disease

USING MACHINE LEARNING AND CLINICAL data from electronic health records, researchers have constructed an in silico marker for coronary artery disease to better measure clinically important characterisations of the disease.

As detailed in a publication in The Lancet, the researchers hope that their model may lead to more targeted diagnosis and better disease management of coronary disease, and claim that their study is the first known research to map its characteristics on a spectrum. Previous studies have focused only on whether or not a patient has coronary disease. This and other common conditions exist on a spectrum of disease; each individual’s mix of risk factors and disease processes determines where they fall on the spectrum. However, most such studies break this disease spectrum into rigid classes of case (patient has disease) or control (patient does not have disease). This may result in missed diagnoses, inappropriate management, and poorer clinical outcomes, say the investigators.

“The information gained from this non-invasive patients who had withdrawn consent after the initial phase of the study or declined participation in extended follow-up. In total, 557 deaths occurred from the start of the trial up to the end of the current study period which ran through to December 2021, roughly doubling the total of 289 deaths that had occurred after the initial study time point.

The objective of the long-term follow-up is to assess whether there are between-group differences and to increase precision around the treatment effect estimates for endpoints including all-cause mortality, cardiovascular [CV] mortality, and non-cardiovascular mortality, Hochman noted in her presentation at AHA 2022. Reporting the interim findings from the study, Hochman relayed that all-cause death occurred in 13.4% of patients in the conservative treatment group, compared to 12.7% in the invasive treatment group (hazard ratio [HR]: 1; 95% confidence interval [CI]: 0.85, 1.18; p=0.741 (log rank).

“For cardiovascular death, we had the hypothesis that spontaneous myocardial infarction (MI) reduction would lead to a reduction in cardiovascular death,” Hochman commented, before going on to note that this “appears to be the case,” with the rate of CV death standing at 8.6% in the conservative treatment arm and 6.4% in the invasive treatment arm (HR: 0.78, 95% CI: 0.63, 0.96; p=0.008). Finally, non-CV death stood at 4.4% in the conservative arm, and 5.6% in the invasive arm (HR: 1.44, 95% CI: 1.08, 1.91; p=0.016).

“Extended follow-up of the ISCHEMIA randomised trial over a median of 5.7 years demonstrated that an initial invasive strategy compared to an initial conservative strategy resulted staging of disease could empower clinicians by more accurately assessing patient status and, therefore, inform the development of more targeted treatment plans,” says senior study author Ron Do (Icahn School of Medicine at Mount Sinai, New York, USA).

“Our model delineates coronary artery disease patient populations on a disease spectrum; this could provide more insights into disease progression and how those affected will respond to treatment. Having the ability to reveal distinct gradations of disease risk, atherosclerosis, and survival, for example, which may otherwise be missed with a conventional binary framework, is critical.”

In the retrospective study, the researchers trained the machine learning model, named in silico score for coronary artery disease (ISCAD), to measure coronary disease on a spectrum using more than 80,000 electronic health records from two large health system-based biobanks, the BioMe Biobank at the Mount Sinai Health System and the UK Biobank.

The model, which the researchers termed a “digital marker,” incorporated hundreds of different clinical features from the electronic health record, including vital signs, laboratory test results, medications, symptoms, and diagnoses, and compared it to both an existing clinical score for coronary disease, which uses only a small number of predetermined features, and a in no difference in all-cause mortality, with nearly twice the number of deaths, lower risk of cardiovascular, but higher risk of non-cardiovascular mortality,” Hochman concluded. “These findings provide evidence for patients with chronic coronary disease and their physicians as they decide whether to add invasive management to guidelinedirected medical therapy.”

Summing up the findings after the presentation, Hochman said: “These longer-term results build on our original findings, which is that patients with daily or even weekly angina will likely see improved quality of life, but we have more robust data that an invasive approach is not going to prolong life.

“Although cardiovascular mortality was reduced, with an invasive approach it was offset by higher noncardiovascular mortality, so all-cause mortality is the same over time in both groups. Either strategy is acceptable, which is good news for patients, who have different preferences.”

Providing a commentary on the results at AHA 2022, Cecilia Bahit (INECO Neurociencias, Rosario, Argentina) said that the findings presented by Hochman appeared to be in line with prior data.

“When we look at the results and we put them into perspective, we see that the findings of all-cause mortality in the ISCHEMIA-EXTEND report are comparable to other previously recorded trials of revascularisation versus medical therapy,” she commented.

“During the extended follow-up investigators were able to identify more death, and were able to identify lower risk of cardiovascular death in the invasive strategy, and this was in benefit of an invasive strategy with a 21% relative risk reduction compared to the conservative arm.” genetic score for coronary artery disease.

The 95,935 participants included participants of African, Hispanic/Latino, Asian, and European ethnicities, as well as a large share of women. Most clinical and machine learning studies on coronary disease have focused on white European ethnicity. Investigators found that the probabilities from the model accurately tracked the degree of coronary stenosis, mortality, and complications.

“Machine learning models like this could also benefit the healthcare industry at large by designing clinical trials based on appropriate patient stratification. It may also lead to more efficient data-driven individualised therapeutic strategies,” says lead author Iain S Forrest, (Icahn Mount Sinai, New York, USA). “Despite this progress, it is important to remember that physician and procedure-based diagnosis and management of coronary artery disease are not replaced by artificial intelligence, but rather potentially supported by ISCAD as another powerful tool in the clinician’s toolbox.”

Next, the investigators envision conducting a prospective large-scale study to further validate the clinical utility and actionability of ISCAD, including in other populations. They also plan to assess a more portable version of the model that can be used universally across health systems.