Year: 2020
Type: editorial
Abstract: can provide a score reflecting the certainty of a prediction, such scores have a high variance, and thus, what may appear successful on average across a cohort may prove less useful for an individual patient.A final note of caution from computer-aided detection in mammography is in order.When the use of such automated methods was evaluated systematically, there was no improvement in performance, and more distressingly, those individuals who interpreted both with and without the computerized systems had worse sensitivity when using it.11 Huang and colleagues 2 should be commended for extending what is possible with deep learning models in medicine.The challenge we continue to face is how we will use them.
Source: Circulation
Author: Rahul C. Deo
Institution: Brigham and Women's Hospital
Language: English
Subfield: Health Informatics
Field: Medicine
Domain: Health Sciences
Open Access status: bronze
OpenAlex ID: W3092722151