Interdisciplinary cardiovascular and thoracic surgery最新文献

Objectives: Large language models (LLMs) are generative-AI which generate text output like a human conversation. We wanted to assess the ability of LLMs to answer patient's questions and benchmark their output using a best evidence topic (BET).

Methods: We asked LLMs whether robot-assisted thoracic surgery (RATS) or video-assisted thoracoscopic surgery (VATS) lobectomy had better perioperative outcomes for postoperative pain, length of hospital stay (LOS) and mortality. A BET was constructed according to a structured protocol for the same questions. An initial search yielded 324 papers, 12 represented the best evidence.

Results: LLM outputs are almost instantaneous while a BET took many hours of searching a database for relevant evidence. However, current iterations and models of LLMs did not provide relevant outputs, suffered from hallucinations, and could be restricted by copyright and paywall issues. The BET, on the other hand, was tailored to the scenario by specialist human oversight and therefore more reliable and nuanced.

Conclusions: There were no major differences between RATS and VATS lobectomy for T1cN0M0 NSCLC apart from shorter LOS following RATS. Current LLMs may not be entirely reliable for answering clinical questions. An LLM-BET protocol could be used as a standardized process to compare LLM outputs for different clinical scenarios, each benchmarked with a BET. It can also be used to analyse outputs of different models of current and future LLMs.

Harvesting the left internal thoracic artery (LITA) during minimally invasive coronary artery bypass grafting (MICS-CABG) can be technically demanding, particularly in patients with a prominent osteophyte at the first costochondral joint. We report the use of an ultrasonic aspirator system, the SONOPET iQ (Stryker), to facilitate safe and effective LITA harvesting in such anatomically challenging cases. The SONOPET iQ was employed when proximal LITA exposure was hindered by the protrusion of the first rib at the costochondral junction. Operating at an ultrasonic frequency of 25 kHz, the device enabled precise bone sculpting with minimal heat generation and bleeding-an essential advantage within the constrained operative field of MICS-CABG. By allowing controlled thinning of the first rib, the SONOPET iQ enhanced visualization and access to the LITA without requiring extensive rib retraction or resection. This adjunctive technique improves surgical safety and exposure by enabling accurate bone modification while minimizing thermal injury and bleeding. The SONOPET iQ offers a simple, reproducible, and effective solution for overcoming anatomical obstacles during LITA harvesting in MICS-CABG.

Objectives: The optimal choice of bioprosthetic valve for mitral valve replacement (MVR) remains controversial, with prior studies reporting inconsistent outcomes. This study aimed to compare mid-term clinical performance between bovine pericardial and porcine bioprostheses using a hierarchical composite end-point reflecting both durability and clinical relevance, including cardiac death, reoperation, and haemodynamic structural valve deterioration (SVD).

Methods: This retrospective study included 304 patients (152 per group) who underwent bioprosthetic MVR between 2007 and 2020, following 1:1 propensity score (PS) matching. The primary outcome was a hierarchical composite of cardiac death, reoperation for SVD, and echocardiographic diagnosis of haemodynamic SVD. Secondary outcomes included overall survival and perioperative complications.

Results: The median clinical follow-up durations were 7.3 years in the bovine group and 5.3 years in the porcine group. The win ratio was 1.49 (95% CI, 1.21-1.83; P < .001), indicating a statistically significant difference favouring the bovine group. While overall survival and freedom from cardiac death were comparable between groups, the cumulative incidence of reoperation for SVD and haemodynamic SVD diagnosis were significantly higher in the porcine group (Gray's P = .031 and 0.037, respectively).

Conclusions: In this propensity-matched analysis, bovine pericardial valves showed a modest, consistent mid-term durability signal on SVD-related components, with similar overall survival. These findings are hypothesis-generating and should inform individualized prosthesis selection rather than dictate device choice; model-specific prospective studies, ideally randomized controlled trials, are needed for definitive guidance.

Objectives: Measuring surgical competency is essential for surgical residents to ensure patient safety. Traditional assessment tools rely on subjective evaluation. This study evaluated whether artificial intelligence (AI)-based hand tracking can more objectively distinguish between levels of surgical competency and predict surgical years of experience versus traditional assessments.

Methods: A total of 44 participants, including medical students, surgical residents, and surgical consultants, performed transcutaneous suturing, intracutaneous suturing, and surgical knot tying. Videos of intracutaneous suturing were scored using the objective structured assessment of technical skills (OSATS). Hand movements were analysed using AI tracking software to extract coordinates to measure velocity, pathlength, and jerk. Linear regression models predicted experience years using procedural time and OSATS in combination with hand tracking metrics.

Results: Hand tracking metrics varied mainly between medical students and more experienced groups. Traditional assessment tools (procedural time, OSATS) could predict experience years during training, with an adjusted coefficient of determination (R2) ranging from 0.537 to 0.638, dependent on procedure type. Hand tracking variables identified multiple significant predictors for years of experience, with an adjusted R2 of 0.540-0.712, which outperformed the traditional tools in each procedure. Combining all assessment tools (time, OSATS, and hand tracking) gave the best predictive value, with an adjusted R2 ranging from 0.540 to 0.809, with velocity, pathlength, jerk, and acceleration as significant predictors.

Conclusions: AI-based hand tracking provides a new method for objective, reproducible measures of surgical skills. Incorporating hand tracking metrics enhances prediction of surgical experience, and supports standardized as well as objective evaluation of skills assessment in surgical training.

Objectives: Although sublobar resection is a standard treatment for small peripheral non-small-cell lung cancer (NSCLC), the optimal margin distance remains under investigation. This study aimed to determine the adequacy of commonly used margin distances by comparing locoregional recurrence (LRR) with lobectomy.

Methods: We retrospectively reviewed data from patients with completely resected ≤3 cm adenocarcinoma, squamous cell carcinoma, or adenosquamous cell carcinoma treated between April 2018 and March 2024. We compared sufficient and insufficient margin sublobar resection and lobectomy in terms of freedom from LRR (FLRR) rate and the cumulative risk of LRR.

Results: Of the 528 included patients, 200 underwent sublobar resection and 328 underwent lobectomy. After excluding 23 patients with pure ground-glass nodules, 505 patients with 19 LRR events were included in the prognostic analysis. In the overall cohort, LRR risk was worse in the insufficient margin group than in the sufficient margin and lobectomy groups, whereas the risk in the sufficient margin group was similar to the lobectomy groups (5-year FLRR: 75.1% vs 95.9% vs 95.8%, respectively). Conversely, in the pure-solid cohort, the risk of LRR was worse in the insufficient margin group and even in the sufficient margin group, compared to the lobectomy group.

Conclusions: Among patients with completely resected ≤3 cm NSCLC, our study indicated that those who underwent sublobar resection with a sufficient margin achieved better local control than those with an insufficient margin, and sublobar resection with a sufficient margin was comparable to lobectomy. However, for pure-solid nodules, conventional margin distance might be insufficient.

Clinical registration number: UMIN Clinical Trials Registry: Registration number: UMIN000058449. https://center6.umin.ac.jp/cgi-bin/ctr/ctr_view_reg.cgi?recptno=R000066821.

Objectives: Selective pulmonary artery perfusion with blood flow occlusion (SPAP-BFO), an experimental endovascular technique, has shown potential to enhance pulmonary drug delivery to the lung. Therefore, it becomes a potential minimally invasive technique for lung cancer and pulmonary metastases. Prior studies predominantly used animal models which do not adequately replicate human vascular anatomy, leaving the clinical feasibility of SPAP-BFO underexplored. To address this gap, we developed a patient-specific 3D model of the human venous system to evaluate the technical feasibility of SPAP-BFO.

Methods: A 1:1 scale 3D model of the human venous system was developed and printed based on CT scans of a patient. This model was connected to a perfusion system to simulate blood flow, enabling testing of the catheterization procedure under realistic clinical conditions. Two commercially available balloon catheters, Coda (Cook) and Reliant (Medtronic), were selected based on length and balloon diameter, and their feasibility of reaching and occluding the left and right pulmonary arteries were assessed.

Results: The model effectively simulated human anatomy and blood flow, allowing for both visual and fluoroscopic assessment of the procedure. Both Coda and Reliant catheters successfully reached the target location, when introduced via the femoral vein, and occluded the left and right pulmonary arteries without physically blocking contralateral flow or extending beyond the first bifurcation.

Conclusions: This patient-specific 3D model provided a valuable platform to evaluate the clinical feasibility of SPAP-BFO. The Coda and Reliant balloon catheters demonstrated effective occlusion of the pulmonary arteries, supporting their potential use in SPAP-BFO procedures.

Objectives: To determine whether CT within 6 months after total arch replacement (TAR) with a frozen elephant trunk (FET) for chronic aortic dissection predicts mid-term outcomes and informs the timing of distal treatment.

Methods: We analysed 56 consecutive patients who underwent TAR with FET at 2 centres (2009-2022) and had evaluable 6-month postoperative CT. Early remodelling was defined as the change from baseline to 6 months in the maximal outer-to-outer diameter of the proximal descending thoracic aorta at Level A (Ishimaru zone 3, 20 mm distal to the left subclavian artery, measured on centreline-orthogonal reconstructions). Patients were classified as early positive remodelling (EPR; no increase or a decrease) or early negative remodelling (ENR; ≥1-mm increase). Prespecified outcomes were distal aortic reintervention, distal stent graft-induced new entry (dSINE), and overall survival.

Results: Mean follow-up was 5.4 years (standard deviation 3.7). Distal reintervention was required in 36/56 patients (64%). At 5 years, freedom from distal reintervention was higher with EPR than with ENR (44.6% vs 6.2%; P = .003). dSINE occurred in 26/56 patients (46.4%); 5-year dSINE-free survival was 65.1% (95% CI, 39.6-81.9) with EPR versus 18.2% (95% CI 5.9-35.2) with ENR (P = .008). Overall, 5-year survival for the cohort was 80.0% (95% CI 64.7-89.2). Among ENR patients, 5-year survival was 0% with conservative management versus 40.5% with distal intervention (P < .001); within EPR, 5-year survival was 65.9% with conservative management versus 85.7% with reintervention (P = .210).

Conclusions: A 6-month CT provides simple, actionable risk stratification after TAR with FET for chronic aortic dissection. Absence of EPR identifies a high-risk subgroup (ENR) that warrants closer surveillance and timely distal intervention, optimizing follow-up intensity and treatment timing.

Objectives: Neoadjuvant therapy (NAT) significantly improves the pathologic complete response (pCR) rates in patients with locally advanced esophageal squamous cell carcinoma (ESCC). Emerging evidence suggests that patients with pCR may experience favourable outcomes and could be considered for active surveillance strategies to delay surgery. This study aims to develop a clinical-radiomics model to predict pCR after NAT in ESCC.

Methods: We retrospectively enrolled 236 patients with locally advanced ESCC who received NAT at our centre and randomly assigned them to training and test cohorts (3:2 ratio). Radiomics features were extracted from tumour regions segmented on post-NAT contrast-enhanced computed tomography (CT) scans. After feature selection, a predictive model integrating radiomics and clinical variables was developed using logistic regression and visualized as a nomogram. Model performance was evaluated using area under the curve (AUC), accuracy, sensitivity, and specificity.

Results: The clinical-radiomics model achieved an AUC of 0.91 (95% confidence interval [CI]: 0.86-0.95), accuracy of 0.84, sensitivity of 0.89, and specificity of 0.81 in the training cohort, and an AUC of 0.84 (95% CI: 0.76-0.92), accuracy of 0.78, sensitivity of 0.84, and specificity of 0.74 in the test cohort. Calibration curves demonstrated good agreement between predicted and observed outcomes, and decision curve analysis confirmed the model's clinical utility.

Conclusions: The clinical-radiomics model accurately predicts pCR following NAT in ESCC and may guide personalized treatment strategies.

Objectives: To investigate the application of a mechanical side-to-side oesophagogastric anastomosis in the reconstruction of the digestive tract of beagles after an oesophagectomy.

Methods: Eighteen beagles were randomly divided into 3 groups: the hand-sewn (HS) group, the linear-stapled anastomosis in the oesophagus and anterior portion of the stomach (LESA) group and the linear-stapled anastomosis in the oesophagus and the posterior portion of the stomach (LESP) group. The gastro-oesophageal reflux, anastomotic area, anastomotic bursting pressure and the breaking strength in the beagles at 1 week and 12 weeks after the operations were compared. The histopathological morphology was observed using haematoxylin-eosin staining and Masson staining, and the expression of the vascular endothelial growth factor (VEGF) was detected by immunohistochemical analysis.

Results: At 1 and 12 weeks after the operation, the percentage of gastro-oesophageal reflux time and the longest reflux time in the HS group and the LESP group were higher than those in the LESA group (P < .05). The anastomotic areas in the HS group were significantly smaller than those in the LESA and LESP groups at 1 and 12 weeks postoperatively (P < .05); there were no differences in the anastomotic areas in the LESA and LESP groups. At 1 and 12 weeks postoperatively, the bursting pressure and breaking strength of the anastomosis, the collagen-fibre area ratio and VEGF positive expression in the LESA group were significantly higher than those in the HS group (P < .05).

Conclusions: The mechanical side-to-side oesophagogastric anastomosis of the oesophagus and the anterior wall of the stomach can reduce the occurrence of gastro-oesophageal reflux, increase the bursting pressure and breaking strength, promote collagen fibre and VEGF expression, to promote healing of the anastomosis.