Lenka Horakova
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Lenka Horakova has a total of 16 co-authors. Here are the top co-authors by number of shared publications:
- Karel Roubík
Faculty of Biomedical Engineering, Czech Technical University in Prague, Kladno, Czech Republic
2 shared publications Josef Skola
Department of Anaesthesiology, Perioperative and Intensive Care Medicine, Masaryk Hospital in Usti nad Labem, Faculty of Health Studies, J. E. Purkinje University in Usti nad Labem, Usti nad Labem, Czech Republic
1 shared publication- Josef Škola
Department of Anaesthesiology, Intensive Care and Emergency Medicine, Bulovka University Hospital and Third Medical Faculty, Charles University, Prague 8, Czechia
1 shared publication Karel Roubik
Czech Technical University in Prague
1 shared publication- Karel Sýkora
Charles University in Prague
1 shared publication Ladislav Bis
Czech Technical University in Prague
1 shared publication
Latest Publications
- Open Access
- 30 September 2026
- Published by Springer Nature in Neurocritical Care
- p. 1-13
Data for Machine Learning Models: Clinical Experience Determines Interrater Reliability in Waveform Anomaly Detection
Background: High-frequency intracranial pressure (ICP) and arterial blood pressure (ABP) waveforms are increasingly utilized for advanced physiological monitoring and as inputs for machine learning (ML) models. The accuracy of derived parameters relies on input data quality, and anomaly contaminated signals may impair model performance. Reliable anomaly annotation is therefore essential. Despite progress in automated labeling, human annotations remain the “ground truth” for ML training approaches. Human labeling is subjective and may vary depending on the signal type and the annotator’s experience. This study aimed to quantify the interrater reliability of anomaly annotations in ICP and ABP waveforms. Methods: Anonymized, synchronized ICP and ABP waveforms were analyzed retrospectively. From a single-center neurocritical care database, 100 h of recordings (32,760 10 s segments) were selected using predefined criteria. Six annotators (three neurocritical care consultants, two residents, one medical student) annotated anomalies under identical conditions. Interrater reliability was assessed for ICP and ABP using multiple metrics [Gwet’s AC2, Fleiss’ κ , prevalence- and bias-adjusted kappa (PABAK), positive/negative agreement, and intraclass correlation coefficient, ICC(2, k )], with segment-level and patient-level cluster-robust 95% confidence intervals. A broader anomaly definition, rather than artifact-based labeling, was used to capture variability in judgments of signal fidelity. Results: Anomaly prevalence was low (3.99% for ICP, 5.83% for ABP). Overall agreement was high for both signals (Gwet’s AC2 ≥ 0.98). However, anomaly specific agreement differed by modality and experience. For ICP, senior annotators achieved higher positive agreement than mixed-experience teams (+ 12.8%), whereas ABP annotation showed no significant experience effect. Across all annotators, ABP exhibited higher anomaly specific concordance, with similar overall agreement. Conclusions: Robust interrater reliability in waveform anomaly annotation is achievable in routine neurocritical care; however, reliability remained signal- and experience-dependent. ICP annotation benefits from expert training and supervision, while ABP can be reliably annotated by mixed-experience teams. These findings support signal-specific annotation strategies and provide an empirical basis for developing ML-based anomaly detection systems.
- 30 September 2026
- Published by Springer Nature in Neurocritical Care
- p. 1-13
Open AccessData for Machine Learning Models: Clinical Experience Determines Interrater Reliability in Waveform Anomaly Detection
Background: High-frequency intracranial pressure (ICP) and arterial blood pressure (ABP) waveforms are increasingly utilized for advanced physiological monitoring and as inputs for machine learning (ML) models. The accuracy of derived parameters relies on input data quality, and anomaly contaminated signals may impair model performance. Reliable anomaly annotation is therefore essential. Despite progress in automated labeling, human annotations remain the “ground truth” for ML training approaches. Human labeling is subjective and may vary depending on the signal type and the annotator’s experience. This study aimed to quantify the interrater reliability of anomaly annotations in ICP and ABP waveforms. Methods: Anonymized, synchronized ICP and ABP waveforms were analyzed retrospectively. From a single-center neurocritical care database, 100 h of recordings (32,760 10 s segments) were selected using predefined criteria. Six annotators (three neurocritical care consultants, two residents, one medical student) annotated anomalies under identical conditions. Interrater reliability was assessed for ICP and ABP using multiple metrics [Gwet’s AC2, Fleiss’ κ , prevalence- and bias-adjusted kappa (PABAK), positive/negative agreement, and intraclass correlation coefficient, ICC(2, k )], with segment-level and patient-level cluster-robust 95% confidence intervals. A broader anomaly definition, rather than artifact-based labeling, was used to capture variability in judgments of signal fidelity. Results: Anomaly prevalence was low (3.99% for ICP, 5.83% for ABP). Overall agreement was high for both signals (Gwet’s AC2 ≥ 0.98). However, anomaly specific agreement differed by modality and experience. For ICP, senior annotators achieved higher positive agreement than mixed-experience teams (+ 12.8%), whereas ABP annotation showed no significant experience effect. Across all annotators, ABP exhibited higher anomaly specific concordance, with similar overall agreement. Conclusions: Robust interrater reliability in waveform anomaly annotation is achievable in routine neurocritical care; however, reliability remained signal- and experience-dependent. ICP annotation benefits from expert training and supervision, while ABP can be reliably annotated by mixed-experience teams. These findings support signal-specific annotation strategies and provide an empirical basis for developing ML-based anomaly detection systems.
- Open Access
- 24 September 2025
- Published by Cambridge University Press (CUP) in Disaster Medicine and Public Health Preparedness
- Vol. 19, e271
From Rapid Manufacturing to Clinical Use: A Retrospective Observational Study of an Emergency Ventilator During the COVID-19 Pandemic
Objective: The COVID-19 pandemic spurred efforts to develop emergency ventilators, though few progressed beyond laboratory testing. This study evaluates CoroVent, a rapidly manufactured ventilator, during real-life deployment in critically ill COVID-19 patients amid ventilator shortages. Methods: This retrospective observational study included patients ventilated with emergency ventilator CoroVent. This device uses a novel and unique way of generating inspiratory flow and gas mixing using fast ON/OFF valves for air and oxygen, producing pneumatic pulses that are then smoothed into continuous flow. Clinical data were collected from 3 hospitals between October 2020 and March 2021, selected from 27 contacted, of which 23 responded and 4 reported clinical use. Results: Eight male patients (mean age 67 years, BMI 37.2 kg·m–2) with COVID-19 or suspected infection were ventilated with CoroVent for 31.3 (10.0–58.5) hours. The mean FiO2 was 71.4%, PEEP 10.6 cmH2O, and Vt 8.9 mL·kg–1 predicted body weight. CoroVent was used as an initial or replacement ventilator during device shortages. No major technical failures occurred. Conclusions: This is the first report of real-life clinical use of a rapidly manufactured emergency ventilator during the COVID-19 pandemic. The results confirm that such devices, if well-designed, can provide effective respiratory support when conventional ventilators are unavailable.
- 24 September 2025
- Published by Cambridge University Press (CUP) in Disaster Medicine and Public Health Preparedness
- Vol. 19, e271
Open AccessFrom Rapid Manufacturing to Clinical Use: A Retrospective Observational Study of an Emergency Ventilator During the COVID-19 Pandemic
Objective: The COVID-19 pandemic spurred efforts to develop emergency ventilators, though few progressed beyond laboratory testing. This study evaluates CoroVent, a rapidly manufactured ventilator, during real-life deployment in critically ill COVID-19 patients amid ventilator shortages. Methods: This retrospective observational study included patients ventilated with emergency ventilator CoroVent. This device uses a novel and unique way of generating inspiratory flow and gas mixing using fast ON/OFF valves for air and oxygen, producing pneumatic pulses that are then smoothed into continuous flow. Clinical data were collected from 3 hospitals between October 2020 and March 2021, selected from 27 contacted, of which 23 responded and 4 reported clinical use. Results: Eight male patients (mean age 67 years, BMI 37.2 kg·m–2) with COVID-19 or suspected infection were ventilated with CoroVent for 31.3 (10.0–58.5) hours. The mean FiO2 was 71.4%, PEEP 10.6 cmH2O, and Vt 8.9 mL·kg–1 predicted body weight. CoroVent was used as an initial or replacement ventilator during device shortages. No major technical failures occurred. Conclusions: This is the first report of real-life clinical use of a rapidly manufactured emergency ventilator during the COVID-19 pandemic. The results confirm that such devices, if well-designed, can provide effective respiratory support when conventional ventilators are unavailable.
- Open Access
- 6 September 2022
- Published by Springer Nature in Scientific Reports
- Vol. 12 (1) , 1-8
Effect of artifacts upon the pressure reactivity index
The pressure reactivity index (PRx) is a parameter for the assessment of cerebrovascular autoregulation, but its calculation is affected by artifacts in the source biosignals—intracranial pressure (ICP) and arterial blood pressure. We sought to describe the most common short-duration artifacts and their effect on the PRx. A retrospective analysis of 935 h of multimodal monitoring data was conducted, and five types of artifacts, characterized by their shape, duration, and amplitude, were identified: rectangular, fast impulse, isoline drift, saw tooth, and constant ICP value. Subsequently, all types of artifacts were mathematically modeled and inserted into undisturbed segments of biosignals. Fast impulse, the most common artifact, did not alter the PRx index significantly when inserted into one or both signals. Artifacts present in one signal exceeded the threshold PRx in less than 5% of samples, except for isoline drift. Compared to that, the shortest rectangular artifact inserted into both signals changed PRx to a value above the set threshold in 55.4% of cases. Our analysis shows that the effect of individual artifacts on the PRx index is variable, depending on their occurrence in one or both signals, duration, and shape. This different effect suggests that potentially not all artifacts need to be removed.
- 6 September 2022
- Published by Springer Nature in Scientific Reports
- Vol. 12 (1) , 1-8
Open AccessEffect of artifacts upon the pressure reactivity index
The pressure reactivity index (PRx) is a parameter for the assessment of cerebrovascular autoregulation, but its calculation is affected by artifacts in the source biosignals—intracranial pressure (ICP) and arterial blood pressure. We sought to describe the most common short-duration artifacts and their effect on the PRx. A retrospective analysis of 935 h of multimodal monitoring data was conducted, and five types of artifacts, characterized by their shape, duration, and amplitude, were identified: rectangular, fast impulse, isoline drift, saw tooth, and constant ICP value. Subsequently, all types of artifacts were mathematically modeled and inserted into undisturbed segments of biosignals. Fast impulse, the most common artifact, did not alter the PRx index significantly when inserted into one or both signals. Artifacts present in one signal exceeded the threshold PRx in less than 5% of samples, except for isoline drift. Compared to that, the shortest rectangular artifact inserted into both signals changed PRx to a value above the set threshold in 55.4% of cases. Our analysis shows that the effect of individual artifacts on the PRx index is variable, depending on their occurrence in one or both signals, duration, and shape. This different effect suggests that potentially not all artifacts need to be removed.
- Open Access
Pulse Oximeter Performance during Rapid Desaturation
The reliability of pulse oximetry is crucial, especially in cases of rapid changes in body oxygenation. In order to evaluate the performance of pulse oximeters during rapidly developing short periods of concurrent hypoxemia and hypercapnia, 13 healthy volunteers underwent 3 breathing phases during outdoor experiments (39 phases in total), monitored simultaneously by five different pulse oximeters. A significant incongruity in values displayed by the tested pulse oximeters was observed, even when the accuracy declared by the manufacturers were considered. In 28.2% of breathing phases, the five used devices did not show any congruent values. The longest uninterrupted congruent period formed 74.4% of total recorded time. Moreover, the congruent periods were rarely observed during the critical desaturation phase of the experiment. The time difference between the moments when the first and the last pulse oximeter showed the typical study endpoint values of SpO2 85% and 75% was 32.1 ± 23.6 s and 24.7 ± 19.3 s, respectively. These results suggest that SpO2 might not be a reliable parameter as a study endpoint, or more importantly as a safety limit in outdoor experiments. In the design of future studies, more parameters and continuous clinical assessment should be included.
Open AccessPulse Oximeter Performance during Rapid Desaturation
The reliability of pulse oximetry is crucial, especially in cases of rapid changes in body oxygenation. In order to evaluate the performance of pulse oximeters during rapidly developing short periods of concurrent hypoxemia and hypercapnia, 13 healthy volunteers underwent 3 breathing phases during outdoor experiments (39 phases in total), monitored simultaneously by five different pulse oximeters. A significant incongruity in values displayed by the tested pulse oximeters was observed, even when the accuracy declared by the manufacturers were considered. In 28.2% of breathing phases, the five used devices did not show any congruent values. The longest uninterrupted congruent period formed 74.4% of total recorded time. Moreover, the congruent periods were rarely observed during the critical desaturation phase of the experiment. The time difference between the moments when the first and the last pulse oximeter showed the typical study endpoint values of SpO2 85% and 75% was 32.1 ± 23.6 s and 24.7 ± 19.3 s, respectively. These results suggest that SpO2 might not be a reliable parameter as a study endpoint, or more importantly as a safety limit in outdoor experiments. In the design of future studies, more parameters and continuous clinical assessment should be included.
- 30 May 2018
- Published by Springer Nature
- p. 711-717
Breathing Experiments into the Simulated Avalanche Snow: Medical and Technical Issues of the Outdoor Breathing Trials
- 30 May 2018
- Published by Springer Nature
- p. 711-717
Breathing Experiments into the Simulated Avalanche Snow: Medical and Technical Issues of the Outdoor Breathing Trials