Explainable machine learning models for predicting 30-day readmission in pediatric pulmonary hypertension: A multicenter, retrospective study.

Background: Short-term readmission for pediatric pulmonary hypertension (PH) is associated with a substantial social and personal burden. However, tools to predict individualized readmission risk are lacking. This study aimed to develop machine learning models to predict 30-day unplanned readmission in children with PH. Methods: This study collected data on pediatric inpatients with PH from the Chongqing Medical University Medical Data Platform from January 2012 to January 2019. Key clinical variables were selected by the least absolute shrinkage and the selection operator. Prediction models were selected from 15 machine learning algorithms with excellent performance, which was evaluated by area under the operating characteristic curve (AUC). The outcome of the predictive model was interpreted by SHapley Additive exPlanations (SHAP). Results: A total of 5,913 pediatric patients with PH were included in the final cohort. The CatBoost model was selected as the predictive model with the greatest AUC for 0.81 (95% CI: 0.77-0.86), high accuracy for 0.74 (95% CI: 0.72-0.76), sensitivity 0.78 (95% CI: 0.69-0.87), and specificity 0.74 (95% CI: 0.72-0.76). Age, length of stay (LOS), congenital heart surgery, and nonmedical order discharge showed the greatest impact on 30-day readmission in pediatric PH, according to SHAP results. Conclusions: This study developed a CatBoost model to predict the risk of unplanned 30-day readmission in pediatric patients with PH, which showed more significant performance compared with traditional logistic regression. We found that age, LOS, congenital heart surgery, and nonmedical order discharge were important factors for 30-day readmission in pediatric PH.

Catheter ablation of premature ventricular contractions originating from periprosthetic aortic valve regions.

BACKGROUND: Little is known about the ablation outcomes of premature ventricular contractions (PVCs) that originate from the periprosthetic aortic valve (PPAV) regions of patients with aortic valve replacement (AVR). METHODS AND RESULTS: Our study had 11 patients who underwent catheter ablation for PVCs arising from the PPAV regions (bioprosthetic aortic valve, n = 5; mechanical aortic valve, n = 6). The PVC characteristics, procedure characteristics, and efficacy of ablation were compared with the control group (n = 33). At baseline, the PPAV group had a lower left ventricular ejection fraction (mean [SD], 41% [12%] vs. 51% [8%]; p = .002). The rate of acute ablation success was 90.9% in the PPAV group. Ablation sites were identified above the left coronary cusp (LCC) and right coronary cusp commissure (LRCC) in one PVC, below the prosthetic valve in eight PVCs (four below LCC and four below LRCC), and within the distal coronary sinus in two PVCs. The mean procedure time, fluoroscopy time, and radiation in the PPAV group were all significantly greater than those in the control group (all p < .05). However, the number of radiofrequency ablation energy deliveries was not different. The PPAV group had a long-term success rate compared with the control group (72.7% vs. 87.9%, p = .48) and an increase of left ventricular ejection fraction from 43% to 49% after successful PVC ablation at follow-up (p < .001). Echocardiography showed no significant change in valve regurgitation after ablation. No new atrioventricular block occurred. CONCLUSION: PVCs arising from PPAV regions can be successfully ablated in patients with prior AVR, without damaging the prosthetic aortic valve and atrioventricular conduction.

Anatomy Versus Physiology-Guided Ablation for Persistent Atrial Fibrillation.

BACKGROUND: Pulmonary vein isolation remains the cornerstone of atrial fibrillation (AF) ablation. However, due to high recurrence rates, especially in patients with persistent AF, PV antral isolation, complemented by linear ablation, autonomic modulation, and ablation of complex fractionated electrograms, have been attempted to increase the odds of success. However, the optimum approach for a complementary strategy in addition to PVI for persistent AF is unknown. METHODS: We performed a prospective randomized trial by assigning 92 patients with persistent AF in 1:1 ratio to pulmonary-vein isolation plus ablation of electrograms showing complex fractionated activity (45 patients), or pulmonary-vein isolation plus additional linear ablation across the left atrial roof and mitral valve isthmus (47 patients). The duration of follow-up was five years. The primary endpoint was freedom from any documented recurrence of atrial fibrillation after a single ablation procedure. RESULTS: At a 12-month follow-up, 9 (23%) patients had AF recurrence in the linear ablation and 8 (21%) patients in the CFAE groups. At a mean follow-up duration of 59±36 months, 48.3% of patients in the linear ablation group and 44.6% of patients in the CFAE group were free from AF (p=0.403). There were no significant differences between the two groups for independent predictors of freedom from AF. The overall procedure time and radiation exposure were higher in the PVI+linear ablation group. There were five adverse events noted, two in the linear group (pericardial effusion not requiring drain) and 3 in the CFAE group (1 pseudoaneurysm, one effusion requiring pericardiocentesis and one effusion nor requiring drain). CONCLUSIONS: Among patients with persistent atrial fibrillation, we found no difference in maintenance of sinus rhythm in either linear ablation or ablation of complex fractionated electrograms was performed in addition to pulmonary vein isolation in short- and long-term follow-up.

Experimental Study on the Effect of a Weifufang on Human Gastric Adenocarcinoma Cell Line BGC-823 Xenografts and PTEN Gene Expression in Nude Mice.

Background: This study aims at investigating the effect of the Weifufang, an effective prescription for the treatment of gastric cancer developed by the Traditional Chinese Medicine (TCM)/Combination of TCM and Western Medicine Department of the Hunan Cancer Hospital, on gastric cancer xenografts in nude mice and its effect on the PTEN gene; it also aims at exploring the possible tumor suppression mechanism. Methods: Nude mice with xenografts were treated with different concentrations of the Weifufang for 2 weeks, and changes in tumor volume were observed. The histopathology of the tumor was detected by hematoxylin and eosin staining; PTEN gene expression in tumor tissues was detected by immunohistochemistry (IHC) and western blot. Results: After 2 weeks of treatment, tumor inhibition rates in the 5-flourouracil (5-FU) group, and in the Weifufang low-, middle-, and high-dose groups were 30.67%, 19%, 49.52%, and 29.36%, respectively. The IOD of the PTEN gene was detected by IHC. The values in the water group, the 5-FU group, and the Weifufang low-, middle-, and high-dose groups were 0.013 ± 0.004, 0.085 ± 0.062, 0.041 ± 0.024, 0.128 ± 0.032, and 0.061 ± 0.052, respectively. Except for the 5-FU group, the differences between the gastric compound middle dose-group and the other groups were statistically significant (p < 0.05). Results of PTEN expression detection by western blot: The expression levels in the water group, 5-FU group, and the Weifufang low-, middle-, and high-dose groups were 0.2240 ± 0.0172, 0.4200 ± 0.0228, 0.2760 ± 0.0163, 0.3840 ± 0.0133, and 0.3040 ± 0.0211, respectively. Except for the 5-FU group, differences between the Weifufang middle-dose group and the other groups were statistically significant (p < 0.05). Conclusion: The Weifufang may inhibit the growth of gastric cancer xenografts by upregulating PTEN gene expression. The middle-dose group had the best effect.

Selective Renal Denervation Guided by Renal Nerve Stimulation in Canine.

Renal nerve stimulation (RNS) can result in substantial blood pressure (BP) elevation, and the change was significantly blunted when repeated stimulation after ablation. However, whether RNS could provide a meaningful renal nerve mapping for identification of optimal ablation targets in renal denervation (RDN) is not fully clear. Here, we compared the antihypertensive effects of selective RDN guided by two different BP responses to RNS and explored the nerve innervations at these sites in Kunming dogs. Our data indicated that ablation at strong-response sites showed a more systolic BP-lowering effect than at weak-response sites (P=0.002), as well as lower levels of tyrosine hydroxylase and norepinephrine in kidney and a greater reduction in plasma norepinephrine (P=0.004 for tyrosine hydroxylase, P=0.002 for both renal and plasma norepinephrine). Strong-response sites showed a greater total area and mean number of renal nerves than weak-response sites (P=0.012 for total area and P<0.001 for mean number). Systolic BP-elevation response to RNS before RDN and blunted systolic BP-elevation to RNS after RDN were correlated with systolic BP changes at 4 weeks follow-up (R=0.649; P=0.012 and R=0.643; P=0.013). Changes of plasma norepinephrine and renal norepinephrine levels at 4 weeks were also correlated with systolic BP changes at 4 weeks (R=0.837, P<0.001 and R=0.927, P<0.001). These data suggest that selective RDN at sites with strong BP-elevation response to RNS could lead to a more efficient RDN. RNS is an effective method to identify the nerve-enriched area during RDN procedure and improve the efficacy of RDN.