Zero-TE MRI: Potential Applications in the Oral Cavity and Oropharynx

Smith, Mark A.; Bambach, Sven; Selvaraj, Bhavani; Ho, Mai‐Lan

doi:10.1097/rmr.0000000000000279

Cited by 4 publications

(3 citation statements)

References 65 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The largest areas of deviation are attributed to missing MR imaging data around regions of hardware and difficult-to-segment anatomic areas, which will guide further investigation. [33][34][35][36] Future comparative effectiveness studies will need to account for the relative risks and benefits of clinical workflow, ionizing radiation exposure, examination duration, anesthesia requirements, diagnostic quality, and treatment outcomes. For example, bone MR imaging represents a key alternative for at-risk patients in whom radiation exposure must be minimized or eliminated, ie, children, pregnant women, and patients with cancer-predisposition syndromes.…”

Section: Model Architectures and Loss Functionsmentioning

confidence: 99%

See 1 more Smart Citation

Deep Learning for Synthetic CT from Bone MRI in the Head and Neck

Bambach

2022

AJNR Am J Neuroradiol

View full text Add to dashboard Cite

BACKGROUND AND PURPOSE: Bone MR imaging techniques enable visualization of cortical bone without the need for ionizing radiation. Automated conversion of bone MR imaging to synthetic CT is highly desirable for downstream image processing and eventual clinical adoption. Given the complex anatomy and pathology of the head and neck, deep learning models are ideally suited for learning such mapping. MATERIALS AND METHODS:This was a retrospective study of 39 pediatric and adult patients with bone MR imaging and CT examinations of the head and neck. For each patient, MR imaging and CT data sets were spatially coregistered using multiple-point affine transformation. Paired MR imaging and CT slices were generated for model training, using 4-fold cross-validation. We trained 3 different encoder-decoder models: Light_U-Net (2 million parameters) and VGG-16 U-Net (29 million parameters) without and with transfer learning. Loss functions included mean absolute error, mean squared error, and a weighted average. Performance metrics included Pearson R, mean absolute error, mean squared error, bone precision, and bone recall. We investigated model generalizability by training and validating across different conditions. RESULTS:The Light_U-Net architecture quantitatively outperformed VGG-16 models. Mean absolute error loss resulted in higher bone precision, while mean squared error yielded higher bone recall. Performance metrics decreased when using training data captured only in a different environment but increased when local training data were augmented with those from different hospitals, vendors, or MR imaging techniques. CONCLUSIONS:We have optimized a robust deep learning model for conversion of bone MR imaging to synthetic CT, which shows good performance and generalizability when trained on different hospitals, vendors, and MR imaging techniques. This approach shows promise for facilitating downstream image processing and adoption into clinical practice. ABBREVIATIONS: DL ¼ deep learning; GRE ¼ gradient recalled-echo; MAE ¼ mean absolute error; MSE ¼ mean squared error; TE ¼ echo time M R imaging is the workhorse of clinical neuroradiology, providing high tissue contrast for the evaluation of CNS structures. However, CT remains the first-line technique for rapid neurologic screening and cortical bone assessment. A novel class of MR imaging techniques uses very short TE to capture weak and short-lived proton signals from dry tissues such as cortical bone. As MR imaging hardware and software have advanced, "black-bone" MR imaging techniques have progressively improved from gradient recalled-echo (GRE) to ultrashort-TE and zero-TE approaches. [1][2][3] TE values are on the order of 1-2 ms for GRE, 50-200 ms for ultrashort-TE, and 0-25 ms for zero-TE (Online Supplemental Data). With shorter TEs, the detectable signal from cortical bone increases, scan times become faster, acoustic noise from gradient switching decreases, and resistance to motion and susceptibility artifacts increases. [4][5] Bone MR imaging offers the potent...

show abstract

Section: Model Architectures and Loss Functionsmentioning

confidence: 99%

“…Therefore, ultra-low-dose CT versus no-dose bone MR imaging may yield a more realistic and equitable image comparison. 7,[33][34][35][36]…”

Section: Model Architectures and Loss Functionsmentioning

confidence: 99%

Deep Learning for Synthetic CT from Bone MRI in the Head and Neck

Bambach

2022

AJNR Am J Neuroradiol

View full text Add to dashboard Cite

show abstract

“…9 and 10 ). Since CT is frequently affected by dental amalgam artefact, this technique can be particularly valuable [ 38 ]. Suzuki et al [ 39 ] evaluated 49 cases of suspected mandibular invasion comparing a conventional 2D FSE and high-resolution 3D VIBE with CT. All three techniques correctly diagnosed the 32 tumours with histopathological evidence of mandibular invasion (sensitivity 100%).…”

Section: Alternatives To Ctmentioning

confidence: 99%

Imaging of Bone in the Head and Neck Region, is There More Than CT?

Eley

Delso

2022

Curr Radiol Rep

View full text Add to dashboard Cite

Purpose of Review The objective of this review is to document the advances in non-ionising imaging alternatives to CT for the head and neck. Recent Findings The main alternative to CT for imaging bone of the head and neck region is MRI, particularly techniques which incorporate gradient echo imaging (Black Bone technique) and ultra-short or zero-echo time imaging. Since these techniques can provide high resolution isometric voxels, they can be used to provide multi-planar reformats and, following post processing, 3D reconstructed images of the craniofacial skeleton. As expected, the greatest advancements in recent years have been focused on enhanced image processing techniques and attempts to address the difficulties encountered at air-bone interfaces. Summary This article will review the imaging techniques and recent advancements which are bringing non-ionising alternatives to CT imaging of the bone of the head and neck region into the realm of routine clinical application.

show abstract