Citations

ABSTRACT Objectives Systematic evidence reviews (SERs) produced by the U.S. Agency for Healthcare Research and Quality (AHRQ) Evidence‐based Practice Center (EPC) Program use contextual questions to provide context and background information on the topic. There is currently no standardized approach to address contextual questions in systematic reviews. This study explored the use of publicly available large language models (LLMs) in addressing contextual questions. Study Design Using a set of 20 published and 5 yet to be published SERs, we selected one contextual question per report and used it as a prompt to elicit answers from an LLM (ChatGPT, Bard, Claude, or Perplexity). Two independent reviewers rated the results using a priori established evaluation criteria ( https://osf.io/4k3cu/ ), comparing the response in the SER to LLM‐generated responses. The study was guided by six research questions addressing feasibility, validity of content, validity of structure, mistakes, congruence between responses, and incremental validity of using LLMs to address contextual questions. Results Using minimal prompt engineering produced relevant responses and documented the feasibility of LLM‐generated answers to contextual questions. Responses differed in content and format and are not reproducible (e.g., LLMs update regularly), but LLMs were able to produce articulate, clinically plausible, and well‐structured responses. We detected few factual errors, contradictions, and no instance of suspected bias, but citations supporting LLM‐generated responses could often not be produced or could not be verified (‘confabulations’). Congruence with human generated responses varied, with LLM‐generated responses providing more background on the topic and SERs providing more nuanced answers in response to the contextual question. Results regarding incremental validity were mixed and may depend on the tool. Conclusion LLMs are potentially helpful in addressing contextual questions in systematic reviews but human expertise remains essential for using the generated information in a meaningful way.

Abstract Background: Hemorrhagic necrotizing pancreatitis (HNP) is a highly lethal form of pancreatitis that lacks mechanism-based therapy. Intra-acinar enzyme activation is fundamental to pancreatitis initiation, yet it does not explain why only a subset of episodes progress to catastrophic hemorrhagic necrosis. Objective: To identify determinants of progression to lethal HNP and test a mechanism-based preventive intervention. Design: Using humanized trypsinogen mouse models and secretagogue regimens, we compared high-dose cerulein with low/moderate-dose cerulein and bombesin. We quantified pancreatic protease activity, performed blinded histopathologic scoring, assessed hemorrhage/vascular disruption, and evaluated enzyme localization and ductal injury by PRSS1 immunostaining. De-identified human pancreatic histology from HNP were examined. We tested a repurposed strategy using clinically approved agents for other indications: secretin to stimulate ductal fluid secretion and isosorbide mononitrate (ISMN) to promote ductal outflow via sphincter of Oddi relaxation. Results: High-dose cerulein induced robust intrapancreatic protease activation but predominantly caused edematous pancreatitis. Paradoxically, bombesin and low/moderate-dose cerulein elicited lower total pancreatic protease activity yet produced severe, frequently lethal HNP with vascular disruption and lobular ischemic necrosis. HNP was associated with intraductal enzyme accumulation and ductal injury. Secretin or ISMN alone showed limited protection, whereas combined secretin+ISMN prevented hemorrhagic necrotizing injury. Conclusion: Pancreatitis severity is determined not only by the magnitude of enzyme activation but also by where enzyme activity is exerted, implicating intraductal enzyme accumulation as a mechanistic determinant and therapeutic target in HNP. Coordinated enhancement of ductal secretion and outflow with clinically approved agents prevents HNP in vivo, supporting a translatable prevention strategy.

160 Background: To our knowledge, no tools are available to guide care in oligometastatic prostate cancer (omPC). Here, we leveraged individual patient-level data pooled from 8 randomized trials in the X-Met collaboration to construct CEREBRO, a tool for predicting patient-specific responses to various treatment scenarios. Methods: Patients with omPC from the STOMP, ORIOLE, SABR COMET, EXTEND continuous ADT, EXTEND intermittent ADT, ARTO, RADIOSA, and RAVENS trials were included. 6 patients (1%) with missing data were excluded. Treatment scenarios were defined as active surveillance (AS), metastasis-directed therapy (MDT), systemic therapy (ST), or MDT+ST. To account for varying baseline hazards, Weibull models were fit for overall survival (OS); progression-free survival (PFS); and castration-resistance free survival (CRFS) for patients with hormone-sensitive disease. Endpoint definitions were harmonized per X-Met framework. Results: Of 600 patients included (treatment scenarios: AS=50; MDT=153; ST=172; MDT+ST=225), 405 (67%) had hormone-sensitive omPC. With median follow-up of 38 months, median PFS, CRFS, and OS of all patients were 18, 70, and 84 months. In order of relative contribution, inputs for predicting outcomes were: treatment scenario, stage, prostate-specific antigen annotated by ADT status, metachronous vs synchronous presentation, imaging type, number of metastases, hormone sensitivity, and age. The model for each endpoint showed robust fit, calibration, superiority against the null model (per the likelihood ratio test), and risk stratification (Table). Median PFS times for model-defined low-, intermediate-, and high-risk strata were 46 months, 18 months, and 7 months ( p <0.0001). In addition, the PFS and OS models also showed strong discrimination via c index and a large interval of risk probabilities with additive clinical utility per decision-curve analysis (Table). A public webpage will be made available to facilitate prospective trial stratification, patient counseling for patient-specific outcome predictions for each potential treatment scenario, and forecasts comparing outcomes between treatment scenarios. Conclusions: CEREBRO is a high-performance tool, trained on robust prospective data from 8 randomized trials with routinely collected clinical variables, for predicting risk and treatment outcomes in omPC. Given the uncertainties in optimal timing of MDT and combination with ST, this tool has been built to facilitate current decision making. External validation is planned. Model summary. Endpoint C index (95% CI) Calibration Likelihood ratio p Risk strata p Net clinical utility observed between thresholds PFS 0.72 (0.70-0.75) 1.04 <0.0001 <0.0001 0.2-1.0 CRFS 0.63 (0.60-0.70) 1.01 0.02 0.004 0.2-0.5 OS 0.81 (0.76-0.86) 1.01 <0.0001 <0.0001 0.05-1.0

496 Background: Approximately 10% of patients with renal cell carcinoma (RCC) have an associated venous tumor thrombus (VTT), the presence of which impacts surgical management and increases perioperative risk. Historical systemic therapy regimens had limited success in downsizing VTT; however, the impact of immune checkpoint inhibitors (ICIs) on VTT shrinkage is unknown. Our primary objective is to assess VTT shrinkage comparing responses to ICI versus non-ICI regimens. Methods: This multicenter retrospective cohort study included patients with RCC and VTT from 2006-2024 who received upfront systemic therapy. Patients were evaluated at tertiary referral centers across Mayo Clinic Rochester, Florida, and Arizona. Systemic therapies were categorized as ICI versus non-ICI regimens. VTT length was quantified by distance relative to the superior border of the renal vein ostium. The primary outcome was change in VTT length on serial imaging as measured by absolute length reduction (cm), RECIST version 1.1 criteria, and Mayo VTT level reclassification. Linear mixed-effect models (LMMs) identified associations with VTT shrinkage and clinicopathologic features. Results: A total of 108 patients with median age 65 years (IQR 57-71) were included, of whom 49 (45%) received ICI regimens. ICI regimens were associated with greater VTT shrinkage (1cm vs 0.1 cm, p=0.025), increased RECIST responses (51% vs 32%, p=0.046), and a higher rate downstaging per Mayo VTT level (33% vs 12%, p=0.008) compared to non-ICI regimens. On multivariable LMM analysis, ICI regimens remained more likely to result in VTT shrinkage compared to non-ICI regimens (-0.12 cm/month, 95% CI –0.24-0.00, p=0.046). Absence of abdominal lymphadenopathy was also associated with VTT shrinkage (-0.18cm/month, 95% CI -0.06- -0.30, p=0.004). Initial Mayo VTT level, IMDC risk, and number of metastatic sites were not associated with VTT shrinkage. Conclusions: Treatment with ICI-regimens resulted in VTT shrinkage and Mayo VTT level reduction. This is a particularly important area of exploration as radical nephrectomy and tumor thrombectomy carry up to a 40% risk of peri-operative major complications. These data support using ICI regimens for patients with RCC with VTT who are not candidates for upfront surgery and prioritizing ICI regimens in neoadjuvant clinical trials.