CARDIAN: a novel computational approach for real-time end-diastolic frame detection in intravascular ultrasound using bidirectional attention networks.

Introduction: Changes in coronary artery luminal dimensions during the cardiac cycle can impact the accurate quantification of volumetric analyses in intravascular ultrasound (IVUS) image studies. Accurate ED-frame detection is pivotal for guiding interventional decisions, optimizing therapeutic interventions, and ensuring standardized volumetric analysis in research studies. Images acquired at different phases of the cardiac cycle may also lead to inaccurate quantification of atheroma volume due to the longitudinal motion of the catheter in relation to the vessel. As IVUS images are acquired throughout the cardiac cycle, end-diastolic frames are typically identified retrospectively by human analysts to minimize motion artefacts and enable more accurate and reproducible volumetric analysis. Methods: In this paper, a novel neural network-based approach for accurate end-diastolic frame detection in IVUS sequences is proposed, trained using electrocardiogram (ECG) signals acquired synchronously during IVUS acquisition. The framework integrates dedicated motion encoders and a bidirectional attention recurrent network (BARNet) with a temporal difference encoder to extract frame-by-frame motion features corresponding to the phases of the cardiac cycle. In addition, a spatiotemporal rotation encoder is included to capture the IVUS catheter's rotational movement with respect to the coronary artery. Results: With a prediction tolerance range of 66.7 ms, the proposed approach was able to find 71.9%, 67.8%, and 69.9% of end-diastolic frames in the left anterior descending, left circumflex and right coronary arteries, respectively, when tested against ECG estimations. When the result was compared with two expert analysts' estimation, the approach achieved a superior performance. Discussion: These findings indicate that the developed methodology is accurate and fully reproducible and therefore it should be preferred over experts for end-diastolic frame detection in IVUS sequences.

Machine learning for atherosclerotic tissue component classification in combined near-infrared spectroscopy intravascular ultrasound imaging: Validation against histology.

BACKGROUND AND AIMS: Accurate classification of plaque composition is essential for treatment planning. Intravascular ultrasound (IVUS) has limited efficacy in assessing tissue types, while near-infrared spectroscopy (NIRS) provides complementary information to IVUS but lacks depth information. The aim of this study is to train and assess the efficacy of a machine learning classifier for plaque component classification that relies on IVUS echogenicity and NIRS-signal, using histology as reference standard. METHODS: Matched NIRS-IVUS and histology images from 15 cadaveric human coronary arteries were analyzed (10 vessels were used for training and 5 for testing). Fibrous/pathological intimal thickening (F-PIT), early necrotic core (ENC), late necrotic core (LNC), and calcific tissue regions-of-interest were detected on histology and superimposed onto IVUS frames. The pixel intensities of these tissue types from the training set were used to train a J48 classifier for plaque characterization (ECHO-classification). To aid differentiation of F-PIT from necrotic cores, the NIRS-signal was used to classify non-calcific pixels outside yellow-spot regions as F-PIT (ECHO-NIRS classification). The performance of ECHO and ECHO-NIRS classifications were validated against histology. RESULTS: 262 matched frames were included in the analysis (162 constituted the training set and 100 the test set). The pixel intensities of F-PIT and ENC were similar and thus these two tissues could not be differentiated by echogenicity. With ENC and LNC as a single class, ECHO-classification showed good agreement with histology for detecting calcific and F-PIT tissues but had poor efficacy for necrotic cores (recall 0.59 and precision 0.29). Similar results were found when F-PIT and ENC were treated as a single class (recall and precision for LNC 0.78 and 0.33, respectively). ECHO-NIRS classification improved necrotic core and LNC detection, resulting in an increase of the overall accuracy of both models, from 81.4% to 91.8%, and from 87.9% to 94.7%, respectively. Comparable performance of the two models was seen in the test set where the overall accuracy of ECHO-NIRS classification was 95.0% and 95.5%, respectively. CONCLUSIONS: The combination of echogenicity with NIRS-signal appears capable of overcoming limitations of echogenicity, enabling more accurate characterization of plaque components.

Detection of lipid-core plaques by intracoronary near-infrared spectroscopy identifies high risk of periprocedural myocardial infarction.

BACKGROUND: Percutaneous coronary intervention (PCI) is associated with periprocedural myocardial infarction (MI) in 3% to 15% of cases (depending on the definition used). In many cases, these MIs result from distal embolization of lipid-core plaque (LCP) constituents. Prospective identification of LCP with catheter-based near-infrared spectroscopy (NIRS) may predict an increased risk of periprocedural MI and facilitate development of preventive measures. METHODS AND RESULTS: The present study analyzed the relationship between the presence of a large LCP (detected by NIRS) and periprocedural MI. Patients with stable preprocedural cardiac biomarkers undergoing stenting were identified from the COLOR Registry, an ongoing prospective observational study of patients undergoing NIRS before PCI. The extent of LCP in the treatment zone was calculated as the maximal lipid-core burden index (LCBI) measured by NIRS for each of the 4-mm longitudinal segments in the treatment zone. A periprocedural MI was defined as new cardiac biomarker elevation above 3× upper limit of normal. A total of 62 patients undergoing stenting met eligibility criteria. A large LCP (defined as a maxLCBI(4 mm) ≥500) was present in 14 of 62 lesions (22.6%), and periprocedural MI was documented in 9 of 62 (14.5%) of cases. Periprocedural MI occurred in 7 of 14 patients (50%) with a maxLCBI(4 mm) ≥500, compared with 2 of 48 patients (4.2%) patients with a lower maxLCBI(4 mm) (P=0.0002). CONCLUSIONS: NIRS provides rapid, automated detection of extensive LCPs that are associated with a high risk of periprocedural MI, presumably due to embolization of plaque contents during coronary intervention.

In vivo validation of a catheter-based near-infrared spectroscopy system for detection of lipid core coronary plaques: initial results of the SPECTACL study.

OBJECTIVES: To determine whether catheter-based near-infrared spectroscopy (NIRS) signals obtained with a novel catheter-based system from coronaries of patients are similar to those from autopsy specimens and to assess initial safety of NIRS device. BACKGROUND: An intravascular NIRS system for detection of lipid core-containing plaques (LCP) has been validated in human coronary autopsy specimens. The SPECTACL (SPECTroscopic Assessment of Coronary Lipid) trial was a parallel first-in-human multicenter study designed to demonstrate the applicability of the LCP detection algorithm in living patients. METHODS: Intracoronary NIRS was performed in patients undergoing percutaneous coronary intervention. Acquired spectra were blindly compared with autopsy NIRS signals with multivariate statistics. To meet the end point of spectral similarity, at least two-thirds of the scans were required to have >80% of spectra similar to the autopsy spectra. RESULTS: A total of 106 patients were enrolled; there were no serious adverse events attributed to NIRS. Spectroscopic data could not be obtained in 17 (16%) patients due to technical limitations, leaving 89 patients for analysis. Spectra from 30 patients were unblinded to test the calibration of the LCP detection algorithm. Of the remaining 59 blinded cases, after excluding 11 due to inadequate data, spectral similarity was demonstrated in 40 of 48 spectrally adequate scans (83% success rate, 95% confidence interval: 70% to 93%, median spectral similarity/pullback: 96%, interquartile range 10%). The LCP was detected in 58% of 60 spectrally similar scans from both cohorts. CONCLUSIONS: This intravascular NIRS system safely obtained spectral data in patients that were similar to those from autopsy specimens. These results demonstrate the feasibility of invasive detection of coronary LCP with this novel system. (SPECTACL: SPECTroscopic Assessment of Coronary Lipid; NCT00330928).

Increased radiation dose to overweight and obese patients from radiographic examinations.

PURPOSE: To estimate the increase in effective radiation dose from diagnostic x-rays for overweight and obese adult patients, as compared with the effective dose for lean reference phantoms. MATERIALS AND METHODS: Relative effective radiation doses (E/E(0)) for the acquisition of chest and abdominal radiographs were calculated by using Monte Carlo computer simulations of effective doses delivered to adult phantoms with (E) and without (E(0)) subcutaneous adipose tissue added to the torso for five fat distributions. Total (anterior plus posterior) fat thicknesses ranged from 0 to 38 cm. RESULTS: For 30 cm of additional fat, E/E(0) values for 120-kVp chest and 80-kVp abdomen radiographs ranged from approximately 2 to 31 and 2 to 83 for male patients, respectively, and from 2 to 45 and 2 to 76 for female patients, respectively, depending on the type of fat distribution and patient orientation in the x-ray beam (anteroposterior or posteroanterior). Orienting the patient such that the thinnest fat layer was facing away from the x-ray tube minimized E/E(0), which was well approximated by using the formula E/E(0) = [B(t)/B(0)] x exp(kt(DF)), where B(t) and B(0) are the antiscatter grid Bucky factors for patient thicknesses of t and t = 20 cm, respectively; k, a constant; and t(DF), the distal (beam exit) fat layer thickness. Reductions in E/E(0) reached 14% and 20% for the thickest phantoms when x-ray tube voltages were increased by 10 and 20 kVp, respectively, for abdominal radiography in the male phantom. CONCLUSION: Effective doses from radiographic examinations in the extremely obese can exceed 100 mSv from only a small number of abdominal examinations and should be minimized to the extent possible and monitored. Exponential dose increases for increased subcutaneous fat thicknesses can be reduced substantially by positioning the patient so that the thinnest fat layer (anterior or posterior) is closest to the image receptor. Increasing the tube voltage also reduces the dose-but to a much smaller extent.

Sustained involvement of a frontal-parietal network for spatial response selection with practice of a spatial choice-reaction task.

With practice, performance on a task typically becomes faster, more accurate, and less prone to interference from competing tasks. Some theories of this performance change suggest it reflects a qualitative reorganization of the cognitive processing required for successful task performance. Other theories suggest this change in performance reflects a more quantitative change in the amount of processing required to perform the task. Neuroimaging research results provide some support for both of these broad theories. This inconsistency may reflect the complex nature of the effect of practice on cognitive and neural processing. Our current experiment addresses this issue by investigating the effect of practice of a relatively easy perceptual-motor task on the frontal-parietal brain network for a specific cognitive process (viz. spatial response selection). Participants were scanned during three functional magnetic resonance imaging sessions on separate days within 4 days while they performed a relatively easy spatial perceptual-motor task. We found sustained activity with practice in right dorsal prefrontal cortex; and sustained but decreasing activity in bilateral dorsal premotor, left superior parietal, and precuneus cortices, supporting a quantitative decrease in difficulty of response selection with practice. Conversely, we found a qualitative change in activity with practice in left dorsal prefrontal cortex. This brain region is outside the response selection network for this task and showed activity only during novel task performance. These results suggest that practice produces both qualitative and quantitative changes in processing. The particular effect of practice depends on the cognitive process in question.