Constanze Schmidt
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Constanze Schmidt has a total of 19 co-authors. Here are the top co-authors by number of shared publications:
- Felix Wiedmann
Department of Cardiology, Angiology and Pneumology, University Hospital Heidelberg Heidelberg, Germany; DZHK (German Center for Cardiovascular Research), Partner Site Heidelberg/Mannheim, University of Heidelberg Heidelberg Germany
3 shared publications - Helge Haarmann
Department of Cardiology and Pneumology University Medical Center Göttingen (UMG) Göttingen Germany
2 shared publications - Leonard Bergau
Department of Cardiology and Pneumology University Medical Center Göttingen (UMG) Göttingen Germany
2 shared publications - Nibras Soubh
Department of Cardiology and Pneumology University Medical Center Göttingen (UMG) Göttingen Germany
2 shared publications - Christoph Beyersdorf
Department for Orthopedics and Trauma Surgery, Medical Faculty, University Hospital Düsseldorf, Heinrich Heine University Düsseldorf, 40225 Dusseldorf, Germany
1 shared publication Eva Rasenack
1 shared publication
Latest Publications
- Open Access
- 3 June 2026
- Published by JMIR Publications Inc. in Journal of Medical Internet Research
- Vol. 28 (1) , e86692
Performance of Vision-Enabled Large Language Models in Image-Based Electrocardiogram Interpretation: Exploratory Evaluation
Background: Vision-enabled large language models (VE-LLMs) have the potential to provide flexible and explainable medical image interpretation. However, their real-world performance on clinical data, such as 12-lead electrocardiograms (ECGs), has not been systematically assessed. Objective: This study aimed to evaluate the diagnostic accuracy and reliability of state-of-the-art generalist VE-LLMs in interpreting real-world ECG images. Methods: We tested 6 generalist VE-LLMs (ChatGPT-5, ChatGPT-4, Gemini 2.5, Copilot, Claude Sonnet-4, and Claude Opus-4.1) using 70 deidentified ECG images. A standardized prompt requested 9 determinations: rhythm, first-degree atrioventricular (AV) block, intraventricular conduction block and pattern, corrected QT (QTc) prolongation, premature atrial and ventricular contractions, ischemic ST-segment deviation, and axis deviation. An expert consensus served as the reference standard. Moreover, 2 image-based ECG-specialized LLMs (PULSE-7B and ECG-Instruct-Llama-3.2-11B-Vision) were tested for exploratory comparison. Model outputs were evaluated using overall and per-category diagnostic metrics. Results: Overall balanced accuracy across generalist models ranged from 50.1% to 61.8% (Cochran Q, P <.001). ChatGPT-5 achieved the highest balanced accuracy (61.8%) but had the slowest response time (median 276, IQR 110-407 s), whereas Copilot responded within a median of 3 (IQR 2-4) seconds. Balanced accuracy for rhythm classification ranged from 38.6% to 55.8%, but sensitivity for atrial fibrillation among generalist models was ≤11.1%, detecting either none or only 1 of the 9 cases. Detection of first-degree AV block (sensitivity 0%‐22%; 0/9 to 2/9) and QTc prolongation (sensitivity 0%‐45.5%; 0/22 to 10/22) was poor. Intraventricular block was identified with up to 67.8% balanced accuracy, but correct subtype assignment was ≤44% (≤11/25). ST-segment deviation sensitivity was <25% for all generalist models (highest 3/14). Agreement with expert interpretation was low, with Cohen κ indicating poor-to-fair concordance (κ≤0.39). Specialized models achieved overall balanced accuracy of 56.5% (ECG-Instruct-Llama-3.2-11B-Vision) and 64.4% (PULSE-7B), with PULSE-7B showing higher task-specific balanced accuracy in rhythm classification and ectopic beats detection (up to 86.3% and 89.2%, respectively). Conclusions: VE-LLMs showed moderate overall performance but mostly low sensitivity and limited agreement with expert ECG interpretation. Current performance remains inconsistent across models and diagnostic categories and is insufficient to support clinical deployment.
- 3 June 2026
- Published by JMIR Publications Inc. in Journal of Medical Internet Research
- Vol. 28 (1) , e86692
Open AccessPerformance of Vision-Enabled Large Language Models in Image-Based Electrocardiogram Interpretation: Exploratory Evaluation
Background: Vision-enabled large language models (VE-LLMs) have the potential to provide flexible and explainable medical image interpretation. However, their real-world performance on clinical data, such as 12-lead electrocardiograms (ECGs), has not been systematically assessed. Objective: This study aimed to evaluate the diagnostic accuracy and reliability of state-of-the-art generalist VE-LLMs in interpreting real-world ECG images. Methods: We tested 6 generalist VE-LLMs (ChatGPT-5, ChatGPT-4, Gemini 2.5, Copilot, Claude Sonnet-4, and Claude Opus-4.1) using 70 deidentified ECG images. A standardized prompt requested 9 determinations: rhythm, first-degree atrioventricular (AV) block, intraventricular conduction block and pattern, corrected QT (QTc) prolongation, premature atrial and ventricular contractions, ischemic ST-segment deviation, and axis deviation. An expert consensus served as the reference standard. Moreover, 2 image-based ECG-specialized LLMs (PULSE-7B and ECG-Instruct-Llama-3.2-11B-Vision) were tested for exploratory comparison. Model outputs were evaluated using overall and per-category diagnostic metrics. Results: Overall balanced accuracy across generalist models ranged from 50.1% to 61.8% (Cochran Q, P <.001). ChatGPT-5 achieved the highest balanced accuracy (61.8%) but had the slowest response time (median 276, IQR 110-407 s), whereas Copilot responded within a median of 3 (IQR 2-4) seconds. Balanced accuracy for rhythm classification ranged from 38.6% to 55.8%, but sensitivity for atrial fibrillation among generalist models was ≤11.1%, detecting either none or only 1 of the 9 cases. Detection of first-degree AV block (sensitivity 0%‐22%; 0/9 to 2/9) and QTc prolongation (sensitivity 0%‐45.5%; 0/22 to 10/22) was poor. Intraventricular block was identified with up to 67.8% balanced accuracy, but correct subtype assignment was ≤44% (≤11/25). ST-segment deviation sensitivity was <25% for all generalist models (highest 3/14). Agreement with expert interpretation was low, with Cohen κ indicating poor-to-fair concordance (κ≤0.39). Specialized models achieved overall balanced accuracy of 56.5% (ECG-Instruct-Llama-3.2-11B-Vision) and 64.4% (PULSE-7B), with PULSE-7B showing higher task-specific balanced accuracy in rhythm classification and ectopic beats detection (up to 86.3% and 89.2%, respectively). Conclusions: VE-LLMs showed moderate overall performance but mostly low sensitivity and limited agreement with expert ECG interpretation. Current performance remains inconsistent across models and diagnostic categories and is insufficient to support clinical deployment.
- Open Access
- 10 December 2025
- Published by Oxford University Press (OUP) in Journal of Bone and Mineral Research
- Vol. 41 (3) , 220-230
Mechanosensitive ion channels as novel targets in osteoporosis
Osteoporosis is the most prevalent metabolic bone disease globally, leading to an increased risk of fractures. Recent advances in ion channel research have shed light on the importance of mechanosensitive ion channels as novel players in these pathophysiological processes. This perspective discusses the involvement of the mechanosensitive ion channels TREK-1, Piezo, and volume-regulated anion channels (VRACs) as potential novel pharmacological targets for the treatment of osteoporosis. TREK-1, a mechanosensitive K2P channel is important for maintaining the resting membrane potential in many cells, including osteoblasts and osteoclasts. K2P channels regulate osteoblast proliferation and differentiation, as well as osteoclast activity, potentially modulating bone remodeling in osteoporosis. Piezo channels influence osteoblast differentiation and osteoclast activity by modulating calcium influx, which is crucial for osteogenic signaling pathways, such as Wnt/β-catenin and ERK1/2. Piezo1 activation promotes bone formation, while its deficiency leads to impaired osteogenesis and increased bone resorption. Volume-regulated anion channels have been shown to be involved in osteoblast adaptation to mechanical stress and macrophage polarization, which indicates their importance for bone homeostasis. Chronic inflammation is a major contributor to osteoporosis progression. Evidence of ion channel involvement in this process has emerged in recent years. Specifically, macrophage function in osteoporosis seems to be linked to ion channel activity. Inflammatory polarization of macrophages is a key player in inflammation-induced bone loss and can be driven by mechanosensitive ion channels. Modulating these ion channels may provide new therapeutic opportunities. Given the complexity of ion channel interactions in bone cells and their regulatory role in bone remodeling, understanding their precise function in osteoporosis is essential. Targeted modulation of mechanosensitive ion channels holds promise as a novel therapeutic approach to mitigate inflammation-driven bone loss and improve bone density. Further research into their role in osteoclasts and macrophage-driven bone degradation will aid in developing innovative osteoporosis treatments.
- 10 December 2025
- Published by Oxford University Press (OUP) in Journal of Bone and Mineral Research
- Vol. 41 (3) , 220-230
Open AccessMechanosensitive ion channels as novel targets in osteoporosis
Osteoporosis is the most prevalent metabolic bone disease globally, leading to an increased risk of fractures. Recent advances in ion channel research have shed light on the importance of mechanosensitive ion channels as novel players in these pathophysiological processes. This perspective discusses the involvement of the mechanosensitive ion channels TREK-1, Piezo, and volume-regulated anion channels (VRACs) as potential novel pharmacological targets for the treatment of osteoporosis. TREK-1, a mechanosensitive K2P channel is important for maintaining the resting membrane potential in many cells, including osteoblasts and osteoclasts. K2P channels regulate osteoblast proliferation and differentiation, as well as osteoclast activity, potentially modulating bone remodeling in osteoporosis. Piezo channels influence osteoblast differentiation and osteoclast activity by modulating calcium influx, which is crucial for osteogenic signaling pathways, such as Wnt/β-catenin and ERK1/2. Piezo1 activation promotes bone formation, while its deficiency leads to impaired osteogenesis and increased bone resorption. Volume-regulated anion channels have been shown to be involved in osteoblast adaptation to mechanical stress and macrophage polarization, which indicates their importance for bone homeostasis. Chronic inflammation is a major contributor to osteoporosis progression. Evidence of ion channel involvement in this process has emerged in recent years. Specifically, macrophage function in osteoporosis seems to be linked to ion channel activity. Inflammatory polarization of macrophages is a key player in inflammation-induced bone loss and can be driven by mechanosensitive ion channels. Modulating these ion channels may provide new therapeutic opportunities. Given the complexity of ion channel interactions in bone cells and their regulatory role in bone remodeling, understanding their precise function in osteoporosis is essential. Targeted modulation of mechanosensitive ion channels holds promise as a novel therapeutic approach to mitigate inflammation-driven bone loss and improve bone density. Further research into their role in osteoclasts and macrophage-driven bone degradation will aid in developing innovative osteoporosis treatments.
- Open Access
- 31 October 2025
- Published by JMIR Publications Inc. in JMIR Preprints
Performance of Vision-Enabled Large Language Models in ECG Interpretation: Exploratory Evaluation (Preprint)
BACKGROUND Vision-enabled large language models (VE-LLMs) have the potential to provide flexible and explainable medical image interpretation. However, their real-world performance on clinical data such as 12-lead electrocardiograms (ECGs) has not been systematically assessed. OBJECTIVE This study aimed to evaluate the diagnostic accuracy and reliability of state-of-the-art VE-LLMs in interpreting real-world ECG images. METHODS We tested eight VE-LLMs (ChatGPT-5, ChatGPT-4, Gemini 2.5 Pro, Copilot, Grok-4, Perplexity, Claude Sonnet-4, and Claude Opus-4.1) using 70 de-identified ECG images. A standardized prompt requested nine determinations: rhythm, first-degree atrioventricular (AV) block, intraventricular conduction block and pattern, corrected QT (QTc) prolongation, premature atrial and ventricular contractions, ischemic ST-segment deviation, and axis deviation. An expert consensus served as the reference standard. Model outputs were evaluated using overall and per-class diagnostic metrics. RESULTS Overall accuracy across models varied significantly from 68.1% to 78.3% (429/630 to 493/630, Cochran’s Q, P<.001). ChatGPT-5 achieved the highest accuracy (78.3%) but had the slowest response time (median 276 seconds), whereas Perplexity and Copilot responded within a median of 2 and 3 seconds, respectively. Rhythm classification reached 72.9%–82.9% accuracy (51/70 to 58/70), but sensitivity for atrial fibrillation was ≤22% (≤2/9). Detection of first-degree AV block was poor (sensitivity 0%–33%; 0/9 to 3/9), and QTc prolongation was also poor (sensitivity 0%–45.5%; 0/22 to 10/22). Intraventricular block was identified with up to 70% accuracy (49/70), but correct subtype assignment was ≤44% (≤11/25). ST-segment deviation sensitivity was <25% for all models (highest 3/14). Agreement with expert interpretation was low, with Cohen’s kappa (κ) indicating poor-to-fair concordance (κ≤.37). CONCLUSIONS VE-LLMs showed moderate overall accuracy but low sensitivity and limited agreement with expert ECG interpretation. Current performance is inconsistent and insufficient for clinical deployment. Future development should focus on domain-specific training and hybrid approaches combining LLM reasoning with established ECG algorithms before use in patient care.
- 31 October 2025
- Published by JMIR Publications Inc. in JMIR Preprints
Open AccessPerformance of Vision-Enabled Large Language Models in ECG Interpretation: Exploratory Evaluation (Preprint)
BACKGROUND Vision-enabled large language models (VE-LLMs) have the potential to provide flexible and explainable medical image interpretation. However, their real-world performance on clinical data such as 12-lead electrocardiograms (ECGs) has not been systematically assessed. OBJECTIVE This study aimed to evaluate the diagnostic accuracy and reliability of state-of-the-art VE-LLMs in interpreting real-world ECG images. METHODS We tested eight VE-LLMs (ChatGPT-5, ChatGPT-4, Gemini 2.5 Pro, Copilot, Grok-4, Perplexity, Claude Sonnet-4, and Claude Opus-4.1) using 70 de-identified ECG images. A standardized prompt requested nine determinations: rhythm, first-degree atrioventricular (AV) block, intraventricular conduction block and pattern, corrected QT (QTc) prolongation, premature atrial and ventricular contractions, ischemic ST-segment deviation, and axis deviation. An expert consensus served as the reference standard. Model outputs were evaluated using overall and per-class diagnostic metrics. RESULTS Overall accuracy across models varied significantly from 68.1% to 78.3% (429/630 to 493/630, Cochran’s Q, P<.001). ChatGPT-5 achieved the highest accuracy (78.3%) but had the slowest response time (median 276 seconds), whereas Perplexity and Copilot responded within a median of 2 and 3 seconds, respectively. Rhythm classification reached 72.9%–82.9% accuracy (51/70 to 58/70), but sensitivity for atrial fibrillation was ≤22% (≤2/9). Detection of first-degree AV block was poor (sensitivity 0%–33%; 0/9 to 3/9), and QTc prolongation was also poor (sensitivity 0%–45.5%; 0/22 to 10/22). Intraventricular block was identified with up to 70% accuracy (49/70), but correct subtype assignment was ≤44% (≤11/25). ST-segment deviation sensitivity was <25% for all models (highest 3/14). Agreement with expert interpretation was low, with Cohen’s kappa (κ) indicating poor-to-fair concordance (κ≤.37). CONCLUSIONS VE-LLMs showed moderate overall accuracy but low sensitivity and limited agreement with expert ECG interpretation. Current performance is inconsistent and insufficient for clinical deployment. Future development should focus on domain-specific training and hybrid approaches combining LLM reasoning with established ECG algorithms before use in patient care.
- 23 October 2024
- Published by Oxford University Press (OUP) in Journal of Experimental Botany
- Vol. 76 (2) , 445-460
Hybrid grain production in wheat benefits from synchronized flowering and high female flower receptivity
The performance of plant hybrids relative to line breeding types is generally associated with higher yields, better adaptation, and improved yield stability. In bread wheat (Triticum aestivum L.), however, a broad commercial success for hybrids has not been accomplished until now largely due to the low efficiency of hybrid grain production, which is highly attributable to its self-pollinating nature. To better understand how hybrid wheat grains can be produced more effectively, we investigated the influence of synchronized flowering between female, i.e. male-sterile, lines and their male cross-pollinator lines as well as of the duration of flowering on hybrid grain production. We found that synchronization of flowering in combination with the longest possible temporal overlap had the largest positive effect on hybrid grain production. However, despite sufficient spatial and temporal synchronization of flowering, we also found that some female lines had lower hybrid grain set than others, suggesting genetic differences in female floral receptivity. To better assess female receptivity, we established a new phenotyping scale of male-sterile wheat flowers that provides the floral basics for effective cross-pollination. Applying this scale in our field and greenhouse trials revealed that better performing female lines remained longer in the pollen-receptive phase.
- 23 October 2024
- Published by Oxford University Press (OUP) in Journal of Experimental Botany
- Vol. 76 (2) , 445-460
Hybrid grain production in wheat benefits from synchronized flowering and high female flower receptivity
The performance of plant hybrids relative to line breeding types is generally associated with higher yields, better adaptation, and improved yield stability. In bread wheat (Triticum aestivum L.), however, a broad commercial success for hybrids has not been accomplished until now largely due to the low efficiency of hybrid grain production, which is highly attributable to its self-pollinating nature. To better understand how hybrid wheat grains can be produced more effectively, we investigated the influence of synchronized flowering between female, i.e. male-sterile, lines and their male cross-pollinator lines as well as of the duration of flowering on hybrid grain production. We found that synchronization of flowering in combination with the longest possible temporal overlap had the largest positive effect on hybrid grain production. However, despite sufficient spatial and temporal synchronization of flowering, we also found that some female lines had lower hybrid grain set than others, suggesting genetic differences in female floral receptivity. To better assess female receptivity, we established a new phenotyping scale of male-sterile wheat flowers that provides the floral basics for effective cross-pollination. Applying this scale in our field and greenhouse trials revealed that better performing female lines remained longer in the pollen-receptive phase.