Medical Assistance in Dying in Oncology Patients: A Canadian Academic Hospital's Experience.

BACKGROUND: Medical assistance in dying (MAID) was legislatively enacted in Canada in June 2016. Most studies of patients who received MAID grouped patients with cancer and non-cancer diagnoses. Our goal was to analyze the characteristics of oncology patients who received MAID in a Canadian tertiary care hospital. METHODS: We conducted a retrospective review of all patients with cancer who received MAID between June 2016 and July 2020 at London Health Sciences Centre (LHSC). We describe patients' demographics, oncologic characteristics, symptoms, treatments, and palliative care involvement. RESULTS: Ninety-two oncology patients received MAID. The median age was 72. The leading cancer diagnoses among these patients were lung, colorectal, and pancreatic. At the time of MAID request, 68% of patients had metastatic disease. Most patients (90%) had ECOG performance status of 3 or 4 before receiving MAID. Ninety-nine percent of patients had distressing symptoms at time of MAID request, most commonly pain. One-third of patients with metastatic or recurrent cancer received early palliative care. The median time interval between the first MAID assessment and receipt of MAID was 7 days. INTERPRETATION: Most oncology patients who received MAID at LHSC had poor performance status and almost all had distressing symptoms. The median time interval between first MAID assessment and receipt of MAID was shorter than expected. Only one-third of patients with metastatic or recurrent cancer received early palliative care. Improving access to early palliative care is a priority in patients with advanced cancer. STUDY REGISTRATION: We received research approval from Western University's Research Ethics Board (REB) with project ID number 115367, and from Lawson's Research Database Application (ReDA) with study ID number 9579.

Machine Learning and Surgical Outcomes Prediction: A Systematic Review.

BACKGROUND: Machine learning (ML) has garnered increasing attention as a means to quantitatively analyze the growing and complex medical data to improve individualized patient care. We herein aim to critically examine the current state of ML in predicting surgical outcomes, evaluate the quality of currently available research, and propose areas of improvement for future uses of ML in surgery. METHODS: A systematic review was conducted in accordance with the Preferred Reporting Items for a Systematic Review and Meta-Analysis (PRISMA) checklist. PubMed, MEDLINE, and Embase databases were reviewed under search syntax "machine learning" and "surgery" for papers published between 2015 and 2020. RESULTS: Of the initial 2677 studies, 45 papers met inclusion and exclusion criteria. Fourteen different subspecialties were represented with neurosurgery being most common. The most frequently used ML algorithms were random forest (n = 19), artificial neural network (n = 17), and logistic regression (n = 17). Common outcomes included postoperative mortality, complications, patient reported quality of life and pain improvement. All studies which compared ML algorithms to conventional studies which used area under the curve (AUC) to measure accuracy found improved outcome prediction with ML models. CONCLUSIONS: While still in its early stages, ML models offer surgeons an opportunity to capitalize on the myriad of clinical data available and improve individualized patient care. Limitations included heterogeneous outcome and imperfect quality of some of the papers. We therefore urge future research to agree upon methods of outcome reporting and require basic quality standards.

A web-based automated machine learning platform to analyze liquid biopsy data.

Liquid biopsy (LB) technologies continue to improve in sensitivity, specificity, and multiplexing and can measure an ever growing library of disease biomarkers. However, clinical interpretation of the increasingly large sets of data these technologies generate remains a challenge. Machine learning is a popular approach to discover and detect signatures of disease. However, limited machine learning expertise in the LB field has kept the discipline from fully leveraging these tools and risks improper analyses and irreproducible results. In this paper, we develop a web-based automated machine learning tool tailored specifically for LB, where machine learning models can be built without the user's input. We also incorporate a differential privacy algorithm, designed to limit the effects of overfitting that can arise from users iteratively developing a panel with feedback from our platform. We validate our approach by performing a meta-analysis on 11 published LB datasets, and found that we had similar or better performance compared to those reported in the literature. Moreover, we show that our platform's performance improved when incorporating information from prior LB datasets, suggesting that this approach can continue to improve with increased access to LB data. Finally, we show that by using our platform the results achieved in the literature can be matched using 40% of the number of subjects in the training set, potentially reducing study cost and time. This self-improving and overfitting-resistant automatic machine learning platform provides a new standard that can be used to validate machine learning works in the LB field.

Presence of an interferon signature in individuals who are anti-nuclear antibody positive lacking a systemic autoimmune rheumatic disease diagnosis.

BACKGROUND: Elevated levels of type I interferons (IFNs) are a characteristic feature of the systemic autoimmune rheumatic diseases (SARDs) and are thought to play an important pathogenic role. However, it is unknown whether these elevations are seen in anti-nuclear antibody-positive (ANA+) individuals who lack sufficient criteria for a SARD diagnosis. We examined IFN-induced gene expression in asymptomatic ANA+ individuals and patients with undifferentiated connective tissue disease (UCTD) to address this question. METHODS: Healthy ANA- control subjects and ANA+ titre (≥1:160 by immunofluorescence) participants meeting no criteria, meeting at least one criterion (UCTD) or meeting SARD classification criteria were recruited. Whole peripheral blood IFN-induced and BAFF gene expression were quantified using NanoString technology. The normalized levels of five IFN-induced genes were summed to produce an IFN5 score. RESULTS: The mean IFN5 scores were increased in all ANA+ participant subsets as compared with healthy control subjects. We found that 36.8% of asymptomatic ANA+ and 50% of UCTD participants had IFN5 scores >2 SD above the mean for healthy control subjects. In all ANA+ subsets, the IFN5 score correlated with the presence of anti-Ro/La antibodies. In the asymptomatic ANA+ subset, this score also correlated with the ANA titre, whereas in the other ANA+ subsets, it correlated with the number of different ANA specificities. Development of new SARD criteria was seen in individuals with normal and high IFN5 scores. CONCLUSIONS: An IFN signature is seen in a significant proportion of ANA+ individuals and appears to be associated with ANA titre and type of autoantibodies, rather than with the presence or development of clinical SARD symptoms.

Conditional Sox9 ablation improves locomotor recovery after spinal cord injury by increasing reactive sprouting.

The absence of axonal regeneration after spinal cord injury (SCI) has been attributed to the up-regulation of axon-repelling molecules, such as chondroitin sulfate proteoglycans (CSPGs) present in the glial scar that forms post-SCI. We previously identified the transcription factor SOX9 as a key up-regulator of CSPG production and also demonstrated that conditional Sox9 ablation leads to decreased CSPG levels and improved recovery of hind limb function after SCI. We herein demonstrate increased neural input onto spinal neurons caudal to the lesion in spinal cord injured Sox9 conditional knock out mice as indicated by increased levels of the presynaptic markers synaptophysin and vesicular glutamate transporter 1 (VGLUT1) compared to controls. Axonal sparing, long-range axonal regeneration and reactive sprouting were investigated as possible explanations for the increase in neural inputs caudal to the lesion and for the improved locomotor outcomes in spinal cord-injured Sox9 conditional knock out mice. Whereas retrograde tract-tracing studies failed to reveal any evidence for increased axonal sparing or for long-range regeneration in the Sox9 conditional knock out mice, anterograde tract-tracing experiments demonstrated increased reactive sprouting caudal to the lesion after SCI. Finally we demonstrate that application of a broad spectrum MMP inhibitor to reduce CSPG degradation in Sox9 conditional knock out mice prevents the improvements in locomotor recovery observed in untreated Sox9 conditional knock out mice. These results suggest that improved recovery of locomotor function in Sox9 conditional knock out mice after SCI is due to increased reactive sprouting secondary to reduced CSPG levels distal to the lesion.