Search

Introduction: Machine learning algorithms are expected to work side-by-side with humans in decision-making pipelines. Thus, the ability of classifiers to make reliable decisions is of paramount importance. Deep neural networks (DNNs) represent the state-of-the-art models to address real-world classification. Although the strength of activation in DNNs is often correlated with the network's confidence, in-depth analyses are needed to establish whether they are well calibrated., Method: In this paper, we demonstrate the use of DNN-based classification tools to benefit cancer registries by automating information extraction of disease at diagnosis and at surgery from electronic text pathology reports from the US National Cancer Institute (NCI) Surveillance, Epidemiology, and End Results (SEER) population-based cancer registries. In particular, we introduce multiple methods for selective classification to achieve a target level of accuracy on multiple classification tasks while minimizing the rejection amount-that is, the number of electronic pathology reports for which the model's predictions are unreliable. We evaluate the proposed methods by comparing our approach with the current in-house deep learning-based abstaining classifier., Results: Overall, all the proposed selective classification methods effectively allow for achieving the targeted level of accuracy or higher in a trade-off analysis aimed to minimize the rejection rate. On in-distribution validation and holdout test data, with all the proposed methods, we achieve on all tasks the required target level of accuracy with a lower rejection rate than the deep abstaining classifier (DAC). Interpreting the results for the out-of-distribution test data is more complex; nevertheless, in this case as well, the rejection rate from the best among the proposed methods achieving 97% accuracy or higher is lower than the rejection rate based on the DAC., Conclusions: We show that although both approaches can flag those samples that should be manually reviewed and labeled by human annotators, the newly proposed methods retain a larger fraction and do so without retraining-thus offering a reduced computational cost compared with the in-house deep learning-based abstaining classifier., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Published by Elsevier Inc.)

The COVID-19 pandemic has impacted the nursing student clinical experience. At the time of this project, hospitals hosting students for clinical experiences had a high population of COVID-19 positive patients. University and hospital policies at the time prohibited students from caring for COVID-19 patients due to the inability to safely ensure that students could access the proper protective equipment (PPE) necessary when caring for patients in airborne precautions. Infection control guidelines for patients in airborne precautions require a N95 respirator to be always worn by healthcare providers. These respirators require yearly fit test following OSHA (Occupational Safety and Health Administration) standards. Hospitals do have the resources to complete N95 fit tests for nursing students. The solution was to provide qualitative N95 mask fit testing, performed by university faculty, staff, and students using OSHA guidelines and standards. Within 3 months, over 50% of the student population had been fit tested, which increased the clinical opportunities for the students., Competing Interests: We have no conflict of interest to disclose., (© 2022 Organization for Associate Degree Nursing. Published by Elsevier Inc. All rights reserved.)

Nutrient sources to coastal waters with coral reefs are not well-characterized. This study documented spatial distributions of nutrients within coastal waters along two developments with coral reefs, and identified nutrient sources through nutrient mixing plots, δ 15 N measurements in macroalgal tissue, and NO 3 - stable isotope mixing models. Nutrients decreased from fresh groundwaters to offshore waters, with some surface waters higher in concentrations than benthic ones. Conservative and non-conservative mixing between fresh and ocean waters occurred, the latter suggestive of local nutrient sources and biological removal. δ 15 N in macroalgal tissue and NO 3 - concurred that fresh groundwater, ocean water, and fertilizers were dominant nutrient sources. Benthic salinity and NO 3 - + NO 2 - concentrations illustrated that submarine groundwater discharge delivered nutrients to reefs in pulses ranging from minutes to days. Information generated from this study is imperative for developing management actions to improve water quality and make coral reefs more resilient to stressors., (Copyright © 2021 Elsevier Ltd. All rights reserved.)

In the last decade, the widespread adoption of electronic health record documentation has created huge opportunities for information mining. Natural language processing (NLP) techniques using machine and deep learning are becoming increasingly widespread for information extraction tasks from unstructured clinical notes. Disparities in performance when deploying machine learning models in the real world have recently received considerable attention. In the clinical NLP domain, the robustness of convolutional neural networks (CNNs) for classifying cancer pathology reports under natural distribution shifts remains understudied. In this research, we aim to quantify and improve the performance of the CNN for text classification on out-of-distribution (OOD) datasets resulting from the natural evolution of clinical text in pathology reports. We identified class imbalance due to different prevalence of cancer types as one of the sources of performance drop and analyzed the impact of previous methods for addressing class imbalance when deploying models in real-world domains. Our results show that our novel class-specialized ensemble technique outperforms other methods for the classification of rare cancer types in terms of macro F1 scores. We also found that traditional ensemble methods perform better in top classes, leading to higher micro F1 scores. Based on our findings, we formulate a series of recommendations for other ML practitioners on how to build robust models with extremely imbalanced datasets in biomedical NLP applications., (Copyright © 2021. Published by Elsevier Inc.)

Rationale: Resective epilepsy surgery is an evidence-based curative treatment option for patients with drug-resistant focal epilepsy. The major preoperative predictor of a good surgical outcome is detection of an epileptogenic lesion by magnetic resonance imaging (MRI). Application of ultra-high field (UHF) MRI, i.e. field strengths ≥ 7 Tesla (T), may increase the sensitivity to detect such a lesion., Methods: A keyword search strategy was submitted to Pubmed, EMBASE, Cochrane Database and clinicaltrials.gov to select studies on UHF MRI in patients with epilepsy. Follow-up study selection and data extraction were performed following PRISMA guidelines. We focused on I) diagnostic gain of UHF- over conventional MRI, II) concordance of MRI-detected lesion, seizure onset zone and surgical decision-making, and III) postoperative histopathological diagnosis and seizure outcome., Results: Sixteen observational cohort studies, all using 7T MRI were included. Diagnostic gain of 7T over conventional MRI ranged from 8% to 67%, with a pooled gain of 31%. Novel techniques to visualize pathological processes in epilepsy and lesion detection are discussed. Seizure freedom was achieved in 73% of operated patients; no seizure outcome comparison was made between 7T MRI positive, 7T negative and 3T positive patients. 7T could influence surgical decision-making, with high concordance of lesion and seizure onset zone. Focal cortical dysplasia (54%), hippocampal sclerosis (12%) and gliosis (8.1%) were the most frequently diagnosed histopathological entities., Significance: UHF MRI increases, yet variably, the sensitivity to detect an epileptogenic lesion, showing potential for use in clinical practice. It remains to be established whether this results in improved seizure outcome after surgical treatment. Prospective studies with larger cohorts of epilepsy patients, uniform scan and sequence protocols, and innovative post-processing technology are equally important as further increasing field strengths. Besides technical ameliorations, improved correlation of imaging features with clinical semiology, histopathology and clinical outcome has to be established., (Copyright © 2021 The Author(s). Published by Elsevier Inc. All rights reserved.)

In Alzheimer's disease (AD), the hippocampus is an early site of tau pathology and neurodegeneration. Histological studies have shown that lesions are not uniformly distributed within the hippocampus. Moreover, alterations of different hippocampal layers may reflect distinct pathological processes. 7 T MRI dramatically improves the visualization of hippocampal subregions and layers. In this study, we aimed to assess whether 7 T MRI can detect volumetric changes in hippocampal layers in vivo in patients with AD. We studied four AD patients and seven control subjects. MR images were acquired using a whole-body 7 T scanner with an eight channel transmit-receive coil. Hippocampal subregions were manually segmented from coronal T2*-weighted gradient echo images with 0.3 × 0.3 × 1.2 mm3 resolution using a protocol that distinguishes between layers richer or poorer in neuronal bodies. Five subregions were segmented in the region of the hippocampal body: alveus, strata radiatum, lacunosum and moleculare (SRLM) of the cornu Ammonis (CA), hilum, stratum pyramidale of CA and stratum pyramidale of the subiculum. We found strong bilateral reductions in the SRLM of the cornu Ammonis and in the stratum pyramidale of the subiculum (p < 0.05), with average cross-sectional area reductions ranging from -29% to -49%. These results show that it is possible to detect volume loss in distinct hippocampal layers using segmentation of 7 T MRI. 7 T MRI-based segmentation is a promising tool for AD research.

Objective: Expeditions organized by commercial companies are becoming increasingly popular. Charity expeditions take inexperienced participants on trips all over the world, with participants being sponsored to raise funds for charitable causes. The incidence of illness or injury while participating in charity expeditions is unknown. The objective of this study is to report the incidence and severity of illness and injuries occurring on worldwide charity expeditions., Methods: Retrospective, observational study reviewing expedition medical reports from 232 expeditions organized by a single commercial expedition company for a 5-year period (January 1, 2004 through December 31, 2008)., Results: Complete expedition medical reports were available for 210 (91%) trips, involving 4077 participants over 1524 expedition days. Expeditions reported a total of 1564 incidents over 42482 participant-days in the field, including days spent traveling to the expedition site. In 1465 (94%) cases "minor" injury or illness was recorded, 79 (5%) "moderate," and 20 (1%) "major" in severity. No deaths were reported. Gastrointestinal upset was the commonest reported minor condition and severe acute mountain sickness the commonest major condition. Overall, the incidence per 1000 participant-days of minor illness or injury was 34.48, moderate illness or injury 1.86, and major illness or injury 0.47., Conclusion: The risk of sustaining major injury or illness on an overseas charity expedition is low. The consequences of becoming injured or unwell in a remote environment can be serious, and appropriate medical care is required., (Copyright © 2010 Wilderness Medical Society. Published by Elsevier Inc. All rights reserved.)

Purpose: Long-term follow-up studies sometimes rely on retrospective identification of the cohort to be followed. Few such studies have been conducted, however, that rely on very old source records, perhaps because of concern that identification and location of a cohort based on very old records would not be feasible., Methods: The Hanford Thyroid Disease Study (HTDS), conducted in the 1990s, identified 5199 cohort members using Washington state birth records from 1940-1946. The limited information contained therein was used to trace cohort members to the present day, nearly 50 years later., Results: We found the best strategy to locate individuals efficiently is to use a combination of methods and resources in a phased approach, beginning with readily available and easy-to-use sources of information before employing more time-intensive strategies that are costly and rely on information that is difficult to obtain. Motor vehicle licensing records and directories were the most useful individual sources of information. Using this approach, the HTDS successfully located 94% of the 5199 cohort members identified., Conclusions: It is feasible to successfully trace and locate a very high proportion of individuals identified from very old records that contain little information.

Cytoarchitectonic fields of the human neocortex are defined by characteristic variations in the composition of a general six-layer structure. It is commonly accepted that these fields correspond to functionally homogeneous entities. Diligent techniques were developed to characterize cytoarchitectonic fields by staining sections of post-mortem brains and subsequent statistical evaluation. Fields were found to show a considerable interindividual variability in extent and relation to macroscopic anatomical landmarks. With upcoming new high-resolution magnetic resonance imaging (MRI) protocols, it appears worthwhile to examine the feasibility of characterizing the neocortical fine-structure from anatomical MRI scans, thus, defining neocortical fields by in vivo techniques. A fixated brain hemisphere was scanned at a resolution of approximately 0.3 mm. After correcting for intensity inhomogeneities in the dataset, the cortex boundaries (the white/grey matter and grey matter/background interfaces) were determined as a triangular mesh. Radial intensity profiles following the shortest path through the cortex were computed and characterized by a sparse set of features. A statistical similarity measure between features of different regions was defined, and served to define the extent of Brodmann's Areas 4, 17, 44 and 45 in this dataset.

PURPOSE: This paper presents a cross-sectional method for deriving age-specific hysterectomy prevalence. The influence of hysterectomy prevalence on uterine and ovarian cancer incidence rates and probabilities is illustrated.METHODS: Hysterectomy prevalence estimates are derived from cross-sectional data using a life table method. Analysis is based on hysterectomy data from the Utah Hospital Discharge Data Base and cancer cases recorded by the Utah Cancer Registry, 1995-1997.RESULTS: Correction for hysterectomy prevalence increases cancer incidence rates, more so at older ages, such that by ages 85 and older the increase is 92% for uterine cancers of the cervix and corpus, and 17% for ovarian cancer. Over the life span, correction for hysterectomy prevalence increases the estimated number per 100,000 developing these diseases by 195 (from 603 to 798) for uterine cervical cancer, 1,553 (from 2,580 to 4,133) for corpus uterine cancer, and 911 (1,674 to 2,585) for ovarian cancer.CONCLUSIONS: The utility of the cross-sectional method for deriving hysterectomy prevalence is that it does not require several years of data to obtain reasonable estimates. Correction for hysterectomy prevalence has a large effect on the cancer rates and probabilities, providing a more accurate representation of the risk and burden of these cancers.

The drawing of blood by orbital sinus puncture (OSP) under ether anesthesia is known to produce a marked increase in serum prolactin (PRL levels in young cycling female rats. The effect of this stressful procedure on PRL release was compared in young and aged female rats. Nonstressed PRL levels were obtained from blood drawn by decapitation. Whereas OSP with a one-minute ether exposure induced a marked increase in PRL levels in young rats on all days of the estrus cycle, older cycling female rats on the day of diestrus -1 and aged rats exhibiting prolonged diestrus (PD) showed virtually no increase above nonstressed levels. However, increasing the ether exposure time to five minutes did produce a rise in PRL levels. Old cycling female rats on the day of estrus and aged rats exhibiting constant estrus (CE) did show a PRL increase comparable to that seen in young animals. Ovariectomy (OVX) completely abolished the stress response seen in aged CE rats. The response, though markedly decreased, was still present in young ovariectomized rats. These experiments show that the stress-induced rise in PRL promoted by OSP under either anesthesia is markedly diminished in aged rats exhibiting a diestrus state. The attenuated response seen in these rats is believed due to factors characteristic of the diestrous state of aging.

Tensile retentive strengths for six polycarbonate crown cementation systems were examined with use of extracted human teeth prepared with and without retentive grooves. From the results of this study, use of unfilled intermediary resins followed by polycarboxylate or zinc phosphate cements is recommended. Retentive grooves did not contribute to overall retention.