Hidden Heart Danger: MRI Strain Model Spots Latent Obstruction in Hypertrophic Cardiomyopathy

Hypertrophic cardiomyopathy, the most common inherited heart muscle disorder, has long posed a deceptively simple question to cardiologists: which patients are truly in danger? A new study published in BMC Medical Imaging suggests that the answer may be hiding in the subtle mechanics of the beating heart, visible only through advanced cardiac magnetic resonance imaging. Researchers led by Yehan Sun and Hao Wu of the Affiliated Zhongshan Hospital of Dalian University in China have developed a three-parameter imaging model that can flag a dangerous but elusive condition known as latent obstructive hypertrophic cardiomyopathy, in which blood flow out of the heart becomes blocked only under physical stress. The work, published on 9 October 2026, offers a non-invasive window into a phenotype that standard resting examinations routinely miss.

The clinical stakes are considerable. In hypertrophic cardiomyopathy, the heart muscle thickens abnormally, and in some patients this thickened muscle obstructs the left ventricular outflow tract, the channel through which blood exits the heart into the aorta. When this obstruction is present at rest, it is easy to detect with echocardiography. But in the latent form, the obstruction emerges only during exertion, when the pumping chamber empties more vigorously and the muscle geometry shifts. Because patients often appear unremarkable during a resting clinic visit, latent obstruction can go unrecognized, even though studies have indicated that its prognosis is comparable to that of overt obstruction. Identifying these patients without resorting to exercise or pharmacological provocation has remained an unmet need in cardiology.

The research team approached the problem by exploiting a technique called cardiac magnetic resonance feature tracking, or CMR-FT. This method analyzes standard cine MRI sequences, the same looping movies of the heart that radiologists routinely acquire, and computationally tracks the movement of tissue features across frames. From that tracking, the software derives myocardial strain, a quantitative measure of how much the heart muscle deforms during each contraction and relaxation. Strain can be resolved into different geometric components: global radial strain, which describes thickening of the muscle wall; global circumferential strain, which captures the wringing motion of the ventricle around its axis; and global longitudinal strain, which reflects the shortening of the heart from base to apex. The team also measured the time to peak longitudinal strain, a marker of how synchronously the heart muscle contracts, and strain rates, which describe the speed of deformation.

Alongside the mechanical measurements, the investigators added a tissue characterization parameter: the extracellular volume fraction, or ECV, derived from T1 mapping. T1 mapping is a quantitative MRI technique that measures the relaxation time of hydrogen protons in tissue, and when combined with gadolinium contrast agent, it allows estimation of the fraction of heart muscle volume occupied by extracellular space. A higher ECV typically signals diffuse fibrosis, the scarring-like expansion of the space between muscle cells that accompanies many forms of heart disease. By pairing mechanics with tissue composition, the study aimed to build a multiparametric fingerprint of the latent obstructive phenotype rather than relying on any single measurement.

The study population was drawn from 159 consecutive patients with hypertrophic cardiomyopathy who underwent cardiac MRI at 3.0 Tesla between September 2024 and December 2025. The researchers applied strict exclusion criteria to eliminate confounding conditions, removing 31 patients with uncontrolled hypertension, infiltrative cardiomyopathy, persistent atrial fibrillation, coronary artery disease, or resting obstructive hypertrophic cardiomyopathy. A further 16 patients were excluded because of poor image quality or missing clinical data, leaving 112 patients in the final analysis: 81 with hypertrophic non-obstructive cardiomyopathy and 31 with latent obstructive disease. All participants underwent cine imaging, T1 mapping, and late gadolinium enhancement scans, followed by feature tracking strain analysis.

The comparisons between groups revealed a consistent pattern. Patients with latent obstruction had a significantly higher left ventricular ejection fraction, the standard measure of pumping efficiency, suggesting a hyperdynamic ventricle. They also showed increased global radial strain and greater absolute global circumferential strain, indicating exaggerated squeezing motions of the heart wall. Perhaps most intriguingly, their time to peak longitudinal strain was prolonged, pointing to mechanical contraction that is not only stronger but also less temporally coordinated. Meanwhile, their extracellular volume fraction was lower than in the non-obstructive group, hinting at differences in tissue composition, potentially including less diffuse fibrotic remodeling, that accompany the provocable obstruction phenotype. All of these differences reached statistical significance.

To convert these observations into a diagnostic tool, the team used univariate and multivariate logistic regression to identify which parameters independently predicted latent obstruction. Three emerged as independent predictors: global radial strain, time to peak longitudinal strain, and extracellular volume fraction. When combined into a single model, these three parameters achieved an area under the receiver operating characteristic curve of 0.888, with a 95 percent confidence interval of 0.814 to 0.940. In practical terms, the model demonstrated a sensitivity of 83.87 percent and a specificity of 88.89 percent, meaning it correctly identified the large majority of patients with latent obstruction while rarely raising false alarms in those without it. Statistical comparison using the DeLong test confirmed that the combined model significantly outperformed any single parameter alone.

The authors are careful to frame the model’s role appropriately. They emphasize that it should not be viewed as a replacement for provocative echocardiography, the current gold standard for unmasking latent obstruction, which involves exercising the patient or administering a drug while imaging the outflow tract. Instead, the multiparametric model is positioned as a non-invasive screening tool, one that could be applied to the vast pool of hypertrophic cardiomyopathy patients who already undergo cardiac MRI as part of routine assessment. Patients flagged by the model could then be prioritized for formal hemodynamic evaluation, concentrating provocative testing where it is most likely to change management. The researchers also note that external validation in independent cohorts is required before the model can be applied clinically, a standard caveat that reflects the single-center nature of the study.

The significance of the work lies in its reframing of what cardiac MRI can reveal about obstruction risk. Traditional MRI assessment of hypertrophic cardiomyopathy focuses on structural measures such as wall thickness, mass, and the presence of late gadolinium enhancement, which marks focal scarring. By shifting attention to functional deformation and quantitative tissue composition, the study suggests that the seeds of provocable obstruction are visible even when the heart is at rest, encoded in the way muscle fibers shorten, twist, and thicken, and in the microscopic architecture of the extracellular matrix. If validated more broadly, this approach could transform a routine diagnostic scan into a risk-stratification instrument, catching silent obstruction before it manifests as symptoms, arrhythmias, or sudden cardiac events.

The research was supported by the Dalian Science and Technology Innovation Fund Program under a major basic research grant, and it was approved by the Ethics Review Committee of Dalian University Affiliated Zhongshan Hospital with written informed consent from all participants. As feature tracking software becomes more widely available and automated, the barrier to adopting such multiparametric models continues to fall, and studies like this one chart a path toward imaging workflows in which every hypertrophic cardiomyopathy MRI yields not just pictures but a quantitative verdict on obstruction risk. For the many patients whose dangerous physiology only reveals itself on the treadmill, that verdict, delivered while lying still inside an MRI scanner, could make the difference between a condition discovered early and one discovered too late.

Subject of Research: Cardiac MRI feature tracking for detecting latent obstructive hypertrophic cardiomyopathy

Article Title: Feature tracking in latent obstructive hypertrophic cardiomyopathy: a multiparametric study based on myocardial strain and tissue characteristics

Article References: Sun, Y., Wu, H., Niu, K., Chen, S., Yu, Q., Zhao, X., Liang, X., Song, D., & Yu, J. (2026). Feature tracking in latent obstructive hypertrophic cardiomyopathy: a multiparametric study based on myocardial strain and tissue characteristics. BMC Medical Imaging. https://doi.org/10.1186/s12880-026-02839-6

Image Credits: AI Generated

DOI: 10.1186/s12880-026-02839-6

Keywords: hypertrophic cardiomyopathy, latent obstruction, cardiac MRI, feature tracking, myocardial strain, extracellular volume fraction, T1 mapping, left ventricular outflow tract, diagnostic model, cardiology, medical imaging, fibrosis