A new blood-based lipodomic-enhanced risk score (LRS) augmenting the Framingham Risk Score (FRS) may be superior to the FRS alone in improving stratification of intermediate-risk patients for primary prevention, according to results of a study published in JACC.

The LRS was developed using a machine learning workflow using data from 10,339 patients in the Australian Diabetes, Obesity and Lifestyle Study (AusDiab), a population based prospective study conducted between 1999 and 2012, and was externally validated using data from 4,492 patients in the Busselton Health Study, a community-based epidemiological study with data collected in 1994/1995, while the predictive ability of coronary artery calcium scoring (CACS) was independently validated using a cohort from the multicenter observational BioHEART study.

Results showed that in the AusDiab and Busselton cohorts respectively, the LRC increased the area under the curve (AUC) for cardiovascular events by 0.114 and 0.077. The net reclassification improvement was 0.36 and 0.33 respectively in the two cohorts. The LRS also achieved a significant improvement in the area under the curve compared with the FRS regarding CACS (0.76 vs 0.74; p<1.0 x 10-5).

Additionally, the researchers found that for every 100 cases initially categorized as intermediate-risk by the FRS, the LRS would reclassify 36 as high risk and 16 as low risk. For every 100 controls initially categorized as intermediate-risk, the LRS would reclassify 18 as high risk and 31 as low risk.

A simplified version of the model focusing on 153 lipids instead of all 705 in the original LRS was found to have comparable performance and that it would be easier to implement clinically.

"With its potential for integration into routine blood tests, the LRS can streamline the selection of individuals for noninvasive imaging such as CACS and pave the way for enhanced precision in [cardiovascular disease] prevention and management," write Jingqin Wu, PhD.

In an accompanying editorial comment, Wen-Liang Song, MD, MTR, et al., call the work an "important proof of principle showing that consideration of the lipidome, comprising hundreds of species, can improve prediction and deepen our understanding of lipids, lipoproteins, and their role in the pathophysiology of [cardiovascular disease]." They add that the LRS must be tested among more diverse cohorts, and that, "ultimately, the clinical utility of lipidomic, and other high-throughput molecular profiling, will be determined by collective consideration of their independent additive value, implementability, and cost-effectiveness."