Predicting immune-related adverse events (irAE) in gastrointestinal malignancies (GI-M): Tumor-specific risk factors (TS-RF), hematological markers (HM), and clinical outcomes.

Abstract

e14652 Background: IrAEs, commonly studied as tumor-agnostic effects of immune checkpoint inhibitors (ICI), require individualization based on baseline (BL) clinical characteristics and tumor type. This study evaluates TS-RF and HM to predict irAE incidence, recovery, and their impact on overall survival (OS) in GI-M. Methods: A single-center retrospective review was conducted on GI-M patients treated with ≥2 ICI doses between 1/2015–7/2021 (upper GI (UGI), lower GI (LGI), and anal (Al)) and 1/2015–6/2023 (hepatocellular (HCC) and biliary tract cancers (BTC)).BL clinical and HM (at irAE onset and recovery) were analyzed using t-tests for continuous factors, chi-square tests for categorical factors, and paired t-tests for changes in HM. Results: Among 218 patients (64 UGI, 46 LGI, 53 HCC, 41 BTC, and 14 Al), 43 (19.7%) experienced irAEs, including 22 (10%) severe (requiring hospitalization). Tumor-specific characteristics are discussed in the table below. Two patients were receiving steroids at the time of data collection. Recovery (Rec-g) and mortality (M-g) rates were 52% and 37%, respectively. A significant (p < 0.05) rise in WBC (by 2.5 K/uL), neutrophil count (NC, by 2.6 K/uL), and neutrophil-lymphocyte ratio (NLR, by 2.9) was noted at irAE onset. NLR rise was significant in Rec-g (3.08, p = 0.003) but not in M-g. NLR drop at recovery (by 2.3, p = 0.06) suggests its utility in predicting irAE outcomes. Distant metastasis at BL reduced recovery likelihood (42% vs. 92%, p = 0.002) of recovery in irAE group. IrAE incidence was associated with a trend toward improved OS in UGI (52 vs. 23 months (m), p = 0.07) and LGI (53 vs. 10m, p = 0.06) but had a detrimental impact in HCC (11 vs. 19m, p = 0.05). Conclusions: This study highlights the need for individualized approaches in managing irAEs across different GI-M, leveraging TS-RF and dynamic changes in HM. Small cohort sizes limit broader generalizability but highlight the need for further tumor-specific research. irAE incidenceall-grade & severe HM BL to irAE incidence, rise in Factor (at BL), irAE vs. no-irAE group,risk-reduction with UGI 9.3% & 6% NC by 5.2 (p=0.07) No significant factors (NSF) LGI 20% & 9% NLR by 1.7 (0.04) High NC (5.7 vs. 3.7, 0.01), WBC (6 vs. 8, 0.04), and alkaline phosphatase (ALP, 10 vs. 90 U/L, 0.07) Al 46% & 30% WBC by 2.8 (0.06) NSF BTC 32% & 12% NC by 4.2 (0.04)& WBC by 4.1 (0.07) Liver mets (18% vs. 53%, p=0.06), low ALP (237 vs. 471 U/L, 0.02), albumin (3.4 vs. 3.7 g/dL, 0.03) & alanine aminotransferase (24 vs. 40 U/L, 0.07), and high thyroid-stimulating hormone (3.4 vs. 2.2 ng/dL, 0.07) HCC 19% & 11% NSF Bevacizumab use (0% vs. 25%, 0.01), ascites (11% vs. 33%, 0.05), and poor liver function (Child-Pugh B& C vs. A, 29% vs. 8%, 0.04) and high total bilirubin (1.5 vs. 0.9 mg/dL, 0.07) WBC and NC units - K/uL.