Assessment of left ventricular contraction patterns using gated SPECT MPI to predict cardiac resynchronization therapy response

Myocardial Perfusion SPECT Imaging Radiomic Features and Machine Learning Algorithms for Cardiac Contractile Pattern Recognition

A U-shaped contraction pattern was shown to be associated with a better Cardiac resynchronization therapy (CRT) response. The main goal of this study is to automatically recognize left ventricular contractile patterns using machine learning algorithms trained on conventional quantitative features (ConQuaFea) and radiomic features extracted from Gated single-photon emission computed tomography myocardial perfusion imaging (GSPECT MPI). Among 98 patients with standard resting GSPECT MPI included in this study, 29 received CRT therapy and 69 did not (also had CRT inclusion criteria but did not receive treatment yet at the time of data collection, or refused treatment). A total of 69 non-CRT patients were employed for training, and the 29 were employed for testing. The models were built utilizing features from three distinct feature sets (ConQuaFea, radiomics, and ConQuaFea + radiomics (combined)), which were chosen using Recursive feature elimination (RFE) feature selection (FS), and then trained using seven different machine learning (ML) classifiers. In addition, CRT outcome prediction was assessed by different treatment inclusion criteria as the study's final phase. The MLP classifier had the highest performance among ConQuaFea models (AUC, SEN, SPE = 0.80, 0.85, 0.76). RF achieved the best performance in terms of AUC, SEN, and SPE with values of 0.65, 0.62, and 0.68, respectively, among radiomic models. GB and RF approaches achieved the best AUC, SEN, and SPE values of 0.78, 0.92, and 0.63 and 0.74, 0.93, and 0.56, respectively, among the combined models. A promising outcome was obtained when using radiomic and ConQuaFea from GSPECT MPI to detect left ventricular contractile patterns by machine learning.

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.

A Mild Dyssynchronous Contraction Pattern Detected by SPECT Myocardial Perfusion Imaging Predicts Super-Response to Cardiac Resynchronization Therapy

Background

Using single photon emission computed tomography myocardial perfusion imaging (SPECT MPI) with phase analysis (PA), we aimed to identify the predictive value of a new contraction pattern in cardiac resynchronization therapy (CRT) response.

Methods

Left ventricular mechanical dyssynchrony (LVMD) was evaluated using SPECT MPI with PA in non-ischemic dilated cardiomyopathy (DCM) patients with left bundle branch block (LBBB) indicated for CRT. CRT super-response was defined as LV ejection fraction (EF) ≥50% or an absolute increase of LVEF >15%. The LV contraction was categorized as the mild dyssynchronous pattern when the phase standard deviation (PSD) ≤ 40.3° and phase histogram bandwidth (PBW) ≤ 111.9°, otherwise it was defined as severe dyssynchronous pattern which was further characterized as U-shaped, heterogeneous or homogenous pattern.

Results

The final cohort comprised 74 patients, including 32 (43.2%) in mild dyssynchronous group, 17 (23%) in U-shaped group, 19 (25.7%) in heterogeneous group, and 6 (8.1%) in homogenous group. The mild dyssynchronous group had lower PSD and PBW than U-shaped, heterogeneous, and homogenous groups (P < 0.0001). Compared to patients with the heterogeneous pattern, the odds ratios (ORs) with 95% confidence intervals (CIs) for CRT super-response were 10.182(2.43–42.663), 12.8(2.545–64.372), and 2.667(0.327–21.773) for patients with mild dyssynchronous, U-shaped, and homogenous pattern, respectively. After multivariable adjustment, mild dyssynchronous group remained associated with increased CRT super-response (adjusted OR 5.709, 95% CI 1.152–28.293). Kaplan-Meier curves showed that mild dyssynchronous group demonstrated a better long-term prognosis.

Conclusions

The mild dyssynchronous pattern in patients with DCM is associated with an increased CRT super-response and better long-term prognosis.

Complete electrical reverse remodeling of native conduction after resynchronization therapies

BACKGROUND

Electrical reverse remodeling of native conduction is associated with better clinical outcome following cardiac resynchronization therapy (CRT). We aimed to describe characteristics, time course and long-term outcome of patients with complete electrical reverse remodeling (CERR) after resynchronization therapies.

METHODS

CRT candidates were treated with bi-ventricular, His bundle or left bundle branch pacing. CERR was defined if native QRS duration post-implantation was narrowed to ≤120 ms.

RESULTS

A total of 322 patients met the inclusion criteria. Among them, 66 were super-responders and 12 exhibited CERR. All 12 patients were diagnosed of non-ischemic cardiomyopathy with left bundle branch block (LBBB) meeting the Strauss criteria. The mean native QRS duration when CERR was achieved was 110.8 ± 10.0 ms, significantly shorter than the baseline (175.0 ± 18.8 ms). The occurrence of CERR varied from several days post-implantation to 18-month follow-up. The persistence of CERR also showed great variations. Eleven patients (91.7%, 11/12) showed echocardiographic super-response. Patients with CERR showed similar baseline characteristics compared to those with echocardiographic super-response but without CERR. Two patients with CERR showed different responses after bi-ventricular pacing was turned off. One patient remained stable with narrow QRS complex and great response. The other patient had reappearance of LBBB and decreased cardiac function, but recovered by turning on the device again.

CONCLUSIONS

Patients with CERR exhibited great response to different resynchronization therapies. The time course of CERR and echocardiographic super-response varied greatly. The variability of native conduction system and accompanied changes of mechanical remodeling suggest the mechanisms of electrical-disorder related cardiomyopathy.

Mechanical contraction to guide CRT left-ventricular lead placement instead of electrical activation in myocardial infarction with left ventricular dysfunction: An experimental study based on non-invasive gated myocardial perfusion imaging and invasive electroanatomic mapping

BACKGROUND

Whether the region of the latest electrical activation (LEA) corresponds with the segment of the latest mechanical contraction (LMC) in ischemic cardiomyopathy (ICM) is uncertain. We aimed to investigate the relationship between the left-ventricular (LV) viable segments with LEA and with LMC after myocardial infarction (MI) and analyze the acute hemodynamic responses (dP/dtmax) after cardiac resynchronization therapy (CRT) pacing at different LV sites.

METHODS AND RESULTS

Bama suckling pigs (n = 6) were subjected to create MI models. Both gated myocardial perfusion imaging (GMPI) and electroanatomic mapping (EAM) were performed successfully before MI and 4 weeks after MI. LMC was assessed by phase analysis of GMPI, while LEA was evaluated by EAM. The dP/dtmax was measured before CRT and when the CRT LV electrode was implanted in viable segments of LMC, viable segments of lateral wall and scar, respectively. The viable segments of LEA were consistent with the sites of LMC for five in six cases. The dP/dtmax increased significantly compared with that before CRT when the CRT LV electrode was implanted in viable segments of LMC (1103.33 ± 195.76 vs 717.83 ± 80.74 mmHg·s-1, P = .001), which was also significantly higher than in viable segments of lateral wall (751.17 ± 105.62 mmHg·s-1, P = .000) and scar (679.50 ± 60.87 mmHg·s-1, P = .001).

CONCLUSIONS

Non-invasive GMPI may be a better option than invasive EAM for guiding LV electrode implantation for CRT in ICM.

Low septal to lateral wall 18F-FDG ratio is highly associated with mechanical dyssynchrony in non-ischemic CRT candidates

Background

In order to better understand the concept of mechanical dyssynchrony, a promising hallmark of cardiac resynchronization therapy (CRT) response, we investigated its effect on regional myocardial metabolism and myocardial blood flow (MBF) in non-ischemic CRT candidates.

Results

Thirty consecutive non-ischemic CRT eligible patients underwent static ¹⁸F-FDG and resting dynamic ¹³N-NH₃ PET/CT. ¹⁸F-FDG uptake and MBF for septal and lateral wall were analysed and septal-to-lateral wall ratios (SLR) were calculated. Based on the presence of mechanical dyssynchrony (septal flash and/or apical rocking) on echocardiography, patients were divided into 2 groups, with (n = 23) and without (n = 7) mechanical dyssynchrony.

Patients with mechanical dyssynchrony had significantly lower ¹⁸F-FDG SUVmean in the septum compared with the lateral wall (5.58 ± 2.65 vs 11.19 ± 4.10, p < 0.0001), while patients without mechanical dyssynchrony had a more homogeneous ¹⁸F-FDG distribution (7.33 ± 2.88 vs 8.31 ± 2.50, respectively, p = 0.30). Similarly, MBF was significantly different between the septal and lateral wall in the dyssynchrony group (0.57 ± 0.11 ml/g/min vs 0.92 ± 0.23 ml/g/min, respectively, p < 0.0001), whereas no difference was observed in the non-dyssynchrony group (0.61 ± 0.23 ml/g/min vs 0.77 ± 0.21 ml/g/min, respectively, p = 0.16). ¹⁸F-FDG SLR, but not MBF SLR, was associated with the presence of mechanical dyssynchrony and showed a significant inverse correlation with volumetric reverse remodeling after CRT (r = − 0.62, p = 0.001).

Conclusions

Non-ischemic heart failure patients with mechanical dyssynchrony demonstrate heterogeneous regional metabolism and MBF compared with patients without dyssynchrony. However, only ¹⁸F-FDG SLR appeared to be highly associated with the presence of mechanical dyssynchrony.

Trial registration

Clinicaltrials, NCT02537782. Registered 2 September 2015.

Multi-modality imaging: Bird's eye view from the 2017 American Heart Association Scientific Sessions

This review summarizes key imaging studies that were presented in the American Heart Association Scientific Sessions 2017 related to the fields of nuclear cardiology, cardiac computed tomography, cardiac magnetic resonance, and echocardiography. The aim of this bird's eye view is to inform readers about multiple studies reported at the meeting from these different imaging modalities. While such a review is most useful for those that did not attend the conference, we find that a general overview may also be useful to those that did since it is often difficult to get exposure to many abstracts at large meetings. The review, therefore, aims to help readers stay updated on the newest imaging studies presented at the meeting and will hopefully stimulate new ideas for future research in imaging.