Citations

Abstract Background: Spatial Reasoning AI (SRAI) systems are emerging as a transformative class of clinical artificial intelligence, enabling real-time interpretation of complex anatomy in imaging, surgery, and chronic disease management. Unlike traditional AI approaches constrained to tabular or 2D inputs, SRAI models incorporate geometric priors, multimodal fusion, and 3D anatomical reasoning to support high-impact tasks such as volumetric segmentation, lesion localization, intraoperative guidance, and prognostic modeling. Objective: This scoping review aimed to (1) map the emerging landscape of SRAI in healthcare, (2) assess its clinical performance across diagnostic, prognostic, and workflow applications, and (3) identify opportunities and barriers for safe translation into practice. Methods: Following PRISMA 2020 guidelines, we systematically searched PubMed, Springer Nature, and IEEE Xplore (including CVPR, ICCV, ISBI, and EMBC) for original studies implementing spatially aware AI models with clinical validation or workflow relevance. Data extraction captured model architecture, input modality, anatomical scope, application domain, and performance outcomes. Results: We identified eligible studies spanning oncology, neurology, cardiology, surgery, and infectious diseases. SRAI systems consistently outperformed conventional methods in segmentation, lesion classification, and risk stratification. A paradigm shift is underway from task-specific CNNs to foundation-level models, including MedSAM and MedCLIP for zero-shot segmentation and retrieval, and NeRF-based pipelines for anatomically faithful 3D reconstruction. Randomized evidence shows that SRAI can improve workflow efficiency without compromising diagnostic accuracy. Conclusion: SRAI models demonstrate strong potential to advance diagnostic precision, prognostic accuracy, and workflow integration across multiple clinical domains. Their safe adoption will require standardized evaluation metrics, prospective multi-institutional validation, and interpretable outputs. As foundation models evolve to better encode spatial priors, SRAI is poised to become a cornerstone of workflow-integrated precision medicine.

Abstract Background: Spatial Reasoning AI (SRAI) systems are reshaping clinical practice by enabling real-time interpretation of complex anatomical structures in medical imaging, surgery, and chronic disease management (Lin et al., 2023, Qiao et al., 2021; Song et al., 2024) . Unlike traditional AI approaches that rely on tabular or 2D data, SRAI models incorporate geometric priors, multimodal fusion, and 3D anatomical reasoning to support high-impact tasks such as volumetric segmentation, disease localization, and intraoperative guidance (Milletari et al., 2016; Hatamizadeh et al., 2022; Zhang et al., 2022) . Objective: This scoping review aimed to (1) characterize the emerging landscape of spatial reasoning models in healthcare, (2) evaluate their clinical performance across diagnostic and prognostic applications, and (3) identify architectural trends that underpin successful translation into practice. Methods: Following PRISMA guidelines, we systematically screened PubMed to identify studies that implemented spatially aware AI models with real-time or clinically validated outcomes (Page et al., 2021) . Data extraction captured model architecture, modality, anatomical scope, clinical application, and performance metrics. Results: We identified 124 eligible studies across the clinical domains of oncology, neurology, cardiology, orthopedic surgery, and infectious diseases (Nie et al., 2019; Hao et al., 2020; Tasken et al., 2023) . SRAI systems consistently outperformed conventional methods in segmentation, lesion classification, and risk stratification. (Lin et al., 2023; Qiao et al., 2021; Mahootiha et al., 2024) . A paradigm shift from task-specific CNNs to foundation models that reason across spatial structures has emerged, beginning with adaptations of generalist models such as MedSAM and MedCLIP for zero-shot segmentation and cross-modal retrieval, and NeRF-based reconstruction for anatomically faithful 3D modeling (Ma et al., 2024; Mildenhall et al., 2020; Radford et al., 2021). Beyond technical performance, randomized evidence shows that SRAI systems can improve workflow efficiency without compromising outcomes (Lang et al., 2023) . Conclusion: SRAI models hold promise for enhancing diagnostic accuracy, prognostic precision, and clinical efficiency across diverse applications. Their successful integration will depend on standardized evaluation metrics, robust validation, and interpretable outputs. As foundation models evolve to better encode spatial priors, SRAI is poised to become a cornerstone of personalized and procedural medicine.

Background: VICTORION-INITIATE (NCT04929249) showed significantly greater LDL-C lowering in patients (pts) with ASCVD with an inclisiran first (IF) strategy (adding inclisiran immediately on failure to achieve LDL-C <70 mg/dL on maximally tolerated statin therapy) vs usual care (UC). Tokenization can be leveraged to link clinical trial data to pt-level real-world data (RWD) to provide a comprehensive overview of the pt journey pre- and post-trial and enhance understanding of clinical trial and long-term pt outcomes. Aim: To utilize tokenization and linked RWD to characterize the pt journey pre-VICTORION-INITIATE enrollment. Methods: Eligible pts from VICTORION-INITIATE were deidentified and tokenized with unique IDs, allowing linking to RWD (EMR, claims, hospital, and lab records) to follow the pt journey in the 3-year pre-trial and 2-year post-trial periods. Here we present demographics, clinical characteristics, and lipid-lowering therapy (LLT) use of pts with linked pre-trial RWD. Results: Of 450 pts randomized, 383 (85.1%) were linked with pre-trial RWD (IF arm, n=193; UC arm, n=190); 26.9% were female, first available recorded mean (SD) age was 64 (9.36) years. Among the 74.2% (n=284) of pts with linked, pre-trial, lipid test RWD, mean (SD) earliest and latest LDL-C values were 100.1 mg/dL (35.70) and 99.3 mg/dL (30.06), respectively, for pts in the IF arm vs 97.3 mg/dL (40.43) and 97.8 mg/dL (27.60) for pts in the UC arm ( Table ). Overall, in the pre-trial period, LLT was dispensed/prescribed to 66.1% of pts, statins to 64.5%, ezetimibe to 4.4%, and PCSK9i to 4.4%. Of pts with ≥1 recorded ASCVD events in the pre-trial period (n=141), 53.2% and 68.8% had available LLT dispensing/prescription records pre- and post-event, respectively. Conclusions: In the 3 years pre-VICTORION-INITIATE, mean LDL-C levels were above recommended goals, non-statin LLT prescription was infrequent, and ~31% of pts with ≥1 recorded ASCVD event had no evidence of post-event LLT use, highlighting a need for more aggressive LDL-C management. Full data visibility for each pt is a limitation of this approach. Tokenization post-trial will allow for long-term evaluation of LLT practices.

Background: Premature ventricular contractions (PVCs) are associated with an increased risk of implanting implantable cardioverter defibrillator (ICD). Therefore, our study aimed to predict the risk factors for implantable cardioverter defibrillator in patients with premature ventricular contractions. Methods: A total of 309 patients (male 57%, mean age 66.73±16.12) in Olmstate Countywho had premature ventricular contractions and received a 24-hour electrocardiography monitoring were included between January 1994 and September 2010 at Mayo Clinic, Rochester, Minnesota. All these patients were continuously followed o Jun 2015. Student’s t-test, Chi-square, univariate and multivariate Logistic regression models were used to estimate the relationship between the risk factors and the incidence of implanting implantable cardioverter defibrillator (ICD) in PVCs patients. Kaplan–Meier curve of survival in PVCs patients with ICD or free-ICD. Results: The average follow-up time was 9.3±3.9 years, in which 43 (13.9%) patients died and 24 (7.8%) patients had received ICD therapy. Sex, age, pulse pressure, early repolarization syndrome (ERS), heart rate, PVC coupling interval, QRS duration, left ventricular ejection fraction (LVEF), left ventricular end-diastolic dimension (LVEDD) and left ventricular end-systolic dimension (LVESD), diabetes were significantly associated with the mortality and the incidence of implanting ICD. Using a multivariate logistic regression model adjusted for potential confounders, showed that LVEF (OR=0.916, 95% CI:0.881 to 0.953, p<0.001) and pulse pressure (OR=1.032, 95% CI:1.001-1.064, p=0.045) were independent factors for ICD implant in PVCs patients. Moreover, implantation of ICD may not reduce mortality in patients with ventricular premature. Conclusions: Increased pulse pressure was associated a higher incidence of ICD in PVCs patients, among many electrocardiographic and clinical variables studied. We also showed that PVs patients with ICD were at high risk of death, and the great pulse pressure was linked with higher morality in PVCs patients.

Abstract Importance: The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infections outbreak in China is now a global issue. There is only a limited understanding of the clinical characteristics of patients with SARS-CoV-2 infections is available.Objective:To describe the characteristics, management strategies, and outcomes of critically ill patients with SARS-CoV-2 infection.Design, Setting, and Patients: This is aretrospective, multi-center case series of 50 critically ill patients with confirmed SARS-CoV-2 infection who were admitted at Zhongnan Hospital of Wuhan University and Wuhan Pulmonary Hospital in Wuhan, China, from January 8 to February 9, 2020.Exposures:Documented Corona Virus Disease, 2019 (COVID-19).Main Outcome Measures: Demographic, clinical, laboratory, imaging data were collected along with management strategies, complications and outcomes of enrolled individuals. Results Fifty critically ill patients with SARS-CoV-2 infections were enrolled. Their median age was 62 (range, 29-92) [IQR,49.5-69.0] years, 68% were male, and 28 (56%) patients had comorbidities, the most common being hypertension. In this cohort, 20(40%) patients survived ,16(32%) patients died, and the rest remained hospitalized. The invasive mechanical ventilator was used in 36(72%) patients with 15(30%) of them requiring prone positioning, and 17(34%) switched to ECMO. The compliance scores of lungs (Cstat)on the day of ICU admission among survivors were higher than those in non-survivors [42.0(18.0-47.0), vs. 19.5(14.0-24.2), p=0.038].The blood IL-6 levels and neutrophils counts at the first day of ICU admission were significantly higher in non-survivors compared to survivors [123.7(85.3-228.8), vs. 20.2(6.8-67.2) ng/ml, p=0.025 for IL-6, and 20.2(6.8-67.2) vs. 4.01(1.99-7.05) × 10⁹/L, p=0.02 for neutrophils counts].The heart rates, PaCO2, lung injury scale (LIS), and positive end-expiratory pressure levels were constantly higher for 10 days in non-survivors than those who survived (p<0.05). The frequency of vasopressor uses and neuromuscular blockers was higher in non-survivors from day 1 to day 10 compared to survivors (p<0.05). In the whole cohort, the most common complications were ARDS (97%), shock (44%), arrhythmia (38%), acute cardiac injury (26%), and acute kidney injury (22%). A secondary bacterial infection was noted in 17(34%) patients. Univariate analysis indicated that lower lung complianceand higher neutrophil counts at the day of ICU admission were related to higher mortality (p-0.03, and 0.04, respectively)ConclusionWe demonstrated that SARS-CoV-2 infection-related critical illness predominantly affected old individuals with comorbidities and characterized by severe hypoxemic respiratory failure, often requiring prolonged mechanical ventilation and rescue therapies. Low lung compliance and persistently elevated PaCO2 indicated poor outcomes.

AbstractBackgroundHuman factors research has suggested benefits of consistent teams yet no surgical team consistency measures have been established for teamwork improvement initiatives.MethodsRetrospective analysis was conducted of teams performing consecutive elective procedures of unilateral primary total knee and hip replacement between June 2008 and May 2010 at a large tertiary medical center. Surgeons who performed fewer than 50 cases of the procedures during the study period were excluded. A team was defined as consistent when its nurse and surgical technologist members were both among the three most frequent working with the surgeon during the study period. Odds ratios for prolonged operative time (in the longest quartile), prolonged hospital stay (longer than median), and 30‐day all‐cause readmissions were adjusted for patient characteristics (sex, age, comorbidity, American Society of Anesthesiology status), surgery characteristics (procedures, time of day), and surgeons.ResultsInconsistent teams performed 61 % of the 1,923 cases with eight surgeons, each of which worked with a median of 43.5 (range, 28–58) nurses and 29 (range, 13–47) technologists. Inconsistent teams were associated with higher likelihood of prolonged operative time [odds ratio 1.52, 95 % confidence interval (CI) 1.20–1.91], higher likelihood of prolonged hospital stay (odds ratio 1.51, 95 % CI 1.23–1.86), and more readmissions (adjusted odds ratio 1.42, 95 % CI 1.07–1.89).ConclusionsTeam consistency was an independent predictor of prolonged operative time, prolonged hospital stay, and 30‐day hospital readmission in elective, primary, unilateral total knee, and hip replacement procedures, after adjusting for patient and surgery characteristics and surgeons.