Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain
the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in
Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles
and JavaScript.
Explainable AI predicts blood-oxygen levels during anaesthesia
This issue highlights machine-learning algorithms that explain hypoxaemia risk under anaesthesia during surgery, that identify polyps in colonoscopy images and videos, and that predict post-surgical adverse pathology in prostate-cancer and breast-cancer tissue samples. It also highlights predictions of tumour uptake and distribution of specific therapeutic agents, and a personalized virtual-heart model for finding radio-frequency ablation targets for infarct-related tachycardia.
The cover illustrates variations in risk factors contributing to hypoxaemia under general anaesthesia, as predicted by machine learning.
Clinical implementations of machine learning that are accurate, robust and interpretable will eventually gain the trust of healthcare providers and patients.
A computational model of the heart that finds the optimal ablation sites to treat infarct-related ventricular tachycardia eliminates the need for invasive electrical mapping.
A deep-learning algorithm enables the real-time video-based recognition of polyps during colonoscopy, with sensitivities and specificities surpassing 90%.
Computational modelling, informed by data from 3D volumetric imaging of transparent and intact whole-tumour samples, predicts blood flow and the spatial distribution of drug uptake in tumours.
A personalized virtual-heart model that determines optimal radio-frequency ablation targets for infarct-related tachycardia is validated in retrospective large-animal and patient studies, and in a prospective study in patients.
A deep-learning algorithm can detect polyps in the colon in real time and with high sensitivity and specificity, according to validation studies with prospectively collected images and videos from colonoscopies performed in 1,138 patients.
An alert system based on machine learning and trained on surgical data from electronic medical records helps anaesthesiologists prevent hypoxaemia during surgery by providing interpretable real-time predictions.
An assay that uses machine-learning algorithms on phenotypic-biomarker data from live primary cells predicts post-surgical adverse pathology in prostate-cancer and breast cancer tissue samples from patients.
The combination of mathematical modelling of tumour tissue, optical imaging of cleared tumours from animal models, and in vivo imaging of vascular perfusion in tumours predicts the tumour uptake and distribution of specific therapeutic agents.