on either sequence, with an absolute agreement of 94.6%. Discordant findings were found in 67 areas, of which in 18 (1.5%) low signal was visible on T1-weighted images alone and in 49 (3.9%) on T2 Dixon fat-only alone. The overall kappa value between the two sequences was 0.39. The agreement was higher for major as compared to minor findings (kappa values of 0.69 and 0.29, respectively) and higher for the older age groups.

CONCLUSION

T2 Dixon fat-only can replace T1-weighted sequence on whole-body MRI for bone marrow assessment in children over the age of nine, thus reducing scan time.