Prediction of cardiovascular markers and diseases using retinal fundus images and deep learning: a systematic scoping review protocol

The document is the protocol of a scoping review.

Objective:

The objective of this scoping review is to identify and characterize studies using retinal fundus images and deep learning to predict cardiovascular risk (including risk markers and hard endpoints).

Introduction:

Cardiovascular risk prediction models or risk scores based on sociodemographic factors and simple clinical measurements have received significant attention from the medical research community. With rapid development in deep learning for image analysis in the last decade and given the well-known association between micro- and macrovascular complications, some recent studies focused on the prediction of cardiovascular risk markers and diseases using retinal fundus images.

Inclusion criteria:

We will include studies published in the medical literature that used deep learning for the analysis of retinal fundus images to predict cardiovascular risk markers or cardiovascular diseases (prevalent or incident). According to our definition, risk markers do not include all risk factors merely associated with cardiovascular risk (e.g. age, sex), but those measurements that are directly derived from the cardiovascular system (e.g. blood pressure, coronary artery calcification, intima-media thickness, etc). Studies that only use pre-defined characteristics of retinal fundus images (e.g. tortuosity, fractal dimension) will not be considered.

Methods:

We searched MEDLINE and Embase for peer-reviewed articles (original articles and brief reports) on 17 November 2023. Abstracts will be screened by at least two independent reviewers. Study characteristics (both clinical and methodological) will be extracted by the first author (LYL) using a data extraction form and verified by the senior author (AH). Results will be presented using descriptive statistics.