Thomas Schmitz
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Thomas Schmitz has a total of 22 co-authors. Here are the top co-authors by number of shared publications:
- Dietmar Jannach
University of Klagenfurt, Klagenfurt, Austria
3 shared publications - Birgit Hofer
Graz University of Technology, Austria
2 shared publications - Franz Wotawa
Graz University of Technology
2 shared publications - Alexander Wolf
Heart Center Wuppertal, Witten-Herdecke University, Wuppertal, Germany (A.W.).
1 shared publication Boris Schmidt
Frankfurt, Germany
1 shared publication- Corrado Tamburino
Centro Alte Specialità e Trapianti (CAST), Azienda Ospedaliero‐Universitaria Policlinico Vittorio Emanuele, University of Catania
1 shared publication
Latest Publications
- Open Access
- 19 April 2022
- Published by Wiley in Clinical Cardiology
- Vol. 45 (6) , 657-663
A deep neural network using audio files for detection of aortic stenosis
Background: Although aortic stenosis (AS) is the most common valvular heart disease in the western world, many affected patients remain undiagnosed. Auscultation is a readily available screening tool for AS. However, it requires a high level of professional expertise.Hypothesis: An AI algorithm can detect AS using audio files with the same accuracy as experienced cardiologists.Methods: A deep neural network (DNN) was trained by preprocessed audio files of 100 patients with AS and 100 controls. The DNN's performance was evaluated with a test data set of 40 patients. The primary outcome measures were sensitivity, specificity, and F1‐score. Results of the DNN were compared with the performance of cardiologists, residents, and medical students.Results: Eighteen percent of patients without AS and 22% of patients with AS showed an additional moderate or severe mitral regurgitation. The DNN showed a sensitivity of 0.90 (0.81–0.99), a specificity of 1, and an F1‐score of 0.95 (0.89–1.0) for the detection of AS. In comparison, we calculated an F1‐score of 0.94 (0.86–1.0) for cardiologists, 0.88 (0.78–0.98) for residents, and 0.88 (0.78–0.98) for students.Conclusions: The present study shows that deep learning‐guided auscultation predicts significant AS with similar accuracy as cardiologists. The results of this pilot study suggest that AI‐assisted auscultation may help general practitioners without special cardiology training in daily practice.
- 19 April 2022
- Published by Wiley in Clinical Cardiology
- Vol. 45 (6) , 657-663
Open AccessA deep neural network using audio files for detection of aortic stenosis
Background: Although aortic stenosis (AS) is the most common valvular heart disease in the western world, many affected patients remain undiagnosed. Auscultation is a readily available screening tool for AS. However, it requires a high level of professional expertise.Hypothesis: An AI algorithm can detect AS using audio files with the same accuracy as experienced cardiologists.Methods: A deep neural network (DNN) was trained by preprocessed audio files of 100 patients with AS and 100 controls. The DNN's performance was evaluated with a test data set of 40 patients. The primary outcome measures were sensitivity, specificity, and F1‐score. Results of the DNN were compared with the performance of cardiologists, residents, and medical students.Results: Eighteen percent of patients without AS and 22% of patients with AS showed an additional moderate or severe mitral regurgitation. The DNN showed a sensitivity of 0.90 (0.81–0.99), a specificity of 1, and an F1‐score of 0.95 (0.89–1.0) for the detection of AS. In comparison, we calculated an F1‐score of 0.94 (0.86–1.0) for cardiologists, 0.88 (0.78–0.98) for residents, and 0.88 (0.78–0.98) for students.Conclusions: The present study shows that deep learning‐guided auscultation predicts significant AS with similar accuracy as cardiologists. The results of this pilot study suggest that AI‐assisted auscultation may help general practitioners without special cardiology training in daily practice.
- Open Access
- 22 December 2018
- Published by Springer Nature in Automated Software Engineering
- Vol. 26 (1) , 203-239
Fragment-based spreadsheet debugging
- 22 December 2018
- Published by Springer Nature in Automated Software Engineering
- Vol. 26 (1) , 203-239
Open AccessFragment-based spreadsheet debugging
- Open Access
- 10 June 2016
- Published by Oxford University Press (OUP) in EP Europace
- Vol. 18 (suppl_1) , i26
22-04: The impact of underlying heart rhythm on the incidence of pericardial effusion after left atrial appendage occlusion using Watchman – Data from the EWOLUTION registry
Introduction: Pericardial effusion (PE) with or without tamponade physiology as well as device embolization (DE) is a rare but serious complication following left atrial appendage closure (LAAC). It may be speculated that LAA contraction during sinus rhythm (SR) exerts mechanical force on the device that eventually lead to PE and/or DE. We sought to determine the incidence and predictors of PE and DE following LAAC using Watchman (Boston Scientific) with special emphasis on the underlying heart rhythm during implant. Methods: From 47 centres in 13 European countries 1005 patients with spontaneous heart rhythm at implant (mean age 73 ± 9 years, 60% male, median CHA2DS2-VASc = 4 (range 0–9)) underwent LAAC. Procedural and follow-up data were collected via electronic case report forms in the EWOLUTION registry database. A Cox-proportional hazard model was calculated adjusting for multiple variables: age, gender, number of recaptures and device oversizing. Results: During implant, 418 (42%) and 587 (58%) patients were in SR and atrial fibrillation (AF), respectively. AF patients were older (75 ± 8 versus 71 ± 9 years; p < 0.0001) and displayed a higher median CHA2DS2-VASc score (5 versus 4; p < 0.0001). The device was successfully implanted in 98.6% of cases without differences between groups. At least one device recapture was performed in 24% and 32% of cases in SR and AF, respectively. Device compression was 18% and 16% in the AF and SR group, respectively (p = 0.002). One DE was observed in each group within the first month of implant. PE/tamponade rate was significantly lower in AF (n = 1; 0.2%) than in SR (n = 5; 1.2%) at day 7 (log-rank p = 0.04) as well as at day 30 post implant (n = 1; 0.2% versus n = 7; 1.7%) (log-rank p = 0.01). No PE requiring intervention occurred in the AF group at day 30 compared to 5 events (1.2%) in the SR group (p = 0.01). The multivariate analysis did not identify SR as an independent predictor of PE/tamponade through 7 days (HR 4.5; p = 0.18) or 30 days (HR 6.9; p = 0.08) after adjusting for age, gender, oversizing and (only at 30 days) number of recaptures. Conclusion: The overall incidence of PE and DE was very low after LAAC using Watchman. Although SR was not identified as an independent predictor of PE/tamponade, all events requiring intervention occurred in patients with SR. It may be advisable to perform an extended echocardiographic follow-up in that patient population.
- 10 June 2016
- Published by Oxford University Press (OUP) in EP Europace
- Vol. 18 (suppl_1) , i26
Open Access22-04: The impact of underlying heart rhythm on the incidence of pericardial effusion after left atrial appendage occlusion using Watchman – Data from the EWOLUTION registry
Introduction: Pericardial effusion (PE) with or without tamponade physiology as well as device embolization (DE) is a rare but serious complication following left atrial appendage closure (LAAC). It may be speculated that LAA contraction during sinus rhythm (SR) exerts mechanical force on the device that eventually lead to PE and/or DE. We sought to determine the incidence and predictors of PE and DE following LAAC using Watchman (Boston Scientific) with special emphasis on the underlying heart rhythm during implant. Methods: From 47 centres in 13 European countries 1005 patients with spontaneous heart rhythm at implant (mean age 73 ± 9 years, 60% male, median CHA2DS2-VASc = 4 (range 0–9)) underwent LAAC. Procedural and follow-up data were collected via electronic case report forms in the EWOLUTION registry database. A Cox-proportional hazard model was calculated adjusting for multiple variables: age, gender, number of recaptures and device oversizing. Results: During implant, 418 (42%) and 587 (58%) patients were in SR and atrial fibrillation (AF), respectively. AF patients were older (75 ± 8 versus 71 ± 9 years; p < 0.0001) and displayed a higher median CHA2DS2-VASc score (5 versus 4; p < 0.0001). The device was successfully implanted in 98.6% of cases without differences between groups. At least one device recapture was performed in 24% and 32% of cases in SR and AF, respectively. Device compression was 18% and 16% in the AF and SR group, respectively (p = 0.002). One DE was observed in each group within the first month of implant. PE/tamponade rate was significantly lower in AF (n = 1; 0.2%) than in SR (n = 5; 1.2%) at day 7 (log-rank p = 0.04) as well as at day 30 post implant (n = 1; 0.2% versus n = 7; 1.7%) (log-rank p = 0.01). No PE requiring intervention occurred in the AF group at day 30 compared to 5 events (1.2%) in the SR group (p = 0.01). The multivariate analysis did not identify SR as an independent predictor of PE/tamponade through 7 days (HR 4.5; p = 0.18) or 30 days (HR 6.9; p = 0.08) after adjusting for age, gender, oversizing and (only at 30 days) number of recaptures. Conclusion: The overall incidence of PE and DE was very low after LAAC using Watchman. Although SR was not identified as an independent predictor of PE/tamponade, all events requiring intervention occurred in patients with SR. It may be advisable to perform an extended echocardiographic follow-up in that patient population.
- 12 February 2014
- Published by Springer Nature in Automated Software Engineering
- Vol. 23 (1) , 105-144
Model-based diagnosis of spreadsheet programs: a constraint-based debugging approach
- 12 February 2014
- Published by Springer Nature in Automated Software Engineering
- Vol. 23 (1) , 105-144
Model-based diagnosis of spreadsheet programs: a constraint-based debugging approach