Interpretation of SPECT wall motion with deep learning.

OBJECTIVES: We sought to develop a novel deep learning (DL) workflow to interpret single-photon emission computed tomography (SPECT) wall motion. BACKGROUND: Wall motion assessment with SPECT is limited by image temporal and spatial resolution. Visual interpretation of wall motion can be subjective and prone to error. Artificial intelligence (AI) may improve accuracy of wall motion assessment. METHODS: A total of 1038 patients undergoing rest electrocardiogram (ECG)-gated SPECT and echocardiography were included. Using echocardiography as truth, a DL-model (DL-model 1) was trained to predict the probability of abnormal wall motion. Of the 1038 patients, 317 were used to train a DL-model (DL-model 2) to assess regional wall motion. A 10-fold cross-validation was adopted. Diagnostic performance of DL was compared with human readers and quantitative parameters. RESULTS: The area under the receiver operating characteristic curve (AUC) and accuracy (ACC) of DL model (AUC: .82 [95% CI: .79-.85]; ACC: .88) were higher than human (AUC: .77 [95% CI: .73-.81]; ACC: .82; P < .001) and quantitative parameter (AUC: .74 [95% CI: .66-.81]; ACC: .78; P < .05). The net reclassification index (NRI) was 7.7%. The AUC and accuracy of DL model for per-segment and per-vessel territory diagnosis were also higher than human reader. The DL model generated results within 30 seconds with operable guided user interface (GUI) and therefore could provide preliminary interpretation. CONCLUSIONS: DL can be used to improve interpretation of rest SPECT wall motion as compared with current human readers and quantitative parameter diagnosis.

The Value of Left Ventricular Mechanical Dyssynchrony and Scar Burden in the Combined Assessment of Factors Associated with Cardiac Resynchronization Therapy Response in Patients with CRT-D.

BACKGROUND: Cardiac resynchronization therapy (CRT) improves the outcome in patients with heart failure (HF). However, approximately 30% of patients are nonresponsive to CRT. The aim of this study was to determine the role of the left ventricular (LV) mechanical dyssynchrony (MD) and scar burden as predictors of CRT response. METHODS: In this study, we included 56 patients with HF and the left bundle-branch block with QRS duration ≥ 150 ms who underwent CRT-D implantation. In addition to a full examination, myocardial perfusion imaging and gated blood-pool single-photon emission computed tomography were performed. Patients were grouped based on the response to CRT assessed via echocardiography (decrease in LV end-systolic volume ≥15% or/and improvement in the LV ejection fraction ≥5%). RESULTS: In total, 45 patients (80.3%) were responders and 11 (19.7%) were nonresponders to CRT. In multivariate logistic regression, LV anterior-wall standard deviation (adjusted odds ratio (OR) 1.5275; 95% confidence interval (CI) 1.1472-2.0340; p = 0.0037), summed rest score (OR 0.7299; 95% CI 0.5627-0.9469; p = 0.0178), and HF nonischemic etiology (OR 20.1425; 95% CI 1.2719-318.9961; p = 0.0331) were the independent predictors of CRT response. CONCLUSION: Scar burden and MD assessed using cardiac scintigraphy are associated with response to CRT.

Effectively Detecting Left Bundle Branch Block False Defects in Myocardial Perfusion Imaging (MPI) with a Convolutional Neural Network (CNN).

Left bundle branch block (LBBB) is a frequent source of false positive MPI reports, in patients evaluated for coronary artery disease. PURPOSE: In this work, we evaluated the ability of a CNN-based solution, using transfer learning, to produce an expert-like judgment in recognizing LBBB false defects. METHODS: We collected retrospectively, MPI polar maps, of patients having small to large fixed anteroseptal perfusion defect. Images were divided into two groups. The LBBB group included patients where this defect was judged as false defect by two experts. The LAD group included patients where this defect was judged as a true defect by two experts. We used a transfer learning approach on a CNN (ResNet50V2) to classify the images into two groups. RESULTS: After 60 iterations, the reached accuracy plateau was 0.98, and the loss was 0.19 (the validation accuracy and loss were 0.91 and 0.25, respectively). A first test set of 23 images was used (11 LBBB, and 12 LAD). The empiric ROC (Receiver operating characteristic) Area was estimated at 0.98. A second test set (18x2 images) was collected after the final results. The ROC area was estimated again at 0.98. CONCLUSION: Artificial intelligence, using CNN and transfer learning, could reproduce an expert-like judgment in differentiating between LBBB false defects, and LAD real defects.

Appropriate use of noninvasive ischemia testing to guide revascularization decision making following acute ST elevation myocardial infarction in Latin American countries: Results from an expert panel meeting of the International Atomic Energy Agency. / Uso adecuado de las pruebas no invasivas de isquemia para guiar la toma de decisión sobre revascularización tras un infarto agudo de miocardio con elevación del segmento ST en países iberoamericanos: Resultados de la reunión de un panel de expertos de la International Atomic Energy Agency.

Across Latin American and Caribbean countries, cardiovascular disease and especially ischemic heart disease is currently the main cause of death both in men and in women. For most Latin American and Caribbean countries, public and community health efforts aim to define care strategies which are both clinically and cost effective and promote primary and secondary prevention, resulting in improved patient outcomes. The optimal approach to deal with acute events such as ST-elevation myocardial infarction (STEMI) is a matter of controversy; however, there is an expanding role for assessing residual ischemic burden in STEMI patients following primary percutaneous coronary intervention. Although randomized clinical trials have established the value of staged fractional flow reserve-guided revascularization, the use of noninvasive functional imaging modalities may play a similar role at a much lower cost. For LAC, available stress imaging techniques could be applied to define residual ischemia in the non-infarct related artery and to target revascularization in a staged procedure after primary percutaneous coronary intervention The use of nuclear cardiac imaging, supported by its relatively wide availability, moderate cost, and robust quantitative capabilities, may serve to guide effective care and to reduce subsequent cardiac events in patients with coronary artery disease. This noninvasive approach may avert potential safety issues with repeat and lengthy invasive procedures, and serve as a baseline for subsequent follow-up stress testing following the index STEMI event. This consensus document was devised from an expert panel meeting of the International Atomic Energy Agency, highlighting available evidence with a focus on the utility of stress myocardial perfusion imaging in post-STEMI patients. The document could serve as guidance to the prudent and appropriate use of nuclear imaging for targeting therapeutic management and avoiding unnecessary invasive procedures within Latin American and Caribbean countries, where resources could be scarce.

Influence of acquisition orbit on phase analysis of gated single photon emission computed tomography myocardial perfusion imaging for assessment of left ventricular mechanical dyssynchrony.

OBJECTIVE: The association between left ventricular (LV) dyssynchrony parameters, given by phase analysis of gated single photon emission computed tomography (SPECT) myocardial perfusion imaging (MPI), and acquisition orbits is unclear. The aim of this study was to assess the dependence of LV dyssynchrony parameters on acquisition orbits. METHODS: Ninety-nine patients who underwent 201Tl-gated SPECT MPI were categorized into minor hypoperfusion or major hypoperfusion groups. Forty-four patients who underwent 99mTc-tetrofosmin-gated SPECT MPI were categorized into minor hypoperfusion or major hypoperfusion groups. The major hypoperfusion group with 201Tl was divided into inferior or non-inferior wall hypoperfusion subgroups, and anteroseptal or non-anteroseptal wall hypoperfusion subgroups. Gated SPECT MPI data over a 360° acquisition orbit (360° images) and a 180° acquisition orbit (180° images) were reconstructed, and histogram bandwidth (HBW) and phase standard deviation (PSD) were compared. RESULTS: Between 360° and 180° images with 201Tl, there were significant differences in HBW and PSD both globally (HBW 34.8 ± 16.6 vs. 29.1 ± 10.2; PSD 8.8 ± 4.9 vs. 7.0 ± 2.3, p < 0.05 for both) and in the inferior wall (HBW 29.5 ± 15.5 vs. 23.3 ± 9.0; PSD 7.6 ± 4.6 vs. 5.6 ± 2.4, p < 0.001 for both) in the major hypoperfusion group, and also in the inferior wall in all subgroups of the major hypoperfusion group. In contrast, no segment had any significant differences in HBW or PSD between 360° and 180° images with 99mTc. CONCLUSION: Differences in acquisition orbit had a significant influence on HBW and PSD with 201Tl-gated SPECT MPI in the inferior wall in patients with major hypoperfusion myocardium.

Prognostic value of rest gated SPECT myocardial perfusion in patients with chronic heart failure / 医学研究生学报

Objective Chronic heart failure ( CHF) is a disease with high morbidity, and mortality.The study was designed to explore the prognostic value of rest gated SPECT myocardial perfusion ( GSMPI) in patients with chronic heart failure. Methods This study was a prospective study, in which 142 CHF patients was included from January 1, 2011 to December 31, 2012.All patients underwent GSMPI, further, summed rest score ( SRS) , left ventricular ejection fraction ( LVEF) , end-diastolic volume ( EDV) , end-systolic volume ( ESV) and phase standard deviation ( PSD) were recorded.All the patients were followed up.Primary endpoint was the major adverse cardiac events ( MACE) .Additionally, we analyzed the independent predictive factor of MACE occurrence by using the Cox proportional hazard model. Results during the follow-up, MACE occurred in 65 cases.The median time to MACE occur-rence was 28(5-48) months.The 3-year MACE free survival rate was 64.3%.Multivariate Cox proportional hazards regression analysis showed that cardiac function grade[HR(95%CI):1.761(1.087~4.824), P=0.005], NT-proBNP[HR(95%CI):1.153(1.125~5.997), P =0.009], LVEF[HR(95%CI): 1.945(1.138 ~ 6.131), P=0.019], SRS[HR(95%CI):1.517(1.108~4.278), P=0.027] and PSD[HR(95%CI):1.442(1.156~5.339), P=0.032] were all independent predictors of MACE. Conclusion GSMPI may be a prognostic factor for the patients with chronic heart failure.

Cut-off values of myocardial perfusion gated-SPECT phase analysis parameters of normal subjects, and conduction and mechanical cardiac diseases.

BACKGROUND: The aim of this study was to determine the cut-off values of gated myocardial perfusion rest SPECT phase analysis parameters of normal subjects, and conduction (CCD) and mechanical cardiac diseases (MCD). METHODS: We prospectively analyzed 455 patients by means of phase analysis using SyncTool™ (Emory Cardiac Toolbox™). Of these, 150 corresponded to the control group (group 1, normal subjects) and 305 corresponded to patients with cardiac diseases (group 2, 63 with only CCD, 121 with only MCD, and 121 with CCD plus MCD). The optimal cut-off (CO) values of the peak phase (P), standard deviation (SD), bandwidth (B), skewness (S), and kurtosis (K) for discriminating between normal and dyssynchrony were obtained. RESULTS: In order to differentiate group 1 from group 2, CO of SD > 18.4 and CO of B > 51 were the most sensitive parameters (75.7%, 95% CI 70.5%-80.4%, and 78.7%, 95% CI 73.7%-83.1%, respectively), and CO of S ≤ 3.2 and CO of K ≤ 9.3 were the most specific (92%, 95% CI 86.4%-95.8%, and 94.7%, 95% CI 89.8%-97.7%, respectively). In order to differentiate patients with CCD and MCD, CO values were SD > 26.1, B > 70, S ≤ 2.89, and K ≤ 10.2. In order to differentiate between patients with (n: 26) and without (n: 216) criteria of cardiac resynchronization therapy, CO values were SD > 40.2, B > 132, S ≤ 2.3, and K ≤ 4.6. CONCLUSIONS: In this pilot study, different CO values of phase histogram parameters were observed between normal subjects and patients with conduction and MCD, and between patients with and without criteria of cardiac resynchronization therapy.