Deep phenotyping to integrate laboratory workflow in hypertrophic and dilated cardiomyopathy: an Italian exploratory study

Heart disease often carries a hidden code within our DNA. Two common forms, known as hypertrophic cardiomyopathy and dilated cardiomyopathy, involve the heart muscle becoming either too thick or too weak and stretched. While doctors know that genetic errors frequently cause these conditions, finding the specific error is not always straightforward. In many cases, genetic tests return results that are unclear, leaving patients and families without a definitive answer. This uncertainty creates a difficult gap in care, as knowing the exact genetic cause helps doctors predict the future course of the disease and guide family members. To bridge this gap, researchers have long wondered if the physical signs of the disease—the way the heart looks on an ultrasound or how it beats on an electrical monitor—could act as a map to predict which patients are most likely to have a clear genetic answer.

A team of scientists across three Italian hospitals set out to test this idea with a large group of patients. They gathered detailed records from 186 individuals with the thickened heart condition and 176 with the stretched heart condition. For each person, they collected a vast amount of information, including age, family history, blood pressure, and specific measurements from heart scans and electrical tests. They then compared these physical details against the results of the patients' genetic tests. The goal was to see if a computer could learn to spot patterns in the physical data that would signal a clear genetic finding versus an unclear one.

The researchers used advanced computer learning tools to sift through the data, looking for connections that a human eye might miss. They found that certain physical traits did stand out. For patients with the thickened heart, factors like the width of the heart's main artery, the thickness of the back wall of the heart, and the duration of the electrical signal across the chest were the strongest clues. For those with the stretched heart, the length of the electrical pause between beats and the age at diagnosis were the most telling signs. However, when the team tested how well these clues could actually predict the outcome for new patients, the results were modest. The computer models could distinguish between clear and unclear genetic results better than random guessing, but they were not powerful enough to be used as a standalone diagnostic tool.

The study suggests that while the physical shape and behavior of the heart hold some hints about the underlying genetics, they are not a complete picture. The researchers found that specific patterns of physical traits consistently appeared alongside clear genetic results, but these patterns were not strong enough to replace the need for genetic testing. The findings indicate that looking at the whole patient—their symptoms, their heart measurements, and their family history—might help doctors decide which patients are most likely to benefit from genetic testing, but it cannot yet tell them the answer on its own. The work highlights that the relationship between a person's physical heart and their genetic code is complex, and while deep observation of the body offers valuable context, it currently serves best as a companion to genetic analysis rather than a replacement for it.