Effects of a comprehensive brain computed tomography deep learning model on radiologist detection accuracy.

OBJECTIVES: Non-contrast computed tomography of the brain (NCCTB) is commonly used to detect intracranial pathology but is subject to interpretation errors. Machine learning can augment clinical decision-making and improve NCCTB scan interpretation. This retrospective detection accuracy study assessed the performance of radiologists assisted by a deep learning model and compared the standalone performance of the model with that of unassisted radiologists. METHODS: A deep learning model was trained on 212,484 NCCTB scans drawn from a private radiology group in Australia. Scans from inpatient, outpatient, and emergency settings were included. Scan inclusion criteria were age ≥ 18 years and series slice thickness ≤ 1.5 mm. Thirty-two radiologists reviewed 2848 scans with and without the assistance of the deep learning system and rated their confidence in the presence of each finding using a 7-point scale. Differences in AUC and Matthews correlation coefficient (MCC) were calculated using a ground-truth gold standard. RESULTS: The model demonstrated an average area under the receiver operating characteristic curve (AUC) of 0.93 across 144 NCCTB findings and significantly improved radiologist interpretation performance. Assisted and unassisted radiologists demonstrated an average AUC of 0.79 and 0.73 across 22 grouped parent findings and 0.72 and 0.68 across 189 child findings, respectively. When assisted by the model, radiologist AUC was significantly improved for 91 findings (158 findings were non-inferior), and reading time was significantly reduced. CONCLUSIONS: The assistance of a comprehensive deep learning model significantly improved radiologist detection accuracy across a wide range of clinical findings and demonstrated the potential to improve NCCTB interpretation. CLINICAL RELEVANCE STATEMENT: This study evaluated a comprehensive CT brain deep learning model, which performed strongly, improved the performance of radiologists, and reduced interpretation time. The model may reduce errors, improve efficiency, facilitate triage, and better enable the delivery of timely patient care. KEY POINTS: • This study demonstrated that the use of a comprehensive deep learning system assisted radiologists in the detection of a wide range of abnormalities on non-contrast brain computed tomography scans. • The deep learning model demonstrated an average area under the receiver operating characteristic curve of 0.93 across 144 findings and significantly improved radiologist interpretation performance. • The assistance of the comprehensive deep learning model significantly reduced the time required for radiologists to interpret computed tomography scans of the brain.

Stimulation of the muscarinic receptor M4 regulates neural precursor cell proliferation and promotes adult hippocampal neurogenesis.

Cholinergic signaling plays a crucial role in the regulation of adult hippocampal neurogenesis; however, the mechanisms by which acetylcholine mediates neurogenic effects are not completely understood. Here, we report the expression of muscarinic acetylcholine receptor subtype M4 (M4 mAChR) on a subpopulation of neural precursor cells (NPCs) in the adult mouse hippocampus, and demonstrate that its pharmacological stimulation promotes their proliferation, thereby enhancing the production of new neurons in vivo. Using a targeted ablation approach, we also show that medial septum (MS) and the diagonal band of Broca (DBB) cholinergic neurons support both the survival and morphological maturation of adult-born neurons in the mouse hippocampus. Although the systemic administration of an M4-selective allosteric potentiator fails to fully rescue the MS/DBB cholinergic lesion-induced decrease in hippocampal neurogenesis, it further exacerbates the impairment in the morphological maturation of adult-born neurons. Collectively, these findings reveal stage-specific roles of M4 mAChRs in regulating adult hippocampal neurogenesis, uncoupling their positive role in enhancing the production of new neurons from the M4-induced inhibition of their morphological maturation, at least in the context of cholinergic signaling dysfunction.

Analysis of Line and Tube Detection Performance of a Chest X-ray Deep Learning Model to Evaluate Hidden Stratification.

This retrospective case-control study evaluated the diagnostic performance of a commercially available chest radiography deep convolutional neural network (DCNN) in identifying the presence and position of central venous catheters, enteric tubes, and endotracheal tubes, in addition to a subgroup analysis of different types of lines/tubes. A held-out test dataset of 2568 studies was sourced from community radiology clinics and hospitals in Australia and the USA, and was then ground-truth labelled for the presence, position, and type of line or tube from the consensus of a thoracic specialist radiologist and an intensive care clinician. DCNN model performance for identifying and assessing the positioning of central venous catheters, enteric tubes, and endotracheal tubes over the entire dataset, as well as within each subgroup, was evaluated. The area under the receiver operating characteristic curve (AUC) was assessed. The DCNN algorithm displayed high performance in detecting the presence of lines and tubes in the test dataset with AUCs > 0.99, and good position classification performance over a subpopulation of ground truth positive cases with AUCs of 0.86-0.91. The subgroup analysis showed that model performance was robust across the various subtypes of lines or tubes, although position classification performance of peripherally inserted central catheters was relatively lower. Our findings indicated that the DCNN algorithm performed well in the detection and position classification of lines and tubes, supporting its use as an assistant for clinicians. Further work is required to evaluate performance in rarer scenarios, as well as in less common subgroups.

Machine Learning Augmented Interpretation of Chest X-rays: A Systematic Review.

Limitations of the chest X-ray (CXR) have resulted in attempts to create machine learning systems to assist clinicians and improve interpretation accuracy. An understanding of the capabilities and limitations of modern machine learning systems is necessary for clinicians as these tools begin to permeate practice. This systematic review aimed to provide an overview of machine learning applications designed to facilitate CXR interpretation. A systematic search strategy was executed to identify research into machine learning algorithms capable of detecting >2 radiographic findings on CXRs published between January 2020 and September 2022. Model details and study characteristics, including risk of bias and quality, were summarized. Initially, 2248 articles were retrieved, with 46 included in the final review. Published models demonstrated strong standalone performance and were typically as accurate, or more accurate, than radiologists or non-radiologist clinicians. Multiple studies demonstrated an improvement in the clinical finding classification performance of clinicians when models acted as a diagnostic assistance device. Device performance was compared with that of clinicians in 30% of studies, while effects on clinical perception and diagnosis were evaluated in 19%. Only one study was prospectively run. On average, 128,662 images were used to train and validate models. Most classified less than eight clinical findings, while the three most comprehensive models classified 54, 72, and 124 findings. This review suggests that machine learning devices designed to facilitate CXR interpretation perform strongly, improve the detection performance of clinicians, and improve the efficiency of radiology workflow. Several limitations were identified, and clinician involvement and expertise will be key to driving the safe implementation of quality CXR machine learning systems.

Cholinergic basal forebrain degeneration due to sleep-disordered breathing exacerbates pathology in a mouse model of Alzheimer's disease.

Although epidemiological studies indicate that sleep-disordered breathing (SDB) such as obstructive sleep apnea is a strong risk factor for the development of Alzheimer's disease (AD), the mechanisms of the risk remain unclear. Here we developed a method of modeling SDB in mice that replicates key features of the human condition: altered breathing during sleep, sleep disruption, moderate hypoxemia, and cognitive impairment. When we induced SDB in a familial AD model, the mice displayed exacerbation of cognitive impairment and the pathological features of AD, including increased levels of amyloid-beta and inflammatory markers, as well as selective degeneration of cholinergic basal forebrain neurons. These pathological features were not induced by chronic hypoxia or sleep disruption alone. Our results also revealed that the cholinergic neurodegeneration was mediated by the accumulation of nuclear hypoxia inducible factor 1 alpha. Furthermore, restoring blood oxygen levels during sleep to prevent hypoxia prevented the pathological changes induced by the SDB. These findings suggest a signaling mechanism whereby SDB induces cholinergic basal forebrain degeneration.

Charting the potential of brain computed tomography deep learning systems.

Brain computed tomography (CTB) scans are widely used to evaluate intracranial pathology. The implementation and adoption of CTB has led to clinical improvements. However, interpretation errors occur and may have substantial morbidity and mortality implications for patients. Deep learning has shown promise for facilitating improved diagnostic accuracy and triage. This research charts the potential of deep learning applied to the analysis of CTB scans. It draws on the experience of practicing clinicians and technologists involved in development and implementation of deep learning-based clinical decision support systems. We consider the past, present and future of the CTB, along with limitations of existing systems as well as untapped beneficial use cases. Implementing deep learning CTB interpretation systems and effectively navigating development and implementation risks can deliver many benefits to clinicians and patients, ultimately improving efficiency and safety in healthcare.

Do comprehensive deep learning algorithms suffer from hidden stratification? A retrospective study on pneumothorax detection in chest radiography.

OBJECTIVES: To evaluate the ability of a commercially available comprehensive chest radiography deep convolutional neural network (DCNN) to detect simple and tension pneumothorax, as stratified by the following subgroups: the presence of an intercostal drain; rib, clavicular, scapular or humeral fractures or rib resections; subcutaneous emphysema and erect versus non-erect positioning. The hypothesis was that performance would not differ significantly in each of these subgroups when compared with the overall test dataset. DESIGN: A retrospective case-control study was undertaken. SETTING: Community radiology clinics and hospitals in Australia and the USA. PARTICIPANTS: A test dataset of 2557 chest radiography studies was ground-truthed by three subspecialty thoracic radiologists for the presence of simple or tension pneumothorax as well as each subgroup other than positioning. Radiograph positioning was derived from radiographer annotations on the images. OUTCOME MEASURES: DCNN performance for detecting simple and tension pneumothorax was evaluated over the entire test set, as well as within each subgroup, using the area under the receiver operating characteristic curve (AUC). A difference in AUC of more than 0.05 was considered clinically significant. RESULTS: When compared with the overall test set, performance of the DCNN for detecting simple and tension pneumothorax was statistically non-inferior in all subgroups. The DCNN had an AUC of 0.981 (0.976-0.986) for detecting simple pneumothorax and 0.997 (0.995-0.999) for detecting tension pneumothorax. CONCLUSIONS: Hidden stratification has significant implications for potential failures of deep learning when applied in clinical practice. This study demonstrated that a comprehensively trained DCNN can be resilient to hidden stratification in several clinically meaningful subgroups in detecting pneumothorax.

Assessment of the effect of a comprehensive chest radiograph deep learning model on radiologist reports and patient outcomes: a real-world observational study.

OBJECTIVES: Artificial intelligence (AI) algorithms have been developed to detect imaging features on chest X-ray (CXR) with a comprehensive AI model capable of detecting 124 CXR findings being recently developed. The aim of this study was to evaluate the real-world usefulness of the model as a diagnostic assistance device for radiologists. DESIGN: This prospective real-world multicentre study involved a group of radiologists using the model in their daily reporting workflow to report consecutive CXRs and recording their feedback on level of agreement with the model findings and whether this significantly affected their reporting. SETTING: The study took place at radiology clinics and hospitals within a large radiology network in Australia between November and December 2020. PARTICIPANTS: Eleven consultant diagnostic radiologists of varying levels of experience participated in this study. PRIMARY AND SECONDARY OUTCOME MEASURES: Proportion of CXR cases where use of the AI model led to significant material changes to the radiologist report, to patient management, or to imaging recommendations. Additionally, level of agreement between radiologists and the model findings, and radiologist attitudes towards the model were assessed. RESULTS: Of 2972 cases reviewed with the model, 92 cases (3.1%) had significant report changes, 43 cases (1.4%) had changed patient management and 29 cases (1.0%) had further imaging recommendations. In terms of agreement with the model, 2569 cases showed complete agreement (86.5%). 390 (13%) cases had one or more findings rejected by the radiologist. There were 16 findings across 13 cases (0.5%) deemed to be missed by the model. Nine out of 10 radiologists felt their accuracy was improved with the model and were more positive towards AI poststudy. CONCLUSIONS: Use of an AI model in a real-world reporting environment significantly improved radiologist reporting and showed good agreement with radiologists, highlighting the potential for AI diagnostic support to improve clinical practice.

Chest radiographs and machine learning - Past, present and future.

Despite its simple acquisition technique, the chest X-ray remains the most common first-line imaging tool for chest assessment globally. Recent evidence for image analysis using modern machine learning points to possible improvements in both the efficiency and the accuracy of chest X-ray interpretation. While promising, these machine learning algorithms have not provided comprehensive assessment of findings in an image and do not account for clinical history or other relevant clinical information. However, the rapid evolution in technology and evidence base for its use suggests that the next generation of comprehensive, well-tested machine learning algorithms will be a revolution akin to early advances in X-ray technology. Current use cases, strengths, limitations and applications of chest X-ray machine learning systems are discussed.

The p75 neurotrophin receptor is required for the survival of neuronal progenitors and normal formation of the basal forebrain, striatum, thalamus and neocortex.

During development, the p75 neurotrophin receptor (p75NTR) is widely expressed in the nervous system where it regulates neuronal differentiation, migration and axonal outgrowth. p75NTR also mediates the survival and death of newly born neurons, with functional outcomes being dependent on both timing and cellular context. Here, we show that knockout of p75NTR from embryonic day 10 (E10) in neural progenitors using a conditional Nestin-Cre p75NTR floxed mouse causes increased apoptosis of progenitor cells. By E14.5, the number of Tbr2-positive progenitor cells was significantly reduced and the rate of neurogenesis was halved. Furthermore, in adult knockout mice, there were fewer cortical pyramidal neurons, interneurons, cholinergic basal forebrain neurons and striatal neurons, corresponding to a relative reduction in volume of these structures. Thalamic midline fusion during early postnatal development was also impaired in Nestin-Cre p75NTR floxed mice, indicating a novel role for p75NTR in the formation of this structure. The phenotype of this strain demonstrates that p75NTR regulates multiple aspects of brain development, including cortical progenitor cell survival, and that expression during early neurogenesis is required for appropriate formation of telencephalic structures.

Downregulation of SNX27 expression does not exacerbate amyloidogenesis in the APP/PS1 Alzheimer's disease mouse model.

There is in vitro evidence that sorting nexin family member 27 (SNX27), a member of the retromer complex, changes the distribution of the amyloid-beta (Aß) precursor protein (APP) to promote its recycling and thereby prevent the production of Aß, the toxic protein associated with Alzheimer's disease (AD). In this study, we analyzed the phenotype of the familial AD APP/PS mouse strain lacking one copy of the SNX27 gene. The reduction in SNX27 expression had no significant effect on the in vivo accumulation of soluble, total, or plaque-deposited Aß, which is overproduced by the familial APP/PS transgenes. Hippocampal structure and cholinergic basal forebrain neuronal health were also unaffected. Nonetheless, mild positive and negative effects of age and/or genotype on spatial navigation performance were observed in SNX27+/- and SNX27+/-APP/PS mice, respectively. These data suggest that downregulation of SNX27 alone does not have long-term negative consequences on spatial memory, but that cognitive dysfunction in the context of high Aß deposition is exacerbated by the cellular or molecular changes induced by reduced SNX27 function.

Regulation of cholinergic basal forebrain development, connectivity, and function by neurotrophin receptors.

Cholinergic basal forebrain (cBF) neurons are defined by their expression of the p75 neurotrophin receptor (p75NTR) and tropomyosin-related kinase (Trk) neurotrophin receptors in addition to cholinergic markers. It is known that the neurotrophins, particularly nerve growth factor (NGF), mediate cholinergic neuronal development and maintenance. However, the role of neurotrophin signalling in regulating adult cBF function is less clear, although in dementia, trophic signalling is reduced and p75NTR mediates neurodegeneration of cBF neurons. Here we review the current understanding of how cBF neurons are regulated by neurotrophins which activate p75NTR and TrkA, B or C to influence the critical role that these neurons play in normal cortical function, particularly higher order cognition. Specifically, we describe the current evidence that neurotrophins regulate the development of basal forebrain neurons and their role in maintaining and modifying mature basal forebrain synaptic and cortical microcircuit connectivity. Understanding the role neurotrophin signalling plays in regulating the precision of cholinergic connectivity will contribute to the understanding of normal cognitive processes and will likely provide additional ideas for designing improved therapies for the treatment of neurological disease in which cholinergic dysfunction has been demonstrated.

Removal of p75 Neurotrophin Receptor Expression from Cholinergic Basal Forebrain Neurons Reduces Amyloid-ß Plaque Deposition and Cognitive Impairment in Aged APP/PS1 Mice.

The degeneration of cholinergic basal forebrain (cBF) neurons in Alzheimer's disease (AD) leads to the cognitive impairment associated with this condition. cBF neurons express the p75 neurotrophin receptor (p75NTR), which mediates cell death, and the extracellular domain of p75NTR can bind to amyloid beta (Aß) and promote its degradation. Here, we investigated the contribution of cBF neuronal p75NTR to the progression of AD by removing p75NTR from cholinergic neurons in the APP/PS1 familial AD mouse strain. Conditional loss of p75NTR slowed cognitive decline and reduced both Aß accumulation into plaques and gliosis. Expression of the amyloid protein precursor and its cleavage enzymes ADAM10 and BACE1 were unchanged. There was also no upregulation of p75NTR in non-cholinergic cell types. This indicates that a direct interaction between cBF-expressed p75NTR and Aß does not contribute significantly to the regulation of Aß load. Rather, loss of p75NTR from cBF neurons, which results in increased cholinergic innervation of the cortex, appears to regulate alternative, more dominant, Aß clearance mechanisms.

Cholinergic basal forebrain neurons regulate fear extinction consolidation through p75 neurotrophin receptor signaling.

Cholinergic basal forebrain (cBF)-derived neurotransmission plays a crucial role in regulating neuronal function throughout the cortex, yet the mechanisms controlling cholinergic innervation to downstream targets have not been elucidated. Here we report that removing the p75 neurotrophin receptor (p75NTR) from cBF neurons induces a significant impairment in fear extinction consolidation. We demonstrate that this is achieved through alterations in synaptic connectivity and functional activity within the medial prefrontal cortex. These deficits revert back to wild-type levels upon re-expression of the active domain of p75NTR in adult animals. These findings demonstrate a novel role for cholinergic neurons in fear extinction consolidation and suggest that neurotrophic signaling is a key regulator of cholinergic-cortical innervation and function.

Targeting mesothelin receptors with drug-loaded bacterial nanocells suppresses human mesothelioma tumour growth in mouse xenograft models.

Human malignant mesothelioma is a chemoresistant tumour that develops from mesothelial cells, commonly associated with asbestos exposure. Malignant mesothelioma incidence rates in European countries are still rising and Australia has one of the highest burdens of malignant mesothelioma on a population basis in the world. Therapy using systemic delivery of free cytotoxic agents is associated with many undesirable side effects due to non-selectivity, and is thus dose-limited which limits its therapeutic potential. Therefore, increasing the selectivity of anti-cancer agents has the potential to dramatically enhance drug efficacy and reduce toxicity. EnGeneIC Dream Vectors (EDV) are antibody-targeted nanocells which can be loaded with cytotoxic drugs and delivered to specific cancer cells via bispecific antibodies (BsAbs) which target the EDV and a cancer cell-specific receptor, simultaneously. BsAbs were designed to target doxorubicin-loaded EDVs to cancer cells via cell surface mesothelin (MSLN). Flow cytometry was used to investigate cell binding and induction of apoptosis, and confocal microscopy to visualize internalization. Mouse xenograft models were used to assess anti-tumour effects in vivo, followed by immunohistochemistry for ex vivo evaluation of proliferation and necrosis. BsAb-targeted, doxorubicin-loaded EDVs were able to bind to and internalize within mesothelioma cells in vitro via MSLN receptors and induce apoptosis. In mice xenografts, the BsAb-targeted, doxorubicin-loaded EDVs suppressed the tumour growth and also decreased cell proliferation. Thus, the use of MSLN-specific antibodies to deliver encapsulated doxorubicin can provide a novel and alternative modality for treatment of mesothelioma.

Effect of Estradiol on Neurotrophin Receptors in Basal Forebrain Cholinergic Neurons: Relevance for Alzheimer's Disease.

The basal forebrain is home to the largest population of cholinergic neurons in the brain. These neurons are involved in a number of cognitive functions including attention, learning and memory. Basal forebrain cholinergic neurons (BFCNs) are particularly vulnerable in a number of neurological diseases with the most notable being Alzheimer's disease, with evidence for a link between decreasing cholinergic markers and the degree of cognitive impairment. The neurotrophin growth factor system is present on these BFCNs and has been shown to promote survival and differentiation on these neurons. Clinical and animal model studies have demonstrated the neuroprotective effects of 17ß-estradiol (E2) on neurodegeneration in BFCNs. It is believed that E2 interacts with neurotrophin signaling on cholinergic neurons to mediate these beneficial effects. Evidence presented in our recent study confirms that altering the levels of circulating E2 levels via ovariectomy and E2 replacement significantly affects the expression of the neurotrophin receptors on BFCN. However, we also showed that E2 differentially regulates neurotrophin receptor expression on BFCNs with effects depending on neurotrophin receptor type and neuroanatomical location. In this review, we aim to survey the current literature to understand the influence of E2 on the neurotrophin system, and the receptors and signaling pathways it mediates on BFCN. In addition, we summarize the physiological and pathophysiological significance of E2 actions on the neurotrophin system in BFCN, especially focusing on changes related to Alzheimer's disease.

Effect of cannabinoids on CGRP release in the isolated rat lumbar spinal cord.

Cannabinoids produce analgesia through a variety of mechanisms. It has been proposed that one mechanism is by modulating the release of CGRP in the spinal cord pain pathways. Previous studies have reported that cannabinoids, particularly CB2 receptor agonists, can modulate CGRP release in the isolated rat spinal cord. In our experiments, the TRPV1 agonist capsaicin evoked CGRP release and this was supressed by the TRPV1 antagonist capsazepine and by the opioid receptor agonist DAMGO. However, none of the cannabinoid receptor agonists that we tested were able to modulate evoked CGRP release; including WIN 55,212-2, methanandamide, and GW405833. These results question the role of spinal cord cannabinoid receptors in the regulation of CGRP signaling.

Estradiol modulation of neurotrophin receptor expression in female mouse basal forebrain cholinergic neurons in vivo.

The neuroprotective effect of estradiol (E2) on basal forebrain cholinergic neurons (BFCNs) has been suggested to occur as a result of E2 modulation of the neurotrophin system on these neurons. The present study provides a comprehensive examination of the relationship between E2 and neurotrophin signaling on BFCNs by investigating the effect of E2 deficiency on the expression levels of neurotrophin receptors (NRs), TrkA, TrkB, and p75 on BFCNs. The number of TrkA receptor-expressing choline acetyltransferase-positive neurons was significantly reduced in the medial septum (MS) in the absence of E2. A significant reduction in TrkB-expressing choline acetyltransferase-positive cells was also observed in ovariectomized mice in the MS and nucleus basalis magnocellularis (NBM). p75 receptor expression was reduced in the NBM and striatum but not in the MS. We also showed that estrogen receptor (ER)-α was expressed by a small percentage of TrkA- and TrkB-positive neurons in the MS (12%) and NBM (19%) and by a high percentage of TrkB-positive neurons in the striatum (69%). Similarly, ERα was expressed at low levels by p75 neurons in the MS (6%) and NBM (9%) but was not expressed on striatal neurons. Finally, ERα knockout using neuron-specific estrogen receptor-α knockout transgenic mice abolished all E2-mediated changes in the NR expression on BFCNs. These results indicate that E2 differentially regulates NR expression on BFCNs, with effects depending on the NR type and neuroanatomical location, and also provide some evidence that alterations in the NR expression are, at least in part, mediated via ERα.