Dataset Information

Machine-learning-based prediction of survival and mitral regurgitation recurrence in patients undergoing mitral valve repair.

ABSTRACT:

Objectives

This study was conducted to assess long-term clinical outcomes after mitral valve repair using machine-learning techniques.

Methods

We retrospectively evaluated 436 consecutive patients (mean age: 54.7 ± 15.4; 235 males) who underwent mitral valve repair between January 2000 and December 2017. Actuarial survival and freedom from significant (≥ moderate) mitral regurgitation (MR) were clinical end points. To evaluate the independent risk factors, random survival forest (RSF), extreme gradient boost (XGBoost), support vector machine, Cox proportional hazards model and general linear models with elastic net regularization were used. Concordance indices (C-indices) of each model were estimated.

Results

The operative mortality was 0.9% (N = 4). Reoperation was required in 15 patients (3.5%). In terms of C-index, the overall performance of the XGBoost (C-index 0.806) and RSF models (C-index 0.814) was better than that of the Cox model (C-index 0.733) in overall survival. For the recurrent MR, the C-index for XGBoost was 0.718, which was the highest among the 5 models. Compared to the Cox model (C-index 0.545), the C-indices of the XGBoost (C-index 0.718) and RSF models (C-index 0.692) were higher.

Conclusions

Machine-learning techniques can be a useful tool for both prediction and interpretation in the survival and recurrent MR. From the machine-learning techniques examined here, the long-term clinical outcomes of mitral valve repair were excellent. The complexity of MV increased the risk of late mitral valve-related reoperation.

SUBMITTER: Kang Y

PROVIDER: S-EPMC10903183 | biostudies-literature | 2023 Nov

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Machine-learning-based prediction of survival and mitral regurgitation recurrence in patients undergoing mitral valve repair.

Kang Yoonjin Y Sohn Suk Ho SH Choi Jae Woong JW Hwang Ho Young HY Kim Kyung Hwan KH

Interdisciplinary cardiovascular and thoracic surgery 20231101 5

<h4>Objectives</h4>This study was conducted to assess long-term clinical outcomes after mitral valve repair using machine-learning techniques.<h4>Methods</h4>We retrospectively evaluated 436 consecutive patients (mean age: 54.7 ± 15.4; 235 males) who underwent mitral valve repair between January 2000 and December 2017. Actuarial survival and freedom from significant (≥ moderate) mitral regurgitation (MR) were clinical end points. To evaluate the independent risk factors, random survival forest ( ...[more]

PMID: 37966944

Similar Datasets

Project description:We aimed to identify predictors of mitral regurgitation recurrence (MR) after percutaneous mitral valve repair (PMVR) in patients with functional mitral regurgitation (FMR). Patients with FMR were enrolled who underwent PMVR using the MitraClip® device. Procedural success was defined as reduction of MR of at least one grade to MR grade ≤ 2 + assessed at discharge. Recurrence of MR was defined as MR grade 3 + or worse at one year after initially successful PMVR. A total of 306 patients with FMR underwent PMVR procedure. In 279 out of 306 patients (91.2%), PMVR was successfully performed with MR grade ≤ 2 + at discharge. In 11.4% of these patients, MR recurrence of initial successful PMVR after 1 year was observed. Recurrence of MR was associated with a higher rate of heart failure rehospitalization during the 12 months follow-up (52.0% vs. 30.3%; p = 0.029), and less improvement in New York Heart Association (NYHA) functional class [68% vs. 19% of the patients presenting with NYHA functional class III or IV one year after PMVR when compared to patients without recurrence (p = 0.001)]. Patients with MR recurrence were characterized by a higher left ventricular sphericity index {0.69 [Interquartile range (IQR) 0.64, 0.74] vs. 0.65 (IQR 0.58, 0.70), p = 0.003}, a larger left atrium volume [118 (IQR 96, 143) ml vs. 102 (IQR 84, 123) ml, p = 0.019], a larger tenting height 10 (IQR 9, 13) mm vs. 8 (IQR 7, 11) mm (p = 0.047), and a larger mitral valve annulus [41 (IQR 38, 43) mm vs. 39 (IQR 36, 40) mm, p = 0.015] when compared to patients with durable optimal long-term results. In a multivariate regression model, the left ventricular sphericity index [Odds Ratio (OR) 1.120, 95% Confidence Interval (CI) 1.039-1.413, p = 0.003)], tenting height (OR 1.207, 95% CI 1.031-1.413, p = 0.019), and left atrium enlargement (OR 1.018, 95% CI 1.000-1.038, p = 0.047) were predictors for MR recurrence after 1 year. In patients with FMR, baseline parameters of advanced heart failure such as spherical ventricle, tenting height and a large left atrium might indicate risk of recurrent MR one year after PMVR.

Project description:ImportanceThere are limited data on the outcomes of transcatheter edge-to-edge mitral valve repair for degenerative mitral regurgitation (MR) in a real-world setting.ObjectiveTo evaluate the outcomes of transcatheter mitral valve repair for degenerative MR.Design, setting, and participantsCohort study of consecutive patients in the Society of Thoracic Surgeons/American College of Cardiology Transcatheter Valve Therapies Registry who underwent nonemergent transcatheter mitral valve repair for degenerative MR in the US from 2014 through 2022.ExposureTranscatheter edge-to-edge mitral valve repair with the MitraClip device (Abbott).Main outcomes and measuresThe primary end point was MR success, defined as moderate or less residual MR and a mean mitral gradient of less than 10 mm Hg. Clinical outcomes were evaluated based on the degree of residual MR (mild or less MR or moderate MR) and mitral valve gradients (≤5 mm Hg or >5 to <10 mm Hg).ResultsA total of 19 088 patients with isolated moderate to severe or severe degenerative MR who underwent transcatheter mitral valve repair were analyzed (median age, 82 years; 48% women; median Society of Thoracic Surgeons predicted risk of mortality with surgical mitral valve repair, 4.6%). MR success was achieved in 88.9% of patients. At 30 days, the incidence of death was 2.7%; stroke, 1.2%; and mitral valve reintervention, 0.97%. MR success compared with an unsuccessful procedure was associated with significantly lower mortality (14.0% vs 26.7%; adjusted hazard ratio, 0.49; 95% CI, 0.42-0.56; P < .001) and heart failure readmission (8.4% vs 16.9%; adjusted hazard ratio, 0.47; 95% CI, 0.41-0.54; P < .001) at 1 year. Among patients with MR success, the lowest mortality was observed in patients who had both mild or less residual MR and mean mitral gradients of 5 mm Hg or less compared with those with an unsuccessful procedure (11.4% vs 26.7%; adjusted hazard ratio, 0.40; 95% CI, 0.34-0.47; P < .001).Conclusions and relevanceIn this registry-based study of patients with degenerative MR undergoing transcatheter mitral valve repair, the procedure was safe and resulted in successful repair in 88.9% of patients. The lowest mortality was observed in patients with mild or less residual MR and low mitral gradients.

Project description:ObjectivesThe Cardiothoracic Surgical Trials Network recently reported no difference in the primary end point of left ventricular end-systolic volume index at 1 year postsurgery in patients randomized to repair (n = 126) or replacement (n = 125) for severe ischemic mitral regurgitation. However, patients undergoing repair experienced significantly more recurrent mitral regurgitation than patients undergoing replacement (32.6% vs 2.3%). We examined whether baseline echocardiographic and clinical characteristics could identify those who will develop moderate/severe recurrent mitral regurgitation or die.MethodsOur analysis includes 116 patients who were randomized to and received mitral valve repair. Logistic regression was used to estimate a model-based probability of recurrence or death from baseline factors. Receiver operating characteristic curves were constructed from these estimated probabilities to determine classification cut-points maximizing accuracy of prediction based on sensitivity and specificity.ResultsOf the 116 patients, 6 received a replacement before leaving the operating room; all other patients had mild or less mitral regurgitation on intraoperative echocardiogram after repair. During the 2-year follow-up period, 76 patients developed moderate/severe mitral regurgitation or died (53 mitral regurgitation recurrences, 13 mitral regurgitation recurrences and death, and 10 deaths). The mechanism for recurrent mitral regurgitation was largely mitral valve leaflet tethering. Our model (including age, body mass index, sex, race, effective regurgitant orifice area, basal aneurysm/dyskinesis, New York Heart Association class, history of coronary artery bypass grafting, percutaneous coronary intervention, or ventricular arrhythmias) yielded an area under the receiver operating characteristic curve of 0.82.ConclusionsThe model demonstrated good discrimination in identifying patients who will survive 2 years without recurrent mitral regurgitation after mitral valve repair. Although our results require validation, they offer a clinically relevant risk score for selection of surgical candidates for this procedure.

Project description:ObjectiveIntraoperative predictors of functional mitral valve (MV) stenosis after surgical repair of mitral regurgitation (MR) caused by prolapse remain poorly characterised. This study evaluated the effect of annuloplasty size on postoperative MV haemodynamics during exercise and evaluated predictors of MV hemodynamics.Methods104 patients were randomly assigned to leaflet resection or preservation for surgical repair of MR in the Canadian Mitral Research Alliance CardioLink-2 study. In this post hoc analysis, we compared MV haemodynamics between the two surgical groups and examined the relationship between annuloplasty size and MV haemodynamics 1 year after repair in the combined groups. Echocardiograms were performed at baseline and intraoperatively. Exercise transthoracic echocardiography was performed 1 year postoperatively. Multivariable linear regression analysis was used to identify predictors of exercise MV gradients at follow-up.ResultsMean age of participants was 65±10 years, and 83% were male. Median annuloplasty size was 34 (IQR 32-36). Dividing by the median, 48 (46%) had annuloplasty size of <34 mm and 56 (54%) had ≥34 mm. Mean and peak exercise gradients at 1 year were 11±5 mm Hg and 22±9 mm Hg in <34, and 6±3 mm Hg and 14±5 mm Hg in ≥34 (p<0.001). Rate of residual MR was similar in both groups. In multivariable analyses, annuloplasty size of ≥34 mm was associated with lower mean and peak exercise gradients at 12 months, after adjustment for repair type, age, sex, heart rate and body surface area (β -4.1, 95% CI -6 to -3, p<0.001, and β -7 95% CI -10 to -4, p<0.001, respectively). Intraoperative mean and peak MV gradients by transesophageal echocardiography independently predicted mean and peak resting and exercise gradients at follow-up (p<0.001). Similar results were obtained in both leaflet resection and preservation.ConclusionAnnuloplasty size of ≥34 mm is associated with a 4 and 7 mm Hg reduction in mean and peak exercise MV gradients, respectively, 1 year post MV repair regardless of the repair strategy used. Intraoperative TEE MV gradients predict exercise MV gradients 1 year post repair.Trial registration numberNCT02552771.

Project description:BackgroundWhen feasible, guidelines recommend mitral valve repair (MVr) over mitral valve replacement (MVR) to treat primary mitral regurgitation (MR), based upon historic outcome studies and transthoracic echocardiography (TTE) reverse remodeling studies. Cardiovascular magnetic resonance (CMR) offers reference standard biventricular assessment with superior MR quantification compared to TTE. Using serial CMR in primary MR patients, we aimed to investigate cardiac reverse remodeling and residual MR post-MVr vs MVR with chordal preservation.Methods83 patients with ≥ moderate-severe MR on TTE were prospectively recruited. 6-min walk tests (6MWT) and CMR imaging including cine imaging, aortic/pulmonary through-plane phase contrast imaging, T1 maps and late-gadolinium-enhanced (LGE) imaging were performed at baseline and 6 months after mitral surgery or watchful waiting (control group).Results72 patients completed follow-up (Controls = 20, MVr = 30 and MVR = 22). Surgical groups demonstrated comparable baseline cardiac indices and co-morbidities. At 6-months, MVr and MVR groups demonstrated comparable improvements in 6MWT distances (+ 57 ± 54 m vs + 64 ± 76 m respectively, p = 1), reduced indexed left ventricular end-diastolic volumes (LVEDVi; - 29 ± 21 ml/m2 vs - 37 ± 22 ml/m2 respectively, p = 0.584) and left atrial volumes (- 23 ± 30 ml/m2 and - 39 ± 26 ml/m2 respectively, p = 0.545). At 6-months, compared with controls, right ventricular ejection fraction was poorer post-MVr (47 ± 6.1% vs 53 ± 8.0% respectively, p = 0.01) compared to post-MVR (50 ± 5.7% vs 53 ± 8.0% respectively, p = 0.698). MVR resulted in lower residual MR-regurgitant fraction (RF) than MVr (12 ± 8.0% vs 21 ± 11% respectively, p = 0.022). Baseline and follow-up indices of diffuse and focal myocardial fibrosis (Native T1 relaxation times, extra-cellular volume and quantified LGE respectively) were comparable between groups. Stepwise multiple linear regression of indexed variables in the surgical groups demonstrated baseline indexed mitral regurgitant volume as the sole multivariate predictor of left ventricular (LV) end-diastolic reverse remodelling, baseline LVEDVi as the most significant independent multivariate predictor of follow-up LVEDVi, baseline indexed LV end-systolic volume as the sole multivariate predictor of follow-up LV ejection fraction and undergoing MVR (vs MVr) as the most significant (p < 0.001) baseline multivariate predictor of lower residual MR.ConclusionIn primary MR, MVR with chordal preservation may offer comparable cardiac reverse remodeling and functional benefits at 6-months when compared to MVr. Larger, multicenter CMR studies are required, which if the findings are confirmed could impact future surgical practice.

Dataset Information

Machine-learning-based prediction of survival and mitral regurgitation recurrence in patients undergoing mitral valve repair.

Objectives

Methods

Results

Conclusions

Publications

Machine-learning-based prediction of survival and mitral regurgitation recurrence in patients undergoing mitral valve repair.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets