Search

Our previous phase II study demonstrated that nivolumab provides modest but durable clinical efficacy in patients with refractory biliary tract cancer, suggesting the significant clinical benefit of nivolumab in selected patients and the necessity of predictive biomarkers. We evaluated clinicopathological characteristics and tumour microenvironment of the patients who were enrolled the trial to identify potential biomarkers.Baseline clinicopathological characteristics and pretreatment tumour samples were collected. The obtained tumour samples were assessed for whole exome sequencing, RNA sequencing and immunohistochemistry. Their correlations with clinical outcome were analysed.Pretreatment tumour evaluation revealed PD-L1 expression on tumour, CD8 T cell infiltration and high ratio of CD8 T cell/regulatory T cell in tumour microenvironment were significantly associated with prolonged progression-free survival (PFS), while PD-1 expression on lymphocytes and CD68 macrophages infiltration in tumour microenvironment had no predictive role. Asian patients (N = 3) had improved PFS and disease control rate compared with non-Asian (N = 54). A six-gene predictive model was constructed by evaluation of total 23,550 candidate genes from RNA sequencing of baseline tumour samples using LASSO-Cox regression analysis, and high score of the six-gene prediction model was associated with prolonged PFS.This study suggests that PD-L1 expression on tumour, CD8 T cell infiltration and high ratio of CD8/regulatory T cells and six-gene expression profile in tumour microenvironment may be potential predictive biomarkers of nivolumab in biliary tract cancers. Further studies are needed to confirm these findings.

Supplement

Outcomes for patients with advanced hepatocellular carcinoma (HCC) remain poor despite recent progress in drug development. Emerging data implicate FGF19 as a potential HCC driver, suggesting its receptor, FGFR4, as a novel therapeutic target. We evaluated fisogatinib (BLU-554), a highly potent and selective oral FGFR4 inhibitor, in a phase I dose-escalation/dose-expansion study in advanced HCC using FGF19 expression measured by IHC as a biomarker for pathway activation. For dose escalation, 25 patients received 140 to 900 mg fisogatinib once daily; the maximum tolerated dose (600 mg once daily) was expanded in 81 patients. Fisogatinib was well tolerated; most adverse events were manageable, grade 1/2 gastrointestinal events, primarily diarrhea, nausea, and vomiting. Across doses, the overall response rate was 17% in FGF19-positive patients [median duration of response: 5.3 months (95% CI, 3.7–not reached)] and 0% in FGF19-negative patients. These results validate FGFR4 as a targetable driver in FGF19-positive advanced HCC.Significance:Fisogatinib elicited clinical responses in patients with tumor FGF19 overexpression in advanced HCC. These results validate the oncogenic driver role of the FGFR4 pathway in HCC and the use of FGF19 as a biomarker for patient selection.See related commentary by Subbiah and Pal, p. 1646.This article is highlighted in the In This Issue feature, p. 1631

Differential tumor response to therapy is partially attributed to tumor heterogeneity. Additional efforts are needed to identify tumor heterogeneity parameters in response to therapy that is easily applicable in clinical practice. We aimed to describe tumor response-speed heterogeneity and evaluate its prognostic value in patients with metastatic colorectal cancer.Individual patient data from Amgen (NCT00364013) and Sanofi (NCT00305188; NCT00272051) trials were retrieved from Project Data Sphere. Patients in the Amgen 5-fluorouracil, leucovorin, oxaliplatin (FOLFOX) arm were used to establish response-speed heterogeneity. Its prognostic value was subsequently validated in the Sanofi FOLFOX arms and the Amgen panitumumab+FOLFOX arm. Kaplan-Meier method and Cox proportional hazards models were used for survival analyses.Patients with high response-speed heterogeneity in the Amgen FOLFOX cohort had significantly shorter (P0.001) median progression-free survival (PFS) of 7.27 months (95% CI, 6.12-7.96 mo) and overall survival (OS) of 16.0 months (95% CI, 13.8-18.2 mo) than patients with low response-speed heterogeneity with median PFS of 9.41 months (95% CI, 8.75-10.89 mo) and OS of 22.4 months (95% CI, 20.1-26.7 mo), respectively. Tumor response-speed heterogeneity was a poor prognostic factor of shorter PFS (hazard ratio, 4.17; 95% CI, 2.49-6.99; P0.001) and shorter OS (hazard ratio, 2.57; 95% CI, 1.64-4.01; P0.001), after adjustment for other common prognostic factors. Comparable findings were found in the external validation cohorts.Tumor response-speed heterogeneity to first-line chemotherapy was a novel prognostic factor associated with early disease progression and shorter survival in patients with metastatic colorectal cancer.

BackgroundWhile the incidence of newly diagnosed early-onset colorectal cancer has been increasing, age-related disparity of survival outcome and treatment-related adverse events in patients with metastatic CRC (mCRC) has been inadequately studied with inconclusive findings. In this study, we aimed to evaluate such age-related disparity in this patient population.MethodsWe used individual patient data from three clinical trials (Study 1:NCT00272051, NCT 00305188 and Study 2:NCT00364013) in Project Data Sphere. All patients were diagnosed with mCRC and received first-line 5-fluorouracil and oxaliplatin. Clinical and genomic data of 763 patients with mCRC from Moffitt Cancer Center were used to assess genomic alterations and serve as an external and real-world validation cohort to evaluate overall survival (OS) disparity. Survival outcomes and treatment-related adverse events were estimated and compared in patients among three age groups: 65 years.ResultsAmong 1223 patients from previous clinical trials, 179 (14.6%) were younger than 50 years. These patients had significantly shorter progression-free survival (PFS) (HR=1.46; 95%CI=1.22–1.76;ppp=0.019), severe abdominal pain (8.4% vs 3.4% vs 3.5%,p=0.018), severe anemia (6.1% vs 1.0% vs 1.5%,pp=0.047), but significantly lower incidence of fatigue, severe diarrhea, severe fatigue, and severe neutropenia. The p=0.012), mucositis (3.6 vs 5.1 vs 5.7 weeks,p=0.051), and neutropenia (8.0 vs 9.4 vs 8.4 weeks,p=0.043), and shorter duration of mucositis (0.6 vs 0.9 vs 1.0 weeks,p=0.006). In the CTNNB1mutation (6.6% vs 3.1% vs 2.3%,p=0.047),ERBB2amplification (5.1% vs 0.6% vs 2.3%,p=0.005), andCREBBPmutation (3.1% vs 0.9% vs 0.5%,p=0.050), but lower prevalence ofBRAFmutation (7.7% vs 8.5% vs 16.7%,p=0.002).ConclusionsPatients with early-onset mCRC had worse survival outcome and unique adverse-event patterns, which could be partially attributed to distinct genomic profiles. Our findings might improve an individualized approach to chemotherapy, counseling, and management of treatment-related adverse events in this patient population.

Biliary tract cancers (BTCs) are a heterogenous group of cancers arising from the biliary tract. The hallmark of these cancers is the advanced stage of presentation and a paucity of durable treatment options. Despite the advances in targeted therapy and immunotherapy in solid tumors, systemic cytotoxic chemotherapy has remained the mainstay for cholangiocarcinomas.With advances in the understanding of the tumor microenvironment, genetic features, and inflammatory milieu, have led to the identification of tumor-infiltrating immune cells as indicators of prognosis and response to treatment in BTC. Through an improved comprehension of immunology, immuno-oncology is becoming another pillar of treatment along with traditional radiation, surgery, cytotoxic chemotherapy, and targeted therapies. This article reviews the evidence for immunotherapy use in cholangiocarcinoma, which still being in infancy, and offers promising new novel options for the management of biliary tract cancers.

PurposeDifferential tumor response to therapy is partially attributed to tumor heterogeneity. Additional efforts are needed to identify tumor heterogeneity parameters in response to therapy that are easily applicable in clinical practice. We aimed to describe tumor response-speed heterogeneity and evaluate its prognostic value in patients with metastatic colorectal cancer (mCRC).Patients and MethodsIndividual patient data from Amgen (NCT00364013) and Sanofi (NCT00305188; NCT00272051) trials were retrieved from Project Data Sphere. Patients in the Amgen 5-fluorouracil, leucovorin, oxaliplatin (FOLFOX) arm were used to establish response-speed heterogeneity. Its prognostic value was subsequently validated in the Sanofi FOLFOX arms and the Amgen panitumumab + FOLFOX arm. Kaplan-Meier method and Cox proportional hazards models were used for survival analyses.ResultsPatients with high response-speed heterogeneity in the Amgen FOLFOX cohort had significantly shorter (PPPConclusionTumor response-speed heterogeneity to first-line chemotherapy was a novel prognostic factor associated with early disease progression and shorter survival in patients with mCRC.Implications for PracticeRoutine clinical decision making heavily relies on radiographic assessment of disease response to therapy. For patients with heterogeneous tumors, the degree and kinetics of individual tumor response to the same therapy can sometimes be vastly different. We explored a novel quantitative parameter to describe response-speed heterogeneity by utilizing individual patient data from previous clinical trials. This parameter was an independent prognostic factor associated with early disease progression and shorter survival. Complementary to existing molecular and radiographic tumor heterogeneity parameters, it may help practicing oncologists describe tumor response disparity and serve as a new prognostic factor for patients with mCRC.

In contrast to mismatch repair deficient (dMMR) colorectal cancer (CRC), mismatch repair proficient (pMMR) CRC is usually unresponsive to anti-PD-1 immunotherapy. Recent preclinical data suggest that regorafenib may enhance the antitumor activity of anti-PD-1 immunotherapy. However, the safety and efficacy of regorafenib plus nivolumab have not been established in patients with refractory metastatic pMMR CRC. This study aimed to evaluate the safety and efficacy of regorafenib plus nivolumab in metastatic pMMR metastatic CRC.This was a phase I/Ib study with standard 3 + 3 design plus dose expansion of the maximum tolerated dose (MTD) in patients with refractory metastatic pMMR CRC. Patients were treated with regorafenib combined with nivolumab. The primary end-points were dose-limiting toxicity (DLT) and MTD. The secondary end-points were objective response rate, safety and overall survival (OS).A total of 52 patients were enrolled, and 51 patients received at least one dose of treatment. Three patients experienced DLT (all grade 3 rash). MTD was regorafenib 80 mg and nivolumab 240 mg every 2 weeks. Most common grade 3/4 treatment-related adverse events were hypertension (16%), rash (10%) and anaemia (6%). Among 40 evaluable patients, four (10%) achieved partial response, including one unconfirmed response, 21 (53%) achieved stable disease, and disease control rate was 63%. The median progression-free survival and OS were 4.3 and 11.1 months, respectively.Regorafenib plus nivolumab appears to be well tolerated with limited anticancer activity in metastatic pMMR CRC.ClinicalTrials.gov identifier: NCT03712943.

188 Background: Our previous phase I/Ib study revealed modest clinical efficacy of rego/nivo in refractory MSS-CRC, implying benefits in selected populations. This has prompted us to investigate predictive biomarkers. Methods: Pre-treatment tumor samples from the previous phase Ib rego/nivo study were obtained and assessed for mRNA sequencing (RNAseq). Response-associated transcripts were identified using DESeq2 method and annotated using gene ontology (GO) and Kyoto encyclopedia of genes and genomes (KEGG) enrichment analysis. Results: A total of 19 pretreatment tumor samples were analyzed including 14 patients (pts) with disease control (DC) (2 partial response and 12 stable disease) and 5 pts with progression disease (PD). We observed significant upregulation of 89 genes including SFTPB ( P

TPS629 Background: The anti–PD-L1 antibody ATEZO prevents PD-L1 from interacting with PD-1 and B7.1, thus reinvigorating antitumor T cell activity. Anti-VEGF BEV can increase dendritic cell maturation, enhance T cell infiltration, and reduce myeloid-derived suppressor cells and regulatory T-cells in tumors. This combination has been FDA approved for first-line treatment of advanced HCC based on the IMbrave150 study. On the other hand, locoregional radiotherapy (e.g., Y90 TARE) enhances the diversity of the intratumoral T cell receptor repertoire. It is the standard of care for pts with intermediate-stage HCC (iHCC). Based on preclinical and clinical data, we hypothesize that the combination of Y90 TARE, BEV, and ATEZO induces synergistic tumor-killing. Methods: This is an open-label, multicenter, randomized phase II study of Y90 TARE and BEV plus ATEZO compared with Y90 TARE alone in pts with unresectable IHCC (NCT04541173). The primary study objective is to assess and compare the progression-free survival (PFS) (per mRECIST 1.1) of pts in each arm. The main secondary objective is to determine the safety and tolerability (CTCAE v5) of TARE combined with ATEZO and BEV in pts with HCC. Exploratory objectives are to assess the role of the immunoscore and PD-L1 expression levels in the prediction of improved clinical outcome in pts receiving Y-90 TARE and BEV plus ATEZO; the composition of tumor-infiltrating immune cell subtypes in predicting response to a chosen therapy; how Y90 therapy affects the proportion of antigen-presenting cells in the tumor; and symptoms experienced by pts receiving TARE and BEV plus ATEZO treatment, using patient-reported outcomes. Eligible pts have either HCC that is not amenable to surgical resection, confirmed by pathology review, or at least BCLC stage B HCC outside of downstaging criteria. Other standard eligibility criteria apply. Pts must have a pretreatment liver biopsy taken and then be randomized 1:1 to TARE (Arm A) or TARE followed by ATEZO and BEV (Arm B). Pts will have TARE mapping during week (wk) 1 and TARE treatment during wk 2. In Arm B, pts will start BEV plus ATEZO 4 wks (±1 wk) after TARE treatment. Pts will have abdominal MRI or CT scans every 12 weeks, CT scans of the chest every 24 wks. Disease progression will be captured by both RECIST 1.1 and mRECIST. We plan to assess the safety of TARE with BEV and ATEZO in the first 10 pts randomized to Arm B for two cycles. If there are no Grade ≥ 3 unexpected toxicities possibly, probably or definitely related to combined TARE plus BEV plus ATEZO, the study will continue to accrue 128 pts in total. Pts will continue study treatment (Arm B) for a total of 24 months from initiation of TARE or until intolerable toxicity or disease progression occur, whichever is earlier. Clinical trial information: NCT04541173 .

228 Background: We previously conducted a phase I/Ib study with regorafenib and nivolumab in patients with refractory metastatic mismatch repair proficient (pMMR) colorectal cancer (CRC). This study aimed to investigate the biomarkers that predict the treatment response. Methods: Out of the 51 patients who received regorafenib and nivolumab, 22 archival pretreatment tumor samples were subjected to the Xerna TME Panel, a machine learning-based RNA-sequencing biomarker assay and were classified into one of four TME biomarker subtypes: Angiogenesis (A), Immune Active (IA), Immune Desert (ID), or Immune Suppressed (IS). Potential predictive biomarkers including the TME subtypes, KRAS (wild type vs mutant), PD-L1 (negative vs. positive, samples with > 1% tumor cells for PD-L1 were considered positive), CD8 expression (low vs. high), and Treg cells (low vs. high) in tumor microenvironment were evaluated for correlation with overall survival (OS), progression free survival (PFS) and disease control rate (DCR, defined as complete response + partial response + stable disease). Results: Among the 22 patients, 16 (72.7%) had liver metastasis and 15 (68.2%) had lung metastasis. KRAS mutation was found in 16 (68.2%) patients. 11/21 (52.4%) were positive for PD-L1. 12 (54.5%) had high CD8 expression, whereas 9/21 (42.9%) had high Treg cells in tumor microenvironment. Ten (45.5%) patients were classified as biomarker-positive (IA + IS subtypes) and 12 (54.5%) were biomarker-negative (A + ID) based on Xerna TME panel. Two (9.1%) patients achieved partial response, 12 (54.5%) had stable disease, and five (22.7%) developed progressive disease. The median PFS was 5.6 months and median OS was 13.1 months. No significant correlation was observed between RAS mutation (p = 0.664, p = 0.609), PD-L1 expression (p = 0.287, p = 0.173), CD8 (p = 0.152, p = 0.456) and PFS or OS. Low Treg was found to be associated with prolonged PFS (median: 9.8 vs. 1.9 months, p = 0.011) but not OS (p = 0.280). Similarly, only low Treg level was related with DCR (83.3% vs. 33.3%, p = 0.032). While not reaching statistical significance, Xerna TME biomarker-positive patients showed trends for higher median PFS (7.9 months vs. 4.1 months, p = 0.254), median OS (15.75 months vs. 11.9 months, p = 0.378), and higher DCR (70% vs. 58%, p = 0.675) compared to biomarker-negative patients. Additionally, the two patients with partial responses were Xerna TME biomarker-positive. Conclusions: Our study demonstrated that low Treg in tumor microenvironment is correlated with better prognosis in patients with refractory metastatic pMMR CRC who were treated with regorafenib plus nivolumab. Xerna TME panel analysis of these patients also showed trends for predictive clinical benefit. Prospective and larger cohort studies are needed to better define predictive biomarkers for this combination in the future.

785 Background: Socioeconomic deprivation has been described as a barrier to cancer clinical trial participation. However, few studies have examined socioeconomic deprivation using patient-level or granular geocoded designations. To overcome this challenge, the Area Deprivation Index (ADI) was used to assess neighborhood socioeconomic deprivation in a cohort of gastrointestinal cancer clinical trial patients at an NCI-Designated Comprehensive Cancer Center. Methods: Patients enrolled in a gastrointestinal cancer clinical trial from 2008 to 2019 with an identifiable ADI national rank were identified. Socioeconomic deprivation was assessed using the ADI, a publicly available, validated dataset that ranks census block groups into percentiles using variables such as income, education, employment, and housing characteristics. For this study, ADI was categorized as quintiles listed in the table. Statistical analyses included Chi-square and Kruskal-Wallis tests. Results: The median age of the cohort (N=1,334) was 62.0 years. Most patients were male (54.3%). Race included White (88.2%), African American (6.8%), and Asian (1.5%) patients. Hispanic/Latinx patients made up 6.7% of the cohort. The median ADI was 46. The proportion of enrollees from lowest to highest ADI quintile was 11.2%, 29.3%, 27.3%, 19.1%, and 13.1%. Trial enrollment differed by ADI and age (p=0.019), gender (p=0.042), race (p

239 Background: The benefit of immunotherapy for patients with mCRC with high tumor mutational burden (TMB-H) has been widely debated. In 2020, the FDA approved the use of pembrolizumab for the treatment of patients with TMB-H unresectable or metastatic solid tumors, with TMB-H defined as ≥10 mutations/Mb on a commercial tissue-based assay, based on KEYNOTE 158 results. However, the clinical value of applying this universal cut off to mCRC needs further investigation. Methods: We queried Moffitt Cancer Center (MCC) databases for patients with MSS mCRC, harboring TMB-H (tested with tissue and/or liquid biopsies) who received immunotherapy between January 2018 and December 2021. Patients with incomplete records were excluded. Clinical data were extracted by trained staff from electronic medical records. Objective response rate was measured by using clinical assessment from chart review. Results: We identified 40 patients with TMB-H MSS mCRC 13 of whom received immune checkpoint inhibitor therapy. Female patients represented 31% (n=4) of the 13 treated patients. Median age for the patients was 60 years (range 34-71. Thirty-eight percent (n=5) of the primary tumors originated in the right side. Thirty-one percent (n=4) presented with stage 4 CRC at diagnosis. Histopathology was adenocarcinoma in 92% (n=12) and one had neuroendocrine differentiation. Tumors were well-differentiated 8% (n=1), moderately differentiated 46% (n=6), or poorly differentiated 23% (n=3). Four patients (31%) had POLE/ POLD-1 mutations. The objective response rate (ORR) was 31% (4/13) with responses limited to tumors only with POLE/ POLD-1 mutations (100%, 4/4). The median progression free survival (mPFS) was 3.8 months for the overall cohort and 28.5 months for patients with POLE/POLD-1 mutated tumors. Agent specific analysis showed mPFS for patients treated with pembrolizumab was 3.5 months (mean 5.2; range 0.6-15.7), and 3.4 months (mean 12.1; range 0.5-12.1) for patients treated with nivolumab. Five patients received immunotherapy in combination with regorafenib. The mPFS for these patients was 3.4 months (mean 3.8; range 0.5-8.0). Of five patients with TMB>30 (38%), 4 (80%) had POLE/POLD-1 mutated tumors. One case without POLE/POLD-1 mutations, but with a TMB>30 did not experience any response. Conclusions: Although TMB-H demonstrated therapeutic significance in the KEYNOTE 158 study, the utility of 10 mutations/MB as a universal cutoff warrants additional evaluation. Here we report that a TMB-H cutoff value of ≥ 10 for patients with MSS CRC was not associated with clinically meaningful response to immunotherapy, but patients with MSS CRC with POLE/ POLD-1 mutations may be more likely to benefit from immunotherapy.

The mainstay pharmacological approaches to patients with hepatocellular carcinoma (HCC) are tyrosine kinase inhibitors, antiangiogenic agents, and immune checkpoint inhibitors in combination therapy. Aberrant signaling of fibroblast growth factor 19 (FGF19) and its corresponding receptor, fibroblast growth factor receptor 4 (FGFR4), are a driver of HCC cell growth and survival. However, the clinical potential of agents targeting aberrant FGF19/FGFR4 signaling has not been adequately explored.We evaluate the existing literature on aberrant signaling of FGF19/FGFR4 in HCC and address the recent preclinical and clinical advances of selective FGFR4 inhibitors in the treatment of advanced HCC. Our literature search was performed in September 2021 on clinical trials and ongoing studies published in journals or presented in conferences for cancer research.Preclinical studies show selective FGFR4 inhibitors to be highly potent. These inhibitors also show promise in clinical trials and demonstrate manageable on-target side effects. An emphasis should be placed on the development of predictive biomarkers and on enhancing the understanding of primary and acquired resistance mechanisms. This will inspire rationale combination therapy strategies for testing in future clinical trials.

BACKGROUND: Recommendations for postoperative surveillance and adjuvant therapy following curative-intent resection for biliary tract cancers—including intrahepatic and extrahepatic cholangiocarcinoma (IHCCA and EHCCA) and primary gallbladder cancer (GBC)—are uniform across primary tumor site. However, these tumors may have distinct patterns of recurrence. METHODS: A retrospective observational cohort study was performed at a specialty cancer center. Patients undergoing resection of IHCCA, EHCCA, and GBC were identified (2005–2020). Recurrence-free survival (RFS) was estimated using Kaplan-Meier and Cox proportional hazard methods. Anatomic patterns of initial site of recurrence were described and compared. RESULTS: There were 142 patients included; 50 IHCCA, 32 EHCCA, and 60 GBC. Median RFS was 30.8 months, which was not significantly different between IHCCA, EHCCA, or GBC in univariate analysis or after adjustment. Nodal positivity was significantly associated with poor RFS (HR 3.92, P≤0.001). The most common initial site of recurrence overall was intrahepatic (n=49/64, 77%), in isolation (n=32) or synchronous with other site of recurrence (n=17). Significant differences in anatomic pattern of recurrence were observed (P=0.049) with IHCCAs more commonly recurring with simultaneous hepatic-pulmonary disease (n=5/22, 23%; EHCCA n=2/19, 10%; GBC n=1/23, 4%), GBC more commonly recurring within the porta (n=7/23, 30%; IHCCA n=0; EHCCA n=1/19, 5%), and EHCCA more commonly recurring within the peritoneum (n=5/19, 26%; IHCCA n=2/22, 9%, GBC n=2/23, 9%). CONCLUSIONS: Patterns of initial recurrence appear to differ between primary tumor site, likely reflecting underlying differences in anatomy and biology. These data could help inform future studies for adjuvant therapy as well as timing and anatomic focus for surveillance imaging.

TPS4177 Background: The anti–PD-L1 antibody atezolizumab (ATEZO) prevents PD-L1 from interacting with PD-1 and B7.1, thus reinvigorating antitumor T cell activity. Anti-VEGF bevacizumab (BEV) increases dendritic cell maturation, enhances T cell infiltration, and reduces myeloid-derived suppressor cells and regulatory T cells in tumors. ATEZO + BEV is FDA approved for first-line treatment of advanced HCC based on the IMbrave 150 study. Locoregional radiotherapy (e.g., Y90 TARE) enhances the diversity of the intratumoral T cell receptor repertoire and increases tumor antigen release. We hypothesize that the Y90 TARE + BEV + ATEZO combination induces synergistic tumor killing and prolongs progression-free survival in patients (pts) receiving Y90 TARE (HR = 0.6 when compared to Y-90 TARE alone). Methods: Eligible pts have HCC that cannot be surgically resected (confirmed by pathology review), is at least BCLC stage B, and is outside Milan criteria. Other requirements include ECOG PS of 0-1 at screening, measurable disease by RECIST 1.1, no prior systemic therapy, and FLR estimated at ≥ 40% post local therapy. Pts must have a pretreatment liver biopsy taken and then be randomized 1:1 to TARE (Arm A) or TARE + ATEZO + BEV (Arm B). Pts will have TARE mapping followed by TARE treatment. In Arm B, pts will begin TARE treatment followed by BEV + ATEZO (4 wks [± 1 wk] later). Pts will have abdominal MRI or CT scans every 12 weeks and CT scans of the chest every 24 wks. The primary study objective is to assess and compare pts' progression-free survival (per mRECIST 1.1) in each study arm. The main secondary objective is to determine the safety and tolerability (CTCAE v5) of pts in Arm B. Exploratory objectives are to define the use of cellular and circulating biomarkers in the prediction of improved clinical outcomes of pts in Arm B. Symptoms experienced by pts in both arms using patient-reported outcomes will be assessed. Disease progression will be captured by both RECIST 1.1 and mRECIST. We plan to assess the safety of Y90 TARE + BEV + ATEZO in the first 10 pts randomized to Arm B for two cycles, and if there are no grade ≥ 3 unexpected toxicities possibly, probably, or definitely related to combined TARE + BEV + ATEZO, continue to accrue 128 pts in total (current enrollment n- = 5). Pts will continue study treatment (Arm B) for a total of 24 months from initiation of TARE or until intolerable toxicity or disease progression occur, whichever is earlier. Enrollment began in September 2020. Clinical trial information: NCT04541173.

3521 Background: Response to antiPD-1 monotherapy is poor in patients (pts) with advanced microsatellite stable (MSS)/mismatch-repair proficient (pMMR) CRC. Combination of chemotherapy with antiPD-1 pembro may potentiate the antitumor immune response and provide greater antitumor activity than either agent alone. Combination of pembro and mFOLFOX7 or FOLFIRI in the ongoing phase 1b multicohort KEYNOTE-651 (NCT03374254) study in metastatic MSS/pMMR CRC showed antitumor activity. We present results with approximately 20 mo of additional follow-up. Methods: Pts had metastatic MSS/pMMR CRC and were previously untreated (cohort B) or received 1 prior line including a fluoropyrimidine + oxaliplatin (cohort D). Pts received pembro 200 mg Q3W + mFOLFOX7 Q2W (cohort B) or pembro 200 mg Q3W + FOLFIRI Q2W (cohort D). Primary end points were safety (DLT) and RP2D. Secondary end point was ORR per RECIST v1.1 by investigator review. DOR, DCR, and PFS per RECIST v1.1, and OS were exploratory end points. Results: Median study follow-up (range) at data cutoff (Oct 15, 2021) was 30.2 mo (25.0-43.6) for cohort B (n = 31) and 33.5 mo (25.2-43.2) for cohort D (n = 32). Treatment was discontinued in 29 pts (94%) in cohort B and 28 pts (88%) in cohort D, mostly because of PD (61% and 66%, respectively). At prior analysis (Feb 10, 2020), RP2D was confirmed as the starting dose in both cohorts; no new DLT had occurred. In cohort B, gr 3/4 TRAEs occurred in 18 pts (58%), most commonly neutropenia and decreased neutrophil count (both n = 5; 16%); 19 pts (61%) discontinued drug because of a TRAE. In cohort D, gr 3/4 TRAEs occurred in 17 pts (53%), most commonly, neutropenia (n = 7; 22%), diarrhea and fatigue (both n = 4; 13%); 3 pts (9%) discontinued drug because of a TRAE. There were no gr 5 TRAEs in either cohort. Efficacy by cohort and KRAS mutation status are below (Table). PD-L1 data and DNA/RNA-based biomarker data including GEP, consensus signatures, TMB, and LoF mutations will be included in the presentation. Conclusions: After 2.5 y of follow-up, the combination of pembro with either mFOLFOX7 or FOLFIRI continued to demonstrate a manageable safety profile with no new safety signals. Efficacy data from these single-arm cohorts appear comparable to historical data for current SOC. Clinical trial information: NCT03374254. [Table: see text]

3573 Background: Response to antiPD-1 monotherapy is poor in microsatellite stable (MSS)/mismatch-repair proficient (pMMR) mCRC; the combination of antiPD-1 pembro + anti-MEK bini, with or without chemo, may improve upon this limited response. KEYNOTE-651 (NCT03374254) is an open-label phase 1b multicenter trial of pembro + bini (cohort A) or pembro + bini + chemo (mFOLFOX7 in cohort C, FOLFIRI in cohort E) in MSS/pMMR mCRC. Preliminary results from the dose-finding phase at 2 dose levels (DL) of bini are presented. Methods: Patients (pts) had MSS/pMMR mCRC and must have been previously treated with fluoropyrimidine, irinotecan, and oxaliplatin in cohort A, or with fluoropyrimidine + oxaliplatin-based regimen in cohort E; pts were previously untreated in cohort C. Pts received pembro 200 mg Q3W + bini 30 mg BID (cohort A, DL1), pembro 200 mg Q3W+ bini 30 mg BID + mFOLFOX7 Q2W (cohort C, DL1) or pembro 200 mg Q3W + bini 30 mg BID + FOLFIRI Q2W (cohort E, DL1). Bini dose escalation to 45 mg BID (DL2) was planned in cohorts A, C, and E, with a target dose-limiting toxicity (DLT) of 30%. Primary end point was safety (DLT). Secondary end point was ORR. DCR, PFS, and OS were exploratory. ORR, DCR, and PFS were assessed by investigator per RECIST v1.1. Results: Median study follow-up at data cutoff (Oct 15, 2021) was 36 mo (range, 32-43) for cohort A, 17 mo (2-24) for cohort C, and 11 mo (2-25) for cohort E. In cohort A, 1/6 pts (17%) had DLT at DL1; no DLT occurred in 14 pts (0%) at DL2. In cohort A, gr 3/4 TRAEs occurred in 3/6 pts (50%) at DL1 and 8/14 pts (57%) at DL2. In cohort C, 3/9 evaluable pts (33%) had DLT at DL1; thus, bini dose was not escalated to DL2. In cohort C, gr 3/4 TRAEs occurred in 9/11 total pts (82%). In cohort E, 1/5 evaluable pts (20%) had DLT at DL1 and 5/10 evaluable pts (50%) had DLT at DL2. Enrollment was stopped in cohort E, DL2 and bini dose was de-escalated to DL1; 2/4 additional pts (50%) had DLT at DL1 (total 3/9 pts [33%] had DLT in cohort E, DL1). In cohort E, gr 3/4 TRAEs occurred in 5/9 pts (56%) at DL1 and 10/11 total pts (91%) at DL2. No gr 5 TRAEs occurred in any cohort. ORR was 0% in cohort A; limited efficacy was seen in cohorts C and E (Table). Conclusions: Bini could be safely combined with pembro in cohort A. However, with bini + pembro + chemo, the 45-mg dose of bini was not well tolerated and required dose reduction to 30 mg. Addition of bini to pembro + chemo did not improve efficacy; therefore, enrollment was prematurely closed in cohorts C and E. Efficacy by KRAS mutation status will be shown. Clinical trial information: NCT03374254. [Table: see text]

2514 Background: Checkpoint inhibitors (CPIs) are usually ineffective in patients (pts) with immune-cold microsatellite stable colorectal cancer (MSS-CRC) or pancreatic cancer (PDAC) and in those who progressed on previously treated with antibodies against PD1 or PD-L1. Here, we report the combination of NT-I7 plus pembrolizumab (pembro) on CPI-naïve MSS-CRC and PDAC cohorts, and patients (pts) with CPI-treated triple-negative breast cancer (TNBC), non-small cell lung cancer (NSCLC), and small cell lung cancer (SCLC) cohorts of this ongoing phase 2a trial. Methods: Pts with CPI-naïve relapsed/refractory (R/R) MSS-CRC and PDAC, and CPI-treated R/R TNBC, NSCLC, and SCLC, were enrolled. NT-I7 (efineptakin alfa) 1200 µg/kg intramuscularly every 6 weeks and 200 mg pembro intravenously every 3 weeks were administered until disease progression/unacceptable toxicity. The primary endpoint of the Phase 2a is the objective response rate (ORR), assessed by RECIST v1.1 and iRECIST. The secondary endpoints are duration of response (DoR), disease control rate, progression-free survival, and overall survival. Results: As of 14 Jan 2022, 92 pts with metastatic or locally advanced cancer who had received a median of 3 prior treatments were enrolled in the study; 32 in PDAC, 28 in MSS-CRC, 22 in NSCLC, 6 in TNBC, and 4 in SCLC. The median age was 62 years [29-81], ECOG PS 0 in 23 (25%), 1 in 68 (74%) and 2 in 1 (1%). Among 71 evaluable pts, the median follow up (months) was 7.7, 5.3, 5.0, 3.7, and 2.4 in TNBC, MSS-CRC, NSCLC, PDAC and SCLC, respectively. The ORR was 50% (1/2) in SCLC, 12% (3/25) in MSS CRC, 8% (2/26) in PDAC 6% (1/16) in NSCLC and 0% (0/2) in TNBC per iRECIST; and 50% (1/2) in SCLC, 4% (1/25) in MSS CRC, and 4% (1/26) in PDAC per RECIST 1.1. All 7 responders are ongoing. The two PDAC pts had DoR over 1.35 months (mos) and 6.64 mos with the best tumor reduction 100% and 72% respectively. The one SCLC pt had DoR over 1.5 mos with the tumor reduction 67%. The one NSCLC pt had DoR over 2.73 mos with the tumor reduction 60%, and the three MSS-CRC pts had DoR 6.34, 2.96, and 0.03 mos with the tumor reduction 60%, 56%, and 43% respectively. A sustained and significant (approximate 3X from baseline) increase of peripheral lymphocytes in all arms was observed as shown in our previous report. Among 92 treated pts, NT-I7-related adverse events (AEs) occurred in 67 (72.8%) pts, including 52 (56.5%) Grade (G)1-2, 13 (14.1%) G3, and 2 (2.2%) G4. There were no NT-I7 related G5 AEs. Additional updated efficacy, safety and biomarker data will be presented. Conclusions: The combination of NT-I7 and pembro demonstrated antitumor activity and manageable toxicity profile in heavily pretreated pts with CPI-naïve MSS-CRC and PDAC and CPI-treated TNBC, NSCLC, and SCLC, suggesting that the addition of NT-I7 to CPI can overcome the primary resistance to CPI in the former group and acquired resistance in the latter. Clinical trial information: NCT04332653.

e15569 Background: Increased molecular profiling ability has resulted in the recognition of important actionable genes that can be used to direct targeted therapies in colorectal cancers (CRC). Neurotrophin receptor tyrosine kinase (NTRK) gene fusions are one of these molecular alterations that are potentially actionable with novel targeted therapeutics. However, other biologically pathogenic NTRK alterations, including point mutations, are not yet actionable in CRC. This study aimed to characterize NTRK alterations, including fusions in metastatic CRC patients. Methods: Molecular characteristics of 917 patients with CRC were collected from the Moffitt Cancer Center Clinical Genomics Action Committee database. Demographic, clinicopathological, and molecular data and treatment history were abstracted from electronic medical records. Mutations with potential oncogenic activity are considered “potentially pathogenic”, while alterations with proven oncogenic activity such as fusions are considered “pathogenic”. Results: There were 917 patients with CRC from November 2013 to December 2021, and 77 of them had NTRK alterations excluding synonymous mutations (8.3%, 77/917). Six patients had potentially pathogenic NTRK alterations, including one NTRK rearrangement (0.6%, 7/917); however, only 1 patient had an actionable pathogenic NTRK1 fusion (0.1%, 1/917), while the others (70) were a variant of unknown significance (VUS). The majority of NTRK alterations were missense mutations (92%; 71/77). The NTRK1 fusion partner was LMNA and occurred in the setting of an MSI-H tumor (3.7%; 1/27). Among patients with pathogenic and potentially pathogenic alterations (N = 7), there was only 1(14%) patient who had low tumor mutation burden (TMB) (< 10 mut/MB). Most patients were older than 50 years (70.1%, 54/77) and male (58.4%, 45/77). Thirty-six patients (46..7%) had right-sided tumor, while 41 patients (53.2%) left-sided. Across all NTRK alterations cohort, MSI-H was found in 10.3% (7/77), and POLE/POLD1 pathogenic mutations were seen in 6.4% (5/77) patients. Conclusions: NTRK gene fusion is a relatively rare event (< 1%) in CRC, including the MSI-H subtype. The majority of NTRK alterations in CRC are VUS, and they are not actionable, and they tend to be seen in the tumors with high TMB (TMB > 10). Hypermutator tumors cause frequent VUS alterations in the NTRK gene with unknown clinical relevance.

e15568 Background: Next-generation sequencing (NGS) based molecular profiling technologies have revealed several oncogenic fusion genes that are actionable with small molecule inhibitors leading to practice change, particularly in lung cancer. The molecular and clinical characteristics of these gene fusions are not well defined in colorectal cancer patients (CRC). In this study, we aimed to define clinical and molecular characteristics of fusion genes in patients with CRC who underwent molecular profiling. Methods: Molecular characteristics of tissue confirmed 917 CRC patients were retrieved from the Moffit Cancer Center Clinical Genomics Action Committee database. Patients’ demographic and clinicopathological features and treatment history were collected from the database. All fusion genes were shown by hybridization-based NGS computational algorithms that determined cancer‐related genes, including single‐nucleotide variations, indels, microsatellite instability (MSI) status. Results: Among a total of 917 patients, 24 patients with CRC (2.6%) were found to have at least one fusion gene with a total number of 26 pathogenic fusions. The gene fusions are shown in Table. The most common, potentially targetable, fusion genes in our cohort were (1) RET fusions 0.5% (5/917), (2) ALK fusions 0.4% (4/917), (3) ROS1 fusions 0.2% (2/917), (4) NTRK1 fusion 0.1% (1/917), (5) NRG1 fusion 0.1% (1/917). Fusion genes were more common in MSI-H CRC (N = 27), and 3 (11.1%) patients with MSI-H CRC were found to have fusion genes [(RET (2) and NTRK(1)]. Fusion genes were present in both RAS wild-type (54%; 13/24) and RAS mutant (46%; 11/24) tumors. Most patients were older than 50 years (75%, 18/24) and had left-sided tumor (61.1%) tumor. Conclusions: Fusion genes are rare events in CRC. While fusion genes seem to be more prevalent in MSI-H CRC, RAS status does not correlate with the frequency of fusion genes. Actionable RET and ALK/ROS gene fusion are more common than NTRK fusion genes in this cohort of CRC patients.[Table: see text]

6560 Background: While the incidence of newly diagnosed early-onset mCRC has been increasing, disparity of treatment-related adverse events (AE) and outcomes of this patient population has been inadequately studied with inconclusive findings. We aimed to evaluate such age-related disparity and explore potential underlying causes. Methods: We used individual patient data from 3 clinical trials in Project Data Sphere where 756 and 467 patients with mCRC received first-line FOLFOX in study 1 (NCT 00305188, NCT00272051) and study 2 (NCT00364013), respectively. Clinical NGS data of 763 patients with mCRC from prospectively maintained Moffitt Clinical Genomics Database were used to assess genomic alterations. Patients were categorized into 3 age groups: 65 years. Continuous and categorical variables were compared with t test and χ2 test, respectively. Kaplan-Meier method and log-rank test were used for survival analysis. Benjamini-Hochberg procedure was used to adjust for multiple comparisons. Results: Among 1986 patients included, 341 (17.2%) in the 65 group had similar baseline characteristics. Outcomes: Patients in the 65 groups (15.5 vs 20.8 months, p=0.004) and shorter median PFS compared to the 50-65 (8.1 vs 9.4 month, p=0.039) and >65 groups (8.1 vs 8.6 months, p=0.07) in study 1. Findings were confirmed in study 2. Toxicity: Compared to other age groups, the

663 Background: GI malignancies continue to be a major cause of cancer related deaths, despite advances in therapy options. In pancreatic cancer, the presence of homologous recombination mutations (HRM) has been shown to confer increased sensitivity to platinum chemotherapy. However, the role of HRM mutations in other GI cancers remains to be determined. The focus of this study is to evaluate the prognostic nature of core and noncore HRM in nonpancreatic GI cancers. Additionally, we aim to understand the predictive nature of these mutations related to platinum exposure. Methods: This was a retrospective review of patients at Moffitt Cancer Center with a primary stage IV nonpancreatic GI cancer treated with platinum therapy. All patients had next generation sequencing and were in the Clinical Genomics Action Committee database between 1/1/2013 and 11/1/2020. All patients had either a core HRM (BRCA1, BRCA2, PALB2) or noncore HRM (such as ATM, ATR, BARD1, BRIP1, FANCA, NBN, and RAD51). Patients were grouped into core HRM vs. noncore HRM. Response was stratified between responders (complete response and partial response) vs. non responders (stable disease and progressive disease), based on RECIST criteria. Progression free survival and overall survival were estimated using Kaplan-Meier method, and the log-rank test was used to assess the difference in progression free survival (PFS) and overall survival (OS) by HRM status and type of platinum therapy used. Results: There were 72 patients included in the study: the majority of patients were male (65.3%) and Caucasian (87.5%). Twenty-one (29.2%) patients had a core HRM and 51 (70.8%) had a noncore HRM. There was no significant difference in response rate between patients with core HRM and patients with noncore HRM for evaluable patients (n = 66): 20.0% of patients with core HRM vs. 21.7% of patients with noncore HRM demonstrated a response to platinum therapy (p = 0.41). An improved OS was noted for patients with core HRM vs. those with noncore HRM at 68.9 vs. 24.3 months (p = 0.019). An improved PFS was also seen in patients with core HRM at 10.4 months vs. 6.3 months with noncore HRM (p = 0.027). There was no difference in PFS based on type of platinum therapy used (oxaliplatin vs. carboplatin vs. cisplatin), while an improved OS was noted for patients treated with oxaliplatin vs. carboplatin vs. cisplatin at 43.1 vs. 28.2 vs. 16.0 months (p = 0.011). Conclusions: Our study demonstrated that in stage IV nonpancreatic GI malignancies treated with platinum therapy, patients with core HRM had a greater OS and PFS compared to those with noncore HRM, suggesting a potential prognostic and predictive role of these mutations. Further studies are needed to confirm our findings.

302 Background: The RAINBOW trial established the standard of care for treatment of metastatic GEAs with ramucirumab and paclitaxel after failure of fluoropyrimidine and platinum-based chemotherapy. While the combination achieved an objective response rate (ORR) of 28%, the incidence of any grade neuropathy was 46%. Therefore, there is an unmet need for novel treatment combinations that minimize the long-term toxicity of neuropathy. In a recently published Asian study, the combination of RAM+TAS-102 showed good disease control and acceptable toxicity profile. Methods: This was a single arm, single institution phase II trial using the combination of TAS-102 plus RAM in refractory GEAs. Patients (pts) received RAM 8mg/kg intravenously on day 1 and 15, and TAS-102 35mg/m2 orally twice daily on days 1-5 and days 8-12 every 28-day cycle. The primary endpoint was 6-months overall survival (OS) rate and secondary endpoints were progression free survival (PFS), ORR and safety profile. The trial was registered at www.clinicaltrials.gov (NCT03686488). Results: At data cut-off of August 15, 2021, 23 pts were enrolled. Baseline demographics are as follows: median age of 62 years (range: 23-74), median lines of prior therapy of 1 (1: 14 pts vs ≥2: 9 pts) and location of primary tumor (GEJ:19 vs GC: 4). 6-month OS rate was 56.2%. Median OS was 6.2 month (95% CI: 5.4-7.0) and median PFS was 4.9 months with median observation of 2.3 months. Of 17 evaluable pts defined as more than one baseline imaging, 1 (6%) had partial response (PR) and 15 (88%) had stable disease (RECIST v1.1). Eleven pts came off the study due to progression of disease, 8 for toxicities and 4 for consent withdrawal. Most common treatment-emergent adverse events (TEAEs) were diarrhea (39%), fatigue (39%) and hypertension (39%). Total 11 pts (48%) experienced grade 3 and 4 TEAEs, and most common Gr3 and 4 TEAEs were neutropenia (17%) and anemia (13%). Conclusions: The combination of RAM and TAS-102 showed very similar disease control rate as the study in Asia. The combination has now shown modest activity in advanced GEAs and should be investigated further now in the context of patients receiving immunotherapy-based treatment as first-line treatment. A randomized phase II study is currently enrolling patients with ramucirumab and TAS-102 or paclitaxel. Clinical trial information: NCT03686488.

174 Background: Colorectal cancer (CRC), while one of the most common cancer diagnoses, can behave heterogeneously based on molecular characteristics. A subset of patients (pts) with CRC are characterized with mismatch repair deficiency (MMR-d), these pts exhibit encouraging responses to immunotherapy. The predictive nature of various factors, such as BRAF status, age, and MMR-D protein loss type, have been investigated in pts with MMR-d CRC. However, the prognostic role of these factors has not been well established. The purpose of this study was to identify characteristics that influence survival in MMR-d mCRC. Methods: This study evaluated pts with MMR-d mCRC in the Flatiron database. Overall survival (OS) was determined from date of diagnosis of stage IV disease to date of death and stratified based on age greater than or less than 50 years, BRAF mutation status, RAS mutation status, and type of MMR gene loss. For statistical analysis, the Chi-Square test was implemented to determine the prognostic significance of clinical and molecular features. Univariate and multivariate analyses were determined through the Cox regression model. Results: There were 1,101 pts in the study. The majority of pts were older than 50 (79.7%), Caucasian (75%), and had ECOG 0-1 (83.4%). Among the 803 pts with known BRAF status, 44.3% (n=356) had BRAF V600E mutation and 55.7% (n=447) were BRAF wildtype. Pts with BRAF V600E mutation had OS of 18.9 months vs. 33.2 months for pts with wild type BRAF (HR 1.52, 95% CI: 1.25-1.86, p