By Sarah Handzel, BSN, RN
Once thought to be a rare cardiovascular condition, hypertrophic cardiomyopathy (HCM) actually occurs in an estimated one out of every 500 people, making it the most common genetic heart disease in the U.S.1 However, because patients are generally mildly symptomatic or even asymptomatic until the disease progresses, HCM remains a significantly underdiagnosed problem.2 Further complicating diagnosis, HCM may be confused with other types of heart disease such as non-ST elevation acute coronary syndrome (NSTEACS), especially since clinical presentation may vary.3
Early identification of HCM patients can be crucial for beginning prompt treatment that prevents the condition from worsening—but clinicians cannot rely on clinical evaluation alone. Instead, they must lean on other tools, such as ECG or echocardiogram, to aid in the diagnostic process. Cardiologists should use the results of ECG when deciding which imaging tests may be most appropriate for diagnostic confirmation; ECG may also allow them to more easily view structural changes that confirm the presence and progression of HCM.
Using both ECG and imaging tests like echocardiography remains the gold standard for confirming a diagnosis—but in the future, machine learning may supplement cardiologists' expertise in reviewing test data, resulting in higher rates of confirmed HCM among the general population.
ECG Patterns Provide First Clues
Because ECG is usually the first test completed when heart disease is suspected, distinguishing HCM from other pathologies requires diligent inspection of ECG test results. However, many ECG changes reflect patient-specific information that may not be related to HCM at all. For example, ECG changes often recorded in athletes may overlap with ECG changes identified in patients with HCM.4 In others, ECG changes may suggest problems like cardiac amyloidosis or the presence of glycogen storage disease.4 Therefore, the cardiologist must rely on expertise to correctly interpret any findings with respect to individual patient factors.
Several ECG changes are already associated with HCM, including higher R waves, higher T wave voltage and peak voltage, and T wave asymmetry.2 These differences may help distinguish HCM from other heart diseases, but more evidence is necessary for a true diagnosis.
To that end, a 2020 study in BMC Cardiovascular Disorders evaluated 12-lead ECG recordings in patients with known HCM and compared them to patients with NSTEACS (Non ST-elevation acute coronary syndrome). Study findings indicate a greater maximum amplitude of the R wave in lead V5 in patients with HCM compared to those without. Additionally, T wave magnitude significantly differed between the two groups, with giant negative T wave (1mv or greater) occurring with greater frequency in the HCM test group.2
The same study also suggests that the magnitude of ST-segment deviation differs significantly between those with HCM and those with NSTEACS; this was true in 10-lead exams excluding both the aVF and V2 leads. HCM was also associated with greater maximal amplitude of ST-segment depression, particularly in leads V3-V6.2
Another 2020 study in Annals of Noninvasive Cardiology indicates that other ECG changes may be present in patients with suspected HCM. This study's authors suggest that the Romhilt-Estes score is actually the most sensitive HCM criterion, with the score having the strongest correlation with myocardial mass (LVM). The authors hypothesize that the Romhilt-Estes score is most sensitive because it incorporates a variety of factors, including voltage criteria, ST-T abnormalities, P-wave features, QRS durations, left axis deviations, and delayed intrinsicoid deflections.5
Machine Learning May Soon Supplement Cardiologists' Expertise
As artificial intelligence (AI) makes its way into more aspects of healthcare, researchers hope to eventually develop machine learning models to help cardiologists better detect and differentiate HCM from other cardiac conditions. A 2022 study in Circulation reports on the results of trained ECG models with respect to ECG interpretation by cardiologists across multiple clinical settings.
In the study, researchers used over 8,000 ECG recordings taken at four separate academic medical centers to train ECG models to identify HCM and differentiate it from other cardiac diseases. When the models trained on internal test data specific to only one medical center, the ability of the model to generalize ECG characteristics specific to HCM on external data sets was limited. However, a federated learning approach trained ECG models to generalize across multiple data sets, leading to greater detection of HCM within test groups.6
When the study authors compared these results with manual detection of HCM by cardiologists, they found that machine-learning models had higher sensitivity for HCM at any given specificity.6 They conclude that the data collected suggests ECG screening with machine-learning models improves sensitivity at the same positive predictive value when compared to the strategy of performing an echocardiogram on all HCM-suspected patients. Ultimately, this may help reduce the number of echocardiograms performed.6
While machine learning models are still being developed, cardiologists cannot yet completely rely on them for diagnostic accuracy. Instead, clinicians must consider ECG results that may indicate the presence of HCM and use those differentiators to direct further testing. Tests like echocardiogram are essential for diagnosis, but some may still be missed—therefore, it is essential that cardiologists thoroughly investigate any ECG changes that may be suggestive of HCM.
Resources
1. Butzner M, Leslie DL, Cuffee Y, et al. Stable rates of obstructive hypertrophic cardiomyopathy in a contemporary era. Front Cardiovasc Med. 2022;8:765876. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8770922/#.
2. Owens AT, Reza N. Diagnosis of hypertrophic cardiomyopathy: what every cardiologist needs to know. American College of Cardiology. Published February 27, 2020. https://www.acc.org/latest-in-cardiology/articles/2020/02/25/06/34/diagnosis-of-hypertrophic-cardiomyopathy#.
3. Tao Y, Xu J, Bako SY, et al. Usefulness of ECG to differentiate apical hypertrophic cardiomyopathy from non-ST elevation acute coronary syndrome. BMC Cardiovascular Disorders. 2020;20(1):306. https://bmccardiovascdisord.biomedcentral.com/articles/10.1186/s12872-020-01592-0.
4. Finocchiaro G, Sheikh N, Biagini E, et al. The electrocardiogram in the diagnosis and management of patients with hypertrophic cardiomyopathy. Heart Rhythm. 2020;17(1):142-151. https://www.sciencedirect.com/science/article/abs/pii/S1547527119306605.
5. Dohy Z, Vereckei A, Horvath V, et al. How are ECG parameters related to cardiac magnetic resonance images? Electrocardiographic predictors of left ventricular hypertrophy and myocardial fibrosis in hypertrophic cardiomyopathy. Noninvasive Electrocardiol. 2020;25(5). https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/anec.12763.
6. Goto S, Solanki D, John JE, et al. Multinational federated learning approach to train ECG and echocardiogram models for hypertrophic cardiomyopathy detection. Circulation. 2022;146(10):755-769. https://www.ahajournals.org/doi/10.1161/CIRCULATIONAHA.121.058696.
