Simple quantitative planimetric measurement of nigrosome-1 for clinical settings

INTRODUCTION

Loss of MRI hyperintense signal in nigrosome-1 (assessed with susceptibility-weighted imaging) is a biomarker for Parkinson's disease (PD). Current clinical practice involves subjectively rating the appearance of nigrosome-1 which is challenging. The study aimed to test and compare a simple method for quantifying nigrosome-1 with the current subjective rating method.

METHODS

Two experienced neuroradiologists measured area of hyperintense signal in nigrosome-1 (quantitative method) and rated nigrosome-1 appearance (as normal, attenuated, or absent; subjective method) in 42 patients encompassing the full spectrum of nigrosome-1 integrity (21 patients aged 55.5 ± 20.9 years with Essential tremor (ET) and a subset of 21 patients aged 69.6 ± 8.6 years with PD). Neuroradiologists were blinded to each other's measurements, clinical notes, and patient group.

RESULTS

Both methods yielded a significant difference between the groups (PD vs ET; p < 0.001). Pooled (across sides) area of nigrosome-1 hyperintense signal was significantly smaller in the PD group (median = 2.1 mm2, range = 0-15.8 mm2) than ET group (median = 8.3 mm2, range = 0-15.7 mm2; p < 0.001). Inter-rater reliability was high to very high for both methods (subjective: weighted kappa = 0.640, p < 0.001; quantitative: W = 0.733, p = 0.004). Our primary hypothesis that area of nigrosome-1 hyperintense signal exhibits higher inter-rater reliability than subjective rating of nigrosome-1 appearance was not supported.

CONCLUSION

The simple quantitative method, used with subjectively rated nigrosome-1 appearance, may improve confidence in longitudinal clinical reporting, when nigrosome-1 is attenuated. However, further work on the incremental diagnostic value of planimetry and bias, repeatability and reproducibility are needed before it can be recommended in clinical practice.

Imaging evaluation of a proposed 3D generative model for MRI to CT translation in the lumbar spine

BACKGROUND CONTEXT

A computed tomography (CT) and magnetic resonance imaging (MRI) are used routinely in the radiologic evaluation and surgical planning of patients with lumbar spine pathology, with the modalities being complimentary. We have developed a deep learning algorithm which can produce 3D lumbar spine CT images from MRI data alone. This has the potential to reduce radiation to the patient as well as burden on the health care system.

PURPOSE

The purpose of this study is to evaluate the accuracy of the synthetic lumbar spine CT images produced using our deep learning model.

STUDY DESIGN

A training set of 400 unpaired CTs and 400 unpaired MRI scans of the lumbar spine was used to train a supervised 3D cycle-Gan model. Evaluators performed a set of clinically relevant measurements on 20 matched synthetic CTs and true CTs. These measurements were then compared to assess the accuracy of the synthetic CTs.

PATIENT SAMPLE

The evaluation data set consisted of 20 patients who had CT and MRI scans performed within a 30-day period of each other. All patient data was deidentified. Notable exclusions included artefact from patient motion, metallic implants or any intervention performed in the 30 day intervening period.

OUTCOME MEASURES

The outcome measured was the mean difference in measurements performed by the group of evaluators between real CT and synthetic CTs in terms of absolute and relative error.

METHODS

Data from the 20 MRI scans was supplied to our deep learning model which produced 20 "synthetic CT" scans. This formed the evaluation data set. Four clinical evaluators consisting of neurosurgeons and radiologists performed a set of 24 clinically relevant measurements on matched synthetic CT and true CTs in 20 patients. A test set of measurements were performed prior to commencing data collection to identify any significant interobserver variation in measurement technique.

RESULTS

The measurements performed in the sagittal plane were all within 10% relative error with the majority within 5% relative error. The pedicle measurements performed in the axial plane were considerably less accurate with a relative error of up to 34%.

CONCLUSIONS

The computer generated synthetic CTs demonstrated a high level of accuracy for the measurements performed in-plane to the original MRIs used for synthesis. The measurements performed on the axial reconstructed images were less accurate, attributable to the images being synthesized from nonvolumetric routine sagittal T1-weighted MRI sequences. It is hypothesized that if axial sequences or volumetric data were input into the algorithm these measurements would have improved accuracy.

Pure retroclival subdural hemorrhage secondary to ruptured posterior communicating artery aneurysm

Approximately 3% of intracranial aneurysm ruptures result in an associated subdural hematoma (SDH). SDH from intracranial aneurysm rupture without radiographic evidence of SAH, however, is rare. We report a case of an isolated retroclival SDH secondary to an intracranial aneurysm rupture.

Federated learning enables big data for rare cancer boundary detection

Although machine learning (ML) has shown promise across disciplines, out-of-sample generalizability is concerning. This is currently addressed by sharing multi-site data, but such centralization is challenging/infeasible to scale due to various limitations. Federated ML (FL) provides an alternative paradigm for accurate and generalizable ML, by only sharing numerical model updates. Here we present the largest FL study to-date, involving data from 71 sites across 6 continents, to generate an automatic tumor boundary detector for the rare disease of glioblastoma, reporting the largest such dataset in the literature (n = 6, 314). We demonstrate a 33% delineation improvement for the surgically targetable tumor, and 23% for the complete tumor extent, over a publicly trained model. We anticipate our study to: 1) enable more healthcare studies informed by large diverse data, ensuring meaningful results for rare diseases and underrepresented populations, 2) facilitate further analyses for glioblastoma by releasing our consensus model, and 3) demonstrate the FL effectiveness at such scale and task-complexity as a paradigm shift for multi-site collaborations, alleviating the need for data-sharing.

Computed tomography scan measurements of the globe and orbit to assess the risks of traumatic complications from medial peribulbar anaesthesia

Complications of peribulbar anaesthesia include retrobulbar haemorrhage, globe perforation and brainstem anaesthesia. Therefore, this study took measurements relating the proximity of medial canthus to the optic nerve and also the safe angle between orbit and globe using 200 multiplanar reconstructed computed tomography (CT) images of the orbit. The principal results show that in 1.5% of the sample, the optic nerve is within 20 mm of the medial canthus, with a minimum distance of 15 mm. One% have a safe angle of 10 degrees or less between bone and globe. None of the demographic data, nor axial length were predictive of these results. We have shown that there are a minority of patients with unusual orbital anatomy. This places them at a theoretical higher risk of complications. These cases are not currently predicted by measured data.

Syringomyelia intermittens: highlighting the complex pathophysiology of syringomyelia. Illustrative case

BACKGROUND

Chiari Type I malformation (CM1) is a disorder recognized by caudal displacement of the cerebellar tonsils through the foramen magnum and into the cervical canal. Syringomyelia is frequently found in patients with CM1, but the pathophysiology of syringomyelia remains an enigma. As a general consensus, symptomatic patients should be treated and asymptomatic patients without a syrinx should not be treated. Mildly symptomatic patients or asymptomatic patients with a syrinx, on the other hand, pose a more challenging dilemma, as the natural evolution is uncertain. For many surgeons, the presence of a syrinx is an indication to offer surgery even if the patient is asymptomatic or mildly symptomatic.

OBSERVATIONS

The authors describe an illustrative case of a 31-year-old female with an incidental finding of a CM1 malformation and cervical syrinx in 2013. Conservative management was advocated as the patient was asymptomatic. Monitoring of the syrinx over a course of 8 years showed resolution, followed by reappearance and finally a complete resolution in 2021. A review of the literature and the possible pathophysiology is discussed.

LESSONS

The unusual course of this patient highlights the importance of guiding treatment by clinical symptoms, not radiological findings. Furthermore it reflects the complexity of the pathophysiology and the uncertain natural history of syringomyelia.

Convolutional neural networks for brain tumour segmentation

The introduction of quantitative image analysis has given rise to fields such as radiomics which have been used to predict clinical sequelae. One growing area of interest for analysis is brain tumours, in particular glioblastoma multiforme (GBM). Tumour segmentation is an important step in the pipeline in the analysis of this pathology. Manual segmentation is often inconsistent as it varies between observers. Automated segmentation has been proposed to combat this issue. Methodologies such as convolutional neural networks (CNNs) which are machine learning pipelines modelled on the biological process of neurons (called nodes) and synapses (connections) have been of interest in the literature. We investigate the role of CNNs to segment brain tumours by firstly taking an educational look at CNNs and perform a literature search to determine an example pipeline for segmentation. We then investigate the future use of CNNs by exploring a novel field-radiomics. This examines quantitative features of brain tumours such as shape, texture, and signal intensity to predict clinical outcomes such as survival and response to therapy.

Central vein sign for multiple sclerosis: A systematic review and meta-analysis

AIMS

To systematically review the diagnostic value of the central vein sign (CVS) in multiple sclerosis (MS) and to meta-analyse the proportion of positive lesions for CVS needed to distinguish MS from non-MS mimics.

MATERIALS AND METHODS

A literature review was performed and a proportion meta-analysis was performed to examine the proportion of the CVS in MS lesions. Studies reporting a threshold of the CVS containing lesions with 100% diagnostic accuracy were included in the meta-analysis. This was compared to MS mimics in order to establish the discriminative value of the CVS.

RESULTS

The CVS was found to be viable at lower field strengths (3 T and 1.5 T) and automated analysis is currently less accurate than manual counting. Five studies were included for the proportional meta-analysis. From the analysis, a proportion of 45% of lesions having the CVS was suggested given that the findings that the weighted proportion was 46.4% (95% confidence interval [CI]: of 40.3%-52.6%) with low heterogeneity (I² = 0.0%; p=0.5).

CONCLUSION

Although the CVS is a clinically relevant and viable sign, further work is needed to integrate this into the existing diagnostic criteria. As manual determination is a time-consuming process, the development of automated methods will be beneficial. With improvements in computational imaging techniques, the CVS will have an important role in the diagnosis and differentiation of MS.

Stroke physician versus stroke neurologist: can anyone thrombolyse?

BACKGROUND/AIM

The aim of this study is to compare the outcomes of thrombolysis under standard clinical settings between subjects treated by a stroke neurologist versus those treated by a non-neurologist stroke physician.

METHODS

Single-centre, observational cohort study of subjects thrombolysed in a calendar year, stratified according to the physician type authorising thrombolysis. Endpoints measured include proportion of subjects with symptomatic intracranial haemorrhage, door-to-needle time, change in National Institute of Health Stroke Scale and discharge destination.

RESULTS

Forty-nine subjects with a mean age 76 ± 16 years underwent thrombolysis, 21 were under the care of a stroke neurologist and 28 by a non-neurologist stroke physician. No symptomatic intracranial haemorrhages were observed. There was no difference in terms of door-to-needle time, proportion of individuals with haemorrhagic transformation, mortality or discharge destination between the two groups.

CONCLUSION

Due to the single-centre, observational nature of this study, the equivalent outcomes between those thrombolysed by a stroke neurologist versus those thrombolysed by a stroke physician must be interpreted with caution pending further studies. Nevertheless, in the current setting, no signal for harm has been detected. This study is unique as it is the first to our knowledge comparing outcomes between a neurologist and non-neurologist following thrombolysis.

Whispering dysphonia (DYT4 dystonia) is caused by a mutation in the TUBB4 gene

OBJECTIVE

A study was undertaken to identify the gene underlying DYT4 dystonia, a dominantly inherited form of spasmodic dysphonia combined with other focal or generalized dystonia and a characteristic facies and body habitus, in an Australian family.

METHODS

Genome-wide linkage analysis was carried out in 14 family members followed by genome sequencing in 2 individuals. The index patient underwent a detailed neurological follow-up examination, including electrophysiological studies and magnetic resonance imaging scanning. Biopsies of the skin and olfactory mucosa were obtained, and expression levels of TUBB4 mRNA were determined by quantitative real-time polymerase chain reaction in 3 different cell types. All exons of TUBB4 were screened for mutations in 394 unrelated dystonia patients.

RESULTS

The disease-causing gene was mapped to a 23cM region on chromosome 19p13.3-p13.2 with a maximum multipoint LOD score of 5.338 at markers D9S427 and D9S1034. Genome sequencing revealed a missense variant in the TUBB4 (tubulin beta-4; Arg2Gly) gene as the likely cause of disease. Sequencing of TUBB4 in 394 unrelated dystonia patients revealed another missense variant (Ala271Thr) in a familial case of segmental dystonia with spasmodic dysphonia. mRNA expression studies demonstrated significantly reduced levels of mutant TUBB4 mRNA in different cell types from a heterozygous Arg2Gly mutation carrier compared to controls.

INTERPRETATION

A mutation in TUBB4 causes DYT4 dystonia in this Australian family with so-called whispering dysphonia, and other mutations in TUBB4 may contribute to spasmodic dysphonia. Given that TUBB4 is a neuronally expressed tubulin, our results imply abnormal microtubule function as a novel mechanism in the pathophysiology of dystonia.

Treatment of metastatic sialoblastoma with chemotherapy and surgery

Tumors of the salivary gland are very uncommon in children. Sialoblastoma is a rare, aggressive, blastomatous, and potentially malignant congenital tumor. Distant metastases are rare. We present a case of sialoblastoma with lung metastases that developed in a 4-year-old girl adjacent to a congenital nevus in the left cheek. The tumor was inoperable at diagnosis but the largest of the pulmonary metastases was removed surgically. The patient responded well to chemotherapy and underwent surgical excision of the primary tumor, followed by three more courses of chemotherapy.