Development and Validation of Artificial Intelligence-Based Algorithms for Predicting the Segments Debulked by Rotational Atherectomy Using Intravascular Ultrasound Images.

We develop and evaluate an artificial intelligence (AI)-based algorithm that uses pre-rotation atherectomy (RA) intravascular ultrasound (IVUS) images to automatically predict regions debulked by RA. A total of 2106 IVUS cross-sections from 60 patients with de novo severely calcified coronary lesions who underwent IVUS-guided RA were consecutively collected. The 2 identical IVUS images of pre- and post-RA were merged, and the orientations of the debulked segments identified in the merged images were marked on the outer circle of each IVUS image. The AI model was developed based on ResNet (deep residual learning for image recognition). The architecture connected 36 fully connected layers, each corresponding to 1 of the 36 orientations segmented every 10°, to a single feature extractor. In each cross-sectional analysis, our AI model achieved an average sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of 81%, 72%, 46%, 90%, and 75%, respectively. In conclusion, the AI-based algorithm can use information from pre-RA IVUS images to accurately predict regions debulked by RA and will assist interventional cardiologists in determining the treatment strategies for severely calcified coronary lesions.

Isomerization and epimerization of fructose in phosphate buffer under subcritical water conditions.

Isomerization and epimerization of fructose to glucose, mannose, allulose, and allose were executed using a subcritical phosphate buffer solution to effectively produce useful monosaccharides. The conversion of the substrate and the yield of products were dependent on the reaction temperature, initial pH, initial substrate concentration, and buffer concentration. A high yield of mannose was achieved under the optimal reaction conditions we identified. We subsequently performed the kinetic analysis based on the proposed reaction network, and evaluated the effects of temperature and pH on the reactions. We then estimated the apparent activation energy values for each reaction.

Frequency and Distribution of Sheet and Nodular Calcification in Coronary Arteries in Japanese Patients.

Whether a nodular calcification (NC), which is the precursor to intracoronary thrombosis, is focally or diffusely distributed in the coronary tree has major implications for ongoing efforts to identify. This study aimed to investigate the frequency and spatial distribution patterns of sheet calcification (SC) and NC in a 3-vessel examination of autopsied human hearts.A total of 323 coronary artery specimens from 110 cadavers were obtained from autopsy cases. After fixation and decalcification, the coronary artery trees were cut every 5 mm into 4-µm transverse cross-sections for histological assessment. An SC was defined as a plate-like calcification of > 1 quadrant of the vessel or > 3 mm in diameter, and NC as nodular calcium deposits separated by fibrin, and a deposit size > 1 mm in diameter.Of the 6,306 histological cross-sections, SCs and NCs were identified in 1,627 (26%) and 233 (4%) cross-sections, respectively. SCs and NCs had a similar distribution pattern in all 3 coronary arteries. In the left anterior descending artery (LAD), NCs were predominantly located in the proximal segment: the first 45 mm from the LAD ostium (72%) and the first 60 mm from the LAD ostium (84%), respectively. However, NCs were evenly distributed throughout the length of the coronary artery in the right coronary artery (RCA) and left circumflex artery (LCX).NCs coexisted with SCs, and tended to cluster in predictable parts within the proximal segments of the LAD, but were evenly distributed throughout the RCA and LCX in coronary arteries from cadavers.

Prediction of medial damage by rotational atherectomy using intravascular ultrasound.

OBJECTIVE: This study evaluated whether intravascular ultrasound (IVUS) images before rotational atherectomy (RA) can predict medial injury caused by RA burr passage in advance. METHODS: Thirty-three patients with de-novo lesions located in the left anterior descending coronary artery (LAD) who underwent IVUS before and immediately after RA were enrolled. The spatial axes of the two matched pre- and post-RA IVUS cross-sectional images were merged. The segment was considered to have medial damage when the continuity of the intimal layer was disrupted and the medial layer was in direct contact with the lumen on post-RA IVUS. RESULTS: Medial injuries on post-RA IVUS were identified in seven segments of five patients. All segments with medial injury were located near the bifurcation of the LAD and the diagonal branch, and the spatial orientations of the medial injury region were mostly distributed in the lateral side with a diagonal branch take-off. The lumen area was significantly smaller in segments with medial injury than in those without medial injury (P < 0.01). The IVUS catheter was in contact with the healthy side of the arterial wall on pre-RA IVUS images for more than 1 mm in length in all segments with medial injury. CONCLUSION: When the guidewire and IVUS catheter are close to the healthy side of the arterial wall on pre-RA IVUS images, there is a higher risk of medial injury due to the RA procedure, especially near the bifurcation of the LAD and diagonal branch.

Diagnostic Accuracy of Optical Frequency Domain Imaging for Identifying Necrotic Cores with Intraplaque Hemorrhage in Advanced Human Carotid Plaques.

This study investigated whether optical frequency domain imaging (OFDI) can identify carotid artery vulnerable plaque characteristics, focusing on lipid-rich necrotic core (NC) and intraplaque hemorrhage (IPH). Fourteen patients scheduled for carotid endarterectomy underwent OFDI scan during preoperative angiography. Atherosclerotic plaque specimens obtained from carotid endarterectomy were cut every 3-4 mm into 4-µm transverse cross-sections and stained with standard methods. Each cross-section was matched with OFDI, and histologically classified into either fibrous, calcific, pathological intimal thickening (PIT), and NC. Of 75 histologic cross-sections, 6 were categorized as fibrous (8%), 18 as calcific (24%), 9 as PIT (12%), and 42 as NC (56%). Tissues categorized as NC had significantly higher OFDI signal attenuation rates than the other tissues (p <0.001), followed by PIT, calcific, and fibrous tissues. The receiver operating characteristic analysis indicated that attenuation rates of >0.023 and >0.031 predicted the presence of NC and IPH with high areas under the curve of 0.91 and 0.88, respectively. OFDI provides potential capability for the detection of NCs with IPH of carotid artery plaques by quantitatively analyzing the attenuation rate.

Automated classification of coronary atherosclerotic plaque in optical frequency domain imaging based on deep learning.

BACKGROUND AND AIMS: We developed a deep learning (DL) model for automated atherosclerotic plaque categorization using optical frequency domain imaging (OFDI) and performed quantitative and visual evaluations. METHODS: A total of 1103 histological cross-sections from 45 autopsy hearts were examined to compare the ex vivo OFDI scans. The images were segmented and annotated considering four histological categories: pathological intimal thickening (PIT), fibrous cap atheroma (FA), fibrocalcific plaque (FC), and healed erosion/rupture (HER). The DL model was developed based on pyramid scene parsing network (PSPNet). Given an input image, a convolutional neural network (ResNet50) was used as an encoder to generate feature maps of the last convolutional layer. RESULTS: For the quantitative evaluation, the mean F-score and IoU values, which are used to evaluate how close the predicted results are to the ground truth, were used. The validation and test dataset had F-score and IoU values of 0.63, 0.49, and 0.66, 0.52, respectively. For the section-level diagnostic accuracy, the areas under the receiver-operating characteristic curve produced by the DL model for FC, PIT, FA, and HER were 0.91, 0.85, 0.86, and 0.86, respectively, and were comparable to those of an expert observer. CONCLUSIONS: DL semantic segmentation of coronary plaques in OFDI images was used as a tool to automatically categorize atherosclerotic plaques using histological findings as the gold standard. The proposed method can support interventional cardiologists in understanding histological properties of plaques.

Additional decrease in fractional flow reserve in the left anterior descending artery after hand-grip exercise during pharmacological coronary hyperemia.

Maximal hyperemia at the time of fractional flow reserve (FFR) measurement is generally induced by vasodilators, even when hyperemia at the onset of angina symptoms is caused by exercise stress. This study was designed to evaluate whether pharmacological hyperemia could be used as a substitute for exercise-induced hyperemia during FFR measurement. Twenty-two patients with angiographically intermediate stenosis in the left anterior descending artery (LAD) were prospectively enrolled. FFR measurements were repeated in the following two conditions while the pressure-wire was positioned in the same segment; (1) during pharmacological hyperemia induced by intracoronary administration of 2 mg nicorandil, (2) immediately after isotonic hand-grip exercise for 90 s (50% of maximum voluntary contraction) followed by intracoronary administration of 2 mg nicorandil. Isotonic hand-grip exercise increased systolic blood pressure (130 ± 19 versus 150 ± 22 mmHg, p < 0.001), heart rate (71 ± 11 versus 79 ± 13 bpm, p < 0.001), and cardiac output (5.1 ± 1.2 versus 5.9 ± 1.5 L/min, p < 0.001), which indicated an increased afterload on the left ventricle. After the hand-grip exercise, FFR significantly decreased from 0.86 ± 0.06 to 0.84 ± 0.06 (p < 0.001). A percent increase in systolic blood pressure and cardiac output after hand-grip exercise strongly correlated with ΔFFR (r = - 0.65, p < 0.001 and r = - 0.55, p < 0.001, respectively). An increase in cardiac output with hand-grip exercise during pharmacological hyperemia could induce an additional decrease in FFR for lesions located in the LAD.

The incidence, natural history, and predictive factors for tissue protrusion after drug-eluting stent implantation.

OBJECTIVES AND BACKGROUND: Although tissue protrusion (TP) between the stent struts after stent implantation has been implicate as a potential factor of stent failure, the incidence, natural history, and predictive factor of TP after stent implantation remains unclear. This prospective study evaluated the fate of TP after drug-eluting stent (DES) deployment using optical coherence tomography (OCT). METHOD AND RESULT: This study analyzed TP for 42 lesions after DES in which three serial OCTs, including preprocedure, postprocedure, and 1-month after the procedure were performed. TP was classified into the five groups: (a) persistent, (b) progressive, (c) healed, (d) regressive, and (e) late-acquired. Immediately after the procedure, 100 TPs in 37 lesions (88%) were identified. Of those, 53 (53%) were persistent, 3 (3%) were progressive, 20 (20%) were healed, and 24 (24%) were regressed at 1-month follow-up. Seven TPs in five patients (13%) were observed only at 1-month follow-up (late-acquired). CONCLUSION: In lesions with late-acquired TP, calcified nodule was identified as an underlying plaque morphology on preprocedural OCT. A serial OCT analysis found TP occurred not only immediately after DES implantation, but also 1-month after DES implantation.

Influence of different physiological hemodynamics on fractional flow reserve values in the left coronary artery and right coronary artery.

BACKGROUND: Although the left coronary artery (LCA) has a flow profile in that most blood flow occurs during diastole rather than systole, the right coronary artery (RCA) has a flow pattern that is less diastolic dominant. This study assessed whether coronary pressure waveforms distal to stenoses with the same fractional flow reserve (FFR) was the same between the LCA and RCA. METHODS: A total of 347 vessels from 318 patients who underwent FFR measurements were included. Conventional FFR was calculated as the ratio of the mean coronary distal pressure (Pd) to the mean aortic pressure (Pa) at maximal hyperemia. The pressure drop ratios in systole (PDRsystole) and diastole (PDRdiastole) were calculated as the sum of (Pa minus Pd) divided by the sum of Pa at the intracoronary diastolic and systolic pressure phases, respectively. RESULTS: Analysis of covariance of the regression line of correlation between conventional FFR and PDRsystole revealed that the slope was significantly greater in the RCA than in the left anterior descending artery (LAD) and left circumflex artery (LCX) (-0.765, -0.578, and -0.589, p < 0.001). On the other hand, the regression line of correlation between conventional FFR and PDRdiastole found that the slope was significantly greater in the LAD and LCX than in the RCA (-1.349, -1.318, and -1.223, p < 0.001). CONCLUSIONS: The pressure waveform distal to the stenosis differs between the LCA and RCA. In the LCA, the decrease in diastolic pressure mainly contributed to the drop in FFR, whereas in the RCA, it was the decrease in systolic pressure.

Diagnostic accuracy of optical coherence tomography for the identification of in-stent fibroatheroma following stent implantation: an ex vivo histological validation study.

The accurate identification of in-stent fibroatheroma by in vivo imaging is clinically important to preventing the late catch-up phenomenon after stent deployment. This study investigated the diagnostic accuracy of optical coherence tomography (OCT) for the detection of "in-stent fibroatheroma" following stent implantation. Fifty stented coronary arteries from the 31 autopsy hearts were examined to compare OCT and histological image findings. A histological in-stent fibroatheroma was defined as a neointima containing an acellular necrotic core generated by macrophage infiltration. OCT-derived in-stent fibroatheroma comprised a heterogeneous pattern with an invisible stent strut behind the low-signal-intensity region. A total of 122 matched OCT and histology cross-sections were evaluated. Using histological findings as the gold standard, the sensitivity, specificity, positive predictive value, and negative predictive value for OCT-derived in-stent fibroatheroma were 100%, 99%, 80%, and 100%, respectively. The only histological finding underlying the false-positive diagnosis of OCT-derived in-stent fibroatheroma was foam cell accumulation without a necrotic core on the neointimal surface. No false-negative diagnosis of OCT for in-stent fibroatheroma was apparent in this analysis. This study demonstrated the potential capability of OCT based on stent strut visualization behind low-signal-intensity regions to discriminate in-stent fibroatheroma from other neointimal tissues.

Prediction of optimal debulking segments before rotational atherectomy based on pre-procedural intravascular ultrasound findings.

This study evaluated whether intravascular ultrasound (IVUS) examination before rotational atherectomy (RA) can predict the optimal route of passage of the RA burr along the vessel. 30 patients with calcified lesions who underwent IVUS before and immediately after RA were enrolled. IVUS analyses were performed at the minimum lumen area (MLA) site and at 0.5 mm intervals. Each IVUS cross-section was divided into 4 quadrants around the center of the lumen, and pre- and post-RA IVUS cross-section images were merged. Of 1140 cross-sections, 498 (44%) contained debulked regions. When the guidewire and IVUS were located within the same quadrant, the debulked region were distributed within the same quadrant in 96% of cross-sections. The debulked region and the guidewire were distributed within the same quadrant in 81% and the debulked region and the IVUS in 72% of cross-sections, in case the guidewire and IVUS were located in different quadrants. When the guidewire and the IVUS was apart > 1.0 mm, the debulked regions were distributed within the same quadrant as the guidewire in 100% and the IVUS in 0% of cross-sections. The position of the guidewire rather than that of the IVUS catheter on pre-RA IVUS images could predict the course of the RA burr's passage, especially when the guidewire and IVUS catheter were located apart from each other.

Tangential signal dropout artefact in optical frequency domain imaging.

BACKGROUND: Tangential signal dropout (TSD), which occurs when the optical frequency domain imaging (OFDI) beam strikes the vessel wall under a glancing angle and travels almost parallel to the vessel wall, is the most important imaging artefact leading to the erroneous diagnosis of lipid-rich plaques. AIMS: This study aimed to evaluate the OFDI artefact of TSD, which mimics the appearance of lipid-rich plaque and macrophage (Mø) infiltration. METHODS: A total of 1,019 histological cross-sections from 23 autopsy hearts were matched with the corresponding OFDI images. Of these, 232 OFDI cross-sections that contained signal-poor regions with diffuse borders were classified as lipid-rich plaques. The angle θ was calculated between the OFDI beam that strikes the edge of the luminal surface of the low-intensity region and that which strikes the surface line of the low-intensity region. RESULTS: On histological evaluation, 182 (78%) cross-sections were classified as histologically lipidic/Mø infiltration, while the remaining 50 (22%) cross-sections were classified as histologically non-lipidic/Mø infiltration. The angle θ was significantly smaller in the non-lipidic/Mø infiltration group than in the lipidic/Mø infiltration group (12±6° versus 37±14°, p<0.001). Receiver operating curve analysis revealed that the optimal cut-off value of the incident angle for predicting TSD was 23° with an area under the curve of 0.98. CONCLUSIONS: When the OFDI imaging beam strikes the tissue at an angle θ<23°, TSD artefact could occur. To eliminate image misinterpretation, our findings suggest that the OFDI catheter geometry should be considered for the accurate diagnosis of lipid-rich plaques and Mø infiltration.

Interobserver variability in assessments of atherosclerotic lesion type via optical frequency domain imaging.

BACKGROUND: To date, there have been no data available regarding the diagnostic performance of optical frequency domain imaging (OFDI) for in vivo histological classification of atherosclerotic lesions. This study investigated whether OFDI can be used to diagnose and classify histological atherosclerotic lesions in the coronary artery by ex vivo histological examinations. METHODS: Three-hundred-fifteen histological cross-sections from 21 autopsy hearts were matched with the OFDI images. Histological cross-sections were classified into six categories: adaptive intimal thickening (AIT), pathological intimal thickening (PIT), fibrous cap atheroma (FA), fibrocalcific plaque (FC), calcified nodule, and healed erosion/rupture. The five observers with different years of experience in the interpretation of OFDI provided a single diagnosis for the OFDI scans of each cross-section according to the aforementioned six histological categories. The diagnostic accuracy and interobserver variability of lesion types for each OFDI observer were determined using histology as the gold standard. RESULTS: The overall agreement rates between OFDI and histopathologic diagnosis for OFDI observers 1-5 were 81%, 70%, 68%, 61%, and 50% (κ values of 0.75, 0.61, 0.58, 0.49, and 0.36), respectively. Although the diagnostic accuracy of OFDI for detecting AIT and FC was excellent for all five observers, the sensitivity, and positive predictive values of OFDI for detecting PIT and FA were low in proportion to years of experience. CONCLUSION: The diagnostic accuracy of atherosclerotic tissue properties from OFDI scans correlated with the observers' years of experience, especially when lesions contained lipid components.

Prognostic Value of Psoas Muscle Mass Index in Patients With Non‒ST-Segment‒Elevation Myocardial Infarction: A Prospective Observational Study.

Background Muscle wasting is an important predictor of long-term outcome in patients with cardiovascular disease, but the prognostic value of muscle wasting in patients with non‒ST-segment‒elevation myocardial infarction is not established. The aim of this study is to investigate the prognostic value of muscle wasting, defined by psoas muscle mass index (PMI), in patients with non‒ST-segment‒elevation myocardial infarction. Methods and Results A total of 132 consecutive patients with non‒ST-segment‒elevation myocardial infarction were prospectively enrolled between 2015 and 2018. Primary end point was incidence of cardiovascular events including cardiovascular deaths, non-fatal myocardial infarction, or non-fatal stroke. Cross-sectional area of the psoas muscle at the L3 vertebral level was obtained by computed tomography and PMI was calculated. The median follow-up period was 2.4 years (interquartile range, 1.1-4.0 years). There were 45 cardiovascular events (34%) during the study periods. The optimal cutoff value of PMI to predict cardiovascular events was 772 mm2/m2, as assessed by receiver operating curve analysis. Patients with reduced PMI (PMI<772 mm2/m2) had significantly higher cardiovascular events than those with preserved PMI (PMI≥772 mm2/m2) (48% versus 21%; log-rank test P<0.001). Multivariate Cox proportional hazards model revealed that reduced PMI was a statistically significant predictor of cardiovascular events (hazard ratio, 3.30; 95% CI, 1.70-6.40; P<0.001). Conclusions Muscle wasting defined as PMI is a simple and useful objective marker to predict future cardiovascular outcome in patients with non‒ST-segment‒elevation myocardial infarction. Registration Information URL: https://www.umin.ac.jp/ctr/; Unique identifier: UMIN000013445.

Loss of skeletal muscle mass predicts cardiac death in heart failure with a preserved ejection fraction but not heart failure with a reduced ejection fraction.

AIMS: Loss of skeletal muscle mass is an important determinant associated with poor long-term prognosis in patients with acute decompensated heart failure (ADHF). However, limited evidence is available. This study investigated the prognostic value of the psoas muscle mass index (PMI) in patients with ADHF. METHODS AND RESULTS: A total of 210 consecutive patients aged ≥60 years with ADHF were enrolled using a prospective database between 2015 and 2017. Primary endpoint was incidence of cardiac death. Cross-sectional psoas muscle area at the L3 vertebral level was obtained by computed tomography, and PMI was calculated by height. Reduced PMI was defined as a PMI below the 25th sex-specific percentile. Patients were also classified by their left ventricular ejection fraction (EF) as having either heart failure with a reduced ejection fraction (HFrEF, EF < 50%) or heart failure with a preserved ejection fraction (HFpEF, EF ≥ 50%). The median follow-up period was 1.8 years. There were 44 cardiac deaths (21%) during the study period. Patients with reduced PMI had significantly higher cardiac death rates than those with preserved PMI (33% vs. 17%, log-rank test P = 0.006). In subgroup analysis, HFpEF patients with reduced PMI had significantly higher cardiac death rates than those with preserved PMI (38% vs. 16%, log-rank test P = 0.006); conversely, HFrEF patients had comparable cardiac death rates regardless of their PMI group (27% for reduced PMI vs. 18% for preserved PMI, log-rank test P = 0.24). Multivariate Cox proportional hazards model revealed that patients with reduced PMI had a 2.3-fold higher risk of cardiac death compared with patients with preserved PMI (95% confidence interval 1.23-4.42, P = 0.01). CONCLUSIONS: Reduced PMI helps to predict long-term outcome in patients with HFpEF but not HFrEF.

Coronary artery calcification as a novel predictive marker of unstable coronary lesion in survivors of out-of-hospital cardiac arrest without ST-segment elevation.

AIM: Acute myocardial infarction (AMI) is the leading cause of out-of-hospital cardiac arrest (OHCA). A highly predictive marker is needed to identify AMI in survivors of OHCA without ST-segment elevation because the appropriate indication for emergency coronary artery angiography in patients without ST-segment segment elevation has not been determined. Accordingly, the aim of this study was to elucidate the clinical significance of coronary artery calcification in identifying survivors of OHCA without ST-segment elevation who could benefit from emergency coronary artery angiography. METHODS: Survivors of OHCA without ST-segment elevation with no obvious extra-cardiac cause who underwent emergency computed tomography and coronary artery angiography were enrolled. Unstable coronary lesion was diagnosed using coronary artery angiography, and presence of coronary artery calcification and coronary artery calcium score were evaluated by non-contrast, non-electrocardiography gated computed tomography. RESULTS: Thirty of 100 consecutive survivors of OHCA were diagnosed to have unstable coronary lesion. Sensitivity and specificity of coronary artery calcification in identifying unstable coronary lesion were 87% and 60%, respectively. Multivariate logistic regression analysis revealed that coronary artery calcification was an independent predictor of unstable coronary lesion (odds ratio: 7.28, 95% confidence interval: 2.00-26.56, p < 0.001). CONCLUSION: Evaluation of coronary artery calcification by computed tomography is useful in identifying patients with unstable coronary lesion who could benefit from emergency coronary artery angiography among survivors of OHCA without ST-segment elevation on post-resuscitation electrocardiography.