APA 2025, San Diego

This final topic in the series combined radiomics with mesenteric inflammation scoring to develop AI-driven predictors for AP severity. Machine-learning models trained on CT imaging were able to identify patterns of mesenteric fat attenuation, vascular changes, and bowel wall involvement that correlate with complications such as necrosis, organ failure, and feeding intolerance. Early AI-based prediction facilitates individualized nutrition plans, targeted anti-inflammatory therapy, and early ICU triaging. APA presenters highlighted that AI integration in AP imaging is rapidly approaching clinical translation.

Building on earlier meta-analytic findings, this session offered pragmatic guidelines for clinicians. LR was reaffirmed as the preferred fluid for routine and high-risk ERCP. Pre-procedure bolus followed by post-procedure infusion provided the best PEP protection. In patients with cardiac or renal compromise, moderate bolus + careful monitoring was advised. The session stressed that hydration should not replace NSAID prophylaxis but complement it. Tailoring hydration volume to patient risk factors was strongly emphasized.

Sophisticated radiomic analyses of CT and MRI can identify subtle imaging patterns that predict early PDAC progression and treatment resistance. This session focused on machine-learning models that detect tumor heterogeneity, stromal density, vascular features, and metabolic signatures. Radiomic phenotypes correlated strongly with CA 19-9 kinetics, genomic instability markers, and response to neoadjuvant chemotherapy. Integrating radiomics with spatial transcriptomics may allow ultra-early detection of PDAC transformation in high-risk groups (CP, genetic syndromes, T3cDM). The APA emphasized that radiomics will become a key decision-support tool in PDAC precision oncology.

Pediatric ERCP is increasingly performed for biliary and pancreatic indications, but PEP remains a significant concern. The APA pediatric study compared two NSAID-based prophylactic strategies: rectal indomethacin vs intravenous ketorolac. Results indicated similar efficacy between the two in reducing PEP incidence, although indomethacin showed a slight trend toward better prevention in high-risk cases. Both medications were well-tolerated, with no major bleeding or renal complications reported. The study emphasized that adult dosing strategies cannot be directly applied to children; dosing must be weight-based and safety-monitored. The session concluded that NSAIDs are effective for pediatric PEP prophylaxis, but more prospective pediatric-specific trials are needed to establish standardized protocols.

This supplementary session expanded on mesenteric-based predictors by evaluating mesenteric fluid collections as indicators of disease severity and feeding tolerance. Mesenteric fluid reflects capillary leak, increased permeability, and intense inflammatory activity. Patients with large mesenteric fluid pockets had poorer tolerance to oral feeds, higher rates of ileus, and increased risk of paralytic bowel. Combining mesenteric fluid assessment with CT Severity Index improved predictive accuracy. The session emphasized that mesenteric biomarkers—imaging + clinical—may soon guide timing of enteral nutrition.

This session addressed whether specific crystalloid fluids—normal saline vs lactated Ringer’s (LR)—influence the incidence of post-ERCP pancreatitis (PEP). The meta-analysis consolidated randomized controlled trials comparing hydration protocols. Findings showed that aggressive hydration with LR significantly reduced PEP rates compared with normal saline. LR’s balanced electrolyte composition may counteract acidosis, stabilize pancreatic microcirculation, and mitigate early inflammatory activation. In contrast, normal saline can induce hyperchloremic metabolic acidosis, potentially worsening pancreatic injury. Combining LR hydration with rectal NSAIDs (indomethacin) achieved the greatest protective effect. The APA emphasized that hydration must be protocol-driven—large volumes without monitoring may risk pulmonary edema in high-risk patients. Overall, LR emerged as the preferred fluid for PEP prophylaxis, reinforcing a shift toward goal-directed hydration strategies in ERCP practice.

Feeding intolerance is common in moderate-to-severe acute pancreatitis (AP), delaying recovery and prolonging hospitalization. APA researchers presented an innovative mesenteric-based imaging classification system that uses CT features to predict which patients will struggle with early oral feeding. The system evaluates mesenteric edema, fat stranding severity, vascular congestion, and bowel wall enhancement patterns—markers of systemic inflammatory response and gut barrier dysfunction. Studies demonstrated a strong association between extensive mesenteric inflammation and delayed return of gut motility, nausea, vomiting, and intolerance to clear liquids. Patients with Grade 3–4 mesenteric involvement had significantly higher rates of persistent organ dysfunction and required prolonged enteral feeding. This classification may help clinicians identify patients needing early enteral tube feeding, proactive anti-inflammatory therapy, and close monitoring for complications. By predicting feeding outcomes early, the model can optimize nutrition plans and improve recovery metrics. The APA emphasized that gut–mesenteric interaction plays a central role in AP pathophysiology, and this imaging tool advances personalized care.

TREM2 (Triggering Receptor Expressed on Myeloid Cells-2) has emerged as a crucial regulator of inflammation and tumor immunity. APA investigators presented compelling data showing that TREM2-positive macrophages accumulate in pancreatitis and PDAC tissues. During acute and chronic pancreatitis, TREM2 modulates macrophage activation, promoting tissue repair but also contributing to fibrotic progression when chronically stimulated. In PDAC, TREM2-expressing tumor-associated macrophages (TAMs) create an immunosuppressive microenvironment by releasing IL-10, TGF-β, and arginase-1, which impair T-cell infiltration and cytotoxicity. High TREM2 expression correlates with worse prognosis and limited response to immunotherapy. Mechanistic studies indicate that TREM2 signaling enhances lipid metabolism in macrophages, enabling them to thrive in PDAC’s dense stromal environment. Blocking TREM2 reprograms TAMs toward a pro-inflammatory, anti-tumor phenotype and improves response to checkpoint inhibitors in preclinical models. Clinically, TREM2 is a promising biomarker for disease severity and a potential therapeutic target to overcome PDAC immune resistance.

This session explored the intersection of metabolic dysfunction and pancreatic cancer biology, focusing on Type 3c diabetes–associated PDAC (T3cD-PDAC). One highlighted pathway was the GPR81–lipogenesis axis. GPR81, a lactate receptor, is overexpressed in PDAC cells and drives tumor survival in nutrient-poor microenvironments. T3c diabetes, characterized by exocrine insufficiency, malnutrition, and altered glucose–lipid metabolism, creates a metabolic milieu that fuels tumor aggressiveness. Lactate from glycolysis binds GPR81, activating downstream lipogenic enzymes (ACC, FASN), supporting membrane synthesis and rapid tumor proliferation. The APA presentation emphasized that this metabolic crosstalk promotes chemoresistance, immune evasion, and metastatic progression. Blocking GPR81, inhibiting fatty-acid synthesis, or targeting lactate transporters (MCT1/4) has shown promising preclinical results. Importantly, T3cD patients demonstrate unique metabolic signatures—reduced incretin response, low insulin, and impaired glucagon counter-regulation—that may accelerate PDAC onset. The session concluded that metabolic pathway inhibitors combined with immunotherapy and chemotherapy could be a future therapeutic avenue for T3cD-associated PDAC.

Fatty pancreas, or pancreatic steatosis, is increasingly recognized as a clinically significant entity with metabolic, inflammatory, and oncologic implications. The APA session highlighted how advanced radiomics—quantitative imaging analytics—has transformed our understanding of this condition. Unlike simple fat accumulation, fatty pancreas involves ectopic lipid deposition within acinar cells and peripancreatic regions, leading to lipotoxicity, oxidative stress, and local inflammation. Radiomics allows extraction of high-dimensional imaging features from CT/MRI, including texture, attenuation, and distribution patterns that correlate with metabolic risk, fibrosis progression, and diabetes development. These features can detect fatty pancreas far earlier than visual assessment, allowing stratification into mild, moderate, and severe categories. Fatty pancreas has been linked to acute pancreatitis severity, worse post-operative outcomes, and increased risk of Type 3c diabetes (T3cDM). Radiomic models can also differentiate fatty pancreas from early chronic pancreatitis and detect subtle inflammatory or fibrotic signatures. Emerging data suggest that fatty infiltration may alter local immune responses and create a pro-tumorigenic environment, contributing to increased PDAC risk. Clinically, radiomics may soon guide screening and prevention strategies in obese, diabetic, or NAFLD patients with subclinical pancreatic involvement. The APA session underscored the need for standardized imaging protocols and integration with metabolic profiling to advance personalized pancreatology.