Identifying treatment heterogeneity in atrial fibrillation using a novel causal machine learning method.

BACKGROUND: Lifelong oral anticoagulation is recommended in patients with atrial fibrillation (AF) to prevent stroke. Over the last decade, multiple new oral anticoagulants (OACs) have expanded the number of treatment options for these patients. While population-level effectiveness of OACs has been compared, it is unclear if there is variability in benefit and risk across patient subgroups. METHODS: We analyzed claims and medical data for 34,569 patients who initiated a nonvitamin K antagonist oral anticoagulant (non-vitamin K antagonist oral anticoagulant (NOAC); apixaban, dabigatran, and rivaroxaban) or warfarin for nonvalvular AF between 08/01/2010 and 11/29/2017 from the OptumLabs Data Warehouse. A machine learning (ML) method was applied to match different OAC groups on several baseline variables including, age, sex, race, renal function, and CHA2DS2 -VASC score. A causal ML method was then used to discover patient subgroups characterizing the head-to-head treatment effects of the OACs on a primary composite outcome of ischemic stroke, intracranial hemorrhage, and all-cause mortality. RESULTS: The mean age, number of females and white race in the entire cohort of 34,569 patients were 71.2 (SD, 10.7) years, 14,916 (43.1%), and 25,051 (72.5%) respectively. During a mean follow-up of 8.3 (SD, 9.0) months, 2,110 (6.1%) of patients experienced the composite outcome, of whom 1,675 (4.8%) died. The causal ML method identified 5 subgroups with variables favoring apixaban over dabigatran; 2 subgroups favoring apixaban over rivaroxaban; 1 subgroup favoring dabigatran over rivaroxaban; and 1 subgroup favoring rivaroxaban over dabigatran in terms of risk reduction of the primary endpoint. No subgroup favored warfarin and most dabigatran vs warfarin users favored neither drug. The variables that most influenced favoring one subgroup over another included Age, history of ischemic stroke, thromboembolism, estimated glomerular filtration rate, Race, and myocardial infarction. CONCLUSIONS: Among patients with AF treated with a NOAC or warfarin, a causal ML method identified patient subgroups with differences in outcomes associated with OAC use. The findings suggest that the effects of OACs are heterogeneous across subgroups of AF patients, which could help personalize the choice of OAC. Future prospective studies are needed to better understand the clinical impact of the subgroups with respect to OAC selection.

Electroanatomical Voltage Mapping to Distinguish Right-Sided Cardiac Sarcoidosis From Arrhythmogenic Right Ventricular Cardiomyopathy.

OBJECTIVES: This study sought to investigate the value of electroanatomical voltage mapping (EAVM) to distinguish cardiac sarcoidosis (CS) from arrhythmogenic right ventricular cardiomyopathy (ARVC) in patients with ventricular tachycardia from the right ventricle (RV). BACKGROUND: CS can mimic ARVC. Because scar in ARVC is predominantly subepicardial, this study hypothesized that the relative sizes of endocardial low bipolar voltage (BV) to low unipolar voltage (UV) areas may distinguish CS from ARVC. METHODS: Patients with CS affecting the RV (n = 14), patients with gene-positive ARVC (n = 13), and a reference group of patients without structural heart disease (n = 9) who underwent RV endocardial EAVM were included. RV region-specific BV and UV cutoffs were derived from control subjects. In CS and ARVC, segmental involvement was determined and low-voltage areas were measured, using <1.5 mV for BV and <3.9 mV, <4.4 mV, and <5.5 mV for UV. The ratio between low BV and low UV area was calculated generating 3 parameters: Ratio3.9, Ratio4.4 and Ratio5.5, respectively. RESULTS: In control subjects, BV and UV varied significantly among RV regions. The basal septum was involved in 71% of CS patients and in none of ARVC patients. Ratio5.5 discriminated CS from ARVC the best. An algorithm including Ratio5.5 ≥0.45 and basal septal involvement identified CS with 93% sensitivity and 85% specificity. This was validated in a separate population (CS [n = 6], ARVC [n = 10]) with 100% sensitivity and 100% specificity. CONCLUSIONS: EAVM provides detailed information about scar characteristics and scar distribution in the RV. An algorithm combining Ratio5.5 (area BV <1.5 mV/area UV <5.5 mV) and bipolar basal septal involvement allows accurate diagnosis of (isolated) CS in patients presenting with monomorphic ventricular tachycardia from the RV.

Stepwise Approach for Ventricular Tachycardia Ablation in Patients With Predominantly Intramural Scar.

OBJECTIVES: The goal of this study was to assess the value of a stepwise, image-guided ablation approach in patients with cardiomyopathy and predominantly intramural scar. BACKGROUND: Few reports have focused on catheter-based ventricular tachycardia (VT) ablation strategies in patients with predominantly intramural scar. METHODS: The study included patients with predominantly intramural scar undergoing VT ablation. A stepwise strategy was performed consisting of a localized ablation guided by conventional mapping criteria followed by a more extensive ablation if VT remained inducible. The extensive ablation was guided by the location and extent of intramural scarring on delayed enhanced-cardiac magnetic resonance imaging. A historical cohort who did not undergo additional extensive ablation was identified for comparison. A novel measurement, the scar depth index (SDI), indicating the percent area of the scar at a given depth, was correlated with outcomes. RESULTS: Forty-two patients who underwent stepwise ablation (median age 61 years [interquartile range: 55 to 69 years], 35 male patients, median left ventricular ejection fraction 36.0% [25.0% to 55.0%], ischemic [n = 4] or nonischemic cardiomyopathy [n = 38]) were followed up for a median of 17 months (8 to 36 months). A stepwise approach resulted in a 1-year freedom from VT, death, or cardiac transplantation of 76% (32 of 42). Patients who underwent additional extensive ablation had a lower risk of events than a clinically similar historical cohort (N = 19) (hazard ratio: 0.30; 95% CI: 0.13 to 0.68; p < 0.004). SDI>5mm was associated with worse long-term outcomes (hazard ratio: 1.03; 95% CI: 1.01 to 1.06%; p = 0.03), SDI>5mm >16.5% was associated with failed ablation (area under the curve: 0.84; 95% CI: 0.71 to 0.97). CONCLUSIONS: Stepwise ablation using delayed enhanced-cardiac magnetic resonance guidance is a novel approach to VT ablation in patients with predominantly intramural scarring. The SDI correlates with immediate procedural and long-term outcomes.

Targeting Noninducible Clinical Ventricular Tachycardias in Patients With Prior Myocardial Infarctions Based on Stored Electrograms.

BACKGROUND: Ablation of postinfarction ventricular tachycardia (VT) has been shown to reduce VT recurrence and decrease mortality. However, VT recurrence can occur despite extensive ablation procedures. The lack of inducibility of clinical VTs during ablation procedures remains problematic and may be in part responsible for VT recurrences. In this prospective study, we targeted documented but noninducible clinical VTs based on stored implantable cardioverter-defibrillator (ICD) electrograms. METHODS: Radiofrequency ablation was performed in a consecutive group of 66 postinfarction patients (mean age, 67.5±9.2 years; men, 61; mean left ventricular ejection fraction, 25.1±10.8%) in whom clinical VTs were not inducible during an ablation procedure. In the first 33 patients (control group), only inducible VTs were targeted, and in the second 33 patients, noninducible clinical VTs were also targeted by pace-mapping based on stored ICD-electrograms (ICD-electrogram-guided ablation group). Procedural and clinical outcomes were compared at 24 months post-ablation. RESULTS: VT recurred in 5 patients (15%) in whom the ICD-electrogram-guided approach was performed and in 13 patients (39%) in the control group. Freedom from recurrent VT was higher (log-rank P=0.04) in the ICD-electrogram-guided group, but there was no difference in ventricular fibrillation or in total mortality between both groups. CONCLUSIONS: Ablation guided by pace-mapping of noninducible postinfarction clinical VTs based on ICD-electrograms is feasible and reduces the risk of recurrent VT.

Postinfarction Myocardial Calcifications on Cardiac Computed Tomography: Implications for Mapping and Ablation in Patients With Nontolerated Ventricular Tachycardias.

BACKGROUND: Postinfarction ventricular tachycardia (VT) generally involves myocardial fibers surrounded by scar. Calcification of scar tissue has been described, but the relationship between calcifications within endocardial scar and VTs is unclear. The purpose of this study was to assess the prevalence of myocardial calcifications as detected by cardiac computed tomography (CT) and the benefit for mapping and ablation focusing on nontolerated VTs. METHODS: Fifty-six consecutive postinfarction patients had a cardiac CT performed before a VT ablation procedure. Another 56 consecutive patients with prior infarction without VT who had cardiac CTs served as a control group. RESULTS: Myocardial calcifications were identified in 39 of 56 patients (70%) in the postinfarction group with VT, compared with 6 of 56 patients (11%) in the control group without VT. Calcifications were associated with VT when compared with a control group. A calcification volume of 0.538 cm3 distinguished patients with calcification-associated VT from patients without calcification-associated VTs (area under the curve, 0.87; sensitivity, 0.87; specificity, 0.88). Myocardial calcifications corresponded to areas of electrical nonexcitability and formed a border for reentry circuits for 49 VTs (33% of all VTs for which target sites were identified) in 24 of 39 patients (62%) with myocardial calcifications. A nonconfluent calcification pattern was associated with VT target sites independent of calcification volume ( P=0.01). CONCLUSIONS: Myocardial calcifications detected by cardiac CT in patients with prior infarction are associated with VT. The calcifications correspond to areas of unexcitability and represent a fixed boundary of reentry circuits that can be visualized by CT. Calcifications correspond to effective ablation sites in >1/3 of patients with postinfarction VT.

Correlation between computer tomography-derived scar topography and critical ablation sites in postinfarction ventricular tachycardia.

BACKGROUND: Myocardial wall thickness (WT) in patients with a prior myocardial infarction has been used to indicate scarring. However, the correlation of WT with sites critical to ventricular tachycardia (VT) has not been previously investigated. The purpose of this study was to correlate electroanatomic mapping data obtained during VT ablation with WT determined by cardiac computed tomography (CT). METHODS AND RESULTS: Cardiac CTs were performed in 15 consecutive patients (mean age 63 ± 10 years, 86% male, left ventricular ejection fraction 27 ± 12%) with a prior infarct referred for VT ablation. The CTs were registered to the electroanatomic maps obtained during the mapping procedure. Pacing was performed throughout the scar at sites with fractionated electrograms and isolated potentials. Ablation sites were identified by pace-mapping or entrainment-mapping and these sites were correlated with WT. Bipolar and unipolar voltage amplitude and bipolar electrogram width correlated with WT (correlation coefficient: 0.63, 0.65, and 0.41, respectively, P < 0.001). Ablation target sites were identified for 58 of 113 inducible VTs. The ablation target sites were located on CT-defined ridges (WT: 4.2 ± 1.2 mm) bordered by areas of thinning (WT: 2.6 ± 1.1 mm, P < 0.0001) in 14 of 15 patients. Ablation targets were found on ridges in 49 of 58 VTs (84%) for which target sites were identified. A total of 70 ridges were localized in the 15 patients. VT became noninducible postablation in 11 of 15 patients (73%). CONCLUSION: WT measured by CT identifies ridges of myocardial tissue that often are critical for postinfarction VT and that can be appropriate target sites for ablation.