Pain trajectories after bilateral orthotopic lung transplantation surgery performed via a clamshell incision.

INTRODUCTION: The nature, intensity, and progression of acute pain after bilateral orthotopic lung transplantation (BOLT) performed via a clamshell incision has not been well investigated. We aimed to describe acute pain after clamshell incisions using pain trajectories for the study cohort, in addition to stratifying patients into separate pain trajectory groups and investigating their association with donor and recipient perioperative variables. METHODS: After obtaining IRB approval, we retrospectively included all patients ≥18 years old who underwent primary BOLT via clamshell incision at a single center between January 1, 2017, and June 30, 2022. We modeled the overall pain trajectory using pain scores collected over the first seven postoperative days and identified separate pain trajectory classes via latent class analysis. RESULTS: Three hundred one adult patients were included in the final analysis. Three separate pain trajectory groups were identified, with most patients (72.8%) belonging to a well-controlled, stable pain trajectory. Uncontrolled pain was either observed in the early postoperative period (10%), or in the late postoperative period (17.3%). Late postoperative peaking trajectory patients were younger (p = .008), and sicker with a higher lung allocation score (p = .005), receiving preoperative mechanical ventilation (p < .001), or VV-ECMO support (p < .001). CONCLUSION: Despite the extensive nature of a clamshell incision, most pain trajectories in BOLT patients had a well-controlled stable pain profile. The benign nature of pain profiles in our patient population may be attributed to the routine institutional practice of early thoracic epidural analgesia for BOLT patients unless contraindicated.

A Comparison of Long-Term Outcomes in Patients Managed With Venovenous Extracorporeal Membrane Oxygenation in the First and Second Waves of the COVID-19 Pandemic in the United Kingdom.

OBJECTIVES: Early studies of venovenous extracorporeal membrane oxygenation (ECMO) in COVID-19 have revealed similar outcomes to historical cohorts. Changes in the disease and treatments have led to differences in the patients supported on venovenous ECMO in the first and second waves. We aimed to compare these two groups in both the acute and follow-up phase. DESIGN: Retrospective single-center cohort study comparing mortality at censoring date (November 30, 2021) and decannulation, patient characteristics, complications and lung function and quality of life (QOL-by European Quality of Life 5 Dimensions 3 Level Version) at first follow-up in patients supported on venovenous ECMO between wave 1 and wave 2 of the COVID-19 pandemic. SETTING: Critical care department of a severe acute respiratory failure service. PATIENTS: Patients supported on ECMO for COVID-19 between wave 1 (March 17, 2020, to August 31, 2020) and wave 2 (January 9, 2020, to May 25, 2021). INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: One hundred twenty-three patients were included in our analysis. Survival at censoring date (χ 2 , 6.35; p = 0.012) and decannulation (90.4% vs 70.0%; p < 0.001) was significantly lower in the second wave, while duration of ECMO run was longer (12.0 d [18.0-30.0 d] vs 29.5 d [15.5-58.3 d]; p = 0.005). Wave 2 patients had longer application of noninvasive ventilation (NIV) prior to ECMO and a higher frequency of barotrauma. Patient age and NIV use were independently associated with increased mortality (odds ratio 1.07 [1.01-1.14]; p = 0.025 and 3.37 [1.12-12.60]; p = 0.043, respectively). QOL and lung function apart from transfer coefficient of carbon monoxide corrected for hemoglobin was similar at follow-up across the waves. CONCLUSIONS: Most patients with COVID-19 supported on ECMO in both waves survived in the short and longer term. At follow-up patients had similar lung function and QOL across the two waves. This suggests that ECMO has an ongoing role in the management of a carefully selected group of patients with COVID-19.

Segmenting Skin Biopsy Images with Coarse and Sparse Annotations using U-Net.

The number of melanoma diagnoses has increased dramatically over the past three decades, outpacing almost all other cancers. Nearly 1 in 4 skin biopsies is of melanocytic lesions, highlighting the clinical and public health importance of correct diagnosis. Deep learning image analysis methods may improve and complement current diagnostic and prognostic capabilities. The histologic evaluation of melanocytic lesions, including melanoma and its precursors, involves determining whether the melanocytic population involves the epidermis, dermis, or both. Semantic segmentation of clinically important structures in skin biopsies is a crucial step towards an accurate diagnosis. While training a segmentation model requires ground-truth labels, annotation of large images is a labor-intensive task. This issue becomes especially pronounced in a medical image dataset in which expert annotation is the gold standard. In this paper, we propose a two-stage segmentation pipeline using coarse and sparse annotations on a small region of the whole slide image as the training set. Segmentation results on whole slide images show promising performance for the proposed pipeline.

The impact of a barrier enclosure on time to tracheal intubation: a randomized controlled trial.

PURPOSE: Novel devices such as the barrier enclosure were developed in hopes of improving provider safety by limiting SARS-CoV-2 transmission during tracheal intubation. Nevertheless, concerns arose regarding a lack of rigorous efficacy and safety data for these devices. We conducted a randomized controlled trial to evaluate the impact of the barrier enclosure on time to tracheal intubation. METHOD: After Research Ethics Board approval, elective surgical patients with normal airway predictors were randomly allocated 1:1 to tracheal intubation with or without a barrier enclosure. The primary outcome was time to tracheal intubation. Secondary outcomes included first-pass success rate, total time of airway manipulation, anesthesiologists' perception of intubation difficulty, likelihood of use in SARS-CoV-2-positive patients, and patients' perception of comfort and acceptability. RESULTS: There were 48 participants in the barrier enclosure group and 46 participants in the control group. The mean (standard deviation [SD]) time to tracheal intubation was 62 (29) sec with barrier closure and 53 (27) sec without barrier enclosure (mean difference, 9 sec; 95% confidence interval, - 3 to 20; P = 0.14). Anesthesiologists rated the difficulty of intubation higher with barrier enclosure (mean [SD] visual analogue scale score, 27 [26] mm vs 9 [17] mm; P < 0.001). There were no significant differences in other secondary outcomes. CONCLUSION: In healthy surgical patients with normal airway predictors, the use of a barrier enclosure during tracheal intubation did not significantly prolong time to intubation or decrease first-pass intubation success. Nevertheless, there was an increase in difficulty of intubation perceived by the anesthesiologists with use of a barrier enclosure. TRIAL REGISTRATION: www.clinicaltrials.gov (NCT04366141); registered 28 April 2020.

RéSUMé: OBJECTIF: De nouveaux dispositifs tels que des boîtes de protection ont été mis au point dans l'espoir d'améliorer la sécurité des fournisseurs de soins en limitant la transmission du SRAS-CoV-2 pendant l'intubation endotrachéale. Néanmoins, des inquiétudes ont été soulevées au sujet d'un manque de données rigoureuses sur l'efficacité et l'innocuité de ces dispositifs. Nous avons réalisé une étude randomisée contrôlée afin d'évaluer l'impact d'une boîte de protection sur le temps de l'intubation endotrachéale. MéTHODE: Après avoir reçu l'approbation du Comité d'éthique de la recherche, des patients de chirurgie élective présentant des prédicteurs des voies aériennes normales ont été aléatoirement répartis à un ratio de 1:1 pour une intubation endotrachéale avec ou sans boîte de protection. Le critère d'évaluation principal était le temps nécessaire à l'intubation endotrachéale. Les critères d'évaluation secondaires comprenaient le taux de réussite à la première tentative, le temps total de manipulation des voies aériennes, la perception par les anesthésiologistes de la difficulté d'intubation, la probabilité d'utilisation chez les patients atteints du SRAS-CoV-2, et la perception de confort et d'acceptabilité des patients. RéSULTATS: Il y avait 48 participants dans le groupe avec boîte et 46 participants dans le groupe témoin. Le temps moyen (écart type [ÉT]) pour l'intubation endotrachéale était de 62 (29) sec avec la boîte et de 53 (27) sec sans la boîte (différence moyenne, 9 sec; intervalle de confiance de 95 %, − 3 à 20; P = 0,14). Les anesthésiologistes ont estimé que la difficulté d'intubation était plus élevée avec une boîte de protection (score moyen sur l'échelle visuelle analogique [ÉT], 27 [26] mm vs 9 [17] mm; P < 0,001). Il n'y avait pas de différences pour les autres critères d'évaluation secondaires. CONCLUSION: Chez les patients chirurgicaux en bonne santé avec des prédicteurs de voies aériennes normales, l'utilisation d'une boîte de protection pendant l'intubation endotrachéale n'a pas prolongé de manière significative le temps d'intubation ni réduit le taux de réussite de l'intubation à la première tentative. Néanmoins, il y avait une augmentation de la difficulté d'intubation perçue par les anesthésiologistes avec l'utilisation d'une boîte de protection. ENREGISTREMENT DE L'éTUDE: www.clinicaltrials.gov (NCT04366141); enregistrée le 28 avril 2020.

Nanotechnology in Retinal Disease: Current Concepts and Future Directions.

The retina is one of the most complex and extraordinary human organs affected by genetic, metabolic, and degenerative diseases, resulting in blindness for ∼1.3 million people in the United States and over 40 million people worldwide. This translates into a huge loss of productivity, especially among younger patients with inherited retinal diseases (IRDs) and diabetic retinopathy. Age-related macular degeneration accounts for 90% of all blindness cases worldwide. The prevalence of this condition is projected to reach over 5 million individuals over the next 3 decades. There are also >20 IRD phenotypes, affecting >2 million people worldwide. Nanobiotechnology uses nanotechnology for biological applications, making use of biological materials either conceptually or directly in the fabrication of new materials. Bionanotechnology, on the other hand, uses molecular biology for the purpose of creating nanostructures (ie, structures with at least 1 dimension <100 nm). Retinal applications of these technologies are developing at a rapid pace. This review includes the most current nanotechnological applications in retinal diagnostics, theranostics, drug delivery, and targeting, including the potential for nonviral vehicles such as liposomes, micelles, and dendrimers, which pose advantages over viral vectors in retinal drug delivery. Furthermore, we discuss current and future applications as surgical adjuncts and in regenerative medicine as they pertain to retinal disease. Structure and function of nanoparticles such as carbon nanotubules, quantum dots, and magnetic nanoparticles, as well as diagnostic technologies such as next-generation DNA sequencing and single-molecule bionanosensing, will also be discussed.

Benzyl N-[(Z)-(1-methyl-2-sulfanyl-propyl-idene)amino]-carbamodithio-ate.

The title compound, C(12)H(16)N(2)S(3), was obtained by the condensation reaction of S-benzyl dithio-carbazate and 3-mercaptobutan-2-one. The phenyl ring and thiol (SH) group are approximately perpendicular [S-C-C-C and N-C-C-S torsion angles = 67.8 (3) and 116.9 (2)°, respectively] to the rest of the mol-ecule. In the crystal, mol-ecules are linked by weak S-H⋯S and N-H⋯S hydrogen bonds, π-π inter-actions between the benzene rings [centroid-centroid distance = 3.823 (2) Å] and C-H⋯π inter-actions.

Cesarean Delivery Outcomes for Patients with Coronavirus Disease-2019 in the USA.

INTRODUCTION: Coronavirus disease-2019 (COVID-19) may have increased morbidity and mortality in patients having Cesarean delivery (CD) in the USA. METHODS: We performed a retrospective cohort study of patients who had CD in 2020 using the national inpatient sample. After stratification by COVID-19 status, demographics, comorbidities, complications, mortality, and costs were compared. RESULTS: There were 31,444,222 hospitalizations in the USA in 2020 with a mortality rate of 2.8%. Among these, 1,453,945 patients had COVID-19 and mortality was 13.2%. There were 1,108,755 patients who had CD and 15,550 had COVID-19. Patients with COVID-19 more frequently had Medicaid and were Hispanic. Patients with COVID-19 had more comorbidities including chronic hypertension, diabetes mellitus, pre-eclampsia, and eclampsia. Mortality in CD patients with COVID-19 was 30 in 10,000 patients, while for non-COVID-19 patients, it was 1 in 10,000 patients, P < 0.001. The crude odds ratio for mortality in COVID-19 patients was 32.1 (95% confidence interval = 22.9 to 44.7), P < 0.001 and the adjusted odds ratio was 29.3 (95% confidence interval = 20.7 to 41.4), P < 0.001. CONCLUSIONS: CD patients with COVID-19 had 30-fold higher mortality before widespread vaccination was available with Hispanic and Medicaid patients disproportionately impacted. Potential explanations for this disparity include reduced access to personal protective equipment (e.g., masks) and testing, as well as socio-economic factors. Further research is needed to understand the factors that contributed to disparities in infection and clinical outcomes among obstetric patients during the COVID-19 pandemic. In future pandemics, enhanced efforts will be needed to protect economically disadvantaged women who are pregnant.

The pandemic's unseen wounds: COVID-19's profound effects on mental health.

Objective: This review aims to explore the impact of the COVID-19 pandemic on mental health, with a focus on the physiological and psychological consequences, including comorbidities. The goal is to understand the direct and indirect populations affected by mental distress and identify potential interventions. Methodology: A comprehensive literature search was conducted using various databases, including Google Scholar, ResearchGate, ScienceDirect, PubMed, PLoS One, and Web of Science. The search utilized relevant keywords to investigate the direct and indirect impacts of COVID-19 on mental health. The selected articles were critically evaluated and analyzed to identify key findings and insights. Main findings: Mental health, being an intrinsic component of overall well-being, plays a vital role in physiological functioning. The COVID-19 pandemic, caused by the emergence of the novel SARS-CoV-2 virus, has had a devastating global impact. Beyond the respiratory symptoms, individuals recovering from COVID-19 commonly experience additional ailments, such as arrhythmia, depression, anxiety, and fatigue. Healthcare professionals on the frontlines face an elevated risk of mental illness. However, it is crucial to recognize that the general population also grapples with comparable levels of mental distress. Conclusion: The COVID-19 pandemic has underscored the significance of addressing mental health concerns. Various strategies can help mitigate the impact, including counselling, fostering open lines of communication, providing mental support, ensuring comprehensive patient care, and administering appropriate medications. In severe cases, treatment may involve the supplementation of essential vitamins and antidepressant therapy. By understanding the direct and indirect impacts of COVID-19 on mental health, healthcare providers and policymakers can develop targeted interventions to support individuals and communities affected by the pandemic. Continued research and collaborative efforts are essential to address this pervasive issue effectively.

Infarct exclusion repair of postmyocardial infarction ventricular septal rupture with a hybrid patch and septal occluder device compared with patch only.

Objective: We developed a hybrid technique for repairing post-myocardial infarction (MI) ventricular septal defect (VSD) that combines infarct exclusion with patch and a nitinol-mesh septal occluder device (SOD) to provide a scaffold to support the damaged septal wall. Here, we compare outcomes of patients with post-MI VSD repaired using patch only or hybrid patch/SOD. Methods: Patients undergoing post-MI VSD repair at our institution from 2013 to 2022 who received patch alone or patch/SOD repair were analyzed. Primary outcome was survival to hospital discharge. Clinical outcomes and echocardiograms were also analyzed. Results: Over a 9-year period, 24 patients had post-MI VSD repair at our institution with either hybrid patch/SOD (n = 10) or patch only repair (n = 14). VSD size was 18 ± 5.8 mm for patch/SOD and 17 ± 4.6 mm for patch only. In the patch/SOD repair cohort, average size of SOD implant was 23.6 ± 5.6 mm. Mild left ventricular dysfunction was present prerepair and was unchanged postrepair in both groups; however, moderate-to-severe right ventricular (RV) dysfunction was common in both groups before repair. RV function worsened or persisted as severe in 10% of hybrid versus 54% of patch-only patients postrepair. Tricuspid annular systolic excursion and RV:left ventricle diameter ratio, quantitative metrics of RV function, improved after patch/SOD repair. No intraoperative mortality occurred in either group. Postoperative renal, hepatic, and respiratory failure requiring tracheostomy was common in both groups. Survival to hospital discharge in both cohorts was 70%. Conclusions: Post-MI VSD repair with patch/SOD has comparable short-term outcomes with patch alone. Addition of a SOD to patch repair provides a scaffold that may enhance the repair of post-MI VSD with patch exclusion.

DiCENet: Dimension-Wise Convolutions for Efficient Networks.

We introduce a novel and generic convolutional unit, DiCE unit, that is built using dimension-wise convolutions and dimension-wise fusion. The dimension-wise convolutions apply light-weight convolutional filtering across each dimension of the input tensor while dimension-wise fusion efficiently combines these dimension-wise representations; allowing the DiCE unit to efficiently encode spatial and channel-wise information contained in the input tensor. The DiCE unit is simple and can be seamlessly integrated with any architecture to improve its efficiency and performance. Compared to depth-wise separable convolutions, the DiCE unit shows significant improvements across different architectures. When DiCE units are stacked to build the DiCENet model, we observe significant improvements over state-of-the-art models across various computer vision tasks including image classification, object detection, and semantic segmentation. On the ImageNet dataset, the DiCENet delivers 2-4 percent higher accuracy than state-of-the-art manually designed models (e.g., MobileNetv2 and ShuffleNetv2). Also, DiCENet generalizes better to tasks (e.g., object detection) that are often used in resource-constrained devices in comparison to state-of-the-art separable convolution-based efficient networks, including neural search-based methods (e.g., MobileNetv3 and MixNet).

End-to-End diagnosis of breast biopsy images with transformers.

Diagnostic disagreements among pathologists occur throughout the spectrum of benign to malignant lesions. A computer-aided diagnostic system capable of reducing uncertainties would have important clinical impact. To develop a computer-aided diagnosis method for classifying breast biopsy images into a range of diagnostic categories (benign, atypia, ductal carcinoma in situ, and invasive breast cancer), we introduce a transformer-based hollistic attention network called HATNet. Unlike state-of-the-art histopathological image classification systems that use a two pronged approach, i.e., they first learn local representations using a multi-instance learning framework and then combine these local representations to produce image-level decisions, HATNet streamlines the histopathological image classification pipeline and shows how to learn representations from gigapixel size images end-to-end. HATNet extends the bag-of-words approach and uses self-attention to encode global information, allowing it to learn representations from clinically relevant tissue structures without any explicit supervision. It outperforms the previous best network Y-Net, which uses supervision in the form of tissue-level segmentation masks, by 8%. Importantly, our analysis reveals that HATNet learns representations from clinically relevant structures, and it matches the classification accuracy of 87 U.S. pathologists for this challenging test set.

Hepcidin in anemia of chronic heart failure.

Anemia is a common finding among patients with chronic heart failure (HF). Although comorbidities, such as kidney failure, might contribute to the pathogenesis of anemia, many patients with HF do not have any other obvious etiology for their anemia. We investigated whether anemia in HF is associated with an elevation in hepcidin concentration. We used time-of-flight mass spectrometry to measure hepcidin concentration in urine and serum samples of patients with HF and in control subjects. We found that the concentration of hepcidin was lower in urine samples of patients with HF compared with those of control subjects. Serum hepcidin was also reduced in HF but was not significantly lower than that in controls. There were no significant differences between hepcidin levels in patients with HF and anemia compared with patients with HF and normal hemoglobin level. We concluded that hepcidin probably does not play a major role in pathogenesis of anemia in patients with chronic HF.

Long-term potency, sterility, and stability of vancomycin, ceftazidime, and moxifloxacin for treatment of bacterial endophthalmitis.

PURPOSE: To determine the long-term potency, sterility, and stability of vancomycin, ceftazidime, and moxifloxacin prepared in single-use polypropylene syringes for intravitreal injection. METHODS: Experimental study. Vancomycin 1 mg/0.1 mL, ceftazidime 2 mg/0.1 mL, and moxifloxacin 160 µg/0.1 mL were compounded and prepared in 1-mL polypropylene syringes and stored at 4 °C, -20 °C, and -80 °C. Antibiotic potency, sterility, pH, osmolality, and concentration were tested at baseline and at 1, 2, 4, 8, 12, and 24 weeks after preparation. RESULTS: Potency, sterility, and stability were preserved for all 3 antibiotics at all temperatures out to 24 weeks, although there was a trend toward reduced potency at Week 24 for vancomycin and ceftazidime stored at 4°C. The largest zones of inhibition for Staphylococcus epidermidis and S. aureus were consistently demonstrated by moxifloxacin. CONCLUSION: Vancomycin, ceftazidime, and moxifloxacin prepared in single-use polypropylene syringes retain potency, sterility, and stability out to 24 weeks when stored at -20 °C or -80 °C. The results of this study may have important implications for the current management of endophthalmitis.

A Comparative Study of the Efficacy of WaveOne and NeoEndo Retreatment File System for the Removal of Gutta Percha from the Root Canal.

BACKGROUND: Teeth with apical periodontitis that either persisted or developed after initial root canal treatment is most often indicated for retreatment procedure. The present study was conducted to compare the WaveOne and NeoEndo retreatment file system for the removal of gutta-percha from the root canal. MATERIALS AND METHODS: Fifty extracted mandibular first premolars with single straight roots were divided into two groups, Group I-WaveOne and Group II-NeoEndo. All samples were scanned with cone-beam computed tomography after the root filling and retreatment procedures. The decrease in the volume of filling material after each retreatment protocol was measured. RESULTS: The mean volume of gutta-percha in Group I before retreatment was 5.6 and in Group II was 5.3, after retreatment in Group I was 2.7, and in Group II was 4.2. The mean percentage reduction in Group I was 45.2% and in Group II was 23.8%. The difference was significant (P < 0.05). CONCLUSION: Both systems failed to completely remove gutta-percha. WaveOne was significantly more effective than the NeoEndo retreatment system.

Scale-Aware Transformers for Diagnosing Melanocytic Lesions.

Diagnosing melanocytic lesions is one of the most challenging areas of pathology with extensive intra- and inter-observer variability. The gold standard for a diagnosis of invasive melanoma is the examination of histopathological whole slide skin biopsy images by an experienced dermatopathologist. Digitized whole slide images offer novel opportunities for computer programs to improve the diagnostic performance of pathologists. In order to automatically classify such images, representations that reflect the content and context of the input images are needed. In this paper, we introduce a novel self-attention-based network to learn representations from digital whole slide images of melanocytic skin lesions at multiple scales. Our model softly weighs representations from multiple scales, allowing it to discriminate between diagnosis-relevant and -irrelevant information automatically. Our experiments show that our method outperforms five other state-of-the-art whole slide image classification methods by a significant margin. Our method also achieves comparable performance to 187 practicing U.S. pathologists who interpreted the same cases in an independent study. To facilitate relevant research, full training and inference code is made publicly available at https://github.com/meredith-wenjunwu/ScATNet.

Analysis of Regions of Interest and Distractor Regions in Breast Biopsy Images.

This paper studies why pathologists can misdiagnose diagnostically challenging breast biopsy cases, using a data set of 240 whole slide images (WSIs). Three experienced pathologists agreed on a consensus reference ground-truth diagnosis for each slide and also a consensus region of interest (ROI) from which the diagnosis could best be made. A study group of 87 other pathologists then diagnosed test sets (60 slides each) and marked their own regions of interest. Diagnoses and ROIs were categorized such that if on a given slide, their ROI differed from the consensus ROI and their diagnosis was incorrect, that ROI was called a distractor. We used the HATNet transformer-based deep learning classifier to evaluate the visual similarities and differences between the true (consensus) ROIs and the distractors. Results showed high accuracy for both the similarity and difference networks, showcasing the challenging nature of feature classification with breast biopsy images. This study is important in the potential use of its results for teaching pathologists how to diagnose breast biopsy slides.

Classifying Breast Histopathology Images with a Ductal Instance-Oriented Pipeline.

In this study, we propose the Ductal Instance-Oriented Pipeline (DIOP) that contains a duct-level instance segmentation model, a tissue-level semantic segmentation model, and three-levels of features for diagnostic classification. Based on recent advancements in instance segmentation and the Mask RCNN model, our duct-level segmenter tries to identify each ductal individual inside a microscopic image; then, it extracts tissue-level information from the identified ductal instances. Leveraging three levels of information obtained from these ductal instances and also the histopathology image, the proposed DIOP outperforms previous approaches (both feature-based and CNN-based) in all diagnostic tasks; for the four-way classification task, the DIOP achieves comparable performance to general pathologists in this unique dataset. The proposed DIOP only takes a few seconds to run in the inference time, which could be used interactively on most modern computers. More clinical explorations are needed to study the robustness and generalizability of this system in the future.

Machine learning techniques for mitoses classification.

BACKGROUND: Pathologists analyze biopsy material at both the cellular and structural level to determine diagnosis and cancer stage. Mitotic figures are surrogate biomarkers of cellular proliferation that can provide prognostic information; thus, their precise detection is an important factor for clinical care. Convolutional Neural Networks (CNNs) have shown remarkable performance on several recognition tasks. Utilizing CNNs for mitosis classification may aid pathologists to improve the detection accuracy. METHODS: We studied two state-of-the-art CNN-based models, ESPNet and DenseNet, for mitosis classification on six whole slide images of skin biopsies and compared their quantitative performance in terms of sensitivity, specificity, and F-score. We used raw RGB images of mitosis and non-mitosis samples with their corresponding labels as training input. In order to compare with other work, we studied the performance of these classifiers and two other architectures, ResNet and ShuffleNet, on the publicly available MITOS breast biopsy dataset and compared the performance of all four in terms of precision, recall, and F-score (which are standard for this data set), architecture, training time and inference time. RESULTS: The ESPNet and DenseNet results on our primary melanoma dataset had a sensitivity of 0.976 and 0.968, and a specificity of 0.987 and 0.995, respectively, with F-scores of .968 and .976, respectively. On the MITOS dataset, ESPNet and DenseNet showed a sensitivity of 0.866 and 0.916, and a specificity of 0.973 and 0.980, respectively. The MITOS results using DenseNet had a precision of 0.939, recall of 0.916, and F-score of 0.927. The best published result on MITOS (Saha et al. 2018) reported precision of 0.92, recall of 0.88, and F-score of 0.90. In our architecture comparisons on MITOS, we found that DenseNet beats the others in terms of F-Score (DenseNet 0.927, ESPNet 0.890, ResNet 0.865, ShuffleNet 0.847) and especially Recall (DenseNet 0.916, ESPNet 0.866, ResNet 0.807, ShuffleNet 0.753), while ResNet and ESPNet have much faster inference times (ResNet 6 s, ESPNet 8 s, DenseNet 31 s). ResNet is faster than ESPNet, but ESPNet has a higher F-Score and Recall than ResNet, making it a good compromise solution. CONCLUSION: We studied several state-of-the-art CNNs for detecting mitotic figures in whole slide biopsy images. We evaluated two CNNs on a melanoma cancer dataset and then compared four CNNs on a public breast cancer data set, using the same methodology on both. Our methodology and architecture for mitosis finding in both melanoma and breast cancer whole slide images has been thoroughly tested and is likely to be useful for finding mitoses in any whole slide biopsy images.

Intravitreal bevacizumab for the treatment of refractory diabetic macular edema.

BACKGROUND AND OBJECTIVE: To report the effect of intravitreal bevacizumab on visual acuity and central retinal thickness (CRT) in refractory diabetic macular edema. PATIENTS AND METHODS: Records of 60 eyes of 54 consecutive patients who underwent intravitreal bevacizumab therapy for refractory diabetic macular edema were reviewed. All eyes received intravitreal bevacizumab 1.25 mg/0.05 mL, and 36 eyes underwent pretreatment and post-treatment optical coherence tomography. Mean follow-up was 7.4 months. RESULTS: Pretreatment mean visual acuity plus or minus standard deviation was 0.71 +/- 0.28 logarithm of the minimum angle of resolution (LogMAR) Snellen letters. At final follow-up, mean visual acuity had improved to 0.66 +/- 0.30 LogMAR (P = .0543). Mean baseline CRT was 440 +/- 106 microm, and follow-up mean CRT was 386 +/- 129 microm (P = .008). Vitrectomized eyes had worse visual acuity and CRT outcomes (P = .002 and P = .028, respectively) compared with nonvitrectomized eyes. CONCLUSION: Intravitreal bevacizumab may provide a functional and anatomic benefit in eyes with persistent diabetic macular edema despite previous treatments.

Does Intolerance of Single-Lung Ventilation Preclude Robotic Off-Pump Totally Endoscopic Coronary Bypass Surgery?

OBJECTIVE: Robotic off-pump totally endoscopic coronary artery bypass (TECAB) usually requires isolated single (right) lung ventilation to adequately expose the surgical site. However, in some patients, persistent oxygen desaturation may occur and conversion to cardiopulmonary bypass (CPB) or sternotomy may be necessary. We reviewed the characteristics and clinical outcomes in patients who did not tolerate single-lung ventilation during TECAB surgery. METHODS: After Institutional Review Board approval we reviewed 440 patients undergoing robotic TECAB at our institution between July 2013 and April 2019. Patients were separated into 2 groups based on their ability to tolerate single-lung ventilation during the procedure. Group 1 included patients able to tolerate single-lung ventilation and Group 2 were patients who required double-lung ventilation to tolerate the procedure. Early and mid-term outcomes were compared. RESULTS: Group 2 (121 patients) had higher Society of Thoracic Surgeons scores, higher body mass index, and more triple-vessel disease than Group 1 (319 patients). Group 2 had more bilateral internal mammary artery use, multivessel grafting, and longer operative times. One patient underwent conversion to sternotomy and 5 required CPB (all in Group 1). Intensive care unit and hospital length of stay were longer in Group 2. Observed/expected mortality did not differ between groups (1.06% in Group 2 vs 0.4% in Group 1; P = 0.215). At mid-term follow-up, cardiac-related/overall mortality and freedom from major adverse cardiac events were similar. CONCLUSIONS: In our cohort, intolerance of single-lung ventilation did not preclude robotic off-pump TECAB. Double-lung ventilation is feasible during the procedure and may prevent conversions to sternotomy or use of CPB, resulting in excellent early and mid-term outcomes.