Effective prediction of response to cardiac resynchronization therapy using a novel program of gated myocardial perfusion single photon emission computed tomography

A method using deep learning to discover new predictors from left-ventricular mechanical dyssynchrony for CRT response

BACKGROUND

Studies have shown that the conventional parameters characterizing left ventricular mechanical dyssynchrony (LVMD) measured on gated SPECT myocardial perfusion imaging (MPI) have their own statistical limitations in predicting cardiac resynchronization therapy (CRT) response. The purpose of this study is to discover new predictors from the polarmaps of LVMD by deep learning to help select heart failure patients with a high likelihood of response to CRT.

METHODS

One hundred and fifty-seven patients who underwent rest gated SPECT MPI were enrolled in this study. CRT response was defined as an increase in left ventricular ejection fraction (LVEF) > 5% at 6 [Formula: see text] 1 month follow up. The autoencoder (AE) technique, an unsupervised deep learning method, was applied to the polarmaps of LVMD to extract new predictors characterizing LVMD. Pearson correlation analysis was used to explain the relationships between new predictors and existing clinical parameters. Patients from the IAEA VISION-CRT trial were used for an external validation. Heatmaps were used to interpret the AE-extracted feature.

RESULTS

Complete data were obtained in 130 patients, and 68.5% of them were classified as CRT responders. After variable selection by feature importance ranking and correlation analysis, one AE-extracted LVMD predictor was included in the statistical analysis. This new AE-extracted LVMD predictor showed statistical significance in the univariate (OR 2.00, P = .026) and multivariate (OR 1.11, P = .021) analyses, respectively. Moreover, the new AE-extracted LVMD predictor not only had incremental value over PBW and significant clinical variables, including QRS duration and left ventricular end-systolic volume (AUC 0.74 vs 0.72, LH 7.33, P = .007), but also showed encouraging predictive value in the 165 patients from the IAEA VISION-CRT trial (P < .1). The heatmaps for calculation of the AE-extracted predictor showed higher weights on the anterior, lateral, and inferior myocardial walls, which are recommended as LV pacing sites in clinical practice.

CONCLUSIONS

AE techniques have significant value in the discovery of new clinical predictors. The new AE-extracted LVMD predictor extracted from the baseline gated SPECT MPI has the potential to improve the prediction of CRT response.

Phase analysis for the assessment of left ventricular dyssynchrony by Gated Myocardial Perfusion SPECT. Importance of clinical and technical parameters

Introducción. El análisis de fase (AF) del ventrículo izquierdo es una herramienta de reciente introducción en los estudios de cardiología nuclear, que permite valorar el sincronismo mecánico de la contracción del ventrículo izquierdo con diferentes aplicaciones clínicas, si bien es poco conocida.Objetivo. Mostrar la factibilidad de la nueva herramienta AF por perfusión miocárdica (Gated-SPECT) para valorar el sincronismo mecánico del ventrículo izquierdo y verificar diferencias entre sus valores, según situaciones clínicas y condiciones técnicas.Materiales y métodos. En el estudio participaron pacientes consecutivos con Gated-SPECT. Las variables principales fueron diferentes condiciones clínicas y técnicas. La valoración del AF se realizó mediante la herramienta FASE del programa cardiodedicado (QPS-QGS, Cedars-Sinai Medical Center, Los Angeles, USA). Se obtuvieron los siguientes parámetros: ancho del histograma (AH), desviación estandar de la fase (DE) y entropía (E). Se realizó análisis descriptivo y analítico de medias o medianas a través de test paramétricos o no paramétircos. El límite de significancia estadísitca fue p<0.05. Se utilizó IBM-SPSS V21®.Resultados. Con un total de 300 pacientes y una media de edad de 65±12.7, en el análisis del AF no existieron diferencias según la fase del estudio (estrés-reposo) [AH (p=0.4), DE (p=0.6), E (p=0.7)], tipo de estrés [AH (p=0.38), DE (p=0.8), E (p=0.84)], dosis utilizada [AH(p=0.19), DE (p=0.05), E (p=0.06)], gammacámara [AH (p=0.02), DE (p=0.06), E (p=0.08)] ni entre antecedente de enfermedad coronaria [AH (p=0.44), DE (p=0.18), E (p=0.17)].Hubo diferencias según trastornos de conducción [AH (p=0.001), DE (p=0.02), E (p=0.001)], fracción de eyección < o >35% [AH (p=0.001), DE (p=0.001), E (p=0.001)], estudio normal o con necrosis [AH (p=0.001), DE (p=0.001), E (p=0.001)] y género [AH (p=0.002), DE (p=0.006), E (p=0.005)].Conclusiones. El uso de la nueva herramienta del AF de medicina nuclear es factible. Sus parámetros no se afectaron por el tipo de estrés producido, dosis administrada o fase del estudio por la gammacámara empleada. Por su parte, sí fueron afectados por género, trastornos de conducción interventricular, necrosis y disfunción sistólica.

Can the Prognosis of Cardiac Resynchronization Therapy Be Predicted by Gated SPECT?

PURPOSE

The aim of this study was to evaluate the role of gated myocardial perfusion SPECT (GMPS) phase parameters (phase SD and histogram bandwidth) in predicting the short-term response to cardiac resynchronization therapy (CRT) as left ventricle (LV) remodeling and the long-term effect as all-cause mortality or cardiac transplantation and hospitalization for heart failure.

PATIENTS AND METHODS

This prospective observational study included 63 consecutive patients undergoing CRT from May 2008 to April 2014. Before CRT, patients underwent evaluation of New York Heart Association functional class, standard 12-lead ECG, Tc-MIBI GMPS, and 2-dimensional echocardiography. The patients were grouped according to complete left bundle-branch block (CLBBB). Follow-up was done every 6 months after pacemaker implantation. The short-term end point was reverse LV remodeling in 6 to 12 months follow-up, and the long-term primary end point was all-cause mortality or cardiac transplantation. The secondary end point was hospitalization for heart failure.

RESULTS

Reverse LV remodeling showed in 60.9% (39/63) patients (74.5% CLBBB [35/47] vs 25.0% non-CLBBB [4/16], P < 0.001). Phase SD less than 55 degrees, CLBBB, and left atrium diameter were independent predictive factors for reverse LV remodeling. At a median follow-up of 39.76 months, there were 17 deaths and 2 transplantations (17.0% CLBBB vs 68.8% non-CLBBB, P < 0.001). Seventeen patients required hospitalization more than once for heart failure, and 2 underwent heart transplantation (14.9% CLBBB vs 62.5% non-CLBBB, P < 0.001). Multivariate logistic regression showed that only CLBBB was an independent predictive factor for both end points.

CONCLUSIONS

The GMPS parameters were not independent predictive factors for all-cause mortality or cardiac transplantation and hospitalization for heart failure.

Radionuclide Assessment of Left Ventricular Dyssynchrony

Phase analysis of gated myocardial perfusion single-photon emission computed tomography is a widely available and reproducible measure of left ventricular (LV) dyssynchrony, which also provides comprehensive assessment of LV function, global and regional scar burden, and patterns of LV mechanical activation. Preliminary studies indicate potential use in predicting cardiac resynchronization therapy response and elucidation of mechanisms. Because advances in technology may expand capabilities for precise LV lead placement in the future, identification of specific patterns of dyssynchrony may have a critical role in guiding cardiac resynchronization therapy.

The clinical dilemma of quantifying mechanical left ventricular dyssynchrony for cardiac resynchronization therapy: segmental or global?

Heart failure (HF) represents a serious clinical and public cause of mortality, morbidity, as well as healthcare expenditures. Guidelines for treatment of HF join in recommending multimedical regimen at targeted doses as the best medical strategy, despite that a significant percentage of patients remain symptomatic. Studies have shown that these patients might benefit from cardiac resynchronization therapy (CRT), particularly those presenting with broad QRS duration, >135 msec. Trials have already shown that CRT results in improved morbidity and survival of these patients particularly those in New York Heart Association class III-IV HF, but almost 30% do not show any symptomatic or survival benefit, hence are classified as nonresponders. Exhaustive efforts have been made in using noninvasive methods of assessing left ventricle (LV) dyssynchrony in predicting nonresponders to CRT, including Doppler echocardiography, magnetic resonance imaging, and even single photon emission computed tomography analysis, but only with modest success. In this report, we aimed to review the available evidence for assessing markers of mechanical LV dyssynchrony by various echocardiographic modalities and their respective strength in predicting favorable response to CRT treatment, comparing global with segmental ones. While the accuracy of segmental markers of dyssynchrony in predicting satisfactory response to CRT remains controversial because of various technical limitations, global markers seem easier to measure, reproducible, and potentially accurate in reflecting overall cavity response and its clinical implications. More studies are needed to qualify this proposal.