Volume 27 Issue 1, January 2021

There has been a lot of hype around the applications of machine learning in medicine. But how is machine learning actually helping bench-to-bedside scientists and clinicians do their jobs?

Is it time to shift to a model of ‘inclusion unless otherwise justified’ in trials of vaccines for pandemics?

Stakeholders in public health must lobby policy makers to make decisions based on evidence, not political expediency, particularly when the studies that hang in the balance are critical to understanding the origins of epidemics.

We argue that deliberative decision making that is inclusive, transparent and accountable can contribute to more trustworthy and legitimate decisions on difficult ethical questions and political trade-offs during the pandemic and beyond.

Shared medical appointments, whereby patients with similar medical conditions consult their medical practitioner together, alleviate pressure on the health system and provide an instant support network for the patient. Why not make them virtual?

The pandemic has thrust many mainstream journalists into unfamiliar grounds, including coverage of expert opinion that is not backed up by peer-reviewed content, reporting on preprints, and assessing high-complexity instant-response science. How did they manage? We asked five journalists from mainstream media about their experience.

An algorithm developed through machine learning on existing radiological data from diverse people has the potential to improve the diagnosis and management of knee pain from osteoarthritis and reduce racial disparities in the assessment of knee pain.

Analysis of glycemic and non-glycemic characteristics in individuals at risk for type 2 diabetes identifies six putative subgroups that differ in terms of risk of diabetes progression and complications.

Liver metastases lead to resistance to immunotherapy through the ‘siphoning’ of tumor antigen–specific CD8⁺ T cells into the liver, which results in a systemic ‘immune desert’ incapable of controlling tumor burden.

The severity of COVID-19 varies by sex and age, and is linked to comorbidities. Differences in immune reactions may go some way to explain this.

Increased understanding of the biology of metastases allows improved targeting and outcomes for patients with both micro- and macrometastases.

A comprehensive census of the dynamics of death and regeneration of cells and tissues provides an estimation of the distribution of cellular turnover in the human body.

Clustering of patients with prediabetes using simple clinical features reveals six distinct groups with differing risk of developing type 2 diabetes and its associated complications.

Retrospective analysis of ¹⁸F-FDG PET/CT scans from over 50,000 patients reveals correlations between presence of brown adipose tissue and lower odds of having cardiometabolic conditions, such as type 2 diabetes, cardiovascular disease and hypertension.

Analysis of 10,900 whole-exome sequences linked to electronic health care records in the Penn Medicine Biobank enabled an exome-wide study of the phenotypic effects of rare loss-of-function gene variants, identifying new gene–disease associations that replicated across other biobanks.

A smartphone app that combines smartwatch and activity tracker data together with self-reported symptoms allows continuous monitoring of SARS-CoV-2 infection.

Functionally competent memory CD8⁺ T cells specific for different viral epitopes are induced by SARS-CoV-2 infection and can be detected in the absence of virus-specific antibodies.

Analysis of five generations of transmission of severe acute respiratory syndrome coronavirus 2 from cases to contacts identifies a high rate of symptomless transmission and risk factors associated with spread of COVID-19.

A modeling study using case and mortality data from the first 8 months of the COVID-19 pandemic in the United States explores five potential future scenarios of social distancing mandates and mask use at the state level, with projections of the course of the epidemic through winter 2021.

New influenza virus vaccines tested in humans elicit broadly cross-reactive antibodies that bind the stalk of the viral hemagglutinin protein and may serve as templates to design a universal influenza vaccine.

A universal method enables high-specificity, unbiased pathogen detection from diverse body fluids using metagenomic sequencing and may accelerate clinical decisions.

An in vitro human tonsil tissue-based system captures key features of a functional germinal center and can be used to study humoral immune responses to vaccines, new antigens and adjuvants.

An algorithmic, machine-learning approach to measuring severe pain from osteoarthritis applied to X-ray images of knees suggests that reported disparities in knee pain in underserved populations can be reduced by comparison with use of standard radiographic measures of disease severity.

Single-cell analysis of gastric cancer samples tracks the cell of origin of metastatic lesions and identifies an independent prognostic signature of the clinical outcome.

Liver metastases reduce clinical and preclinical immune-checkpoint inhibitor efficacy through hepatic siphoning of circulating activated CD8⁺ T cells, but therapeutic benefit can be improved by combining immunotherapy with liver-directed radiotherapy.

Cabozantinib, an inhibitor of multiple receptor tyrosine kinases, has efficacy in a mouse model of neurofibromatosis type I and has clinical activity in reducing plexiform neurofibroma volume in a phase II trial of patients with NF1.

The functional magnetic resonance imaging connectivity pattern of tonic experimental orofacial pain can be used as a quantitative and unbiased biomarker of clinical pain.