BLBálint Károly Lakatos

Bálint Károly Lakatos

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Bálint Károly Lakatos has a total of 221 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Hematocrit-corrected pulmonary vascular resistance: a prognostic marker in left ventricular assist device candidates

    by Bálint Károly Lakatos, Zsuzsanna Ladányi, Tímea Turschl, Alexandra Fábián, Andrea Ferencz, Eszter Pál, Veronika Párkányi, Zsolt Túróczi, Gergely Richárd Csikós, Márton Tokodi, et al.

  • Open Access

    Deep learning models for automated view and orientation classification and cardiac cycle phase detection in echocardiography

    by Ádám Szijártó, Thomas Akos Szeier, Tetsuji Kitano, Yosuke Nabeshima, Bálint Károly Lakatos, Alexandra Fábián, Zsuzsanna Ladányi, Luca Szávai, Máté Tolvaj, Andrea Ferencz, et al.

    Background: Deep learning (DL) is increasingly used for the automated analysis of echocardiographic videos. These end-to-end pipelines often incorporate models targeting common preprocessing tasks to ensure the selection of videos and frames suitable for their primary analytic objective. However, publicly available models supporting these core steps remain limited. Objectives: We aimed to develop three DL models for distinct preprocessing tasks: Model-V for view classification, Model-O for orientation classification, and Model-CC for identifying end-diastolic (ED) and end-systolic (ES) frames. Methods: Model-V was trained and internally validated on 13,864 videos and externally evaluated on 5,650 videos. Model-O and Model-CC were trained and internally validated on 5,513 and 3,108 apical 4-chamber view (A4C) videos, respectively, and both were externally tested on the same set of 1,200 A4C videos. Results: Model-V discriminated A4C from non-A4C with balanced accuracies of 0.991 (0.987–0.994) and 0.895 (0.884–0.907), and Model-O distinguished Stanford from Mayo orientation with balanced accuracies of 0.995 (0.990–1.000) and 0.960 (0.948–0.971) during internal and external testing, respectively. Model-CC identified ED frames with mean absolute errors (MAEs) of 4.041 (3.878–4.210) and 3.117 (3.044–3.193) frames and ES frames with MAEs of 2.799 (2.708–2.894) and 4.827 (4.738–4.917) frames in the internal and external test sets, while failing to identify only 7.3 and 4.2% of ED frames and 1.2 and 3.6% of ES frames, respectively. Conclusions: The proposed models demonstrated good performance across key preprocessing tasks; thus, they constitute readily integrable components for end-to-end DL pipelines in automated echocardiographic analysis.

  • Open Access

    Addendum to: A machine learning framework for performing binary classification on tabular biomedical data

    by Ádám Szijártó, Alexandra Fábián, Bálint Károly Lakatos, Máté Tolvaj, Béla Merkely, Attila Kovács, Márton Tokodi

    "Addendum to: A machine learning framework for performing binary classification on tabular biomedical data" published on 30 Apr 2025 by Akadémiai Kiadó.

  • Open Access

    Investigation of Body Composition and Cardiac Sports Adaptation in Elite Water Polo Players

    by Mark Zamodics, Mate Babity, Gusztav Schay, Agnes Bucsko-Varga, Eva Kovacs, Marton Horvath, Kinga Grebur, Marcell Janos Laszlo, Alexandra Fabian, Balint Karoly Lakatos, et al.

    The effects of physical activity on skeletal muscle mass and cardiac function are well-documented, but there is limited information on the relationship between the two. Furthermore, differentiating between the ‘athlete’s heart’ and pathological cardiac conditions often presents challenges. We aimed to analyze resting echocardiographic parameters in elite water polo athletes, considering sex, anthropometrics, and body composition. We examined 161 youth and adult athletes (age: 19.7 ± 5.6 years, male: 50.9%). Data analysis was performed with R software (version 4.2), using multivariate linear regression models. Confounders besides the main predictor were sex, age, and height. Male players had higher weight (87.55 ± 12.83 vs. 69.77 ± 9.8 kg), height (188.59 ± 6.82 vs. 173.47 ± 6.76 cm), skeletal muscle mass (SMM, 43.87 ± 5.50 vs. 30.38 ± 3.95 kg), and fat-free mass (FFM, 76.60 ± 9.23 vs. 54.52 ± 6.68 kg) and lower percentage of body fat values (12.14 ± 4.00.vs 21.51 ± 4.76%) compared to the female players. Youth players had lower height (178.51 ± 9.53 vs. 186.74 ± 9.27 kg) and weight (74.34 ± 12.12. vs. 88.23 ± 14.66 kg) compared to adults. Left ventricular end-diastolic and end-systolic diameters correlated positively with SMM (Est: 0.38, StE: 0.08, p < 0.001 and Est: 0.42, StE: 0.11, p < 0.001) and FFM (Est: 0.25, StE: 0.05, p < 0.001 and Est: 0.25, StE: 0.06, p < 0.001). Right ventricular end-diastolic diameter correlated positively with SMM (Est: 0.18, StE: 0.08, p < 0.05) and FFM (Est: 0.12, StE: 0.05, p < 0.05). Interventricular septal wall thickness showed positive correlation with SMM (Est: 0.16, StE: 0.04, p < 0.001) and FFM (Est: 0.10, StE: 0.02, p < 0.001). Left ventricular posterior wall thickness correlated with SMM, with a stronger correlation in females (Est: 0.17, StE: 0.05, p < 0.001) than in males (Est:0.7, StE: 0.04, p < 0.05). The close relationship between body composition and cardiac dimensions provides an opportunity for professionals to distinguish between athlete’s heart and pathological conditions.


  • Open Access

    Mechanical Adaptation of the Right Ventricle to Secondary Tricuspid Regurgitation and Its Association With Patient Outcomes

    by Zsuzsanna Ladányi, Bálint Károly Lakatos, Alexandra Clement, Michele Tomaselli, Alexandra Fábián, Noela Radu, Tímea Katalin Turschl, Andrea Ferencz, Béla Merkely, Elena Surkova, et al.


  • Open Access

    Evaluation of treadmill cardiopulmonary exercise testing and field measurement results in women’s youth and adult national team water polo players

    by Mark Zamodics, Mate Babity, Attila Mihok, Csaba Bognar, Agnes Bucsko-Varga, Panka Kulcsar, Dora Boroncsok, Regina Benko, Alexandra Fabian, Balint Lakatos, et al.


  • The Perioperative Period of Heart Transplantation Is Affected by Thyroid Hormone Status

    by Balázs Szécsi, Richárd Sinkó, Alexandra Vereb, Dmitry Khochanskiy, Kálmán Benke, Tamás Radovits, Bálint Lakatos, Andrea Kőszegi, Eszter Losoncz, Szilvia Kugler, et al.

    Background: Orthotopic heart transplantation (HTx) is a long-term surgical therapeutic approach for patients with end-stage heart failure. The objective of the present study was to uncover associations between altered thyroid hormone (TH) status and adverse outcomes after HTx. Methods: In this prospective, single-center cohort study, 283 patients underwent HTx between 2013 and 2020 at the Heart and Vascular Center of Semmelweis University in Hungary. We measured serum free triiodothyronine (fT3), free thyroxine (fT4), and thyrotropin (TSH) pre- and postoperatively. TaqMan qPCR was used to measure type 2 deiodinase and type 3 deiodinase mRNA ( Dio2 and Dio3 , respectively) levels from the diseased heart bioptates. To assess the local TH action of the heart, mRNA levels of Hcn2 and Myh7 were measured in a subgroup of patients receiving extracorporeal membrane oxygenation (ECMO) postoperatively. Groups were compared using nonparametric tests. Cox regression analysis and logistic regression test were used to investigate the outcomes. The connection between serum TH parameters and cardiac gene expressions was assessed using linear regression. Results: Serum TSH ( p = 0.009), fT3 ( p < 0.001), and fT4 ( p < 0.001) levels were lower after HTx than preoperatively. Levothyroxine (LT4) administered to donors was associated with better survival after 30 days ( p = 0.049). LT4 replacement given to recipients after HTx was associated with better survival after 30 days ( p = 0.018), 1 year ( p = 0.002), and 2 years ( p = 0.001). Dio3 mRNA level was significantly increased in patients who were treated with ECMO ( p = 0.026), left ventricular assist device (LVAD) ( p = 0.008), and biventricular assist device (BiVAD) ( p = 0.013) preoperatively, and ECMO ( p = 0.042) postoperatively, compared with those who did not require any type of mechanical circulatory support (MCS). We found no significant difference in the expression of the Hcn2 and Myh7 marker genes between patients on postoperative ECMO and those without MCS, and neither did they correlate with serum hormone levels ( p = 0.519 and p = 0.056, respectively). Conclusions: We conclude that TH status plays an important role in HTx patients, and monitoring of TH status in the perioperative period may contribute to improved treatment outcomes. Our findings require independent confirmation in a randomized controlled clinical trial.


  • Open Access

    Trajectory of Diastolic Function after Heart Transplantation as Assessed by Left Atrial Deformation Analysis

    by Borbála Edvi, Alexandra Assabiny, Tímea Teszák, Máté Tolvaj, Alexandra Fábián, István Hartyánszky, Miklós Pólos, Bálint Károly Lakatos, Hajnalka Vágó, Balázs Sax, et al.

    Diastolic dysfunction (DD) is a prevalent and clinically significant complication after heart transplantation (HTX). We aimed to characterize the diastolic function of HTX recipients with both short-term and long-term follow-ups by applying left atrial (LA) deformation analysis. We consecutively enrolled and followed up with 33 HTX patients. Three assessments were performed one month, 3–5 months, and 3–5 years after surgery. Beyond conventional echocardiographic measurements, apical four-chamber views optimized for speckle tracking analysis were acquired and post-processed by dedicated software solutions (TomTec AutoStrain LA and LV). Left atrial phasic functions were characterized by reservoir, conduit, and contraction strains. We categorized diastolic function according to current guidelines (normal diastolic function, indeterminate, DD). At the first assessment, nine (27%) patients were in the DD category, and eleven (33%) were indeterminate. At the second assessment, only one patient (3%) remained in the DD category and six (18%) were indeterminate. At the third assessment, 100% of patients were categorized as having normal diastolic function. LA reservoir strain gradually increased over time. LA contraction strain significantly improved from the second to the third assessment. We found a correlation between the LA reservoir strain and NT-proBNP (r = 0.40, p < 0.05). DD is prevalent immediately after HTX but rare until the end of the first postoperative quarter. Speckle tracking analysis enables the characterization of LA phasic functions that may reflect both short- and long-term changes in diastolic function and correlate with NT-proBNP.

  • Open Access

    Atrioventricular Interval Modulation (AVIM) therapy for ambulatory systolic blood pressure (aSBP) reduction with a dual chamber pacemaker: comparison of conduction system vs. right ventricular pacing

    by I Osztheimer, M Ruppert, M Nemeth, B Veres, G Y Apponyi, B Lakatos, A Kiraly, R Papp, G Der, Z Tarjanyi, et al.

    Background: AVIM is a pacing algorithm incorporated into dual-chamber pacemakers to treat hypertension. AVIM delivers repeating sequences of short and longer atrioventricular delays leading to reduced preload and afterload. The therapy lowers blood pressure while simultaneously modulating the autonomic nervous system. The marked blood pressure reducing capability of AVIM has been already confirmed by a previous randomized clinical trial using conventional right ventricular (RV) pacing. Nevertheless, whether a similar blood pressure reduction could be also achieved by AVIM when conduction system pacing (CSP) is applied instead of RV pacing has not been investigated yet. Considering the facts that CSP using Left Bundle Branch Area Pacing (LBBAP) has been found to elicit more synchronous ventricular contraction and improved hemodynamics compared to RV pacing and AVIM is associated with a signficant burden of ventricular pacing, it is tempting to hypothesize that AVIM with CSP could be particularly beneficial. Purpose: To compare the effects of AVIM therapy when delivered via an LBBAP location with the effects when delivered using a traditional RV lead location. Methods: This single-center study evaluated subjects indicated for a dual chamber pacemaker who also had uncontrolled hypertension despite a stable dose of ³ 1 antihypertensive drug. Leads were placed in the RV or the LBBAP position. Subjects entered a run-in period of 4 weeks in which baseline office (oSBP)and 24-hour aSBP were measured. Subjects with baseline aSBP >130 mmHg and oSBP>140 mmHg at week 4 were included in the study. AVIM therapy was activated and a 24-hour aSBP was performed immediately after activation. Results: Using a dedicated CSP pacemaker lab setup, CSP was performed in 5 patients with the Select Secure electrode (model 3830-69 cm) delivered via a fixed curve sheath (C315HIS). During the run in phase, in 3 out of the 5 CSP-AVIM patients, high blood pressure did not reach the level of inclusion and therefore these patients were not used for further analysis. In the remaining two patients, AVIM was activated. These two patients had a V6 stimulus-peak time of 69 ms and 61 ms, while V1 stimulus peak time was: 121ms and 109ms respectively. The effect of AVIM on 24-hour aSBP in the activated CSP-AVIM patients was compared to the average effect seen in patients from the same center enrolled in the prior randomized study with leads in the RV. Table 1 shows individual aSBP values for the two patient groups. Results suggest that delivery of AVIM therapy through leads in either position generates a significant reduction in blood pressure: -24.1 mmHg vs. -15.4 mmHg when the lead is placed in LBBAP or RV positions respectively. Conclusion These results demonstrate a significant reduction in aSBP in both lead locations. Additional studies evaluating larger sample sizes are required to assess if specific lead positions are preferable.Table 1


  • Open Access

    Long-Term Durability of High- and Very High-Power Short-Duration PVI by Invasive Remapping: The HPSD Remap Study

    by Nándor Szegedi, Zoltán Salló, Vivien Klaudia Nagy, István Osztheimer, István Hizoh, Bálint Lakatos, Melinda Boussoussou, Gábor Orbán, Márton Boga, Arnold Béla Ferencz, et al.

    BACKGROUND: High-power short-duration ablation has shown impressive efficacy and safety for pulmonary vein isolation (PVI); however, initial efficacy results with very high power short-duration ablation were discouraging. This study compared the long-term durability of PVI performed with a 90- versus 50-W power setting. METHODS: Patients were randomized 1:1 to undergo PVI with the QDOT catheter using a power setting of 90 or 50 W. Three months after the index procedure, patients underwent a repeat electrophysiology study to identify pulmonary vein reconnections. Patients were followed for 12 months to detect AF recurrences. RESULTS: We included 46 patients (mean age, 64 years; women, 48%). Procedure (76 versus 84 minutes; P =0.02), left atrial dwell (63 versus 71 minutes; P =0.01), and radiofrequency (303 versus 1040 seconds; P P =0.01). There was no difference in first-pass isolation (83% versus 82%; P =1.0) or acute reconnection (4% versus 14%; P =0.3) rates between 90 and 50 W. Forty patients underwent a repeat electrophysiology study. Durable PVI on a per PV basis was present in 72/78 (92%) versus 68/77 (88%) PVs in the 90- and 50-W energy setting groups, respectively; effect size: 72/78−68/77=0.040, lower 95% CI=−0.051 (noninferiority limit=−0.1, ie, noninferiority is met). No complications occurred. There was no difference in 12-month atrial fibrillation-free survival between the 90- and 50-W groups ( P =0.2). CONCLUSIONS: Similarly high rates of durable PVI and arrhythmia-free survival were achieved with 90 and 50 W. Procedure, left atrial dwell, and radiofrequency times were shorter with 90 W compared with 50 W. The sample size is too small to conclude the safety and long-term efficacy of the high and very high-power short-duration PVI; further studies are needed to address this topic. REGISTRATION: URL: https://www.clinicaltrials.gov ; Unique identifier: NCT05459831.