Citations

567 Background: Selecting adjuvant therapy after nephrectomy for renal cell carcinoma (RCC) requires reviewing scattered electronic health records (EHR) and performing accurate eligibility and risk assessments. Large Language Models (LLMs) may assist physicians by extracting key clinical data and generating evidence-based recommendations between active surveillance and adjuvant pembrolizumab. Methods: We selected RCC patients who underwent nephrectomy at Mayo Clinic in Arizona, Florida, and Minnesota between December 2021 and September 2025. Data extraction from EHRs was performed by a HIPAA-compliant LLM (GPT-5 and Gemini 2.0-Flash-001) and compared to human labels. Data on patients’ demographics, medical history, pathology, laboratory results, and imaging were collected. GPT-5 generated treatment recommendations using a structured prompt that incorporated patients’ features and relevant clinical information that is available to clinicians. Concordance between the human labels and LLM was evaluated for extraction. For treatment recommendations, agreement between the LLM’s output and physician’s documented recommendation in clinical notes was assessed, with disagreements reviewed independently by two blinded genitourinary oncologists. Results: The dataset included 85 patients, median (range) age of 65 (34-84) years; 24.7% were females. Patients who declined adjuvant treatment for personal reasons were excluded. LLM achieved an extraction accuracy of 95% for presenting symptoms, 97% for BMI and ECOG, and > 99% for pathology variables. Concordance between LLM and physician treatment recommendations was 84.7% (72/85). Among 13 discrepant cases, one oncologist agreed with the LLM in 9 cases and the other in 8. The main reasons for disagreement between the physicians and the LLM were that the LLM favored active surveillance for patients with declining renal function, whereas physicians still considered adjuvant pembrolizumab appropriate in such cases. Additionally, the LLM demonstrated hesitancy to recommend pembrolizumab for patients with low performance status. Overall, the 13 discrepant cases were nuanced and challenging, with disagreement observed between the two oncologists themselves in three instances. Conclusions: LLMs demonstrated high accuracy in extracting clinically relevant variables and generating objective, evidence-based adjuvant treatment recommendations for RCC after nephrectomy. Importantly, LLM did not generate any inappropriate or potentially harmful recommendations. Our pipeline can serve as a supportive decision tool to enhance consistency in post-nephrectomy management and assist physicians in choosing between active surveillance and adjuvant pembrolizumab. However, final treatment decisions should remain the responsibility of the treating physician after comprehensive patient evaluation.

464 Background: Based on LITESPARK-005 trail, Belzutifan was FDA-approved to treat advanced RCC following ICI and a VEGF-TKI. We conducted a retrospective cohort study to report the real-world outcomes of Belzutifan in pre-treated RCC patients. Methods: A retrospective chart review of all patients who received Belzutifan at the Mayo Clinic Comprehensive Cancer Center between Dec 2023 and Sep 2025 was conducted. Patients treated for sporadic RCC were eligible; those with von Hippel–Lindau disease, non-RCC pathologies (e.g., Pheochromocytoma, ovarian cancer) were excluded. Radiologic response was assessed by trained curators using RECIST-defined criteria, approximated through systematic review of radiology reports with clinician adjudication when needed. Progression-free survival (PFS), defined as the time from treatment initiation to radiographic progression or death, was estimated using the Kaplan–Meier method. Non-progressors were censored at the last radiology assessment. Patients who discontinued therapy for non-progression reasons were censored at the last radiology assessment prior to discontinuation. Any-cause hypoxia and anemia were reported using CTCAEv5 definitions. Results: A total of 158 patients were identified; 83 (52.5%) who received Belzutifan for sporadic RCC were included. The remaining were excluded for VHL disease (68, 43%) or other pathologies (7, 4.4%). Median follow up was 12 months, 67% were male, median age was 70 (range 39, 94) with 57 patients (69%) were ≥ 65. 28 patients (34%) had a poor IMDC score at baseline. The cohort received a median of 3 previous lines of therapy (range 1,8). Median treatment duration was 3 months (IQR: 1-9.5). Median PFS was 5.7 months (95% CI, 3.8–8.7). At 12 and 18 months, 23.1% and 13.9% of patients, respectively, remained alive and progression-free. Corresponding OS rates at these time points were 60.2% and 51.5%. A confirmed ORR occurred in 12 patients (14.4%). Hypoxia occurred in 44 (53%) patients, median onset 31 days (range 1–270) and 29 (34.9%) required hospitalizations. Median hemoglobin (Hb) levels at baseline were 12.65 g/dl (range, 7.2-15.5), and 68 (82%) patients developed Hb decline on Belzutifan, 27 (32.5%) required transfusions. Conclusions: In an older, high-risk cohort with higher number of prior lines of therapy, Belzutifan yielded PFS/OS and ORR landmarks comparable to the trial supporting its effectiveness in the real world. Outcomes compared with LITESPARK-005 trial. Variable Real world cohort (Mayo Clinic) LITESPARK-005, Belzutifan arm Median PFS (months, 95% CI) 5.7 (3.8-8.7) 5.6 (3.9-7) 12-month PFS rate (%, 95% CI) 23.1% (14.4-33.4) 33.4% (28.4-33.5) 18-month PFS rate (%, 95% CI) 13.9% (7.7-23.9) 24% (19-29.4) 12-month OS rate (%, 95% CI) 60.2% (48.9-70.8) 67.9% (62.9-72.4) 18-month OS rate (%, 95% CI) 51.5% (40.6-62.9) 55.2% (50-60.1) ORR (%, 95% CI) 14.4% (7.7-23.9) 21.9% (17.8-26.5)

Key Points Atrial fibrillation after kidney transplantation is common and is associated with poor prognosis. Artificial intelligence applied to 12-lead electrocardiograms pretransplant accurately predicts the risk of new-onset atrial fibrillation in our cohort. High-risk artificial intelligence electrocardiography score was independently associated with increased risk of mortality and allograft failure. Background Incident atrial fibrillation (AF) is common after kidney transplantation (KTx) and is associated with worse clinical outcomes. Artificial intelligence electrocardiography (AI-ECG) algorithms have demonstrated efficacy in predicting risk of new-onset AF in the general population; however, their prognostic value in KTx recipients is relatively unknown. Methods Retrospective analysis was conducted on KTx recipients without AF, with at least one pretransplant ECG between 2011 and 2021 across three tertiary centers in the United States (Mayo Clinic sites in MN, AZ, and FL). A previously validated AI-ECG algorithm estimated the probability of incident AF for each patient. Based on AI-ECG probabilities, patients were categorized into high-risk and low-risk groups, with the optimal AI-ECG score cutoff determined. The incidence of new-onset AF, allograft failure, and mortality was compared between groups. Results Overall, 6246 patients (age 53.5±13.8 years; 58.9% male) were included. Pretransplant AI-ECG probability of AF ≥5% was the optimal cutoff for high risk of incident AF (sensitivity 72%, specificity 62%). High-risk scores were associated with true new-onset AF at 30 days (adjusted hazard ratio [aHR], 2.89; 95% confidence intervals [CI], 2.05 to 4.09; P < 0.001), 3 years (aHR, 2.54; 95% CI, 1.99 to 3.26, P < 0.001), and 5 years post-transplant (aHR, 2.48; 95% CI, 1.99 to 3.09, P < 0.001). High-risk AI-ECG scores were also associated with increased mortality (aHR, 1.56; 95% CI, 1.30 to 1.88, P < 0.001) and overall allograft failure (aHR, 1.50; 95% CI, 1.30 to 1.75, P < 0.001) through the 5-year follow-up. Conclusions This pretransplant AI-ECG parameter identified patients at increased risk of new-onset AF post-KTx and provided prognostic utility. Overall, this easy to obtain tool allows for risk stratification of patients who may benefit from closer monitoring, targeted risk factor modification, and early intervention.

BACKGROUND AND OBJECTIVES: External ventricular drain (EVD) placement is the most commonly performed bedside emergent neurosurgical procedure. However, catheter misplacement occurs in approximately 25% of cases, potentially resulting in neurological injury or the need for repositioning, highlighting the need for safer and more accurate placement techniques. This review aims to (1) summarize technological innovations developed to enhance the accuracy and safety of bedside EVD placement, (2) evaluate their reported benefits and limitations, and (3) identify key barriers to their clinical implementation in critical care settings. METHODS: We conducted a systematic review of PubMed, Embase, and Scopus combining terms related to EVD placement and assistive technologies. All the studies evaluated EVD placement using simulations, phantoms, cadavers, and patient procedures conducted at the bedside and in the operating room. RESULTS: We identified 3898 records, and 76 studies met the inclusion criteria. Of these, 30 studies compared stereotactic systems, ultrasound guidance, frameless neuronavigation, and augmented or mixed reality technologies to the freehand technique. These studies consistently reported improved accuracy with advanced systems although complication rates and procedural times were variably reported. An additional 35 and 12 studies evaluated nonimmersive and immersive technologies, respectively, in noncomparative settings. Most of these demonstrated high first-pass success and mean tip deviations under 3 mm. Variability in the definitions of accuracy and procedural metrics limited data synthesis, but the overall findings suggest a promising trajectory for technology-assisted EVD placement. CONCLUSION: Assistive technologies for EVD placement show potential to improve accuracy and safety over the freehand technique. These early successes signal potential to improve bedside neurosurgical care. Realizing this promise will require future studies to standardize accuracy metrics, report time, and safety outcomes consistently and validate performance in real-world critical care settings.

Background: Transthyretin (TTR) cardiac amyloidosis is a progressive cardiomyopathy with high mortality; however, the role of implantable cardioverter-defibrillators (ICDs) in this population remains unclear. Methods: This retrospective cohort study included patients with confirmed TTR cardiac amyloidosis, with or without ICDs, from January 1, 2001, to December 31, 2024, across all three Mayo Clinic sites (Arizona, Florida, and Minnesota). Diagnosis was confirmed by endomyocardial biopsy or abnormal technetium pyrophosphate (PYP) scintigraphy. A 1:4 propensity score-matched cohort of non-ischemic cardiomyopathy (NICM) patients with ICDs served as a control group. The primary outcome was all-cause mortality, comparing transthyretin cardiac amyloidosis (TTR-CA) patients by ICD status and against matched NICM patients. Secondary analyses evaluated predictors of mortality, including the use of tafamidis and the indication for ICD (primary vs. secondary prevention). Kaplan–Meier and Cox regression analyses were used to assess predictors of survival and mortality. Results: A total of 463 patients with confirmed TTR cardiac amyloidosis were included. The median follow-up duration was 7.4 years (interquartile range (IQR): 5.3–9.2 years) for the non-ICD group and 6.8 years (IQR: 4.5–9.0 years) for the ICD group. The median age was 74.5 years (IQR: 68.0–80.0 years), and 92.9% of patients were male. Among them, 206 (44.5%) received ICDs and 257 (55.5%) did not. ICD recipients were younger (71.0 vs. 77.0 years; p = 0.001) and had higher rates of hypertension (62.6% vs. 45.6%; p = 0.001), chronic kidney disease (CKD) (62.6% vs. 44.4%; p = 0.001), and diabetes (30.1% vs. 21.8%; p = 0.043). Median left ventricular ejection fraction was lower in the ICD groups (43% vs. 54%; p = 0.007), and N-terminal pro-B-type natriuretic peptide (NT-proBNP) levels were higher in the ICD group (2259.0 pg/mL vs. 1503.0 pg/mL; p = 0.007). Among ICD recipients, 157 (76.2%) received the device for primary prevention, while 48 (23.3%) received the ICD for secondary prevention. Appropriate shocks were delivered in 22 patients (10.6%), primarily for ventricular tachycardia (n = 18) and ventricular fibrillation (n = 4). Inappropriate shocks occurred in six patients (3.0%), and 12 patients (5.8%) experienced device-related complications. Over 10 years of follow-up, ICD implantation did not confer a survival benefit for patients with TTR-CA compared to those without an ICD (p = 0.74). In contrast, a 1:4 propensity-matched NICM cohort with ICDs, which had a median follow-up of 7.1 years (IQR: 4.6–8.8 years), showed significantly improved survival than TTR-CA patients with ICDs (p = 0.034). Among the TTR-CA patients with ICDs, neither the use of tafamidis (p = 0.10) nor the ICD indication (primary vs. secondary prevention; p = 0.85) influenced mortality. In the Cox regression analysis, predictors of mortality in TTR-CA patients included older age (hazard ratio (HR) 1.048; p = 0.001), CKD (HR 1.637; p = 0.029), troponin T >50 ng/L (HR 1.594; p = 0.031), NT-proBNP >3000 pg/mL (HR 1.514; p = 0.050), and ejection fraction <40% (HR 1.935; p = 0.003). ICD implantation was not associated with improved survival (HR 0.932; p = 0.763). Conclusions: In conclusion, our data suggest that ICD therapy may not provide a significant overall survival benefit in older TTR-CA patients with impaired pump function; thus, prospective studies are warranted before any changes to clinical practice are considered. Key predictors of mortality included reduced ejection fraction and elevated cardiac biomarkers. Additional prospective studies are needed to clarify the role of ICDs in treatment strategies for patients with TTR-CA.

Introduction: Optimizing patient selection and time-to-intervention will be crucial to the success of future clinical trials seeking to implement therapies and improve intracerebral hemorrhage (ICH) patient outcomes. Automated and accurate ICH and perihematomal edema (PHE) volumetry carries significant potential to strengthen trials targeting ICH and PHE expansion. Aims: The Quantification of Hematoma and Perihematomal Edema Volumes in Intracerebral Hemorrhage (QUANTUM) trial was designed with aims of (1) validating a fully automated artificial intelligence (AI) method for supratentorial, spontaneous ICH and PHE volumetry against the manual and semiautomated quantification methods and (2) establishing a model framework for the translation of AI technology through conduct of a clinical trial under the equivalence hypothesis paradigm. Methods: This trial was a prospective, randomized, clinical study with blinded outcome assessment. Six independent raters were randomized in a 1:1 ratio to ICH and PHE volumetry from non-contrast CT scans in the Virtual International Stroke Trials Archive with either the semiautomated or manual quantification method. The primary outcomes were equivalence between (1)ICH volume measurements on admission (<24h from symptom onset) for the manual versus the fully automated or the sem-automated quantification methods, and (2)PHE volume measurements on day 3 (72±12h from symptom onset) for the manual versus the fully automated or the semiautomated quantification methods. The required sample size was 126 ICH and 126 PHE measurements, based upon a significance level of 0.025 to account for the two equivalence tests. Results: The geometric mean relative discrepancies between the fully automated and the manual ICH and PHE volume measurements did not cross the pre-specified 10% equivalence margin. The geometric mean relative discrepancies between the semiautomated and the manual ICH and PHE volume measurements did cross the pre-specified 10% equivalence margin. The fully automated method (mean 10.0±2.7s/scan) was significantly faster than the manual (mean 189±199 s/scan; P <0.001) and semiautomated (mean 79±68s/scan; P <0.001) methods. Conclusions: ICH and PHE volumes quantified by AI could replace clinician measurements within a 10% error margin and with substantially greater efficiency. The semiautomated method could not replace manual measurements. This trial demonstrated feasibility of the equivalence hypothesis framework for AI validation studies.

Abstract Aims Pre-capillary pulmonary hypertension (PHTN), typically due to intrinsic pulmonary pathology, requires right heart catheterization (RHC) for diagnosis. ePLAR, the echocardiographic pulmonary to left atrial ratio [calculated: tricuspid regurgitation maximum velocity (TRVmax) divided by mitral valve E to e′ wave ratio (E/e′)], has been demonstrated as a promising non-invasive surrogate for characterization of pre-capillary PHTN. As such, this study aimed to further evaluate the diagnostic utility of ePLAR in a large cohort of patients with PHTN. Methods and results All RHCs between 2018 and 2023 across three tertiary centres were retrospectively reviewed. Patients with PHTN were characterized as pre-capillary or post-capillary (isolated or combined). Concurrent ePLAR was calculated. ROC analysis evaluated the accuracy of ePLAR in detecting pre-capillary PHTN. Overall, 2235 patients had PHTN by RHC, mean age 63.5 ± 14.7 years, 53.5% male. Of these, 43.3% had pre-capillary PHTN. ROC analysis demonstrated ePLAR to have good accuracy in detection of pre-capillary PHTN (AUC 0.760), with a cutoff of 0.20 m/s demonstrating a sensitivity and specificity of 77% and 60% respectively. This persisted on sensitivity analysis for contemporaneous investigations and for right ventricular function. Conclusion In this large cohort, ePLAR was validated as an accurate tool for identifying patients with pre-capillary PHTN. This widely available, non-invasive, and reproducible parameter can have an important role in the work up of PHTN and may guide the need for invasive testing.

OBJECTIVE Achieving submillimetric accuracy in stereotactic neurosurgery remains critical for safely targeting deep brain structures. Current workflows rely on intraoperative CT to register stereotactic frames to preoperative imaging, but this introduces additional radiation exposure, cost, and workflow complexity. Three-dimensional surface scanning (3DSS) potentially offers a fast radiation-free alternative capable of capturing detailed craniofacial anatomy and frame geometry. Therefore, the aim of this cadaveric study was to evaluate the feasibility, accuracy, and navigational compatibility of integrating 3DSS into stereotactic workflows as a potential replacement for intraoperative CT. METHODS A human cadaveric head was imaged with both thin-slice CT and structured-light 3DSS after mounting stereotactic frame components. CT, 3DSS, and computer-aided design (CAD) models were registered using a multistep point cloud alignment pipeline involving the fast point feature histogram, the random sample consensus method, and the iterative closest point algorithm. Accuracy was assessed using surface deviation error, root mean square error (RMSE), and fiducial registration error within a commercial neuronavigation platform. RESULTS Fusion of the 3D facial scan with CT provided a mean surface deviation error of approximately 0.4 mm. Registration of the 3DSS-derived N-bar localizer to its CAD reference model produced a frame alignment RMSE of 1.3 mm. Within the navigation system, the 3DSS-based workflow achieved a mean fiducial registration error of 0.14 mm (range 0.02–0.20 mm), outperforming the conventional CT-based method (mean error 0.20 mm, range 0.09–0.40 mm). CONCLUSIONS The 3DSS-based method enabled precise submillimetric stereotactic registration without the need for intraoperative CT, reducing radiation exposure and operative complexity. This workflow is fully compatible with existing navigation systems and could serve as a practical radiation-free alternative in stereotactic neurosurgery. Future work will focus on automating frame detection, incorporating artificial intelligence–driven fusion methods, and validating this approach in live surgical settings.

OBJECTIVES: This systematic review evaluates artificial intelligence (AI)-based predictive models developed for early sepsis detection in adult hospitalized patients. It explores model types, input features, validation strategies, performance metrics, clinical integration, and implementation challenges. DATA SOURCES: A systematic search was conducted across PubMed, Scopus, Web of Science, Google Scholar, and CENTRAL for studies published between January 2015 and March 2025. STUDY SELECTION: Eligible studies included those developing or validating AI models for adult inpatient sepsis prediction using electronic health record data and reporting at least one performance metric (area under the curve [AUC], sensitivity, specificity, or F1 score). Studies focusing on pediatric populations, lacking quantitative evaluation, or unpublished in peer-reviewed journals were excluded. DATA EXTRACTION: Data extraction followed preferred reporting items for systematic reviews and meta-analyses guidelines. Extracted variables included study design, patient population, model type, input features, validation approach, and performance outcomes. DATA SYNTHESIS: A total of 52 studies met the inclusion criteria. Most used retrospective designs, with limited prospective or real-time clinical validation. Commonly used algorithms included random forests, neural networks, support vector machines, and deep learning architectures (long short-term memory, convolutional neural network). Input data varied from structured sources (vital signs, laboratory values, demographics) to unstructured clinical notes processed via natural language processing. Reported AUC values ranged from 0.79 to 0.96, indicating strong predictive performance across models. CONCLUSIONS: AI models demonstrate significant promise for early sepsis detection, outperforming conventional scoring systems in many cases. However, generalizability, interpretability, and clinical implementation remain major challenges. Future research should emphasize externally validated, explainable, and scalable AI solutions integrated into real-time clinical workflows.