Search

Modern climate projections lack adequate spatial and temporal resolution due to computational constraints. A consequence is inaccurate and imprecise prediction of critical processes such as storms. Hybrid methods that combine physics with machine learning (ML) have introduced a new generation of higher fidelity climate simulators that can sidestep Moore's Law by outsourcing compute-hungry, short, high-resolution simulations to ML emulators. However, this hybrid ML-physics simulation approach requires domain-specific treatment and has been inaccessible to ML experts because of lack of training data and relevant, easy-to-use workflows. We present ClimSim, the largest-ever dataset designed for hybrid ML-physics research. It comprises multi-scale climate simulations, developed by a consortium of climate scientists and ML researchers. It consists of 5.7 billion pairs of multivariate input and output vectors that isolate the influence of locally-nested, high-resolution, high-fidelity physics on a host climate simulator's macro-scale physical state. The dataset is global in coverage, spans multiple years at high sampling frequency, and is designed such that resulting emulators are compatible with downstream coupling into operational climate simulators. We implement a range of deterministic and stochastic regression baselines to highlight the ML challenges and their scoring. The data (https://huggingface.co/datasets/LEAP/ClimSim_high-res) and code (https://leap-stc.github.io/ClimSim) are released openly to support the development of hybrid ML-physics and high-fidelity climate simulations for the benefit of science and society.

5531 Background: The International Federation of Obstetrics and Gynecology staging system for cervical cancer (FIGO 2008) was revised in 2018 to incorporate lymph node involvement (FIGO 2018). OUTBACK is an international, randomized phase 3 trial of adjuvant chemotherapy versus observation after standard of care treatment with chemoradiation for women with locally advanced cervical cancer. OUTBACK found no benefit from the addition of adjuvant chemotherapy. We evaluated the effects of classifying participants with these 2 staging systems in the OUTBACK trial population. Methods: OUTBACK recruited April 2011 to June 2017 and staged participants according to FIGO 2008. Lymph node status, smoking status, age, race and histological subtype were documented at trial entry as important prognostic factors. We assessed the effects of stage grouping into stage I, II, and III/IVa with FIGO 2008 versus FIGO 2018, on progression-free survival (PFS) and overall survival (OS) at 5 years using Kaplan-Meier estimates, and in univariable proportional-hazards regression analyses, and in multivariable analyses adjusting for important prognostic factors and randomly allocated treatment. Results: All 919 study participants had complete data for staging according to the 2 staging systems and most prognostic factors for adjustment. Among all participants, the 5-year outcomes were PFS = 62% and OS = 72%. Classification according to FIGO 2018 rather than FIGO 2008 yielded higher 5-year PFS and OS in each stage group (see table for numbers of participants, PFS and OS for each stage group). Predictors of PFS in multivariable analysis included squamous vs non-squamous histology (HR 0.71 for FIGO 2008 and 0.74 for FIGO 2018), but not nodal involvement when FIGO 2018 was used. Both staging systems were the only independently significant prognostic factors in both univariable and multivariable analyses (all p < 0.0001) for both PFS and OS. Conclusions: Compared to FIGO 2008, reclassifying pts by FIGO 2018 staging resulted in more pts being classified as stage 3 due to the incorporation of nodal status. Staging locally advanced cervical cancer using FIGO 2018 rather than FIGO 2008 resulted in higher PFS and OS in each stage grouping that reflected stage migration, not a true improvement in outcomes. FIGO stage remains the strongest predictor of overall survival after CRT but survival outcomes by stage in trials using the old vs new staging system are not comparable. Clinical trial information: ACTRN12610000732088. [Table: see text]

We introduce Codex, a GPT language model fine-tuned on publicly available code from GitHub, and study its Python code-writing capabilities. A distinct production version of Codex powers GitHub Copilot. On HumanEval, a new evaluation set we release to measure functional correctness for synthesizing programs from docstrings, our model solves 28.8% of the problems, while GPT-3 solves 0% and GPT-J solves 11.4%. Furthermore, we find that repeated sampling from the model is a surprisingly effective strategy for producing working solutions to difficult prompts. Using this method, we solve 70.2% of our problems with 100 samples per problem. Careful investigation of our model reveals its limitations, including difficulty with docstrings describing long chains of operations and with binding operations to variables. Finally, we discuss the potential broader impacts of deploying powerful code generation technologies, covering safety, security, and economics., Comment: corrected typos, added references, added authors, added acknowledgements

Supplemental material, sj-pdf-1-uro-10.1177_2051415821995107 for Correlation of Likert scores III–V with increasingly worse pathology in radical prostatectomy specimens significant only for men aged 0.15, with age

We outline a perspective of an entirely new research branch in Earth and climate sciences, where deep neural networks and Earth system models are dismantled as individual methodological approaches and reassembled as learning, self-validating, and interpretable Earth system model-network hybrids. Following this path, we coin the term "Neural Earth System Modelling" (NESYM) and highlight the necessity of a transdisciplinary discussion platform, bringing together Earth and climate scientists, big data analysts, and AI experts. We examine the concurrent potential and pitfalls of Neural Earth System Modelling and discuss the open question whether artificial intelligence will not only infuse Earth system modelling, but ultimately render them obsolete., Comment: Perspective paper submitted to Nature Machine Intelligence, 23 pages, 3 figures

Supplemental material, sj-pdf-4-uro-10.1177_2051415821995107 for Correlation of Likert scores III–V with increasingly worse pathology in radical prostatectomy specimens significant only for men aged 0.15, with age

Supplemental material, sj-pdf-2-uro-10.1177_2051415821995107 for Correlation of Likert scores III–V with increasingly worse pathology in radical prostatectomy specimens significant only for men aged 0.15, with age

Additional file 1. Table S1: Coefficient of variation of plasma biomarkers. Table S2: Admission Diagnoses within Cohorts. Table S3: Day 1 plasma biomarkers at study enrollment between COVID-19 and non-COVID-19. Table S4: Plasma biomarkers restricted to patients with a primary ICU diagnosis of pneumonia or respiratory failure. Table S5: Trend in plasma biomarkers over Day 1 and 3 between COVID-19 and non-COVID-19. Table S6: Association of baseline biomarkers with ARDS. Table S7: Association of baseline biomarkers with severe AKI. Figure S1: Similar set of plasma biomarkers are different between critically ill patients with COVID-19 compared to without COVID-19 adjusting for receipt of corticosteroids.

Supplemental material, sj-pdf-4-uro-10.1177_2051415821995107 for Correlation of Likert scores III–V with increasingly worse pathology in radical prostatectomy specimens significant only for men aged 0.15, with age

Supplemental material, sj-pdf-5-uro-10.1177_2051415821995107 for Correlation of Likert scores III–V with increasingly worse pathology in radical prostatectomy specimens significant only for men aged 0.15, with age

Supplemental material, sj-pdf-3-uro-10.1177_2051415821995107 for Correlation of Likert scores III–V with increasingly worse pathology in radical prostatectomy specimens significant only for men aged 0.15, with age

Supplemental material, sj-pdf-5-uro-10.1177_2051415821995107 for Correlation of Likert scores III–V with increasingly worse pathology in radical prostatectomy specimens significant only for men aged 0.15, with age

Supplemental material, sj-pdf-3-uro-10.1177_2051415821995107 for Correlation of Likert scores III–V with increasingly worse pathology in radical prostatectomy specimens significant only for men aged 0.15, with age

Background Clostridium difficile infection (CDI) recurs after initial treatment in approximately one in four patients. A single-centre pilot study suggested that this could be reduced using ‘follow-on’ rifaximin treatment. We aimed to assess the efficacy of rifaximin treatment in preventing recurrence. Methods A multisite, parallel group, randomised, placebo controlled trial recruiting patients aged ≥18 years immediately after resolution of CDI through treatment with metronidazole or vancomycin. Participants received either rifaximin 400 mg three times a day for 2 weeks, reduced to 200 mg three times a day for a further 2 weeks or identical placebo. The primary endpoint was recurrence of CDI within 12 weeks of trial entry. Results Between December 2012 and March 2016, 151 participants were randomised to either rifaximin or placebo. Primary outcome data were available on 130. Mean age was 71.9 years (SD 15.3). Recurrence within 12 weeks was 29.5% (18/61) among participants allocated to placebo compared with 15.9% (11/69) among those allocated to rifaximin, a difference between groups of 13.7% (95% CI −28.1% to 0.7%, p=0.06). The risk ratio was 0.54 (95% CI 0.28 to 1.05, p=0.07). During 6-month safety follow-up, nine participants died in each group (12%). Adverse event rates were similar between groups. Conclusion While ‘follow-on’ rifaximin after CDI appeared to halve recurrence rate, we failed to reach our recruitment target in this group of frail elderly patients, so the estimated effect of rifaximin lacks precision. A meta-analysis including a previous trial suggests that rifaximin may be effective; however, further, larger confirmatory studies are needed.

The atmosphere is chaotic. This fundamental property of the climate system makes forecasting weather incredibly challenging: it's impossible to expect weather models to ever provide perfect predictions of the Earth system beyond timescales of approximately 2 weeks. Instead, atmospheric scientists look for specific states of the climate system that lead to more predictable behaviour than others. Here, we demonstrate how neural networks can be used, not only to leverage these states to make skillful predictions, but moreover to identify the climatic conditions that lead to enhanced predictability. Furthermore, we employ a neural network interpretability method called ``layer-wise relevance propagation'' to create heatmaps of the regions in the input most relevant for a network's output. For Earth scientists, these relevant regions for the neural network's prediction are by far the most important product of our study: they provide scientific insight into the physical mechanisms that lead to enhanced weather predictability. While we demonstrate our approach for the atmospheric science domain, this methodology is applicable to a large range of geoscientific problems., Comment: conference paper from NeurIPS 2020 AI for Earth Sciences

Multidecadal variability in the North Atlantic jet stream in general circulation models (GCMs) is compared with that in reanalysis products of the twentieth century. It is found that almost all models exhibit multidecadal jet stream variability that is entirely consistent with the sampling of white noise year-to-year atmospheric fluctuations. In the observed record, the variability displays a pronounced seasonality within the winter months, with greatly enhanced variability toward the late winter. This late winter variability exceeds that found in any GCM and greatly exceeds expectations from the sampling of atmospheric noise, motivating the need for an underlying explanation. The potential roles of both external forcings and internal coupled ocean–atmosphere processes are considered. While the late winter variability is not found to be closely connected with external forcing, it is found to be strongly related to the internally generated component of Atlantic multidecadal variability (AMV) in sea surface temperatures (SSTs). In fact, consideration of the seasonality of the jet stream variability within the winter months reveals that the AMV is far more strongly connected to jet stream variability during March than the early winter months or the winter season as a whole. Reasoning is put forward for why this connection likely represents a driving of the jet stream variability by the SSTs, although the dynamics involved remain to be understood. This analysis reveals a fundamental mismatch between late winter jet stream variability in observations and GCMs and a potential source of long-term predictability of the late winter Atlantic atmospheric circulation.

PURPOSE: In brain tumours, brain metastases or advanced cancer; treatment with corticosteroids, side effects can add to symptoms. These are best assessed by patients, complementing clinical assessment. We assessed the feasibility and validity of the Dexamethasone Symptom Questionnaire-Chronic (DSQ-Chronic), patient and caregiver versions. METHODS: A longitudinal cohort study was conducted, collecting clinician-rated toxicity, performance status, dexamethasone dose and DSQ-Chronic (patient and caregiver versions) at baseline, then 2, 4 and 8 weeks later. Patients had a primary malignant brain tumour, brain metastases, or advanced cancer; Karnofsky Performance Status >/=40 and predicted survival >/=8 weeks. Analysis included questionnaire completion rates, frequency and severity of dexamethasone-attributable side effects, agreement between patient and caregiver ratings, comparison with clinician-rated toxicity and correlation with performance status. RESULTS: Sixty-six patients were recruited (mean age 60 years), with their caregivers. Completion of questionnaires was over 90 % for the dyad at baseline but dropped over time, with caregiver completion rates higher at all timepoints. Agreement between patients and proxies was fair to moderate, and while proxies systematically overestimated symptom severity on DSQ-chronic total scores, the bias was less than 10 points. Patient and clinician agreement was higher for more objective symptoms. CONCLUSION: The DSQ-Chronic is feasible when the patient is relatively well. As capacity to complete the DSQ-Chronic diminishes, caregivers can be proxy-raters. Clinicians capture corticosteroid toxicities, which may not be obvious to the patient. The DSQ-Chronic, patient and caregiver versions, are useful tools to be used with clinician assessment.

Climate extremes such as drought and heat waves can cause substantial reductions in terrestrial carbon uptake. Advancing projections of the carbon uptake response to future climate extremes depends on (1) identifying mechanistic links between the carbon cycle and atmospheric drivers, (2) detecting and attributing uptake changes, and (3) evaluating models of land response and atmospheric forcing. Here, we combine model simulations, remote sensing products, and ground observations to investigate the impact of climate variability on carbon uptake in the Texas-northern Mexico region. Specifically, we (1) examine the relationship between drought, carbon uptake, and variability of El Niño-Southern Oscillation (ENSO) and the North Atlantic Oscillation (NAO) using the Joint UK Land-Environment Simulator (JULES) biosphere simulations from 1950-2012, (2) quantify changes in carbon uptake during record drought conditions in 2011, and (3) evaluate JULES carbon uptake and soil moisture in 2011 using observations from remote sensing and a network of flux towers in the region. Long-term simulations reveal systematic decreases in regional-scale carbon uptake during negative phases of ENSO and NAO, including amplified reductions of gross primary production (GPP) (-0.42 ± 0.18 Pg C yr-1) and net ecosystem production (NEP) (-0.14 ± 0.11 Pg C yr-1) during strong La Niña years. The 2011 megadrought caused some of the largest declines of GPP (-0.50 Pg C yr-1) and NEP (-0.23 Pg C yr-1) in our simulations. In 2011, consistent declines were found in observations, including high correlation of GPP and surface soil moisture (r = 0.82 ± 0.23, p = 0.012) in remote sensing-based products. These results suggest a large-scale response of carbon uptake to ENSO and NAO, and highlight a need to improve model predictions of ENSO and NAO in order to improve predictions of future impacts on the carbon cycle and the associated feedbacks to climate change.

The Brooker classification is a commonly used tool to characterize the extent of heterotopic ossification (HTO), which is the dystrophic formation of mature lamellar bone in non-osseous tissues such as muscles, connective tissue or nerves. We aim to provide illustrative cases to describe the limitations and challenges of this system.A retrospective review was conducted in all patients who had retrieved hip and/or pelvic radiotherapy for HTO prophylaxis at Sunnybrook Health Sciences Centre between July 1998 and August 2016. An independent musculoskeletal radiologist who was blinded to the condition of the patient and the clinical outcome evaluated the imaging studies based on the Brooker classification.22 patients were identified for analysis. Three patient cases were presented to illustrate the discrepancies between the Brooker classification and radiographic images, CT images and different radiographical projections. Two patient cases were presented to demonstrate the inability of the Brooker classification to account for volumetric differences in HTO and arthroplasty lengths.Although the Brooker classification for HTO is a widely used quantitative and qualitative assessment tool given its simplicity and familiarity, it possesses several limitations. Utilization of other radiographic modalities, such as CT, and orthogonal projections may reduce ambiguities from using the HTO system. Development of a classification method that can appropriately correspond to the clinical outcomes such as functional capabilities to further HTO development is warranted. Advances in knowledge: The present study critically evaluated the Brooker classification system, and identified areas in which improvement is warranted on characterizing HTO. This is important in further research that aims to improve the accuracy of HTO classification guidelines.

An audit conducted at UHSM identified the need for a pain assessment tool to help identify pain for people with communication difficulties which led to this small study to test the usefulness of the Bolton Pain Assessment Tool (BPAT). BPAT was developed for patients with communication difficulties and includes a section for the family to be involved in the assessment and provide information specifically relating to the individual.The Bolton Pain Assessment Tool has been introduced and used within four clinical areas of the Trust that were identified as having a proportion of patients with communication difficulties, including dementia.This study helps to address some of the recommendations from the National Dementia Strategy (2009) to improve the care provided in Hospital by addressing the need to recognise, assess and manage pain for all patients admitted to hospital.The aim of the proof of concept study was to assess if the BPAT is useful and helps to identify pain in acute clinical areas. Findings are: • All the staff involved in the evaluation agreed that it was easy to use• The time taken was between half a minute and 5 minutes. Mode was 2 minutes • The use of the BPAT facilitated the process of a review and/ or the administration of analgesia to ensure appropriate / adequate pain management for patients with communication difficulties. • There was limited involvement of the family or close carer of an individual with communication difficulties, but when they were asked they provided relevant information to assist in the assessment and management of pain.• Clinical staff in acute care rated the scale as 8 out of 10 and the majority rated it better or a lot better than other pain assessment tools.Future Plans The intention is to extend the use of BPAT across the organisation and collect further data to support its use.