Search

Content-based medical image retrieval is an important diagnostic tool that improves the explainability of computer-aided diagnosis systems and provides decision-making support to healthcare professionals. A common approach to content-based image retrieval is learning a distance metric by transforming images into a feature space where the distance between samples is a similarity measure. Proxy metric learning methods are effective at learning this transformation due to the use of proxy feature vectors that enable efficient learning. Training with a distance-based classification loss enables a single proxy model to be suitable for both retrieval and classification. However, these methods are designed only for single-label data, making them unsuitable for multimorbidity medical images. Addressing this, we propose a novel multi-label proxy metric learning method for content-based image retrieval and classification. Unlike existing proxy-based methods, training samples assign to multiple proxies that span multiple class labels. This results in a feature space that encodes the complex relationships between diseases. We introduce negative proxies to better encode the relationships between samples without detected diseases. The efficacy of our approach is demonstrated experimentally on two multimorbidity radiology datasets. Results show that our method outperforms state-of-the-art image retrieval systems and baseline approaches. Our method is clinically significant as it improves on two key factors shown to affect medical professionals’ willingness to use computer-aided diagnosis systems: accuracy and interpretability.

Abstract Objectives The aim of this study is to evaluate the efficacy and safety of human anti-SARS-CoV-2 convalescent plasma in hospitalized adults with severe SARS-CoV-2 infection. Trial Design This is a prospective, single-center, phase 2, randomized, controlled trial that is blinded to participants and clinical outcome assessor. Participants Eligible participants include adults (≥ 18 years) with evidence of SARS-CoV-2 infection by PCR test of nasopharyngeal or oropharyngeal swab within 14 days of randomization, evidence of infiltrates on chest radiography, peripheral capillary oxygen saturation (SpO2) ≤ 94% on room air, and/or need for supplemental oxygen, non-invasive mechanical ventilation, or invasive mechanical ventilation, who are willing and able to provide written informed consent prior to performing study procedures or who have a legally authorized representative available to do so. Exclusion criteria include participation in another clinical trial of anti-viral agent(s)* for coronavirus disease-2019 (COVID-19), receipt of any anti-viral agent(s)* with possible activity against SARS-CoV-2

An amendment to this paper has been published and can be accessed via the original article.

An amendment to this paper has been published and can be accessed via the original article.