Objective Conventional prediction models fail to integrate the constantly evolving nature of critical illness. Alternative modelling approaches to study dynamic changes in critical illness progression are needed. We compare static risk prediction models to dynamic probabilistic models in early critical illness. Design We developed models to simulate disease trajectories of critically ill COVID-19 patients across different disease states. Eighty per cent of cases were randomly assigned to a training and 20% of the cases were used as a validation cohort. Conventional risk prediction models were developed to analyse different disease states for critically ill patients for the first 7 days of intensive care unit (ICU) stay. Daily disease state transitions were modelled using a series of multivariable, multinomial logistic regression models. A probabilistic dynamic systems modelling approach was used to predict disease trajectory over the first 7 days of an ICU admission. Forecast accuracy was assessed and simulated patient clinical trajectories were developed through our algorithm. Setting and participants We retrospectively studied patients admitted to a Cleveland Clinic Healthcare System in Ohio, for the treatment of COVID-19 from March 2020 to December 2022. Results 5241 patients were included in the analysis. For ICU days 2–7, the static (conventional) modelling approach, the accuracy of the models steadily decreased as a function of time, with area under the curve (AUC) for each health state below 0.8. But the dynamic forecasting approach improved its ability to predict as a function of time. AUC for the dynamic forecasting approach were all above 0.90 for ICU days 4–7 for all states. Conclusion We demonstrated that modelling critical care outcomes as a dynamic system improved the forecasting accuracy of the disease state. Our model accurately identified different disease conditions and trajectories, with a <10% misclassification rate over the first week of critical illness.

BACKGROUND: Significant controversy continues to confound patient selection and referral for revascularization and mitral valve intervention in patients with ischemic cardiomyopathy (ICM). Cardiac magnetic resonance (CMR) enables comprehensive phenotyping with gold-standard tissue characterization and volumetric/functional measures. Therefore, we sought to determine the impact of CMR-enriched phenomapping patients with ICM to identify differential outcomes following surgical revascularization and surgical mitral valve intervention (sMVi). METHODS: Consecutive patients with ICM referred for CMR between 2002 and 2017 were evaluated. Latent class analysis was performed to identify phenotypes enriched by comprehensive CMR assessment. The primary end point was death, heart transplant, or left ventricular assist device implantation. A multivariable Cox survival model was developed to determine the association of phenogroups with overall survival. Subgroup analysis was performed to assess the presence of differential response to post-magnetic resonance imaging procedural interventions. RESULTS: A total of 787 patients were evaluated (63.0±11.2 years, 24.8% women), with 464 primary events. Subsequent surgical revascularization and sMVi occurred in 380 (48.3%) and 157 (19.9%) patients, respectively. Latent class analysis identified 3 distinct clusters of patients, which demonstrated significant differences in overall outcome ( P <0.001). Latent class analysis identified differential survival benefit of revascularization in patients as well as patients who underwent revascularization with sMVi, based on phenogroup classification, with phenogroup 3 deriving the most survival benefit from revascularization and revascularization with sMVi (hazard ratio, 0.61 [0.43–0.88]; P =0.0081). CONCLUSIONS: CMR-enriched unsupervised phenomapping identified distinct phenogroups, which were associated with significant differential survival benefit following surgical revascularization and sMVi in patients with ICM. Phenomapping provides a novel approach for patient selection, which may enable personalized therapeutic decision-making for patients with ICM.

Background: Immunocompromised patients, especially those receiving B–cell–depleting therapy (BCDT), remain at high risk for severe coronavirus disease 2019 (COVID–19). Although monoclonal antibodies showed benefit pre–Omicron, the real–world effectiveness of the oral antivirals in BCDT recipients is unclear. Methods: We emulated a target trial using the Cleveland Clinic electronic health record. Adult patients (≥18 years) with immune–mediated inflammatory diseases who received rituximab or ocrelizumab and tested positive for severe acute respiratory syndrome coronavirus 2 between 19 December 2021 and 30 June 2025 were eligible. The primary endpoint was hospitalization or death within 21 days; the secondary endpoint was death within 90 days. Propensity scores adjusted for confounding, and Weighted Cox regression models were used to evaluated treatment effects. Results: Among 255 853 COVID–19 cases, 876 met criteria for the nirmatrelvir/ritonavir cohort (411 treated, 465 untreated) and 566 for the molnupiravir cohort (91 treated, 475 untreated). In the nirmatrelvir/ritonavir cohort, 18 (4.4%) of treated versus 43 (9.2%) of untreated patients were hospitalized or died within 21 days. After weighting, nirmatrelvir/ritonavir reduced the hazard by 44% (hazard ratio [HR], 0.56; 95% confidence interval [CI], 0.31–0.99). In the molnupiravir cohort, 6 (6.6%) of treated and 46 (9.7%) of untreated patients met the primary endpoint; the weighted HR was 0.56 (95% CI, 0.24–1.33), not statistically significant. Conclusions: Early outpatient nirmatrelvir/ritonavir significantly lowers the risk of COVID–19–related hospitalization or death in patients on BCDT. Molnupiravir showed a nonsignificant trend toward benefit. These findings support prioritizing nirmatrelvir/ritonavir for high–risk immunocompromised populations.

Due to the heterogeneity of sarcoidosis, there is a need to define clinical phenotypes to allow for tailoring of clinical care and identification of more homogenous populations to facilitate research. We utilized data from a prospectively collected registry of sarcoidosis patients seen at a single quaternary referral center between January 2019 and February 2021. We used multiple correspondence analysis (MCA) and k-means clustering to investigate if the clusters previously identified in the GenPhenReSa study were reproducible in a US population. We also investigated if these clusters were stable when the population was stratified by race. We replicated 3 of the 5 clusters seen in the GenPhenReSa study in our cohort. We likewise identified similar clusters between White and Black patients with sarcoidosis. Differences in organ manifestations associations between White and Black patients were seen primarily in relation to cardiac, neurologic, and ocular involvement. The organ clusters of liver-spleen, isolated pulmonary, and musculoskeletal-skin were reproducible in a US cohort, and in both Black and White patients.

**Author Listing - Lead/ Submitting Author (we will be communicating with this individual): Name, Institution, City and State where the work was | POSTER

**Author Listing - Lead/ Submitting Author (we will be communicating with this individual): Name, Institution, City and State where the work was | POSTER