Actual applications of magnetic resonance imaging in dentomaxillofacial region

Salivary gland tumor detection from saliva to theranostic application of surface-enhanced Raman spectroscopy

Surface-enhanced Raman spectroscopy (SERS) can be considered a rapid, label-free, nondestructive analytical measurement for tumor detection and theranostic applications, beginning from diagnosis as well as tumor treatment and recovery. SERS of saliva samples collected from patients with salivary gland tumors and healthy controls were used to establish a new tool for fast diagnosis before surgery and in follow-up surgery results. The Partial Least Squares Regression (PLSR) method divided the two analyzed data sets, namely the saliva of control patients and those with salivary gland tumors, with 96 % of explained variables in the first three consecutive factors. The outcome indicates the prediction ability of the analyzed model as the low value of root mean square error (cross-validation; RMSE(CV) = 0.11) and high values of R-squared (cross-validation; R2(CV) = 0.95) were obtained. The calibration models were created and optimized using other supervised methods, e.g., partial least squares-discriminant analysis, support vector machine classification, and linear discriminant analysis-principal component analysis. Then, their classification abilities were tested with external samples, achieving impressive accuracy. The study showed that the SERS spectra of the two analyzed classes related to the patient's disease state showed significant differences, allowing the discrimination between them and identifying the external sample.

3D brain tumor segmentation based on a novel nettree merging

Brain tumors rank among the most devastating diseases in the human body, with their growth potentially resulting in impaired brain tissue function and even life-threatening scenarios, profoundly impacting patients physical and mental well-being. Presently, segmentation methods for three-dimensional (3D) brain tumors can better reflect the position and relationship of the entire brain tissue, providing more reference information. However, challenges persist with existing 3D segmentation techniques, including the lack of direct manipulation on 3D images and suboptimal performance on clinical datasets. Addressing this, this paper proposes a 3D brain tumor segmentation method based on nettree merging. This approach leverages a topological determination between body targets to establish relationships, constructs a nettree structure based on pathological topology, and subsequently merges topological block according to its structure and intensity, thereby segmenting the Gross Tumor Volume (GTV) of 3D MRI brain tumor images. The proposed method is validated using Dice, HD95, Precision, and Recall, and the values of these metrics on the clinical dataset are approximately 0.824, 10.981, 0.829, 0.834, while on the BraTS dataset, they are close to 0.873, 4.902, 0.871, 0.864. Experimental validation on both public and clinical datasets substantiates the effectiveness of the proposed method.

Detecting the articular disk in magnetic resonance images of the temporomandibular joint using YOLO series

The purpose of this study was to construct an artificial intelligence object detection model to detect the articular disk from temporomandibular joint (TMJ) magnetic resonance (MR) images using YOLO series. The study included two experiments using datasets from different MR imaging machines. A total of 536 MR images were retrospectively examined. The performance of YOLOv5 and YOLOv8 in detecting the TMJ articular disk in both normal and displaced conditions was evaluated. The impact of image-processing techniques, such as histogram equalization (HE) and contrast-limited adaptive HE (CLAHE) on model performance, was also examined. The results showed that the YOLO series could detect the articular disk regardless of displacement, with superior performance on images of normal disk position. The results suggest the applicability of object detection models in improving the diagnosis of TMJ disorders.

Deep learning-based automated diagnosis of temporomandibular joint anterior disc displacement and its clinical application

Introduction

This study aimed to develop a deep learning-based method for interpreting magnetic resonance imaging (MRI) scans of temporomandibular joint (TMJ) anterior disc displacement (ADD) and to formulate an automated diagnostic system for clinical practice.

Methods

The deep learning models were utilized to identify regions of interest (ROI), segment TMJ structures including the articular disc, condyle, glenoid fossa, and articular tubercle, and classify TMJ ADD. The models employed Grad-CAM heatmaps and segmentation annotation diagrams for visual diagnostic predictions and were deployed for clinical application. We constructed four deep-learning models based on the ResNet101_vd framework utilizing an MRI dataset of 618 TMJ cases collected from two hospitals (Hospitals SS and SG) and a dataset of 840 TMJ MRI scans from October 2022 to July 2023. The training and validation datasets included 700 images from Hospital SS, which were used to develop the models. Model performance was assessed using 140 images from Hospital SS (internal validity test) and 140 images from Hospital SG (external validity test). The first model identified the ROI, the second automated the segmentation of anatomical components, and the third and fourth models performed classification tasks based on segmentation and non-segmentation approaches. MRI images were classified into four categories: normal (closed mouth), ADD (closed mouth), normal (open mouth), and ADD (open mouth). Combined findings from open and closed-mouth positions provided conclusive diagnoses. Data augmentation techniques were used to prevent overfitting and enhance model robustness. The models were assessed using performance metrics such as precision, recall, mean average precision (mAP), F1-score, Matthews Correlation Coefficient (MCC), and confusion matrix analysis.

Results

Despite lower performance with Hospital SG's data than Hospital SS's, both achieved satisfactory results. Classification models demonstrated high precision rates above 92%, with the segmentation-based model outperforming the non-segmentation model in overall and category-specific metrics.

Discussion

In summary, our deep learning models exhibited high accuracy in detecting TMJ ADD and provided interpretable, visualized predictive results. These models can be integrated with clinical examinations to enhance diagnostic precision.

Magnetic resonance imaging phantom-based S1 vertebral scores are indicators of fat-water-like osteoporotic changes in postmenopausal women: a pilot study

STUDY DESIGN

A prospective study.

PURPOSE

To assess fat-water-like tissue changes on the 1st sacral vertebra using novel magnetic resonance imaging (MRI) phantombased F- and W-scores and evaluate their diagnostic performances in osteoporosis detection.

OVERVIEW OF LITERATURE

Using an uncommonly advanced MRI technique, previous studies have found that fat-water changes were consistent with osteoporosis. The role of routine MRI sequences can be extended in this regard. The S1 vertebra is considered a crucial anatomical site in spine surgeries because it seldom suffers from fractures. Thus, S1 could indicate osteoporotic fat-water changes.

METHODS

Forty-two female volunteers (aged 62.3±6.3 years) underwent spine examination with both MRI (including a phantom) and dual-energy X-ray absorptiometry (DXA) following ethical approval. MRI phantom-based F- and W-scoreS1 were defined by normalizing S1 vertebral signal intensities (SIs) by coconut oil and water SIs of the phantom on T1- and T2-weighted imaging, respectively. Using receiver operating characteristic analysis, the diagnostic performances of the new scores for evaluating osteoporosis and vertebral fractures were investigated against standard areal bone mineral density measured with DXA (DXA-aBMD).

RESULTS

The F-scoreS1 and W-scoreS1 were greater (4.11 and 2.43, respectively) in patients with osteoporosis than those without osteoporosis (3.25 and 1.92, respectively) and achieved areas under the curve (AUCs) of 0.82 and 0.76 (p<0.05), respectively, for osteoporosis detection. Similarly, the mean F-scoreS1 and W-scoreS1 were higher (4.11 and 2.63, respectively) in patients with vertebral fractures than in those without fractures (3.30 and 1.82, respectively) and had greater AUCs (0.90 for W-scoreS1 and 0.74 for F-scoreS1) than DXA-aBMD (AUC, 0.26; p<0.03). In addition, the F- and W-scoreS1 demonstrated a strong correlation (r=0.65, p<0.001).

CONCLUSIONS

The new S1 vertebral-based MRI scores were developed to detect osteoporotic changes and demonstrated improvements over DXA-aBMD in differentiating patients with vertebral fractures.

Image preprocessing with contrast-limited adaptive histogram equalization improves the segmentation performance of deep learning for the articular disk of the temporomandibular joint on magnetic resonance images

OBJECTIVES

The objective was to evaluate the robustness of deep learning (DL)-based encoder-decoder convolutional neural networks (ED-CNNs) for segmenting temporomandibular joint (TMJ) articular disks using data sets acquired from 2 different 3.0-T magnetic resonance imaging (MRI) scanners using original images and images subjected to contrast-limited adaptive histogram equalization (CLAHE).

STUDY DESIGN

In total, 536 MR images from 49 individuals were examined. An expert orthodontist identified and manually segmented the disks in all images, which were then reviewed by another expert orthodontist and 2 expert oral and maxillofacial radiologists. These images were used to evaluate a DL-based semantic segmentation approach using an ED-CNN. Original and preprocessed CLAHE images were used to train and validate the models whose performances were compared.

RESULTS

Original and CLAHE images acquired on 1 scanner had pixel values that were significantly darker and with lower contrast. The values of 3 metrics-the Dice similarity coefficient, sensitivity, and positive predictive value-were low when the original MR images were used for model training and validation. However, these metrics significantly improved when images were preprocessed with CLAHE.

CONCLUSIONS

The robustness of the ED-CNN model trained on a dataset obtained from a single device is low but can be improved with CLAHE preprocessing. The proposed system provides promising results for a DL-based, fully automated segmentation method for TMJ articular disks on MRI.

Assessing osteoporosis in postmenopausal women: preliminary results using a novel lumbar spine phantom-based MRI scoring method

OBJECTIVE

To develop a novel magnetic resonance imaging (MRI) phantom for producing F-score (for fat) and W-score (for water) and to evaluate the performance of these scores in assessing osteoporosis and related vertebral fractures.

MATERIALS AND METHODS

First, a real-time phantom consisting of oil and water tubes was manufactured. Then, 30 female volunteers (age: 62.3 ± 6.3 years) underwent lumbar spine examination with MRI (using a novel phantom) and dual-energy X-ray absorptiometry (DXA), following ethical approval. MRI phantom-based F-score and W-score were defined by normalizing the vertebral signal intensities (SIs) by the oil and water SIs of the phantom on T1- and T2-weighted images, respectively. The diagnostic performances of the new scores for assessing osteoporosis and vertebral fractures were examined using receiver operating characteristic analysis and compared with DXA-measured areal bone mineral density (DXA-aBMD).

RESULTS

The F-score and W-score were greater in the osteoporotic patients (3.93 and 2.29) than the non-osteoporotic subjects (3.05 and 1.79) and achieved AUC values of 0.85 and 0.74 (p < 0.05), respectively, when detecting osteoporosis. Similarly, F-score and W-score had greater values for the fracture patients (3.94 and 2.53) than the non-fracture subjects (3.14 and 1.69) and produced better AUC values (0.90 for W-score and 0.79 for F-score) compared to DXA-aBMD (AUC: 0.27, p < 0.05). In addition, the F-score and W-score had a strong correlation (r = 0.77; p < 0.001).

CONCLUSION

A novel real-time lumber spine MRI phantom was developed, based upon which newly defined F-score and W-score were able to detect osteoporosis and demonstrated an improved ability over DXA-aBMD in differentiating patients with vertebral fractures.

Surveying the landscape of diagnostic imaging in dentistry's future: Four emerging technologies with promise

BACKGROUND

Advances in digital radiography for both intraoral and panoramic imaging and cone-beam computed tomography have led the way to an increase in diagnostic capabilities for the dental care profession. In this article, the authors provide information on 4 emerging technologies with promise.

TYPES OF STUDIES REVIEWED

The authors feature the following: artificial intelligence in the form of deep learning using convolutional neural networks, dental magnetic resonance imaging, stationary intraoral tomosynthesis, and second-generation cone-beam computed tomography sources based on carbon nanotube technology and multispectral imaging. The authors review and summarize articles featuring these technologies.

RESULTS

The history and background of these emerging technologies are previewed along with their development and potential impact on the practice of dental diagnostic imaging. The authors conclude that these emerging technologies have the potential to have a substantial influence on the practice of dentistry as these systems mature. The degree of influence most likely will vary, with artificial intelligence being the most influential of the 4.

CONCLUSIONS AND PRACTICAL IMPLICATIONS

The readers are informed about these emerging technologies and the potential effects on their practice going forward, giving them information on which to base decisions on adopting 1 or more of these technologies. The 4 technologies reviewed in this article have the potential to improve imaging diagnostics in dentistry thereby leading to better patient care and heightened professional satisfaction.

Magnetic resonance imaging for jawbone assessment: a systematic review

PURPOSE

To evaluate the accuracy of magnetic resonance imaging (MRI) for jawbone assessment compared to reference-standard measurements in the literature.

MATERIALS AND METHODS

An electronic database search was conducted in PubMed, EMBASE, Scopus, Web of Science, and the Cochrane Library in June 2022, and updated in August 2023. Studies evaluating the accuracy of MRI for jawbone assessment compared with reference-standard measurements (histology, physical measurements, or computed tomography) were included. The outcome measures included bone histomorphometry and linear measurements. The risk of bias was assessed by the Quality Assessment Tool for Diagnostic Accuracy Studies (QUADAS-2). The review was registered in the PROSPERO database (CRD42022342697).

RESULTS

From 63 studies selected for full-text analysis, nine manuscripts were considered eligible for this review. The studies included assessments of 54 participants, 35 cadavers, and one phantom. A linear measurement error ranging from 0.03 to 3.11 mm was shown. The accuracy of bone histomorphometry varies among studies. Limitations of the evidence included heterogeneity of MRI protocols and the methodology of the included studies.

CONCLUSION

Few studies have suggested the feasibility of MRI for jawbone assessment, as MRI provides comparable results to those of standard reference tests. However, further advancements and optimizations are needed to increase the applicability, validate the efficacy, and establish clinical utility of these methods.

Non-destructive visualization of internal structural changes in humidified magnesium oxide tablets using X-ray computed tomography

Detailed examinations of the internal structure of tablets are imperative for comprehending their formulation, physical attributes, and ensuring their safe utilization. While X-ray computed tomography (CT) is valuable for noninvasively analyzing internal structural changes, the influence of humidity on these structural changes remains unexplored. Accordingly, we aimed to assess the viability of X-ray CT in non-destructively evaluating the internal structure of humidified magnesium oxide (MgO) tablets. MgO tablets were subjected to conditions of 40 °C and 75% humidity for 7 days, weighed pre- and post-humidification, and subsequently stored at room temperature (22-27 °C) until day 90. Their internal structure was evaluated using X-ray CT. We observed a substantial increase in the weight of MgO tablets concomitant with moisture absorption, with minimal changes observed upon storage at room temperature. The skewness reduced immediately post-moisture absorption, remained almost the same post-storage at room temperature, and failed to revert to pre-humidification levels during the storage period. These findings highlight the utility of X-ray CT as an effective tool for non-destructive, three-dimensional, and detailed evaluation of internal structural transformations in MgO tablets.

MRI-based cephalometrics: a scoping review of current insights and future perspectives

OBJECTIVE

This review aims to explore the current status of magnetic resonance imaging (MRI) as a cephalometric tool, summarize the equipment design and methods, and propose recommendations for future research.

METHODS

A systematic search was conducted in electronic databases, including PubMed, Ovid MEDLINE, Scopus, Embase, Web of Science, EBSCOhost, LILACS, and Cochrane Library, using broad search terms. The articles published in any language till June 2022 were considered. Cephalometric studies conducted using the MRI dataset on human participants, phantom or cadaver were included. Two independent reviewers assessed the final eligible articles using the quality assessment score (QAS).

RESULTS

Nine studies were included in the final assessment. Studies used various methods, including 1.5 T or 3 T MRI systems and 3D or 2D MRI datasets. Among the imaging sequences, T1-weighted, T2-weighted and black bone MR images were used for cephalometric analysis. In addition, the reference standards varied among studies, such as traditional 2D cephalogram, cone-beam CT and phantom measurements. The mean QAS of all the included studies was 79% (± 14.4%). The main limitation of most studies was the small sample size and the heterogeneity of the methods, statistical tools used, and metric outcomes assessed.

CONCLUSIONS

Despite the heterogeneity and lack of metrological evidence on the effectiveness of MRI-based cephalometric analysis, the preliminary results demonstrated by in vivo and in vitro studies are encouraging. However, future studies exploring MRI sequences specific to cephalometric diagnosis are required for wider adoption of this technique in routine orthodontic practice.

Performance of PROPELLER FSE T2WI in reducing metal artifacts of material porcelain fused to metal crown: a clinical preliminary study

This study aimed to compare MRI quality between conventional fast spin echo T₂ weighted imaging (FSE T₂WI) with periodically rotated overlapping parallel lines with enhanced reconstruction (PROPELLER) FSE T₂WI for patients with various porcelain fused to metal (PFM) crown and analyze the value of PROPELLER technique in reducing metal artifacts. Conventional FSE T₂WI and PROPELLER FSE T₂WI sequences for axial imaging of head were applied in participants with different PFM crowns: cobalt-chromium (Co–Cr) alloy, pure titanium (Ti), gold–palladium (Au–Pd) alloy. Two radiologists evaluated overall image quality of section in PFM using a 5-point scale qualitatively and measured the maximum artifact area and artifact signal-to-noise ratio (SNR) quantitatively. Fifty-nine participants were evaluated. The metal crown with the least artifacts and the optimum image quality shown in conventional FSE T₂WI and PROPELLER FSE T₂WI were in Au–Pd alloy, Ti, and Co–Cr alloy order. PROPELLER FSE T₂WI was superior to conventional FSE T₂WI in improving image quality and reducing artifact area for Co-Cr alloy (17.0 ± 0.2% smaller artifact area, p < 0.001) and Ti (11.6 ± 0.7% smaller artifact area, p = 0.005), but had similar performance compared to FSE T₂WI for Au–Pd alloy. The SNRs of the tongue and masseter muscle were significantly higher on PROPELLER FSE T₂WI compared with conventional FSE T₂WI (tongue: 29.76 ± 8.45 vs. 21.54 ± 9.31, p = 0.007; masseter muscle: 19.11 ± 8.24 vs. 15.26 ± 6.08, p = 0.016). Therefore, the different PFM crown generate varying degrees of metal artifacts in MRI, and the PROPELLER can effectively reduce metal artifacts especially in the PFM crown of Co-Cr alloy.

Bilateral parotid glands aplasia: a case report and literature review

OBJECTIVES

Bilateral parotid gland aplasia is a rare congenital anomaly that almost consistently leads to xerostomia and caries. It is often associated with other congenital craniofacial abnormalities. The objective was to describe a case with asymptomatic bilateral parotid gland aplasia and to review previously reported cases.

METHODS

Panoramic radiograph, computed tomography and magnetic resonance imaging were obtained and an in-depth assessment of patient's dental status and sequence analysis of FGF10, FGFR2 and FGFR3 genes were performed. Previous reports of bilateral parotid gland aplasia were assessed.

RESULTS

In a 64-year-old woman with extensive basal cell carcinoma of nasal skin an incidental bilateral parotid gland aplasia was noted during radiotherapy treatment planning. Dental status revealed surprisingly numerous (n = 15) teeth without active caries lesions. No other craniofacial abnormalities were identified. To rule out most probable syndromes associated with parotid gland aplasia, sequence analysis of FGF10, FGFR2 and FGFR3 genes was performed showing no pathogenic variants. With a literature review, we identified 148 cases of salivary gland aplasia in which median age at diagnosis was 21 years and one third were asymptomatic. In only 10 of these cases, the patients presented with bilateral aplasia of parotid glands without other craniofacial abnormalities.

CONCLUSIONS

Absence of salivary glands can have a debilitating effect on oral health and is often accompanied by other craniofacial abnormalities. However, relatively frequent asymptomatic course suggests that this rare malformation is probably underdiagnosed. Therefore, we propose systematic reporting of salivary gland aplasia to assess its true prevalence in general population.

Automated segmentation of articular disc of the temporomandibular joint on magnetic resonance images using deep learning

Temporomandibular disorders are typically accompanied by a number of clinical manifestations that involve pain and dysfunction of the masticatory muscles and temporomandibular joint. The most important subgroup of articular abnormalities in patients with temporomandibular disorders includes patients with different forms of articular disc displacement and deformation. Here, we propose a fully automated articular disc detection and segmentation system to support the diagnosis of temporomandibular disorder on magnetic resonance imaging. This system uses deep learning-based semantic segmentation approaches. The study included a total of 217 magnetic resonance images from 10 patients with anterior displacement of the articular disc and 10 healthy control subjects with normal articular discs. These images were used to evaluate three deep learning-based semantic segmentation approaches: our proposed convolutional neural network encoder-decoder named 3DiscNet (Detection for Displaced articular DISC using convolutional neural NETwork), U-Net, and SegNet-Basic. Of the three algorithms, 3DiscNet and SegNet-Basic showed comparably good metrics (Dice coefficient, sensitivity, and positive predictive value). This study provides a proof-of-concept for a fully automated deep learning-based segmentation methodology for articular discs on magnetic resonance images, and obtained promising initial results, indicating that the method could potentially be used in clinical practice for the assessment of temporomandibular disorders.