Sung-Hoon Kim

Background/Objectives: CO₂ pneumoperitoneum (PnP) alters hemodynamics during laparoscopic surgery. We examined whether PnP modifies mean arterial pressure (MAP) and hypotension prediction index (HPI) responses to controlled, stepwise hemorrhage and retransfusion in a porcine model. Methods: This exploratory secondary analysis included recordings from three PnP-only pilot animals and nine animals with analyzable data from a randomized crossover experiment involving both PnP and No-PnP conditions. Blood was withdrawn and retransfused in 100 mL increments up to 600 mL per phase under both PnP (15 mmHg) and No-PnP conditions. MAP and HPI were averaged within protocol windows and analyzed using linear mixed-effects models to assess changes across protocol windows, condition-by-window interactions, and paired within-animal differences. Results: Twenty-one hemorrhage–transfusion sets from 12 animals were analyzed, comprising 12 PnP and nine No-PnP sets. MAP decreased and HPI increased overall during hemorrhage, with the opposite pattern observed during retransfusion in both conditions (all protocol-window effects, p < 0.001). No significant condition-by-window interactions were detected during hemorrhage (MAP, p = 0.407; HPI, p = 0.925) or retransfusion (MAP, p = 0.999; HPI, p = 0.910). Within-animal differences did not vary significantly across protocol windows in either phase. Conclusions: No significant PnP-related differences in the patterns of MAP and HPI change were detected in this porcine model. These findings provide experimental context for interpreting MAP and HPI during blood loss in laparoscopic surgery; clinical studies are needed to determine whether HPI adds predictive information beyond MAP.

Background: We evaluated phase-specific associations between perioperative hemodynamic variables and 30-day mortality following aortic surgery requiring cardiopulmonary bypass (CPB) and to assess the utility of post-CPB variables in risk stratification. Methods: We analyzed 691 patients who underwent aortic surgery requiring CPB. Hemodynamic variables were assessed during the preoperative, pre-incision, pre-CPB, and post-CPB phases and included heart rate (HR), systolic blood pressure (SBP), mean blood pressure (MBP), diastolic blood pressure (DBP), shock index (SI; HR/SBP), and modified shock index (MSI; HR/MBP). The primary outcome was 30-day mortality. Holm correction was applied to the between-group comparisons of hemodynamic variables. Adjusted Cox models used ridge penalization and included surgery group, CPB time, and mean post-CPB vasoactive-inotropic score. Results: Thirty patients (4.3%) died within 30 days. After Holm correction, significant differences between survivors and non-survivors remained only for post-CPB SBP, MBP, SI, and MSI. In adjusted Cox models, 30-day mortality was associated with higher post-CPB SI (hazard ratio [95% CI], 1.38 [1.17 to 1.64] per 0.1-unit increase; p < 0.001) and MSI (1.29 [1.12 to 1.47] per 0.1-unit increase; p < 0.001), as well as lower SBP (1.60 [1.24 to 2.05] per 10-mmHg decrease; p < 0.001) and MBP (2.00 [1.32 to 3.03] per 10-mmHg decrease; p = 0.001); in contrast, HR and DBP did not show a significant association with 30-day mortality. Kaplan–Meier analyses confirmed poorer survival across ROC-derived post-CPB thresholds for HR, SBP, MBP, SI, and MSI (all log-rank p ≤ 0.013). Conclusions: Post-CPB SBP, MBP, SI, and MSI showed significant adjusted associations with 30-day mortality and may assist in early postoperative risk stratification. These exploratory findings require external validation, and the identified thresholds should not be interpreted as treatment targets.

Background and Objectives: Routine preoperative chest radiography (CXR) has limited value for perioperative risk prediction when interpreted qualitatively. The Automated Diagnosis of Cardiovascular abnormalities (ADC) model enables automated quantification of cardiomediastinal vascular border (CVB) parameters on CXR. This study evaluated the associations of AI-derived CVB metrics with postoperative major adverse cardiovascular events (MACE) and their incremental predictive value beyond clinical factors. Materials and Methods: This retrospective cohort study included patients who underwent surgery under general anesthesia at a tertiary academic center. The ADC model quantified CVB parameters as raw measurements and age- and sex-adjusted z-scores. The primary outcome was postoperative in-hospital MACE. Associations were assessed using multivariable logistic regression, and discrimination was evaluated using receiver operating characteristic curve analysis. Incremental predictive performance was assessed by adding selected CVB parameters to a clinical reference model. Results: Among 101,531 patients, MACE occurred in 1655 patients (1.6%). All ADC-derived CVB parameters remained independently associated with MACE after multivariable adjustment. For cardiothoracic ratio, the adjusted OR was 1.64 (95% CI, 1.58–1.70; p < 0.001), with an AUC of 0.789. For the composite CVB z-score, the adjusted OR was 1.88 (95% CI, 1.79–1.97; p < 0.001), with an AUC of 0.762. Adding either the CT ratio or composite CVB z-score to the clinical reference model increased the AUC from 0.894 to 0.906 (p < 0.001 for both), with similar improvement in temporal validation. Conclusions: AI-derived CVB parameters from preoperative CXR were independently associated with postoperative in-hospital MACE and provided modest incremental predictive information beyond routinely available clinical factors. Automated CVB analysis may provide additional cardiovascular risk information from a routinely obtained preoperative CXR without additional imaging or patient burden, although further external validation is required to establish its clinical utility.

Background/Objectives: Postoperative pain after robot-assisted gynecologic surgery delays recovery and prolongs hospitalization, yet evidence on the role of ultrasound-guided rectus sheath block (RSB) in this setting is limited. We investigated whether preoperative ultrasound-guided RSB was associated with a shorter length of hospital stay (LOS) after robot-assisted gynecologic surgery. Methods: This single-center retrospective cohort study included 266 consecutive female patients who underwent robot-assisted gynecologic surgery between November 2023 and April 2024. RSB was introduced in January 2024; 113 patients from the RSB-eligible era who received the block and 153 patients before RSB introduction served as the RSB and comparator groups, respectively. The primary outcome was LOS. Secondary outcomes included rescue intravenous fentanyl and rescue antiemetic use in the post-anesthesia care unit (PACU) and rescue analgesic administration on the general ward through postoperative day (POD) 2. Multivariable quasi-Poisson regression was used to adjust for potential confounders. Results: LOS was significantly shorter in the RSB group (median 3 [IQR 3–3] vs. 4 [3–5] days; p < 0.001; adjusted IRR 0.78, 95% CI 0.72–0.85). The RSB group also required less PACU rescue fentanyl (0.86 [0.68–1.38] vs. 1.17 [0.85–1.69] μg/kg; p < 0.001), fewer rescue antiemetics (3.5% vs. 11.8%; p = 0.029), and fewer ward rescue analgesics on POD 0 (52.2% vs. 68.6%; p = 0.009) and POD 1 (13.3% vs. 34.2%; p < 0.001). Conclusions: Preoperative ultrasound-guided RSB was associated with shorter LOS, reduced PACU opioid and antiemetic requirements, and fewer early ward rescue analgesics. Randomized trials are warranted to confirm these benefits.

Despite various available methods for monitoring a patient's respiratory system, conventional monitors provide limited ventilatory function information. This study explored the feasibility of intraoperative lung sound patterns to discriminate preoperatively diagnosed ventilatory dysfunction. Forty-five patients who had undergone preoperative pulmonary function testing were enrolled for analysis, comprising 15 patients per normal, obstructive, and restrictive group. High-fidelity lung sounds were recorded intraoperatively using esophageal stethoscopes equipped with digital microphone devices. After signal processing, morphological features of the acoustic data, including the inhale/exhale peak ratio (I/Ep), were extracted. Their discriminative abilities for obstructive and restrictive types were assessed and compared with conventional monitoring parameters. I/Ep showed strong discriminative performance, with an area under the receiver operating characteristic curve of 0.950 (95% CI: 0.887-0.991) for obstructive and 0.950 (95% CI: 0.867-0.995) for restrictive types. The median values of I/Ep were 2.9 in the restrictive, 2.2 in the normal, and 1.5 in the obstructive group (P < 0.001). Conventional ventilatory parameters (compliance, peak inspiratory pressure, and slope of end-tidal CO2) did not significantly differ among the groups. Our study demonstrated that the I/Ep derived from intraoperative acoustic data differed according to preoperatively diagnosed ventilatory dysfunction. Hence, acoustic inhale and exhale patterns contain clinically useful information not captured by conventional ventilator parameters. Further studies are warranted to explore the clinical application of acoustic feature analysis for real-time intraoperative monitoring.

Objective . We developed stroke volume variation (SVV) Net, a deep learning-based model for estimating SVV, and validated its performance and clinical applicability through both large-scale retrospective modeling and prospective real-time deployment within intraoperative settings. Approach . A SVVNet was trained on arterial blood pressure (ABP) waveforms from 2849 patients undergoing general anesthesia at Asan Medical Center (AMC). For retrospective validation, we utilized AMC cohorts and a publicly accessible database to assess agreement, trending ability, noise robustness (quantifies via signal quality index), and inter-individual deviation against ABP-based pulse wave analysis (ABP-PWA) and other DL models. For prospective validation, the optimized model was deployed onto a tablet and prospectively evaluated in real-time during 67 surgeries. Main results . In the retrospective external validation, SVVNet achieved higher agreement mean absolute error (MAE = 0.70, Pearson ρ = 0.92), compared to ABP-PWA algorithm (MAE = 1.30, Pearson ρ = 0.80) and lower inter-individual deviation. It was more robust to signal noise (MAE = 4.33 vs 30.87). In the prospective real-time deployment, our deployed model achieved an MAE of 1.33 and a Pearson correlation of 0.87 against the ground truth. Significance . This study emphasizes the advantages of the DL-based model compared to ABP-PWA-based algorithms and presents prospective results that support the clinical applicability of SVVNet in the operating room. Our model has the potential to replace pre-existing monitoring systems and to enhance the complex surgical environment.

The acoustic variability index (AVI), derived from continuous phonocardiographic analysis of heart sounds via an esophageal stethoscope, is a novel minimally invasive method for intraoperative hemodynamic monitoring. We evaluated the ability of AVI to predict fluid responsiveness in patients undergoing hepatectomy and compared its performance with that of the conventional indices. Forty adult patients who underwent hepatectomy were enrolled in this prospective, single-center observational study. After major surgical resection and stabilization, a 500 mL crystalloid fluid challenge was administered. Hemodynamic parameters including AVI, central venous pressure (CVP), stroke volume variation (SVV), and pulse pressure variation (PPV) were recorded before and after volume expansion. Patients with ≥ 10% increase in cardiac output (CO) were defined as responders. Thirty-seven patients were included in the final analysis, of which 12 (32.4%) were classified as responders. After fluid loading, responders showed a significant decrease in AVI (11.4% ± 2.5% to 7.8% ± 2.9%, p = 0.004), while non-responders showed no significant change (7.1% ± 3.2% to 6.3% ± 3.0%, p = 0.356). AVI demonstrated predictive performance comparable to SVV and PPV, with an optimal cutoff of > 9.8% and an area under the receiver operating characteristic curve of 0.873 (95% CI: 0.743–0.967). Real-time intraoperative monitoring of the AVI shows potential as a predictor for fluid responsiveness in patients undergoing hepatectomy. AVI offers a promising, minimally invasive approach for guiding fluid therapy. Further research is warranted to validate its utility in broader surgical populations.