TPS4271 Background: Gastrointestinal (GI) malignancies are associated with poor survival and novel therapies are urgently needed. Sacituzumab govitecan (SG) is an antibody-drug conjugate (ADC) that targets the tumor-associated antigen Trop-2 (trophoblast cell-surface antigen 2). SG is covalently bound to the topoisomerase I inhibitor SN-38, which is the active metabolite of irinotecan. Trop-2 is overexpressed in most GI malignancies, including colorectal, gastric, and pancreatic cancers, and is associated with poor prognosis. SG has been approved for the treatment of second-line triple-negative breast cancer and pre-treated hormone-receptor positive metastatic breast cancer but its role in GI cancers is not determined. Furthermore, its combination with capecitabine has never been studied. Methods: We are conducting a single arm, single institution, dose escalation phase 1 trial with a 3+3 design of combination SG plus capecitabine. SG dosing will be at three levels: Level -1 at 5mg/kg; Level 0 at 7.5mg/kg; and Level +1 at 10mg/kg. Dosing will start at dose level 0. SG will be administered as an intravenous infusion on Days 1 and 8 of a 21-day cycle. Capecitabine dose is fixed at 825mg/m 2 and will be delivered orally twice a day for 14 days on and 7 days off, starting on Day 1, of the 21-day cycle. Key eligibility criteria include histologically documented metastatic adenocarcinoma of GI origin, including gastroesophageal, colorectal, and pancreaticobiliary, that has failed standard therapy; age ≥18years; ECOG performance status 0-1; and adequate end organ function. Key exclusion criteria include previous receipt of topoisomerase 1 inhibitors. The primary endpoint is the Recommended Phase 2 Dose (RP2D). Secondary endpoints include objective response rates, duration of response, progression-free and overall survival. An exploratory endpoint is the correlation between Trop-2 expression in collected archival tissue and clinical outcomes. Optional biopsies will be offered for patients with missing or insufficient archival tissue. This study has been registered under NCT06065371 and is actively enrolling. The trial will enroll up to 20 patients. The trial is funded by Gilead Sciences, Inc. Clinical trial information: NCT06065371 .

787 Background: NRG GI 011, a study randomizing patients with locally advanced pancreatic cancer (LAPC) with and without ablative radiation therapy, has strict entry criteria regarding performance status, induction chemotherapy, and CA 19-9 to best select patients who may benefit from ablative radiation. This study evaluates the effect of those patient characteristics on survival outcomes of patients with LAPC treated with stereotactic MR guided adaptive radiation therapy (SMART) at a single urban medical center. Methods: A retrospective review was conducted of 106 LAPC patients treated with SMART between 2018 and 2024. All patients received SMART with 50Gy in 5 fractions, with or without induction chemotherapy. Overall survival (OS) was measured from diagnosis and from the end of RT and estimated using Kaplan-Meier (KM) analysis, and KM curves were compared using log-rank tests. Multivariate Cox proportional hazards models tested the impact of performance status (PS), largest tumor size at diagnosis (TS), induction chemotherapy (yes vs no), carbohydrate 19-9 (CA19-9), and age on survival outcomes. Acute and late grade ≥3 toxicities were graded per CTCAE v5.0. Results: The median OS from diagnosis was 16.5 months (95% CI: 15.1-18.0), and 9.1 months from the end of RT (95% CI: 7.7-10.4). In univariate analysis, lower CA19-9, younger age, and a PS of 0-1 with induction chemotherapy was significantly associated with improved OS from diagnosis (p < 0.05) while tumor size was not. Multivariate analysis confirmed lower CA19-9 and PS 0-1 with induction chemotherapy were independent predictors for improved OS from diagnosis. The 38 patients who meet the NRG GI 011 enrollment criteria had a median OS of 25.1 months and a 3-year survival of 35.2%. For OS from the end of RT, only PS 0-1 with induction chemotherapy remained significant for improved OS (p = 0.058) in the multivariate model. Grade ≥3 acute toxicities occurred in 2.8% of patients and grade ≥3 late toxicities occurred in 13.2% of patients. Conclusions: SMART for LAPC demonstrates excellent OS from both diagnosis and from the end of RT with low rates of grade ≥3 acute and late toxicities. Elevated CA19-9, worse PS, and lack of induction chemotherapy emerged as key negative prognostic factors in treating LAPC with SMART, confirming the enrollment criteria for NRG GI 011.

e16302 Background: Biliary tract cancers (BTCs) are a heterogeneous group of malignancies, collectively associated with poor prognosis. Treatment typically consists of curative intent surgery for eligible candidates. Based on the landmark BILCAP trial, capecitabine is currently used as an adjuvant therapy for pts with BTCs. However, the trial included a small number of pts with early-stage disease. Additionally, the staging system used for the trial was updated since. Methods: This retrospective review was based on the National Cancer Data Base (NCDB) participant user file (PUF) from 2014 to 2023. Patients (pts) were identified using codes for intrahepatic biliary, extrahepatic biliary, or gallbladder adenocarcinoma who underwent curative intent surgical resection and had documented chemotherapy status. Only node-negative (N0) disease, pathological stage T1 or 2 were included. Pts were categorized by receipt of adjuvant chemotherapy versus no chemotherapy. Overall survival was evaluated using Kaplan–Meier survival analysis. Analysis of the NCDB PUF is exempt from IRB review. Results: A total of 11,111 pts were identified. 79.3% were White; 44.3% were male; mean age was 67.9years. 2,975 (26.8%) had intrahepatic cholangiocarcinoma (CCA), 2,687 (24.5%) extrahepatic CCA, and 5,449 (49.0%) had gallbladder adenocarcinoma. 9,237 (85.5%) had negative margins; 6,093 (75.7%) had negative LVI. As for size, 3,525 (38.2%) had tumors 2cm or less, 3,953 (42.8%) had tumors between 2 and 5cm, and 1,760 (19.0%) had tumors larger than 5cm. 6,932 pts (62.4%) received adjuvant chemotherapy. A trend toward improved survival was observed across most groups. Statistical significance was associated for T1a and T2a with negative LVI; T1a, T2a, and T2b with negative margins; and T1a and T1b intrahepatic CCA, T2a and b extrahepatic CCA, T2a and b gallbladder adenocarcinoma. The remaining groups had low numbers of pts, and the survival analyses were either not performed due to low numbers or were not statistically significant. Conclusions: Despite the limited retrospective nature of this analysis, adjuvant chemotherapy appears to be associated with improved survival in pts with early-stage CCAs, including those with favorable tumor features. Pt and tumor specific features should be considered individually.

759 Background: Pancreatic adenocarcinomas (PDAC) harboring deficiencies in BRCA1/2 or PALB2 are more susceptible to platinum (Pl) chemotherapy regimens as well as PARP inhibitors. The same is not fully elucidated for PDAC harboring alterations in other genes within the DNA damage repair (DDR) pathway. In this study, we aim to compare outcomes of patients (pts) with advanced PDAC harboring mutations in DDR pathway in genes other than BRCA 1/2 and PALB2 with Pl- versus non-Pl -including chemotherapy regimens in first-line (1L). Methods: We used Tempus Lens, a platform used to query multimodal data from millions of de-identified patient records in the Tempus Database, to identify pts with advanced PDAC without BRCA1/2 and PALB2 mutations who had Tempus xT (solid tumor) or xF (liquid biopsy) testing (N = 3,175). Pts with clinically reportable copy number losses (0 copies) or gains (8 or more copies), or pathogenic/likely pathogenic single nucleotide variations or indels in one of 67 other DDR genes were classified as DDR-mutated (DDR-mut, N = 783). 1L treatment regimens were categorized as Pl if any agent in the regiment was categorized as a Pl compound, or Non-Pl otherwise. Real-world (rw) objective response rate (rwORR) was defined as the proportion of pts with a documented complete or partial response after 1L start. Rw time to next treatment (rwTTNT) was defined as the time from 1L start to start of next treatment or death and rw overall survival (rwOS) as the time from 1L start to death or loss to follow up. Median rwOS and rwTTNT were estimated using Kaplan-Meier curves and compared with Cox proportional hazards likelihood ratio tests. Results: Among the DDR-mut pts, median age at diagnosis was 66 years (range: 29-86), 55% were White, 7% were Black, 2% were Asian and 36% unknown/other. Mutations in ATM were most prevalent (25%), followed by CHEK2 (8.7%). Of DDR-mut pts included in rwOS, 55% (N = 256) received Pl regimens compared to 45% (N = 212) who received non-Pl regimens. Of those receiving Pl regimens, the majority received a triplet regimen rather than a doublet (84% vs 7%). Although no statistically significant difference was observed in rwOS between Pl and non-Pl regimens in the DDR-mut group, there was a trend toward improved survival starting around 5 months of treatment (median rwOS 11.7 vs 9.8 months, p=0.471). Interestingly, rwTTNT was longer in pts receiving non-Pl (8.3 months vs 6.4, p=0.115). rwORR was not different between the two groups (40% vs 37%, p=0.6). Conclusions: This is one of the largest cohorts comparing outcomes in DDR-mut PDAC with Pl vs non-Pl regimens. Despite a trend toward improved survival with Pl regimens, this trend was not statistically significant. Furthermore, rwTTNT in the non-Pl regimens was longer. The results of this study should be regarded as investigational and do not include length of treatment, performance status, but suggest that future studies capture this information.

742 Background: KRAS, TP53, CDKN2, and SMAD4 mutations are implicated in PDAC development and progression. Methylthioadenosine phosphorylase (MTAP) is a key enzyme in the methionine salvage pathway, and its loss may be associated with a shorter overall survival (OS). Early clinical trials in patients with MTAP-deficient malignancies are currently underway with encouraging results. However, there are only few studies characterizing patients with MTAP-deficient PDAC. Our study is a retrospective case series interrogating the molecular profiles, demographics, and clinical courses of patients diagnosed with MTAP-deficient PDAC in 3 institutions. Methods: This is a multi-center retrospective cohort study of patients with metastatic PDAC from January 2019 to August 2025. MTAP gene loss and associated genomic changes were determined by next generation sequencing (NGS) of tumor tissue. Patient demographics, tumor characteristics, treatment history, and survival data were collected. Plasma ctDNA based NGS was done for select patients. Results: 21 patients were identified. 85% of patients were white, 10% African American, 5% Asian. 57% of patients were female. The median age at diagnosis was 63 years, with a range of 49 to 81 years of age. Liver metastasis was present in 80% of patients. TP53 and SMAD4 mutations were present in 95%, and 25% of patients, respectively. CDKN2A and B mutations were present in 80% and 85% of patients, respectively. 10% of patients had neither CDKN2A nor B mutation. All patients had KRAS mutations; G12D and Q61H/R mutations were each present in 29% of patients, while G12V and G12R mutations were seen in 20% and 15% of patients, respectively. MTAP deletion was not detected by plasma ctDNA analysis undergone by 9 patients. Five patients (25%) were initially diagnosed with localized disease and underwent surgical resection prior to developing metastatic disease. Three patients (15%) had rapidly progressive disease and died before starting treatment. Two patients recently began treatment and are yet to be assessed for response. Of the 16 patients with an evaluable treatment response, 25% experienced disease progression on first-line therapy and 75% experienced disease stability or partial response by RECIST criteria. Mean response duration was 8.5 months. Median survival of the entire cohort was 15.5 months. Conclusions: Patient demographics and clinical courses did not vary significantly from the general metastatic PDAC population. KRAS mutation subtype frequency is different than expected, with the higher frequency of Q61H/R mutations. Further study of the relationship between MTAP deletion and specific KRAS mutation subtypes may influence treatment decision-making and future clinical trial designs. Lack of detection in liquid biopsy may relate to type of assay used.

AIMS:To evaluate the effect of UGT1A1*28 homozygosity on the safety profile of NALIRIFOX (liposomal irinotecan + 5-fluorouracil/leucovorin + oxaliplatin) in patients with metastatic pancreatic ductal adenocarcinoma (mPDAC) in NAPOLI 3 (NCT04083235). METHODS:This pre-specified exploratory analysis evaluated safety outcomes by UGT1A1*28 status in patients with mPDAC receiving NALIRIFOX or gemcitabine plus nab-paclitaxel in NAPOLI 3. All patients receiving NALIRIFOX initiated liposomal irinotecan at the full starting dose. RESULTS:Among 749 treated patients, 83 were homozygous for UGT1A1*28. In the NALIRIFOX arm, grade ≥3 and serious treatment-emergent adverse events (TEAEs) related to liposomal irinotecan occurred in 61.5% and 33.3% of patients with homozygous UGT1A1*28 (n = 39), respectively, compared with 63.7% and 22.9% of patients with other genotypes (n = 328). The most common any-grade TEAEs for NALIRIFOX were diarrhea (59.0%), nausea (56.4%), vomiting (46.2%), and anemia (41.0%) in the homozygous group and diarrhea (71.6%), nausea (59.8%), vomiting (38.4%), and decreased appetite (38.1%) in the other genotypes group. Rates of TEAE-related treatment discontinuation and dose reduction/interruption in the NALIRIFOX arm were similar across genotypes. CONCLUSIONS:UGT1A1*28 homozygosity was not associated with increased toxicity in patients receiving NALIRIFOX in NAPOLI 3. These findings support the use of full NALIRIFOX dosing irrespective of UGT1A1*28 status. CLINICAL TRIAL REGISTRATION:www.clinicaltrials.gov identifier is NCT04083235; EudraCT 2018-003585-14.

4012 Background: Preclinical studies have demonstrated that PARP inhibitors modulate the immune microenvironment by increasing genomic instability, PD-L1 expression and activating the immune inflammatory stimulator of interferon genes (STING) pathway. The POLAR trial demonstrated promising results combining pembrolizumab with olaparib as maintenance therapy in patients (pts) with homologous recombination deficiency (HRD) mutations. S2001 is a randomized phase II trial evaluating the addition of pembrolizumab to olaparib as maintenance therapy in patients with gBRCA1/2. Methods: Eligible pts had mPDA with gBRCA1/2 mutations and had received a minimum of 4 months of platinum-based chemotherapy without progression. Pts were randomized 1:1 to olaparib 300 mg twice per day plus pembrolizumab 200 mg every 3 weeks or olaparib alone. The primary endpoint was progression-free survival (PFS) with a target hazard ratio (HR) of 0.6 (median PFS 7.0 vs 11.7 months) with 1-sided alpha=0.1 and 80% power. Secondary outcomes were safety and tolerability, overall survival (OS), overall response rate (ORR), and disease control rate (DCR). The accrual goal was 88 pts to have 78 eligible for analysis. Differences in PFS by treatment arm were assessed via stratified log-rank test, with first line platinum-based chemotherapy, Zubrod PS, and disease status after first line treatment as stratification factors. Tissue and blood samples for correlative studies were banked. Results: The study enrolled 73 patients, with 68 eligible, and was closed for futility after a planned interim analysis showed that PFS was not improved in the pembrolizumab + olaparib arm (p=0.82), despite higher ORR (p=0.20) (Table). Treatment was well tolerated, with no grade 4-5 treatment-related adverse events and similar toxicity profiles across arms. Autoimmune toxicities on the experimental arm included adrenal insufficiency, hyperthyroidism, and hypothyroidism. Grade 3 adverse events in both arms included neutropenia, anemia, thrombocytopenia, gastrointestinal toxicity and fatigue. Conclusions: S2001 is the first randomized controlled trial evaluating the addition of pembrolizumab to olaparib in pancreas cancer. Despite doubling the ORR and improvement in DCR and PFS in the experimental arm, the high bar of improving PFS to 11.7 months was not met. Activity seen in the experimental arm of S2001 is comparable to the HRD cohort of the POLAR trial which demonstrated ORR 35%, DCR 90% and PFS of 8.2 months. Further follow-up for mature survival outcomes and translational analysis of DNA and RNA are planned. Funding: NIH/NCI/NCTN grants U10CA180888, U10CA180819 with additional support from by Merck & Co, Inc. Clinical trial information: NCT04548752 . Pembrolizumab + Olaparib (n=34) Olaparib (n=34) Median PFS 8.2 months 6.4 months Median OS 20.0 months 22.1 months ORR 28.1% 14.3% DCR 84.4% 64.3%

786 Background: There is ongoing interest in randomized studies utilizing stereotactic MR guided adaptive radiation therapy (SMART) in patients with medically inoperable pancreatic cancer. This study evaluates overall survival (OS) from diagnosis and from the end of radiation treatment (RT) of patients with medically inoperable pancreatic cancer. Methods: A retrospective review analyzed medically inoperable pancreatic cancer patients treated with 50 Gy in 5 fractions of SBRT at a single urban medical center. OS was calculated from diagnosis and end of RT to the date of death or last follow-up. Kaplan-Meier estimates and 95% confidence intervals (CI) were reported, and the Kaplan-Meier curves were compared using log rank tests. Cox’s proportional hazard model was used to determine the combined impact of Zubrod performance status (PS) (0-1 vs. 2-3) and tumor size (TS) at diagnosis (<2.35 cm or ≥2.35 cm) with OS (diagnosis) and OS (RT). Multivariate models were adjusted for age at diagnosis and induction chemotherapy (yes vs no). Hazard ratio (HR) and 95% CI were reported. Acute (<=90 days) and late grade ≥3 toxicities were graded per CTCAE v5.0. Results: 62 patients were included; the median OS was 10.4 (95% CI: 7.3-13.5) months from diagnosis. The receipt of induction chemotherapy was the only independent predictor of OS from diagnosis; 16 months from diagnosis compared to 8.8 months for no induction group (p = 0.021). TS was the only independent significant predictor for OS from RT. Patients with PS 2-3 and TS ≥2.35 cm showed the worst OS from diagnosis and end of RT compared to patients with PS of 0-1 and TS <2.35 cm (Table 1). Late Grade 3 upper GI bleeding occurred in one patient (1.85%) and acute Grade 3 abdominal pain or diarrhea occurred in two other patients (3.7%). Conclusions: Medically inoperable pancreatic cancer patients treated with induction chemotherapy and SMART have favorable survival outcomes. Delivery of SMART to patients with poor performance status and large tumor size may not be impactful. These findings support the importance of patient selection in medically inoperable pancreatic cancer patients receiving SMART. OS from diagnosis and RT versus PS+TS. Variable OS from diagnosis (months) OS from end of RT (months) PS of 0-1 and TS <2.35 cm Median = 11.9 Median = 10.4 PS 2-3 and TS ≥2.35 cm Median = 8.7 HR = 2.1 Median = 5.4 HR = 2.4 * *p<0.05.

717 Background: SN-38, the active metabolite of irinotecan, is cleared by the liver enzyme uridine diphosphate glucuronosyltransferase 1A1 (UGT1A1) and biliary excretion. UGT1A1*28 (genotype 7/7) homozygosity is associated with reduced activity of UGT1A1 and a recommendation for dose adjustment of non-liposomal irinotecan, a component of the FOLFIRINOX regimen. In this post-hoc analysis, we analyzed the impact of UGT1A1*28 homozygosity on the incidence and profile of treatment-emergent adverse events (TEAEs) among patients enrolled in NAPOLI 3 (NCT04083235). Methods: Patients (N = 770) with confirmed untreated mPDAC were randomized (1:1) to receive liposomal irinotecan 50 mg/m 2 + oxaliplatin 60 mg/m 2 + leucovorin 400 mg/m 2 + 5-fluorouracil 2400 mg/m 2 (NALIRIFOX) on days 1 and 15 of a 28-day cycle or nab-paclitaxel 125 mg/m 2 and gemcitabine 1000 mg/m 2 (Gem+NabP) on days 1, 8 and 15 of a 28-day cycle. UGT1A1*28 status was evaluated in all patients before enrollment. All patients, regardless of UGT1A1*28 status, received the full starting dose of liposomal irinotecan and followed the same dose reduction rules. This exploratory analysis of TEAEs by UGT1A1*28 status was descriptive; statistical analyses were not performed. Results: Overall, 83 patients (11.0%) were homozygous for UGT1A1*28 , among whom the incidence of TEAEs was similar to patients with non-homozygous status (Table). The most frequently reported TEAEs (≥ 40%) among the homozygous group (vs the overall NAPOLI 3 population) were diarrhea (59.0% vs 70.5%), nausea (56.4% vs 59.5%), vomiting (46.2% vs 39.7%) and anemia (41.0% vs 26.2%) in the NALIRIFOX arm and nausea (45.5% vs 42.7%), fatigue (45.5% vs 37.7%), diarrhea (45.5% vs 36.7%) and anemia (40.9% vs 40.4%) in the Gem+NabP arm. Conclusions: The incidence of TEAEs, including TEAEs leading to death, was similar between patients with homozygous and non-homozygous UGT1A1*28 status, in both treatment arms. The profile of TEAEs was consistent with that of the overall NAPOLI 3 population. UGT1A1*28 status did not substantially influence the safety profile or subsequent need for dose reductions of liposomal irinotecan in NAPOLI 3. Clinical trial information: NCT04083235 . UGT1A1*28 homozygous UGT1A1*28 non-homozygous TEAE, n (%) NALIRIFOX (n = 39) Gem+NabP (n = 44) NALIRIFOX (n = 328) Gem+NabP (n = 331) Any 39 (100.0) 44 (100.0) 327 (99.7) 328 (99.1) Related to any drug 39 (100.0) 41 (93.2) 310 (94.5) 308 (93.1) Grade ≥ 3 31 (79.5) 34 (77.3) 288 (87.8) 288 (87.0) Serious 24 (61.5) 24 (54.5) 176 (53.7) 168 (50.8) Leading to treatment discontinuation 15 (38.5) 10 (22.7) 103 (31.4) 100 (30.2) Leading to dose reduction of any drug 23 (59.0) 14 (31.8) 182 (55.5) 174 (52.6) Leading to interruption of any drug 1 (2.6) 1 (2.3) 15 (4.6) 3 (0.9) Leading to death 2 (5.1) 4 (9.1) 20 (6.1) 19 (5.7)

IntroductionPancreatic tumors and cell lines derived from them exhibit elevated expression of 5-lipoxygenase (5-Lox), whereas non-tumor glands or normal cells do not exhibit this overexpression. Arachidonic acid stimulates pancreatic cancer cell growth via metabolic conversion through the 5-Lox pathway, and inhibition of 5-Lox activity decreases the viability of pancreatic cancer cells. However, the downstream signaling mechanisms through which 5-Lox exerts its effects on the survival of pancreatic cancer cells remain to be elucidated.MethodsThe effects of 5-Lox inhibition on cell proliferation, apoptosis, and invasive potential were investigated in pancreatic cancer cells. The protein expression was analyzed by Western blot. Apoptosis was analyzed by Annexin-V binding assay and by detecting the degradation of chromatin-DNA to nucleosomal fragments. The protein kinase C-epsilon (PKCε) activity was measured by an immunoprecipitation-kinase assay. The in vivo effects of MK591 were evaluated in pancreatic tumor xenograft model.ResultsMK591, a specific inhibitor of 5-Lox activity, killed pancreatic cancer cells via induction of apoptosis, involving externalization of phosphatidylserine, cleavage of PARP (poly-ADP ribose polymerase) and degradation of chromatin DNA to nucleosomes. MK591 effectively blocked in vitro invasion and soft-agar colony formation by pancreatic cancer cells and decreased pancreatic tumor growth in nude mice xenografts. Furthermore, inhibition of 5-Lox downregulated K-Ras and inhibited phosphorylation of c-Raf and ERKs. Interestingly, 5-Lox inhibition induced apoptosis in pancreatic cancer cells without the inhibition of Akt but the protein level of PKCε was dramatically downregulated. Furthermore, inhibition of 5-Lox decreased the phosphorylation of Stat3 at Serine-727. Pre-treatment of pancreatic cancer cells with peptide activators of PKCε prevented apoptosis induced by 5-Lox inhibition, suggesting that the mechanism by which 5-Lox inhibition causes cell death in pancreatic cancer involves downregulation of PKCε. The combination of low doses of MK591 and gemcitabine synergistically reduced the oncogenic phenotype and killed pancreatic cancer cells by inducing apoptosis.DiscussionThese findings indicate that inhibition of 5-Lox interrupts an Akt-independent, PKCε-dependent survival mechanism in pancreatic cancer cells and suggest that metabolism of arachidonic acid through the 5-Lox pathway plays an integral part in the survival of pancreatic cancer cells via signaling through PKCε, an oncogenic, pro-survival serine/threonine kinase.

4136 Background: NAPOLI 3 (NCT04083235, N = 770), a global, randomized, open-label phase 3 study, demonstrated that NALIRIFOX significantly improved overall survival (OS, primary endpoint) and progression-free survival compared with nab-paclitaxel and gemcitabine in patients with untreated mPDAC. Here, we report an updated analysis of OS. Methods: Eligible patients with histopathologically/cytologically confirmed untreated mPDAC were randomized (1:1) to receive liposomal irinotecan 50 mg/m 2 + oxaliplatin 60 mg/m 2 + leucovorin 400 mg/m 2 + 5-fluorouracil 2400 mg/m 2 (NALIRIFOX, n = 383) on days 1 and 15 of a 28-day cycle or nab-paclitaxel 125 mg/m 2 and gemcitabine 1000 mg/m 2 (Gem+NabP, n = 387) on days 1, 8 and 15 of a 28-day cycle. Randomization was stratified by ECOG performance status, geographic region and presence/absence of liver metastases. An updated OS analysis (data cutoff: 3 October 2023) was conducted. Kaplan–Meier methods were used to estimate median (95% confidence interval [CI]) OS and hazard ratios (HRs [95% CI]) were estimated using stratified Cox proportional hazard models. The difference between groups was tested using a stratified log-rank test. Results: The median follow-up for OS was 28.7 months in the NALIRIFOX arm and 29.7 months in the Gem+NabP arm. At data cutoff, 11 patients were still receiving study treatment, all of whom were in the NALIRIFOX arm. With a total of 673 OS events (NALIRIFOX, n = 328; Gem+NabP, n = 345), median OS was 11.1 months (95% CI 10.0–12.1) in the NALIRIFOX arm compared with 9.2 months (8.3–10.6) in the Gem+NabP arm (HR 0.84 [95% CI 0.72–0.98]; nominal p = 0.026). At 12 months and 18 months, survival rates were 45.6% (95% CI 40.5–50.5) and 26.6% (22.2–31.1), respectively, in the NALIRIFOX arm and 39.6% (34.7–44.5) and 20.0% (16.1–24.1), respectively, in the Gem+NabP arm. Conclusions: In this 29-month follow-up of NAPOLI 3, NALIRIFOX continued to demonstrate improved OS compared with Gem+NabP, with 11 patients still receiving the NALIRIFOX regimen. These data confirm NALIRIFOX as a new possible standard of care and reference regimen for the first-line treatment of patients with mPDAC. Clinical trial information: NCT04083235 .

Abstract Pancreatic Cancer has a 12% 5-year survival rate due to late-stage detection and lack of many chemotherapy or targeted therapy options. TIGIT is an immune checkpoint inhibitor being explored in clinical trials in pancreatic cancer due to the implications of its ligand (CD155) promoting immune evasion. TIGIT is a marker of T cell exhaustion and plays a key role in the inhibition of anti-tumor immune responses. We hypothesize that targeted anti-TIGIT therapy, in conjunction with other therapies targeting the tumor microenvironment, could reverse immune suppression that is characteristic of pancreatic ductal adenocarcinoma (PDAC). As such, here we aim to quantify TIGIT expression using automated RNAscope technology and to correlate it with overall patient survival. We performed RNAscope in situ hybridization and immunohistochemistry on 25 primary and 4 metastatic (liver) formalin-fixed paraffin-embedded (FFPE) tissue sections from patients with histologically confirmed PDAC. A nuclear counterstain combined with an RNAscope probe specific for the human TIGIT mRNA was utilized. Slides were scanned at 40X magnification with an automated slide scanner and quantified using the ISH-IHC module of the HALO-v3.5 software. After cells were segmented based on nuclear recognition, the presence of TIGIT probe within the cytoplasm of each cell was determined and quantified. Percent positive cell values were then exported for statistical analysis in SPSS. De-identified clinical metadata was obtained from REDCap, a cloud-based HIPAA-compliant database used to compile patient data for research purposes. Gender, survival months, ethnicity, race, tumor histology, stage, cancer history, treatment status, and comorbidities were evaluated. We analyzed a cohort of 29 histologically-confirmed surgically resected PDACs, comprised of 25 primary pancreatic tumor samples and 4 metastatic liver core biopsies. The mean percentage of TIGIT positive cells was 63%. From this, the TIGIT high versus low threshold was determined to be 70%. High TIGIT was associated with significantly lower overall survival (p=.013), suggesting that expression of TIGIT and T cell exhaustion may predict overall worse survival. The median overall survival months for the TIGIT low group was 321 months versus 106 months for the TIGIT high group. This data leads us to hypothesize that expression of TIGIT and T cell exhaustion may predict worse overall survival. Anti-TIGIT medications are currently in clinical trials for metastatic PDAC. Here we demonstrate that high TIGIT expression is associated with an overall worse prognosis and present a novel, automated TIGIT detection protocol that may be used to determine if patients are candidates for anti-TIGIT therapy. In a clinical setting, this novel, automated biomarker TIGIT detection protocol could be used to select candidates for anti-TIGIT therapy. This may lead to improved overall patient survival. Citation Format: Madison George, Julie Clark, Kendyll Gartrelle, Georges Nassif, Donald Rempinski, Daniel Long, Hui-Ju Wen, Simone Benitz, Samuel Zwernik, Rupen Shah, Hakmin Park, David Kwon, Philip Philip, Gazala Khan, Howard Crawford, Brian Theisen, Nina Steele. TIGIT expression correlates to worse overall survival in primary and metastatic pancreatic ductal adenocarcinoma [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Pancreatic Cancer; 2023 Sep 27-30; Boston, Massachusetts. Philadelphia (PA): AACR; Cancer Res 2024;84(2 Suppl):Abstract nr C012.

Abstract Neuroendocrine tumors (NETs) are rare and diverse, with increasing incidence but poorly understood prognostic variables. While African American or Black patients typically have poorer outcomes across various tumor types, this is not well documented for NETs. Our study aimed to investigate racial disparities in NET outcomes. Patients who had a diagnosis of neuroendocrine tumor between 2014 and 2022 were identified and analyzed in this retrospective chart analysis, with race determined as Caucasian (CA) or Black African American (BAA). Histology was stratified as high grade or those without high grade histology. Grade of tumor was identified as well differentiated, moderately or poorly differentiated, and other/unknown. Log-rank test was conducted to compare the survival curves between CA and AA groups. Cox proportional hazards model was built to adjust for potential confounders, including age at diagnosis and gender (if applicable) while comparing racial disparities. The CA vs. AA hazard ratio (HR) and p-values were calculated. Among the 655 patients analyzed, 63.4% were Caucasian (CA) and 36.6% were Black African American (BAA). The average age at diagnosis across the cohort was 61.5 years. The gender distribution was approximately equal, with 50.2% females and 49.8 % males. High-grade histology was identified in 114 patients, of whom 28.1% were BAA and 71.9% were CA. Within this subgroup, males were more prevalent than females (61.4% vs. 38.6%). Overall survival (OS) for CA patients with high-grade histology was 57.4 months, compared to 67.5 months for BAA patients (p=0.7). Additionally, no statistically significant difference in OS was observed between males and females with high-grade histology (60.8 months vs. 68.1 months, respectively; p=0.3). Analysis of tumor grade revealed that well- differentiated tumors were present in 122 patients (18.63%). Among these patients, there was no statistically significant difference in survival between BAA and CA. The primary tumor site was the colon in 89 patients (13.6%), with 77.5% being CA and 22.5% BAA. Pancreatic tumors were present in 119 patients (18.2%), with 72% being CA and 28% BAA. The next most common primary sites were the small intestine (24.9%), rectum (17.6%), and stomach (9.9%). In comparing survival outcomes between racial groups, CA patients had a significantly worse OS (76.9 months) compared to BAA patients (84.6 months) (p=0.05). Regardless of race, females demonstrated better survival (84.2 months) compared to males (75.7 months) (p=0.04). In our retrospective analysis, Caucasian (CA) patients had worse overall survival (OS) than Black African American (BAA) patients, likely due to a higher proportion of CA patients with high- grade histology and primary tumors in the colon. Females showed better OS than males across all subtypes. The reasons for the higher incidence of worse histology in CA patients are unclear. Larger studies are needed to confirm these findings, understand racial disparities, and develop strategies to mitigate them. Citation Format: Radhika Gutta, Wan-Ting su, Ruicond She, Muhammad Shahid, Gazala Khan. Real world treatment patterns and patient outcomes of neuroendocrine tumors: A single institution study [abstract]. In: Proceedings of the 17th AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2024 Sep 21-24; Los Angeles, CA. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2024;33(9 Suppl):Abstract nr A070.

Abstract Pancreatic ductal adenocarcinoma (PDAC) has a dismal 12% 5-year survival rate (SEER) due to a lack of early detection biomarkers and resistance to standard therapeutic options (surgery, chemotherapy, radiation). Black African Americans (BAA) have 20% increased incidence of PDAC compared to those with European Ancestry (EA). TIGIT, an immune checkpoint receptor, is a marker of T cell exhaustion and plays a key role in the inhibition of anti-tumor immune responses. Recent studies have demonstrated that immune checkpoint receptor expression (PD-1 and TIGIT) on specific T cell populations correlates to worse overall survival (OS). TIGIT inhibitors are being explored in clinical trials in pancreatic cancer due to implications of its ligand (CD155) promoting immune evasion. We hypothesize that targeted anti-TIGIT therapy, in conjunction with other therapies targeting the tumor microenvironment, could reverse the immune suppression that is characteristic of PDAC. We performed RNAscope in situ hybridization (ISH) with a probe specific for human TIGIT mRNA (combined with a nuclear counterstain) on 79 tissue samples. The cohort of tissue samples included 8 biopsies (endoscopic-guided fine needle biopsies at time of diagnosis), 66 primary (from surgical resection), and 5 metastatic (liver core biopsies from patients with a primary PDAC diagnosis) formalin-fixed paraffin-embedded (FFPE) tissue sections from patients with histologically confirmed PDAC. After cells were segmented based on nuclear recognition, the presence of TIGIT probe within the cytoplasm of each cell was determined and quantified. Percent positive cell values were then exported for statistical analysis in R. De-identified clinical metadata was obtained from REDCap, a cloud-based HIPAA-compliant database used to compile patient data for research purposes. We tested for associations between %TIGIT present and clinical covariates, using linear regression for continuous outcomes, and logistic regression for binary outcomes. ScRNAseq revealed that TIGIT mRNA is enriched in, but not exclusive to the T/NK cellular compartments in PDAC. Staining analysis showed that TIGIT expression did not differ significantly between racial groups (comparing BAA to non-BAA). The mean percentage of TIGIT positive cells was 64.0%. High expression of TIGIT was associated with clinical stage, where an increase in stage was associated with increasing %TIGIT (p < 0.05).The TIGIT biomarker assay can be conducted at time of diagnosis, time of surgical resection, and time of metastatic biopsy. If patients’ samples contain high levels of TIGIT expression, these patients may be candidates for anti-TIGIT drug therapy. Considering that TIGIT expression correlates with advancing clinical stage, patients with more advanced staging at diagnosis (stage IIB, III, IV) especially may benefit from anti-TIGIT therapy. Citation Format: Madison George, Julie Clark, Kendyll Gartrelle, Georges Nassif, Kailee Hartway, Daniel Long, Daniel Salas-Escabillas, Allison Wombwell, Thais Pichardo, Hui-Ju Wen, Simone Benitz, Samuel Zwernik, Rupen Shah, Hakmin Park, Philip, Gazala Khan, Howard Crawford, David Kwon, Brian Theisen, Nina Steele. TIGIT expression increases with advancing clinical stages and does not differ across racial groups in resected pancreatic cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 5176.

e16314 Background: While ground has been gained, pancreatic ductal adenocarcinoma (PDAC) continues to have a low 5-year survival of 13%. This is owed partially to a lack of early detection biomarkers and resistance to standard therapeutic options. TIGIT, an immune checkpoint receptor, is a marker of T-cell exhaustion and plays a key role in the inhibition of anti-tumor immune responses. TIGIT inhibitors are being explored in clinical trials in PDAC. Here we evaluate TIGIT expression in a cohort of PDAC patients and correlate level and intensity of expression with clinical parameters. We also examine changes in expression as the disease progresses from primary to metastatic disease. Methods: We performed RNAscope in situ hybridization (ISH) with a probe specific for human TIGIT mRNA on 82 formalin-fixed paraffin-embedded (FFPE) tissue samples. The cohort of tissue samples included 9 biopsies, 67 primary resections and 6 metastatic lesions. We evaluated the total TIGIT expression (%) as well as the intensity of TIGIT expression (% of cells with 3+ punctae, signifying putative immune cells), and compared these values between samples. Utilizing linear regression for continuous outcomes and logistic regression for binary outcomes, we tested for associations between TIGIT expression and clinical covariates. Results: Staining analysis showed that TIGIT expression did not differ significantly between racial groups. The mean percentage of TIGIT positive cells was 64.0%. High expression of TIGIT was associated with more advanced clinical stage (p < 0.05). Evaluation of three longitudinal samples from the same patient revealed decreased TIGIT expression from the initial biopsy (41.6%) to resection (33.4%) and metastasis (2.8%). In these specimens, 3+ TIGIT expression also declined (2.1%, 1.2% and 0.02%, respectively). Conclusions: Anti-TIGIT therapy has potential to reverse immune suppression and, with other therapeutic modalities, may provide survival benefit. Here we demonstrate that increased TIGIT expression correlates with more advanced stage at diagnosis and also present data demonstrating that overall TIGIT expression, as detected by RNAscope ISH, may decrease as PDAC progresses to metastasis. As such, anti-TIGIT therapy may have important implications for evading an important mechanism of cancer progression.

648 Background: The survival of metastatic pancreatic cancer (mPC) is still disappointing though advancement in recent regimens. Antroquinonol, a new chemical entity, has been proposed for the treatment of neoplasms. In this phase I/II trial, we investigated the dose and efficacy of antroquinonol combined with gemcitabine and nab-paclitaxel (Gem/Nab-P) on mPC patients. Methods: Patients with chemo-naive, metastatic PDAC were enrolled. In the phase I, run-in drug-drug interaction (DDI) and dose escalation in a 3 + 3 designed to determine the maximal tolerated dose (MTD) of antroquinonol for phase II study. Gem/Nab-P (gemcitabine 1000 mg/m2 and nab-paclitaxel 125 mg/m2 on days 1, 8, and 15 every 4 weeks) was given from cycle 0 in phase I. The dose of antroquinonol was escalated from 200mg orally three times a day since the first cycle of Gem/Nab-P. The primary end points were median PFS and 6-month PFS rate. This trial is registered at ClinicalTrials.gov: NCT03310632. Results: In the phase I study of 15 patients, the MTD of antroquinonol was 300mg tid. In the phase II study of 40 patients, the median PFS was 5.3 (95% CI: 3.7–7.5) months and 6-month PFS rate was 40% (95% CI: 21%–57%), whereas median OS was 12.6 (95% CI: 8.8–15.8) months and12-month OS rate was 59.9% (95% CI: 37.8%–76.4%), respectively. The adverse events including hematological and non-hematological classes were decreased in the antroquinonol plus Gem/Nab-P, the GI discomforts were increased but manageable. Conclusions: In this phase I/II trial, antroquinonol plus Gem/Nab-P showed good efficacy in survival and less adverse events than a first-line strategy of Gem/Nab-P for mPC patients. Clinical trial information: NCT03310632 . [Table: see text]

Supplementary Table 1. Summary of Treatment-Related Adverse Events