Deep learning to predict long-term mortality from plain chest X-ray in patients referred for suspected coronary artery disease.

Background: The hypothesis that a deep learning (DL) model can produce long-term prognostic information from chest X-ray (CXR) has already been confirmed within cancer screening programs. We summarize our experience with DL prediction of long-term mortality, from plain CXR, in patients referred for angina and coronary angiography. Methods: Data of patients referred to an Italian academic hospital were analyzed retrospectively. We designed a deep convolutional neural network (DCNN) that, from CXR, could predict long-term mortality. External validation was performed on patients referred to a Dutch academic hospital. Results: A total of 6,031 were used for model training (71%; n=4,259) and fine-tuning/validation (10%; n=602). Internal validation was performed with the remaining patients (19%; n=1,170). Patients' stratification followed the DL-CXR risk score quartiles division. Median follow-up was 6.1 years [interquartile range (IQR), 3.3-8.7 years]. We observed an increment in estimated mortality with the increase of DL-CXR risk score (low-risk 5%, moderate 17%, high 29%, very high 46%; P<0.001). The DL-CXR risk score predicted median follow-up outcome with an area under the curve (AUC) of 0.793 [95% confidence interval (CI): 0.759-0.827, sensitivity 78%, specificity 68%]. Prediction was better than that achieved using coronary angiography findings (AUC: 0.569, 95% CI: 0.52-0.61, P<0.001) and age (AUC: 0.735, 95% CI: 0.69-0.77, P<0.004). At Cox regression, the DL-CXR risk score predicted follow-up mortality (P<0.005, hazard ratio: 3.30, 95% CI: 2.35-4.64). External validation confirmed the DL-CXR risk score performance (AUC: 0.71, 95% CI: 0.49-0.92; sensitivity 0.838; specificity 0.338). Conclusions: In patients referred for coronary angiogram because of angina, the DL-CXR risk score could be used to stratify mortality risk and predict long-term outcome better than age and coronary artery disease status.

Sutureless Aortic Valve Replacement vs. Transcatheter Aortic Valve Implantation in Patients with Small Aortic Annulus: Clinical and Hemodynamic Outcomes from a Multi-Institutional Study.

OBJECTIVE: This study aimed to compare hemodynamic performances and clinical outcomes of patients with small aortic annulus (SAA) who underwent aortic valve replacement by means of sutureless aortic valve replacement (SUAVR) or transcatheter aortic valve implantation (TAVI). METHODS: From 2015 to 2020, 622 consecutive patients with SAA underwent either SUAVR or TAVI. Through a 1:1 propensity score matching analysis, two homogeneous groups of 146 patients were formed. Primary endpoint: all cause-death at 36 months. Secondary endpoints: incidence of moderate to severe patient-prosthesis mismatch (PPM) and incidence of major adverse cardiovascular and cerebrovascular events (MACCEs). RESULTS: All-cause death at three years was higher in the TAVI group (SUAVR 12.2% vs. TAVI 21.0%, P=0.058). Perioperatively, comparable hemodynamic performances were recorded in terms of indexed effective orifice area (SUAVR 1.12 ± 0.23 cm2/m2 vs. TAVI 1.17 ± 0.28 cm2/m2, P=0.265), mean transvalvular gradients (SUAVR 12.9 ± 5.3 mmHg vs. TAVI 12.2 ± 6.2 mmHg, P=0.332), and moderate-to-severe PPM (SUAVR 4.1% vs. TAVI 8.9%, P=0.096). TAVI group showed a higher cumulative incidence of MACCEs at 36 months (SUAVR 18.1% vs. TAVI 32.6%, P<0.001). Pacemaker implantation (PMI) and perivalvular leak ≥ 2 were significantly higher in TAVI group and identified as independent predictors of mortality (PMI: hazard ratio [HR] 3.05, 95% confidence interval [CI] 1.34-6.94, P=0.008; PPM: HR 2.72, 95% CI 1.25-5.94, P=0.012). CONCLUSION: In patients with SAA, SUAVR and TAVI showed comparable hemodynamic performances. Moreover, all-cause death and incidence of MACCEs at follow-up were significantly higher in TAVI group.

Evaluation of the Nautilus Smart Extracorporeal Membrane Oxygenation in Patients With Hemostatic Alteration: A Case Series.

The intricate management of hemostatic disorders in extracorporeal membrane oxygenation (ECMO) assisted patients poses challenges, particularly when procoagulant administration is necessary. We hereby report the performance of the Nautilus* Smart ECMO Module in three patients with hemostatic disorders. We collected data from ECMO procedures with Nautilus* Smart ECMO Module and analyzed the performance: the operating pressures and resistance of the device in addition to the coagulation status of the patients. During the three procedures, partial pressure oxygen post-oxygenator (paO2) stayed above 100 mm Hg and partial pressure carbon dioxide post-oxygenator (paCO2) did not exceed 45 mm Hg. Membrane FiO2 (fractional inspired O2) did not exceed 75% and air flow remained within a 1:1 ratio with blood flow in veno-arterial ECMO (V-A) and within 1:2 in veno-venous ECMO (V-V). There was no evidence of excessive operating pressure for the device, with a pressure drop consistently below 28 mm Hg and a maximum peak resistance of 7 Δmm Hg/L/min. The Nautilus* Smart ECMO Module showed good performance in patients with hemostatic disorders despite the implications associated with procoagulant administration.

Efficacy and Safety of Botulinum Toxin Type A for the Prevention of Postoperative Atrial Fibrillation.

BACKGROUND: Postoperative atrial fibrillation (POAF) is associated with increased morbidity and mortality. Epicardial injection of botulinum toxin may suppress POAF. OBJECTIVES: This study sought to assess the safety and efficacy of AGN-151607 for the prevention of POAF after cardiac surgery. METHODS: This phase 2, randomized, placebo-controlled trial assessed the safety and efficacy of AGN-151607, 125 U and 250 U vs placebo (1:1:1), for the prevention of POAF after cardiac surgery. Randomization was stratified by age (<65, ≥65 years) and type of surgery (nonvalvular/valve surgery). The primary endpoint was the occurrence of continuous AF ≥30 seconds. RESULTS: Among 312 modified intention-to-treat participants (placebo, n = 102; 125 U, n = 104; and 250 U, n = 106), the mean age was 66.9 ± 6.8 years; 17% were female; and 64% had coronary artery bypass graft (CABG) only, 12% had CABG + valve, and 24% had valve surgery. The primary endpoint occurred in 46.1% of the placebo group, 36.5% of the 125-U group (relative risk [RR] vs placebo: 0.80; 95% CI: 0.58-1.10; P = 0.16), and 47.2% of the 250-U group (RR vs placebo: 1.04; 95% CI: 0.79-1.37; P = 0.78). The primary endpoint was reduced in the 125-U group in those ≥65 years of age (RR: 0.64; 95% CI: 0.43-0.94; P = 0.02) with a greater reduction in CABG-only participants ≥65 years of age (RR: 0.49; 95% CI: 0.27-0.87; P = 0.01). Rehospitalization and rates of adverse events were similar across the 3 groups. CONCLUSIONS: There were no significant differences in the rate of POAF with either dose compared with placebo; however, there was a lower rate of POAF in participants ≥65 years undergoing CABG only and receiving 125 U of AGN-151607. These hypothesis-generating findings require investigation in a larger, adequately powered randomized clinical trial. (Botulinum Toxin Type A [AGN-151607] for the Prevention of Post-operative Atrial Fibrillation in Adult Participants Undergoing Open-chest Cardiac Surgery [NOVA]; NCT03779841); A Phase 2, Multi-Center, Randomized, Double-Blind, Placebo-Controlled, Dose Ranging Study to Evaluate the Efficacy and Safety of Botulinum Toxin Type A [AGN 151607] Injections into the Epicardial Fat Pads to Prevent Post-Operative Atrial Fibrillation in Patients Undergoing Open-Chest Cardiac Surgery; 2017-004399-68).

Surgical options for atrial fibrillation treatment during concomitant cardiac procedures.

Current guidelines recommend concomitant surgical ablation (SA) of atrial fibrillation (AF) in the context of mitral valve disease. A variety of energy sources have been tested for SA to perform effective transmural lesions reliably. To date, only radiofrequency and cryothermy energies are considered viable options. The gold standard for SA is the Cox-Maze ablation set, especially for non-paroxysmal AF (nPAF), with the aim of interrupting macro-reentrant drivers perpetuating AF, without hampering the sinus node activation of both atria, and to maintain the atrioventricular synchrony. Although the efficacy of SA in terms of early and late sinus rhythm restoration has been clearly demonstrated over the years, concomitant AF ablation is still underperformed in patients with AF undergoing cardiac surgery. From a surgical standpoint, concerns have been raised about whether a single (left) or double atriotomy would be justified in AF patients undergoing a "non-atriotomy" surgical procedure, such as aortic valve or revascularization surgery. Thus, an array of simplified lesion sets have been described in the last decade, which have unavoidably hampered procedural efficacy, somewhat jeopardizing the standardization process of ablation surgery. As a matter of fact, the term "Maze" has improperly become a generic term for SA. Surgical interventions that do not align with the principles of forming conduction-blocking lesions according to the Maze pattern, cannot be classified as Maze procedures. In this complex scenario, a tailored approach according to the different AF patterns has been proposed: for patients with concomitant nPAF, a biatrial Cox-Maze ablation is recommended. Conversely, it might be reasonable to limit lesions to the left atrium or the pulmonary veins in patients with paroxysmal AF (PAF) in some clinical scenarios. The aim of this review is to provide an overview of the current ablation strategies for patients with AF undergoing concomitant cardiac surgery.

Impact of coronary-subclavian steal after surgical myocardial revascularization with internal thoracic artery in chronic hemodialysis patients: A meta-analysis.

Patients in hemodialysis with an arm arteriovenous fistula undergoing coronary artery bypass grafting (CABG) with an internal thoracic artery have been reported to suffer from coronary-subclavian steal (CSS) during dialysis session. However, its occurrence is still debated. A systematic literature review was performed to identify all studies investigating the occurrence of a CSS event in this subset of patients. The primary endpoint was the analysis of CSS and the following early and late survival outcomes. Independent determinants of CSS and the impact of the distance between the arteriovenous fistula (upper arm vs forearm) and the ipsilateral internal thoracic artery graft on CSS events and mortality were studied. Early and late survival outcomes were analyzed by comparing ipsilateral versus contralateral arteriovenous fistula. Of the 1,383 retrieved articles, 10 were included (n = 643 patients). The pooled event rate of CSS was 6.46% [95%CI=2.10-18.15], while of symptomatic CSS incidence was 3.99% [95%CI=0.95-15.25]. No survival differences were noted when comparing ipsilateral to contralateral arteriovenous fistula-internal thoracic artery combinations. On meta-regression, the upper arm was associated with more CSS events, while the forearm to lower late mortality rates. Independently from arteriovenous fistula-internal thoracic artery combination, CSS was not associated to higher mortality rates. Particular attention is warranted when selecting the type of conduits for CABG in patients with an arteriovenous fistula or if highly expected to need one in the near future after surgery. A contralateral arteriovenous fistula-internal thoracic artery combination is preferable. If this is not possible, a forearm arteriovenous fistula position should be preferred.

Mitral valve surgery in acute infective endocarditis: long-term outcomes of mitral valve repair versus replacement.

AIMS: Timing and surgical strategies in acute infective endocarditis are still questionable. We sought to investigate clinical outcomes of patients undergoing mitral valve repair (MVR) compared with mitral valve replacement [mitral valve prosthesis (MVP)] for acute infective endocarditis. METHODS: From 2004 to 2019, 109 consecutive patients with acute mitral valve infective endocarditis were retrospectively investigated. Patients were divided into two groups according to surgical strategy: MVR 53/109 (48.6%) versus MVP 56/109 (51.4%). Primary end points were in-hospital mortality and overall survival at 10 years. Secondary end point was the freedom from infective endocarditis relapse. RESULTS: Our institutional surgical approach for infective endocarditis allowed us to achieve MVR in 48.6% of patients. Hospital mortality was comparable between the two groups [MVR: 1/53 (1.9%) versus MVP: 2/56 (3.6%), P  = 1.000]. Overall 10-year survival was 80.0 ±â€Š14.1 and 77.2 ±â€Š13.5% for MVR and MVP, respectively ( P  = 0.648). MVR showed a lower incidence of infective endocarditis relapse compared with MVP (MVR: 93.6 ±â€Š7.1 versus MVP: 80.9 ±â€Š10.8%, P  = 0.041). At Cox regression, infective endocarditis relapse was an independent risk factor for death (hazard ratio 4.03; 95% confidence interval 1.41-11.52; P  = 0.009). CONCLUSION: The tendency to postpone surgery in stable patients with mitral infective endocarditis allowed achievement of MVR in almost 50% of patients. Although repair remains the approach of choice in our institution, no differences between MVR and MVP were reported in terms of early/late survival. However, MVP had a higher incidence of infective endocarditis relapse that represents an independent risk of mortality.

Sutureless Aortic Valve Replacement vs. Transcatheter Aortic Valve Implantation in Patients with Small Aortic Annulus: Clinical and Hemodynamic Outcomes from a Multi-Institutional Study

ABSTRACT Objective: This study aimed to compare hemodynamic performances and clinical outcomes of patients with small aortic annulus (SAA) who underwent aortic valve replacement by means of sutureless aortic valve replacement (SUAVR) or transcatheter aortic valve implantation (TAVI). Methods: From 2015 to 2020, 622 consecutive patients with SAA underwent either SUAVR or TAVI. Through a 1:1 propensity score matching analysis, two homogeneous groups of 146 patients were formed. Primary endpoint: all cause-death at 36 months. Secondary endpoints: incidence of moderate to severe patient-prosthesis mismatch (PPM) and incidence of major adverse cardiovascular and cerebrovascular events (MACCEs) Results: All-cause death at three years was higher in the TAVI group (SUAVR 12.2% vs. TAVI 21.0%, P=0.058). Perioperatively, comparable hemodynamic performances were recorded in terms of indexed effective orifice area (SUAVR 1.12 ± 0.23 cm2/m2 vs. TAVI 1.17 ± 0.28 cm2/m2, P=0.265), mean transvalvular gradients (SUAVR 12.9 ± 5.3 mmHg vs. TAVI 12.2 ± 6.2 mmHg, P=0.332), and moderate-to-severe PPM (SUAVR 4.1% vs. TAVI 8.9%, P=0.096). TAVI group showed a higher cumulative incidence of MACCEs at 36 months (SUAVR 18.1% vs. TAVI 32.6%, P<0.001). Pacemaker implantation (PMI) and perivalvular leak ≥ 2 were significantly higher in TAVI group and identified as independent predictors of mortality (PMI: hazard ratio [HR] 3.05, 95% confidence interval [CI] 1.34-6.94, P=0.008; PPM: HR 2.72, 95% CI 1.25-5.94, P=0.012). Conclusion: In patients with SAA, SUAVR and TAVI showed comparable hemodynamic performances. Moreover, all-cause death and incidence of MACCEs at follow-up were significantly higher in TAVI group.

Del Nido cardioplegia in adult cardiac surgery: Clinical outcomes in a single center all-comer study.

INTRODUCTION: The use of Del Nido Cardioplegia (DNC) has been extended in the latest years from pediatrics to adult cardiac surgery with encouraging results. We sought to investigate clinical and biochemical outcomes in adult patients who underwent cardiac surgery with different degrees of complexity who received DNC for myocardial protection. METHODS: Data on one-thousand patients were retrospectively collected from 2020 to 2022. The only exclusion criteria was off-pump adult cardiac surgery. Surgical procedures weight was categorized according EuroSCORE II in six groups: Single-CABG(G1), isolated non-CABG(mitral) (G2), isolated non-CABG(aortic) (G3), isolated non-CABG(any) (G4), 2-procedures(G5), 3/more-procedures(G6). Primary endpoint was to identify a binomial correlation between hs-TnT/CK-MB and the cross-clamp time (X-Clamp). A secondary endpoint was the comparison between the treatment groups of the vasoactive-inotropic score (VIS) and the need of mechanical circulatory support (MCS). RESULTS: A linear correlation was identified between hs-TnT and X-clamp in the overall population (rho:0.447, p< .001) and in the treatment groups (G1:rho=0.357, p< .001/G2:rho=0.455, p< .001/G3:rho=0.307, p= .001/G4:rho=0.165, p= .257/G5:rho=0.157, p= .031/G6:rho=0.226, p= .015). Similarly, a linear correlation between CK-MB and X-clamp in the overall population (rho=0.457, p< .001) and treatment group (G1:rho=0.282, p< .001/G2:rho=0.287, p= .025/G3:rho=0.211, p= .009/G4:rho=0.0878, p= .548/G5:rho=0.309, p< .001/G6: rho=0.212, p= .024) was identified. As regard for the secondary endpoint, no differences were reported between the treatment groups in terms of VIS and MCS (VIS G1:7; G2:4; G3:7; G4:7, G5:5.5, G6:6, p-value= .691) (MCS G1: 4.5%; G2:4.8%; G3:3.3%; G4:3.1%; G5:1.4%; G6:5.3%; p-value= .372). CONCLUSIONS: Del Nido Cardioplegia is a safe and useful tool in adult cardiac surgery allowing operators to achieve a stable and durable cardioplegic arrest. Despite accounting with different types of surgery, the six subgroups of our study population showed similar perioperative results.

Left Atrium Volume Reduction Procedure Concomitant With Cox-Maze Ablation in Patients Undergoing Mitral Valve Surgery: A Meta-Analysis of Clinical and Rhythm Outcomes.

BACKGROUND: The management of an enlarged left atrium (LA) in mitral valve (MV) disease with atrial fibrillation (AF) is still being debated. It has been postulated that a reduction in LA size may improve patient outcomes. This meta-analysis aimed to assess rhythm and clinical outcomes of combined surgical AF treatment with or without LA volume reduction (LAVR) in patients undergoing MV surgery. METHODS: A systematic review was performed and all available literature to May 2022 was included. The primary endpoint was analysis of early and late mortality and rhythm outcomes. Secondary outcomes included early and late cerebrovascular accident (CVA) and permanent pacemaker implantation. RESULTS: The search strategy yielded 2,808 potentially relevant articles, and 19 papers were eventually included. The pooled estimated rate of 30-day mortality was 3.76% (95% CI 2.52-5.56). The incidence rate of late mortality and late cardiac-related mortality was 1.75%/year (95% CI 0.63-4.84) and 1.04%/year (95% CI 0.31-3.53), respectively. At subgroup analysis when comparing the surgical procedure with and without AF ablation, the ablation subgroup showed a significantly lower rate of postoperative CVA (p<0.0001) and higher restoration to sinus rhythm at discharge (p=0.0124), with only a trend of lower AF recurrence at 1 year (p=0.0608). At univariable meta-regression, reintervention was significantly associated with higher late mortality (p=0.0033). CONCLUSION: In enlarged LA undergoing MV surgery, LAVR combined with AF ablation showed a trend of improved rhythm outcomes when compared with AF ablation without LAVR. Each LAVR technique has its advantages and disadvantages, which must be managed accordingly.

Thoracoscopic Surgical Ablation of Lone Atrial Fibrillation: Long-term Outcomes at 7 Years.

BACKGROUND: Antiarrhythmic drugs and transcatheter ablation in atrial fibrillation (AF) provide suboptimal rhythm control with a not negligible rate of failure in paroxysmal AF (PAF) and nonparoxysmal AF (n-PAF) at midterm and long-term follow-up. This study evaluated the safety profile and long-term efficacy of thoracoscopic ablation in patients with lone AF. METHODS: A consecutive 153 patients with lone AF were prospectively enrolled and underwent thoracoscopic surgical ablation. Inclusion criteria were symptomatic AF refractory to pharmacologic therapy (Vaughan-Williams class I-III), age >18 years, and absence of left atrial thrombosis. Exclusion criteria were long-standing AF >5 years, left atrial diameter >55 mm, and contraindication to oral anticoagulation. The "box lesion set" (encircling of pulmonary veins) was always used. Exclusion of the left atrial appendage was performed only in selected cases. The primary study end point was freedom from AF. Secondary end points were overall survival and cumulative incidence function of cardiac event-related death, cerebrovascular accidents, and pacemaker implantation. RESULTS: There was no in-hospital mortality. Early postoperative complications were pacemaker implantation (4/153 [2.6%]), cerebrovascular accident (2/153 [1.3%]) with full recovery of both, and bleeding requiring surgical revision (2/153 [1.3%]). Overall freedom from AF at 7 years was 86% ± 4% (76.9% in n-PAF, 96.1% in PAF). Survival freedom from AF in patients without antiarrhythmic drugs in PAF and n-PAF groups was 79.1% and 52.2%, respectively. CONCLUSIONS: Thoracoscopic surgical ablation of lone AF by means of an isolated left atrial box lesion provided an excellent long-term rhythm outcome, even in long-standing persistent AF. The isolated left atrial ablation showed an excellent safety profile with low incidence of pacemaker implantation and postoperative complications.

Cerebral Perfusion and Neuromonitoring during Complex Aortic Arch Surgery: A Narrative Review.

Complex ascending and aortic arch surgery requires the implementation of different cerebral protection strategies to avoid or limit the probability of intraoperative brain damage during circulatory arrest. The etiology of the damage is multifactorial, involving cerebral embolism, hypoperfusion, hypoxia and inflammatory response. These protective strategies include the use of deep or moderate hypothermia to reduce the cerebral oxygen consumption, allowing the toleration of a variable period of absence of cerebral blood flow, and the use of different cerebral perfusion techniques, both anterograde and retrograde, on top of hypothermia, to avoid any period of intraoperative brain ischemia. In this narrative review, the pathophysiology of cerebral damage during aortic surgery is described. The different options for brain protection, including hypothermia, anterograde or retrograde cerebral perfusion, are also analyzed, with a critical review of the advantages and limitations under a technical point of view. Finally, the current systems of intraoperative brain monitoring are also discussed.

Brixia system of screws: a titanium modular system with wide applicability for reinforced sternal closure in cardiac and thoracic surgery.

The use of cannulated screws and titanium plates to reinforce the sternal closure or to treat sternal dehiscence after median sternotomy has already been suggested in several articles. The system proposed here has some important advantages over those already described. Moreover, thanks to its characteristics, this system can also be used to treat pathologies affecting the entire rib cage. The system consists of a first threaded cannulated screw that is inserted in the bone or chondral cartilage and accommodates a cap screw that is tightened into the first screw and fixes a plate according to the following scheme: a threaded cannulated screw/plate/cap screw (Brixia system of screws). This system allows the plates to be fixed on the anterior face of the ribs and/or sternum without the need to enlarge dissection of the tissue, thereby lowering the danger of haemorrhage and injury to the thoracic organs. For this reason, it is particularly suitable for treating post-sternotomy sternal dehiscence, but it can be used to reinforce the primary sternal closure (after median or transversal sternotomy) in high-risk patients with sternal dehiscence. Owing to the modular nature of the system, singular components can also be utilized independently.

Rhythm outcomes of minimally-invasive off-pump surgical versus catheter ablation in atrial fibrillation: A meta-analysis of reconstructed time-to-event data.

BACKGROUND: Mid- and long-term rhythm outcomes of catheter ablation (CA) for atrial fibrillation (AF) are reported to be suboptimal. Minimally invasive surgical off-pump ablation (MISOA), including both thoracoscopic and trans-diaphragmatic approaches, has been developed to reduce surgical invasiveness and overcome on-pump surgery drawbacks. We sought to compare the efficacy and safety of MISOA and CA for AF treatment. METHODS: A systematic review and meta-analysis of the literature was performed including studies comparing MISOA and CA. The primary endpoint was survival freedom from AF at follow-up after a 3-month blanking period. Subgroup analysis of the primary endpoint was performed according to the type of surgical incision and hybrid approach. RESULTS: Freedom from AF at 4 years was 52.1% ± 3.2% vs 29.1% ± 3.5%, between MISOA and CA respectively (log-rank p < 0.001; Hazard Ratio: 0.60 [95%Confidence Interval (CI):0.50-0.72], p < 0.001). At landmark analysis, a significant improvement in rhythm outcomes was observed in the MISOA group after the 5th month of follow-up (2 months from the blanking period). The Odds Ratio between MISOA and CA of postoperative cerebrovascular accident incidence and postoperative permanent pacemaker implant (PPM) were 2.00 (95%CI:0.91-4.40, p = 0.084) and 1.55 (95%CI:0.61-3.95, p = 0.358), respectively. The incidence rate ratio of late CVA between MISOA and CA was 0.86 (95%CI:0.28-2.65, p = 0.787), while for late PPM implant was 0.45 (95%CI:0.11-1.78, p = 0.256). CONCLUSIONS: The current meta-analysis suggests that MISOA provides superior rhythm outcomes when compared to CA in terms of sinus rhythm restoration. Despite the rhythm outcome superiority of MISOA, it is associated to higher postoperative complications compared to CA.

Deep learning to detect significant coronary artery disease from plain chest radiographs AI4CAD.

BACKGROUND: The predictive role of chest radiographs in patients with suspected coronary artery disease (CAD) is underestimated and may benefit from artificial intelligence (AI) applications. OBJECTIVES: To train, test, and validate a deep learning (DL) solution for detecting significant CAD based on chest radiographs. METHODS: Data of patients referred for angina and undergoing chest radiography and coronary angiography were analysed retrospectively. A deep convolutional neural network (DCNN) was designed to detect significant CAD from posteroanterior/anteroposterior chest radiographs. The DCNN was trained for severe CAD binary classification (absence/presence). Coronary angiography reports were the ground truth. Stenosis severity of ≥70% for non-left main vessels and ≥ 50% for left main defined severe CAD. RESULTS: Information of 7728 patients was reviewed. Severe CAD was present in 4091 (53%). Patients were randomly divided for algorithm training (70%; n = 5454) and fine-tuning/model validation (10%; n = 773). Internal clinical validation (model testing) was performed with the remaining patients (20%; n = 1501). At binary logistic regression, DCNN prediction was the strongest severe CAD predictor (p < 0.0001; OR: 1.040; CI: 1.032-1.048). Using a high sensitivity operating cut-point, the DCNN had a sensitivity of 0.90 to detect significant CAD (specificity 0.31; AUC 0.73; 95% CI DeLong, 0.69-0.76). Adding to the AI chest radiograph interpretation angina status improved the prediction (AUC 0.77; 95% CI DeLong, 0.74-0.80). CONCLUSION: AI-read chest radiographs could be used to pre-test significant CAD probability in patients referred for suspected angina. Further studies are required to externally validate our algorithm, develop a clinically applicable tool, and support CAD screening in broader settings.

Predicting early mortality following single-stage coronary artery or valve surgery and carotid endarterectomy.

BACKGROUND: Surgical management of coexisting cardiac disease and extra-cranial carotid artery disease is a controversial area of debate. Thus, in this challenging scenario, risk stratification may play a key role in surgical decision making. AIM: To report the results of single-stage coronary/valve surgery (CVS) and carotid endarterectomy (CEA), and to identify predictive factors associated with 30-day mortality. METHODS: This was a multicenter, retrospective study of prospectively maintained data from three academic tertiary referral hospitals. For this study, only patients treated with single-stage CVS, meaning coronary artery bypass surgery or valve surgery, and CEA between March 1, 2000 and March 30, 2020, were included. Primary outcome measure of interest was 30-day mortality. Secondary outcomes were neurologic events rate, and a composite endpoint of postoperative stroke/death rate. RESULTS: During the study period, there were 386 patients who underwent the following procedures: CEA with isolated coronary artery bypass graft in 243 (63%) cases, with isolated valve surgery in 40 (10.4%), and combination of coronary artery bypass grafting and valve surgery in 103 (26.7%). Postoperative neurologic event rate was 2.6% (n = 10) which includes 5 (1.3%) transient ischemic attacks and 5 (1.3%) strokes (major n = 3, minor n = 2). The 30-day mortality rate was 3.9% (n = 15). Predictors of 30-day mortality included preoperative left heart insufficiency (odds ratio [OR]: 5.44, 95% confidence interval [CI]: 1.63-18.17, p = .006), and postoperative stroke (OR: 197.11, 95% CI: 18.28-2124.93, p < .001). No predictor for postoperative stroke and for composite endpoint was identified. CONCLUSIONS: Considering that postoperative stroke rate and mortality was acceptably low, single-stage approach is an effective option in such selected high-risk patients.