1
|
van de Leur RR, de Brouwer R, Bleijendaal H, Verstraelen TE, Mahmoud B, Perez-Matos A, Dickhoff C, Schoonderwoerd BA, Germans T, Houweling A, van der Zwaag PA, Cox MGPJ, Peter van Tintelen J, Te Riele ASJM, van den Berg MP, Wilde AAM, Doevendans PA, de Boer RA, van Es R. ECG-only explainable deep learning algorithm predicts the risk for malignant ventricular arrhythmia in phospholamban cardiomyopathy. Heart Rhythm 2024:S1547-5271(24)00210-8. [PMID: 38403235 DOI: 10.1016/j.hrthm.2024.02.038] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 02/15/2024] [Accepted: 02/15/2024] [Indexed: 02/27/2024]
Abstract
BACKGROUND Phospholamban (PLN) p.(Arg14del) variant carriers are at risk for development of malignant ventricular arrhythmia (MVA). Accurate risk stratification allows timely implantation of intracardiac defibrillators and is currently performed with a multimodality prediction model. OBJECTIVE This study aimed to investigate whether an explainable deep learning-based approach allows risk prediction with only electrocardiogram (ECG) data. METHODS A total of 679 PLN p.(Arg14del) carriers without MVA at baseline were identified. A deep learning-based variational auto-encoder, trained on 1.1 million ECGs, was used to convert the 12-lead baseline ECG into its FactorECG, a compressed version of the ECG that summarizes it into 32 explainable factors. Prediction models were developed by Cox regression. RESULTS The deep learning-based ECG-only approach was able to predict MVA with a C statistic of 0.79 (95% CI, 0.76-0.83), comparable to the current prediction model (C statistic, 0.83 [95% CI, 0.79-0.88]; P = .054) and outperforming a model based on conventional ECG parameters (low-voltage ECG and negative T waves; C statistic, 0.65 [95% CI, 0.58-0.73]; P < .001). Clinical simulations showed that a 2-step approach, with ECG-only screening followed by a full workup, resulted in 60% less additional diagnostics while outperforming the multimodal prediction model in all patients. A visualization tool was created to provide interactive visualizations (https://pln.ecgx.ai). CONCLUSION Our deep learning-based algorithm based on ECG data only accurately predicts the occurrence of MVA in PLN p.(Arg14del) carriers, enabling more efficient stratification of patients who need additional diagnostic testing and follow-up.
Collapse
Affiliation(s)
- Rutger R van de Leur
- Department of Cardiology, University Medical Center Utrecht, Utrecht, The Netherlands.
| | - Remco de Brouwer
- Department of Cardiology, University Medical Center Groningen, Groningen, The Netherlands
| | - Hidde Bleijendaal
- Department of Cardiology, Amsterdam UMC location University of Amsterdam, Amsterdam, The Netherlands; European Reference Network for Rare, Low-Prevalence, or Complex Diseases of the Heart (ERN GUARD-Heart); Department of Clinical Epidemiology, Biostatistics and Bioinformatics, Amsterdam University Medical Center, University of Amsterdam, Amsterdam, The Netherlands
| | - Tom E Verstraelen
- Department of Cardiology, Amsterdam UMC location University of Amsterdam, Amsterdam, The Netherlands; European Reference Network for Rare, Low-Prevalence, or Complex Diseases of the Heart (ERN GUARD-Heart)
| | - Belend Mahmoud
- Department of Cardiology, University Medical Center Groningen, Groningen, The Netherlands
| | - Ana Perez-Matos
- Department of Cardiology, St Antonius Hospital, Sneek, The Netherlands
| | | | - Bas A Schoonderwoerd
- Department of Cardiology, Medical Centre Leeuwarden, Leeuwarden, The Netherlands
| | - Tjeerd Germans
- Department of Cardiology, Noordwest Hospital Group, Alkmaar, The Netherlands
| | - Arjan Houweling
- Department of Human Genetics, Amsterdam University Medical Center, Amsterdam, The Netherlands
| | - Paul A van der Zwaag
- Department of Genetics, University Medical Center Groningen, Groningen, The Netherlands
| | - Moniek G P J Cox
- Department of Cardiology, University Medical Center Groningen, Groningen, The Netherlands
| | - J Peter van Tintelen
- European Reference Network for Rare, Low-Prevalence, or Complex Diseases of the Heart (ERN GUARD-Heart); Department of Genetics, University Medical Center Utrecht, Utrecht, The Netherlands
| | | | - Maarten P van den Berg
- Department of Cardiology, University Medical Center Groningen, Groningen, The Netherlands
| | - Arthur A M Wilde
- Department of Cardiology, Amsterdam UMC location University of Amsterdam, Amsterdam, The Netherlands; European Reference Network for Rare, Low-Prevalence, or Complex Diseases of the Heart (ERN GUARD-Heart)
| | - Pieter A Doevendans
- Department of Cardiology, University Medical Center Utrecht, Utrecht, The Netherlands; European Reference Network for Rare, Low-Prevalence, or Complex Diseases of the Heart (ERN GUARD-Heart); Netherlands Heart Institute, Utrecht, The Netherlands; Central Military Hospital, Utrecht, The Netherlands
| | - Rudolf A de Boer
- Department of Cardiology, University Medical Center Groningen, Groningen, The Netherlands; Department of Cardiology, Erasmus Medical Center, Rotterdam, The Netherlands
| | - René van Es
- Department of Cardiology, University Medical Center Utrecht, Utrecht, The Netherlands
| |
Collapse
|
2
|
Abstract
Mutants of Escherichia coli and Klebsiella aerogenes that are deficient in glutamate synthase (glutamate-oxoglutarate amidotransferase [GOGAT]) activity have difficulty growing with nitrogen sources other than ammonia. Two models have been proposed to account for this inability to grow. One model postulated an imbalance between glutamine synthesis and glutamine degradation that led to a repression of the Ntr system and the subsequent failure to activate transcription of genes required for the use of alternative nitrogen sources. The other model postulated that mutations in gltB or gltD (which encode the subunits of GOGAT) were polar on a downstream gene, gltF, which is necessary for proper activation of gene expression by the Ntr system. The data reported here show that the gltF model is incorrect for three reasons: first, a nonpolar gltB and a polar gltD mutation of K. aerogenes both show the same phenotype; second, K. aerogenes and several other enteric bacteria lack a gene homologous to gltF; and third, mutants of E. coli whose gltF gene has been deleted show no defect in nitrogen metabolism. The argument that accumulated glutamine represses the Ntr system in gltB or gltD mutants is also incorrect, because these mutants can derepress the Ntr system normally so long as sufficient glutamate is supplied. Thus, we conclude that gltB or gltD mutants grow slowly on many poor nitrogen sources because they are starved for glutamate. Much of the glutamate formed by catabolism of alternative nitrogen sources is converted to glutamine, which cannot be efficiently converted to glutamate in the absence of GOGAT activity. Finally, GOGAT-deficient E. coli cells growing with glutamine as the sole nitrogen source increase their synthesis of the other glutamate-forming enzyme, glutamate dehydrogenase, severalfold, but this is still insufficient to allow rapid growth under these conditions.
Collapse
Affiliation(s)
- T J Goss
- Department of Biology, The University of Michigan, Ann Arbor 48109-1048, USA
| | | | | |
Collapse
|
3
|
Janes BK, Pomposiello PJ, Perez-Matos A, Najarian DJ, Goss TJ, Bender RA. Growth inhibition caused by overexpression of the structural gene for glutamate dehydrogenase (gdhA) from Klebsiella aerogenes. J Bacteriol 2001; 183:2709-14. [PMID: 11274137 PMCID: PMC95194 DOI: 10.1128/jb.183.8.2709-2714.2001] [Citation(s) in RCA: 17] [Impact Index Per Article: 0.7] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/20/2022] Open
Abstract
Two linked mutations affecting glutamate dehydrogenase (GDH) formation (gdh-1 and rev-2) had been isolated at a locus near the trp cluster in Klebsiella aerogenes. The properties of these two mutations were consistent with those of a locus containing either a regulatory gene or a structural gene. The gdhA gene from K. aerogenes was cloned and sequenced, and an insertion mutation was generated and shown to be linked to trp. A region of gdhA from a strain bearing gdh-1 was sequenced and shown to have a single-base-pair change, confirming that the locus defined by gdh-1 is the structural gene for GDH. Mutants with the same phenotype as rev-2 were isolated, and their sequences showed that the mutations were located in the promoter region of the gdhA gene. The linkage of gdhA to trp in K. aerogenes was explained by postulating an inversion of the genetic map relative to other enteric bacteria. Strains that bore high-copy-number clones of gdhA displayed an auxotrophy that was interpreted as a limitation for alpha-ketoglutarate and consequently for succinyl-coenzyme A (CoA). Three lines of evidence supported this interpretation: high-copy-number clones of the enzymatically inactive gdhA1 allele showed no auxotrophy, repression of GDH expression by the nitrogen assimilation control protein (NAC) relieved the auxotrophy, and addition of compounds that could increase the alpha-ketoglutarate supply or reduce the succinyl-CoA requirement relieved the auxotrophy.
Collapse
Affiliation(s)
- B K Janes
- Department of Biology, The University of Michigan, Ann Arbor, Michigan 48109-1048, USA
| | | | | | | | | | | |
Collapse
|