Deep Learning and Machine Learning Algorithms for Retinal Image Analysis in Neurodegenerative Disease: Systematic Review of Datasets and Models.

Purpose: Retinal images contain rich biomarker information for neurodegenerative disease. Recently, deep learning models have been used for automated neurodegenerative disease diagnosis and risk prediction using retinal images with good results. Methods: In this review, we systematically report studies with datasets of retinal images from patients with neurodegenerative diseases, including Alzheimer's disease, Huntington's disease, Parkinson's disease, amyotrophic lateral sclerosis, and others. We also review and characterize the models in the current literature which have been used for classification, regression, or segmentation problems using retinal images in patients with neurodegenerative diseases. Results: Our review found several existing datasets and models with various imaging modalities primarily in patients with Alzheimer's disease, with most datasets on the order of tens to a few hundred images. We found limited data available for the other neurodegenerative diseases. Although cross-sectional imaging data for Alzheimer's disease is becoming more abundant, datasets with longitudinal imaging of any disease are lacking. Conclusions: The use of bilateral and multimodal imaging together with metadata seems to improve model performance, thus multimodal bilateral image datasets with patient metadata are needed. We identified several deep learning tools that have been useful in this context including feature extraction algorithms specifically for retinal images, retinal image preprocessing techniques, transfer learning, feature fusion, and attention mapping. Importantly, we also consider the limitations common to these models in real-world clinical applications. Translational Relevance: This systematic review evaluates the deep learning models and retinal features relevant in the evaluation of retinal images of patients with neurodegenerative disease.

Update on the Management of Diabetic Retinopathy: Anti-VEGF Agents for the Prevention of Complications and Progression of Nonproliferative and Proliferative Retinopathy.

Diabetic retinopathy (DR) is a microvascular disease caused by poorly controlled blood glucose, and it is a leading cause of vision loss in people with diabetes. In this review we discuss the current management of DR with particular focus on the use of intraocular anti-vascular endothelial growth factor (anti-VEGF) agents. Intraocular anti-VEGF agents were first studied in the 1990s, and now several of these agents are either FDA approved or used off-label as first-line treatments for DR. Recent evidence shows that anti-VEGF agents can halt the progression of markers of DR severity, reduce the risk of DR worsening, and reduce the onset of new macular edema. These significant benefits have been demonstrated in patients with proliferative DR and the milder nonproliferative DR (NPDR). A wealth of evidence from recent trials and meta-analyses has detailed the intraoperative and postoperative benefits of adjunctive anti-VEGF therapy prior to pars plana vitrectomy (PPV) for proliferative DR with vitreous hemorrhage. In this review, we also discuss literature comparing various anti-VEGF injection regimens including monthly, quarterly, as-needed, and treat and extend protocols. Combination protocols with panretinal photocoagulation (PRP) or PPV are also discussed. Current evidence suggests that anti-VEGF therapies are effective therapy for NPDR and PDR and may also provide significant benefits when used adjunctively with other DR treatment modalities such as PRP or PPV.

Comparison of netarsudil and latanoprostene bunod as adjuncts to maximum medical therapy in primary open-angle glaucoma.

OBJECTIVE: Compare the efficacy of new agents netarsudil 0.02% (NET) and latanoprostene bunod 0.024% (LB) ophthalmic solutions as adjuncts to traditional 4-class maximum medical therapy (MMT) in primary open angle glaucoma (POAG). DESIGN: Single-centre retrospective cohort study using records from a university glaucoma clinic from 2017 to 2021 with follow-up at 30-90 days. PARTICIPANTS: Patients with POAG already taking 4-class MMT who either added NET (n = 24) or exchanged a currently prescribed prostaglandin analogue (PGA) for LB (n = 11) with no prior surgery except for selective laser trabeculoplasty or cataract extraction >1 year prior. METHODS: Either addition of NET or exchange of PGA for LB and otherwise continuing MMT. Outcome measures were absolute intraocular pressure reduction (IOPR) in mm Hg, percent IOPR, and proportion of patients achieving >10% IOPR. RESULTS: Data from 35 eyes in 35 patients were analyzed. Intraocular pressure reduction after adding NET was significantly greater than after exchanging a PGA for LB. Percent IOPR by NET also was significantly greater than after exchanging PGA for LB. The proportion of patients reaching therapeutic threshold after the addition of NET was significantly greater than after exchange of PGA for LB. CONCLUSIONS: In patients with POAG on MMT, addition of NET was associated with significantly greater magnitude of IOPR and a significantly greater proportion of patients reaching the >10% IOPR threshold compared with exchange of PGA for LB.

Creating Cell Model 2.0 Using Patient Samples Carrying a Pathogenic Mitochondrial DNA Mutation: iPSC Approach for LHON.

Leber's Hereditary Optic Neuropathy is the most prevalent mitochondrial neurological disease caused by mutations in mitochondrial DNA encoded respiratory complex I subunits. Although the genetic origin for Leber's hereditary optic neuropathy was identified about 30 years ago, the underlying pathogenesis is still unclear primarily due to the lack of a relevant system or cell model. Current models are limited to lymphoblasts, fibroblasts, or cybrid cell lines. As the disease phenotype is limited to retinal ganglion cells, induced pluripotent stem cells will serve as an excellent model for studying this tissue-specific disease, elucidating its underlying molecular mechanisms, and identifying novel therapeutic targets. Here, we describe a detailed protocol for the generation of retinal ganglion cells, and also cardiomyocytes for proof of iPSC pluripotency.

Acute inflammatory presentation of orbital metastasis from urothelial carcinoma: a case report.

Orbital metastasis of urothelial carcinoma is very rare; only 22 cases have been documented. In this case report, we describe a patient 1 month status post transurethral resection of urothelial carcinoma who presented with a clinical picture suggestive of orbital cellulitis. However, neither broad-spectrum antibiotics nor a subsequent trial of methylprednisolone was effective at relieving the patient's symptoms. CT imaging of the head, chest, abdomen, pelvis, and lower extremity showed no signs of metastatic disease. Six days after presentation, punch biopsy of the mass was performed and confirmed urothelial carcinoma metastatic to the orbit. The patient died 3 months later due to multiple sites of distant metastasis. This case report suggests that a high index of suspicion for orbital metastasis is important for patients with a history of urothelial carcinoma with new and acute onset of ocular symptoms and emphasizes the need for urgent systemic evaluation and treatment.

Safety, efficacy, and timing of Nd:YAG laser goniopuncture after nonpenetrating deep sclerectomy for glaucoma: A retrospective cohort study.

BACKGROUND: Neodymium-doped yttrium aluminum garnet laser goniopuncture is an adjuvant procedure for nonpenetrating deep sclerectomy. We investigated optimal laser goniopuncture timing and the effect of laser iridoplasty on success rates. METHODS: This single-center retrospective cohort study compared intraocular pressure control in patients with early versus late laser goniopuncture after nonpenetrating deep sclerectomy and evaluated the effects of laser iridoplasty pretreatment. A 3-month cut-off was used to define early versus late laser goniopuncture. The primary outcome was the proportion of patients maintaining intraocular pressure control according to definitions of complete (no medications) and qualified (with medications) success at 15, 18, and 21 mmHg thresholds. Data were analyzed using right-censored Kaplan-Meier estimation and log-rank testing. RESULTS: A total of 124 eyes of 124 patients were analyzed. Complete success rates after 3 years were 9.2%, 14.6%, and 23.3% for early laser goniopuncture and 21.8%, 26.0%, and 55.4% for late laser goniopuncture for 15, 18, and 21 mmHg, respectively (all p < .01). Qualified success rates after 3 years were 16.6%, 24.8%, and 40.9% for early laser goniopuncture and 21.5%, 56.1%, and 69.6% for late laser goniopuncture for 15, 18, and 21 mmHg, respectively (p = .096, .0026, .0061). Late laser goniopuncture was associated with decreased risk of iris incarceration and bleb collapse. Iridoplasty pretreatment was not associated with improved outcomes. CONCLUSION: Late laser goniopuncture (3-month cut-off) was associated with better intraocular pressure control and less adverse events than early laser goniopuncture.

Mitochondrial chaperones in human health and disease.

Molecular chaperones are a family of proteins that maintain cellular protein homeostasis through non-covalent peptide folding and quality control mechanisms. The chaperone proteins found within mitochondria play significant protective roles in mitochondrial biogenesis, quality control, and stress response mechanisms. Defective mitochondrial chaperones have been implicated in aging, neurodegeneration, and cancer. In this review, we focus on the two most prominent mitochondrial chaperones: mtHsp60 and mtHsp70. These proteins demonstrate different cellular localization patterns, interact with different targets, and have different functional activities. We discuss the structure and function of these prominent mitochondrial chaperone proteins and give an update on newly discovered regulatory mechanisms and disease implications.

Acute Retinal Necrosis in an Immunocompetent Patient Treated With Intravitreal Ganciclovir.

Acute retinal necrosis (ARN) is an inflammatory syndrome of high clinical concern; untreated or misdiagnosed cases may progress to optic neuropathy or retinal detachment, leading to irreversible blindness. ARN affects men and women equally and is often seen in immunocompromised patients but is also known to present in immunocompetent patients. It is usually due to systemic viral infection with secondary vitreoretinal inflammation. Most commonly, the first-line management of ARN is oral or intravenous antiviral therapy. Here, we report the case of an immunocompetent patient presenting with necrotizing retinopathy secondary to ARN. This patient was treated with oral valacyclovir and then intravenous acyclovir with no improvement. However, intravitreal injection of ganciclovir successfully halted the progression of ARN and led to the preservation of vision in the patient.

A sensitivity analysis of probability maps in deep-learning-based anatomical segmentation.

PURPOSE: Deep-learning-based segmentation models implicitly learn to predict the presence of a structure based on its overall prominence in the training dataset. This phenomenon is observed and accounted for in deep-learning applications such as natural language processing but is often neglected in segmentation literature. The purpose of this work is to demonstrate the significance of class imbalance in deep-learning-based segmentation and recommend tuning of the neural network optimization objective. METHODS: An architecture and training procedure were chosen to represent common models in anatomical segmentation. A family of 5-block 2D U-Nets were independently trained to segment 10 structures from the Cancer Imaging Archive's Head-Neck-Radiomics-HN1 dataset. We identify the optimal threshold for our models according to their Dice score on the validation datasets and consider perturbations about the optimum. A measure of structure prominence in segmentation datasets is defined, and its impact on the optimal threshold is analyzed. Finally, we consider the use of a 2D Dice objective in addition to binary cross entropy. RESULTS: We observe significant decreases in perceived model performance with conventional 0.5-thresholding. Perturbations of as little as ±0.05 about the optimum threshold induce a median reduction in Dice score of 11.8% for our models. There is statistical evidence to suggest a weak correlation between training dataset prominence and optimal threshold (Pearson r = 0.92 and p ≈ 10 - 4 ). We find that network optimization with respect to the 2D Dice score itself significantly reduces variability due to thresholding but does not unequivocally create the best segmentation models when assessed with distance-based segmentation metrics. CONCLUSION: Our results suggest that those practicing deep-learning-based contouring should consider their postprocessing procedures as a potential avenue for improved performance. For intensity-based postprocessing, we recommend a mixed objective function consisting of the traditional binary cross entropy along with the 2D Dice score.

Cholesterol-lowering activity of natural mono- and sesquiterpenoid compounds in essential oils: A review and investigation of mechanisms using in silico protein-ligand docking.

Mono- and sesquiterpenoids are the main chemical constituents of essential oils. Essential oils and their constituents have received increasing attention for lipid-lowering properties in both cell and animal models. Despite the chemical diversity of essential oil compounds, the effects of many of these compounds on cholesterol metabolism are highly similar. In this report, we review the literature regarding the effects of essential oils and their terpenoid constituents on cholesterol homeostasis, and explore likely mechanisms using protein-ligand docking. We identified 98 experimental and seven clinical studies on essential oils, isolated compounds, and blends; 100 of these described improvements either in blood cholesterol levels or in sterol metabolic pathways. Our review and docking analysis confirmed two likely mechanisms common to many essential oil compounds: (1) direct agonism of peroxisome-proliferator-activated receptors, and (2) direct interaction with sterol-sensing domains, motifs found in key sterol regulatory proteins including sterol regulatory element binding protein cleavage activating protein and HMG-CoA reductase. Notably, these direct interactions lead to decreased transcription and accelerated degradation of HMG-CoA reductase. Our work suggests that terpene derivatives in essential oils have cholesterol-lowering activity and could potentially work synergistically with statins, however, further high quality studies are needed to establish their clinical efficacy.

Diagnostic value of circulating cell-free mtDNA in patients with suspected thyroid cancer: ND4/ND1 ratio as a new potential plasma marker.

Thyroid cancer is the most common endocrine malignancy, and its incidence continues to rise. For clinicians with cancer patients, choosing and interpreting diagnostic laboratory studies has become increasingly important. Previously, changes in plasma free mitochondrial DNA levels have been found in colorectal, breast, lung, and urinary cancers, and have demonstrated diagnostic value. In this study, we investigated whether the occurrence and development of thyroid cancer might be predicted using mtDNA copy number (ND1), mtDNA integrity (ND4/ND1) and levels of cell-free nDNA (GAPDH). We analyzed ND1, ND4, and GAPDH levels in plasma and blood cells from 75 patients with thyroid cancer, 40 patients with nodular goiter, and 107 normal controls using real-time PCR. Although both the thyroid nodule and thyroid cancer patients had significantly increased ND1 levels, the ND4/ND1 ratio in the thyroid cancer group was higher than the thyroid nodule group (P < 0.05), and significantly higher than the normal control group (P < 0.01). Plasma levels of nuclear DNA (GAPDH) in the thyroid cancer group were also higher compared to normal (P < 0.05). These results indicate that increased intactness of plasma free mtDNA is associated with increased levels of plasma cell-free nDNA, and that the ND4/ND1 ratio has the potential to be a new detection indicator in thyroid cancer. Furthermore, we classified thyroid cancer patients according to clinical data including age, tumor size, and metastasis. We found significantly higher levels of GAPDH in malignant tissues. Because ND4/ND1 correlated with plasma GAPDH in the plasma studies, this also suggests a potential relationship between ND4 intactness and thyroid tumor tissue size. Taken together, our findings suggest a tumor-specific process involving increased release of intact mtDNA, detectable in the plasma, which differentiates normal patients from patients with thyroid cancer.

Deep learning-based survival analysis for brain metastasis patients with the national cancer database.

PURPOSE: Prognostic indices such as the Brain Metastasis Graded Prognostic Assessment have been used in clinical settings to aid physicians and patients in determining an appropriate treatment regimen. These indices are derivative of traditional survival analysis techniques such as Cox proportional hazards (CPH) and recursive partitioning analysis (RPA). Previous studies have shown that by evaluating CPH risk with a nonlinear deep neural network, DeepSurv, patient survival can be modeled more accurately. In this work, we apply DeepSurv to a test case: breast cancer patients with brain metastases who have received stereotactic radiosurgery. METHODS: Survival times, censorship status, and 27 covariates including age, staging information, and hormone receptor status were provided for 1673 patients by the NCDB. Monte Carlo cross-validation with 50 samples of 1400 patients was used to train and validate the DeepSurv, CPH, and RPA models independently. DeepSurv was implemented with L2 regularization, batch normalization, dropout, Nesterov momentum, and learning rate decay. RPA was implemented as a random survival forest (RSF). Concordance indices of test sets of 140 patients were used for each sample to assess the generalizable predictive capacity of each model. RESULTS: Following hyperparameter tuning, DeepSurv was trained at 32 min per sample on a 1.33 GHz quad-core CPU. Test set concordance indices of 0.7488 ± 0.0049, 0.6251 ± 0.0047, and 0.7368 ± 0.0047, were found for DeepSurv, CPH, and RSF, respectively. A Tukey HSD test demonstrates a statistically significant difference between the mean concordance indices of the three models. CONCLUSION: Our results suggest that deep learning-based survival prediction can outperform traditional models, specifically in a case where an accurate prognosis is highly clinically relevant. We recommend that where appropriate data are available, deep learning-based prognostic indicators should be used to supplement classical statistics.

Redox regulation by SOD2 modulates colorectal cancer tumorigenesis through AMPK-mediated energy metabolism.

Colorectal cancer (CRC) is a common malignancy. Many reports have implicated aberrant mitochondrial activity in the progression of CRC, with particular emphasis on the dysregulation of redox signaling and oxidative stress. In this study, we focused on manganese superoxide dismutase (MnSOD/SOD2), a key antioxidant enzyme, which maintains intracellular redox homeostasis. Current literature presents conflicting mechanisms for how SOD2 influences tumorigenesis and tumor progression. Here, we explored the role of SOD2 in CRC specifically. We found high levels of SOD2 expression in CRC tissues. We carried out a series of experiments to determine whether knockdown of SOD2 expression in CRC cell lines would reverse features of tumorigenesis. We found that reduced SOD2 expression decreased cell proliferation, migration, and invasion activity in CRC cells. Results from an additional series of experiments on mitochondrial function implicated a dual role for SOD2 in promoting CRC progression. First, proper level of SOD2 helped CRC cells maintain mitochondrial function by disposal of superoxide (O2.- ). Second, over-expression of SOD2 induced H2 O2 -mediated tumorigenesis by upregulating AMPK and glycolysis. Our results indicate that SOD2 may promote the occurrence and development of CRC by regulating the energy metabolism mediated by AMPK signaling pathways.

Emerging model systems and treatment approaches for Leber's hereditary optic neuropathy: Challenges and opportunities.

Leber's hereditary optic neuropathy (LHON) is a mitochondrial disease mainly affecting retinal ganglion cells (RGCs). The pathogenesis of LHON remains ill-characterized due to a historic lack of effective disease models. Promising models have recently begun to emerge; however, less effective models remain popular. Many such models represent LHON using non-neuronal cells or assume that mutant mtDNA alone is sufficient to model the disease. This is problematic because context-specific factors play a significant role in LHON pathogenesis, as the mtDNA mutation itself is necessary but not sufficient to cause LHON. Effective models of LHON should be capable of demonstrating processes that distinguish healthy carrier cells from diseased cells. In light of these considerations, we review the pathophysiology of LHON as it relates to old, new and future models. We further discuss treatments for LHON and unanswered questions that might be explored using these new model systems.

The Effects of Various Essential Oils on Epilepsy and Acute Seizure: A Systematic Review.

Many essential oils (EOs) have anticonvulsant activity and might benefit people with epilepsy. Lemongrass, lavender, clove, dill, and other EOs containing constituents such as asarone, carvone, citral, eugenol, or linalool are good candidates for evaluation as antiepileptic drugs. On the other hand, some EOs have convulsant effects and may trigger seizures in both epileptic and healthy individuals. Internal use of EOs like sage, hyssop, rosemary, camphor, pennyroyal, eucalyptus, cedar, thuja, and fennel can cause epileptic seizures because they contain thujone, 1,8-cineole, camphor, or pinocamphone, which have been identified as convulsive agents. While more research is needed to confirm their mechanisms of action, it appears that the convulsant or anticonvulsant properties of essential oils are largely due to (1) their ability to modulate the GABAergic system of neurotransmission and (2) their capacity to alter ionic currents through ion channels. This review presents a systematic analysis of the current research on EOs and epilepsy, including human case studies, animal models, and in vitro studies.

Effects of a massage-like essential oil application procedure using Copaiba and Deep Blue oils in individuals with hand arthritis.

BACKGROUND AND PURPOSE: Existing research suggests that both massage and essential oils may have analgesic and anti-inflammatory benefits. We investigate the benefits of the AromaTouch Hand Technique® (ATHT), a procedure that combines a moderate pressure touch with the application of essential oils to the hand, in individuals with hand arthritis. METHODS AND MATERIALS: Thirty-six participants with rheumatoid arthritis, osteoarthritis, and/or chronic inflammation received ATHTs with either a 50/50 preparation of Deep Blue® and Copaiba oil or a coconut oil placebo twice daily for 5 consecutive days. Changes in maximum flexion in finger and thumb joints, items from the Arthritis Hand Function Test, and hand pain scores were evaluated. RESULTS: Participants treated with the essential oil preparation required significantly less time to complete dexterity tasks and showed about 50% decrease in pain scores, increased finger strength, and significantly increased angle of maximum flexion compared to subjects treated with coconut oil. CONCLUSION: The ATHT with Copaiba and Deep Blue may have ameliorative effects on hand arthritis.

Black pepper (Piper nigrum) essential oil demonstrates tissue remodeling and metabolism modulating potential in human cells.

Very few studies have investigated the biological activities of black pepper essential oil (BPEO) in human cells. Therefore, in the current study, we examined the biological activities of BPEO in cytokine-stimulated human dermal fibroblasts by analyzing the levels of 17 important protein biomarkers pertinent to inflammation and tissue remodeling. BPEO exhibited significant antiproliferative activity in these skin cells and significantly inhibited the production of Collagen I, Collagen III, and plasminogen activator inhibitor 1. In addition, we studied the effect of BPEO on the regulation of genome-wide expression and found that BPEO diversely modulated global gene expression. Further analysis showed that BPEO affected many important genes and signaling pathways closely related to metabolism, inflammation, tissue remodeling, and cancer signaling. This study is the first to provide evidence of the biological activities of BPEO in human dermal fibroblasts. The data suggest that BPEO possesses promising potential to modulate the biological processes of tissue remodeling, wound healing, and metabolism. Although further research is required, BPEO appears to be a good therapeutic candidate for a variety of health conditions including wound care and metabolic diseases. Research into the biological and pharmacological mechanisms of action of BPEO and its major active constituents is recommended.