Colorectal cancer (CRC) is the third most commonly diagnosed malignancy worldwide and a leading cause of cancer-related mortality. This study aims to investigate an automatic detection pipeline for identification and localization of the primary CRC in portal venous phase contrast-enhanced CT scans, which is a crucial first step for downstream CRC staging, prognostication, and treatment planning. We propose a deep learning-based automated detection pipeline using YOLOv11 as the baseline architecture. A ResNet50 module was incorporated into the YOLOv11 backbone to enhance image feature extraction. Additionally, a scale-adaptive loss function, which introduces an adaptive coefficient and a scaling factor to adaptively measure the Intersection over Union (IoU) and center point distance for improving box regression performance, was designed to further improve detection performance. The proposed pipeline achieved a recall of 0.8092, precision of 0.8187, and F-1 score of 0.8139 for CRC detection on our in-house dataset at the patient level (inter-patient evaluation) and a recall of 0.9949, precision of 0.9894, and F-1 score of 0.9921 at the slice level (intra-patient evaluation). Validation on an external public dataset demonstrated that our pipeline, when trained on a patient-level in-house dataset, obtained a recall of 0.8283, precision of 0.8414, and F-1 score of 0.8348 and, when trained on a slice-level in-house dataset, achieved a recall of 0.6897, precision of 0.7888, and F-1 score of 0.7358, outperforming existing representative detection methods. The superior CRC detection performance on the in-house CT dataset and state-of-the-art generalization performance on the public dataset (with a 31.97 %age point improvement in detection sensitivity (recall) over the next closest state-of-the-art method), highlight the potential translational value of our pipeline for CRC clinical decision support, conditional upon validation in larger cohorts.

Abstract Background Multiple drugs targeting the human epidermal growth factor receptor-2 (HER2) have shown clinical activity in gastrointestinal (GI) cancers. How to best integrate these various options in clinical practice alongside conventional therapy, however, is unclear. Methods We assessed the clinical outcomes of 103 patients with advanced HER2-altered GI malignancies who received ≥ 1 line of anti-HER2 therapy at our institution between 2010 and 2023. Results Next-generation sequencing (NGS) detected an ERBB2 amplification in 66% (27/41) of tumors that were HER2 + by IHC/ISH. Conversely, all patients with ERBB2 amplification were HER2 + by IHC. In the second-line and beyond, trastuzumab plus pertuzumab had the longest median time to treatment discontinuation (7.9 months), followed by trastuzumab deruxtecan (5.0 months). Among 25 patients who received ≥ 2 anti-HER2 agents, 24% had more durable clinical benefit to their second-line HER2 therapy compared to their first, as measured by the growth modulation index. Conclusions Multimodal HER2 testing is needed to accurately identify candidates for HER2-targeted treatment, as NGS alone misses a significant proportion of cases. Patients who develop resistance to one anti-HER2 agent may still achieve benefit from subsequent HER2-directed regimens. Early and serial treatment with HER2-directed agents should be considered for patients with HER2 + GI cancers.

BACKGROUND:Patients with metastatic pancreatic ductal adenocarcinoma (mPDAC) often respond to cytotoxic therapy, but early disease progression is typical. Responses to immunotherapy alone are rare. Recent advances in chemoimmunotherapy combinations offer promise. We report results from cohorts A and B of REVOLUTION, an adaptive platform trial designed to evaluate the safety and antitumor activity of chemoimmunotherapy combinations in untreated mPDAC. METHODS:REVOLUTION (NCT04787991) is an open-label, exploratory platform trial. Patients were assigned to enrolling cohorts in a non-randomized fashion. All patients received gemcitabine (1,000 mg/m2), nab-paclitaxel (125 mg/m2), and two doses of ipilimumab (1 mg/kg), administered intravenously. In addition, cohort A received nivolumab (360 mg intravenously every 3 weeks) and cohort B received hydroxychloroquine (600 mg orally two times a day). The primary endpoint was safety. Secondary endpoints included objective response rate (ORR), disease control rate, duration of response, progression-free survival, and overall survival (OS). Exploratory endpoints included pharmacodynamic changes and associations between biomarkers and clinical outcomes. RESULTS:Both cohorts enrolled 15 patients. Grade 3-4 treatment-related adverse events occurred in 60% and 53% of patients in cohorts A and B, respectively. One grade 5 event occurred in cohort B, which exhibited more frequent dose modifications and non-compliance. Cohort A demonstrated an ORR of 33% (5/15) and a 12-month OS rate of 65.5% (95% CI 35.7% to 84.0%), with higher baseline levels of programmed cell death protein-1 (PD-1)+CD39+ central memory CD4+ T cells associated with prolonged survival. Cohort B demonstrated an ORR of 40% (6/15) and a 12-month OS rate of 53.9% (95% CI 24.3% to 76.3%). Cohort A showed increases in activated and proliferating CD4+ and CD8+ T cells, regulatory T cells, and circulating soluble PD-1 and Th1-associated cytokines. Cohort B exhibited delayed but sustained increases in activated CD4+ T cells and pharmacodynamic evidence of autophagy inhibition. CONCLUSIONS:REVOLUTION cohorts A and B demonstrated encouraging antitumor activity in patients with mPDAC. In cohort B, hydroxychloroquine-related tolerability issues contributed to early discontinuations and reduced drug exposure. These findings highlight the potential and limitations of current chemoimmunotherapy approaches. Although neither cohort will be expanded, the results reinforce the continued promise of chemoimmunotherapy in mPDAC and the importance of refining these strategies.

Importance:Colorectal cancer is the third most commonly diagnosed cancer and the third leading cause of cancer-associated deaths in the US. Hispanic and non-Hispanic Black patients experience higher colorectal cancer mortality rates compared with non-Hispanic White patients. More data are needed to understand the role of cancer biology in colorectal cancer survival disparities among racial and ethnic minority groups. Objective:To evaluate racial and ethnic differences in KRAS variant frequency and the association of presence of a KRAS variant with colorectal cancer-specific survival. Design, Setting, and Participants:This population-based cross-sectional study used data from the Surveillance, Epidemiology, and End Results Program and included patients diagnosed with colorectal cancer from 2010 through 2015, with follow-up through December 31, 2018. Data were analyzed between December 2023 and August 2024. Exposure:Racial and ethnic differences in KRAS variant frequency. Main Outcomes and Measures:Outcomes of interest were cumulative incidence of colorectal cancer-specific death, assessed using cumulative incidence functions, and subdistribution hazard ratio (sHR) for colorectal cancer-specific death, assessed using Fine-Gray regression models. Results:A total of 21 354 patients (mean [SD] age at diagnosis, 62.54 [13.78] years; 9653 females [45.2%]; median [IQR] follow-up, 2.67 [1.25-4.17] years) were included in the analysis, including 1680 Asian or Pacific Islander patients (7.8%), 2459 Hispanic patients (11.5%), 2761 non-Hispanic Black patients (12.9%), and 14 454 non-Hispanic White patients (67.7%). Hispanic patients and non-Hispanic Black patients had higher KRAS variant frequencies than non-Hispanic Asian or Pacific Islander patients and non-Hispanic White patients (44.2% and 48.3% vs 37.5% and 39.3%, respectively). Among patients with KRAS wild-type tumors, the unadjusted cumulative incidence of colorectal cancer-specific death was highest for Hispanic patients (59.5%; 95% CI, 55.4%-63.3%; P < .001); among patients with KRAS variant tumors, colorectal cancer-specific death was highest for non-Hispanic Black patients (67.3%; 95% CI, 63.3%-70.9%; P < .001). Among patients with KRAS wild-type tumors, Hispanic patients showed a significantly increased risk of colorectal cancer-specific death (sHR, 1.11; 95% CI, 1.01-1.22; P = .03). Among patients with KRAS variant tumors, non-Hispanic Black patients had a significantly increased risk of colorectal cancer-specific death (sHR, 1.18; 95% CI, 1.07-1.29; P < .001). Conclusions and Relevance:In this cross-sectional study of patients with colorectal cancer, Hispanic patients and non-Hispanic Black patients had higher KRAS variant prevalence than non-Hispanic White patients. Among patients with a KRAS variant, non-Hispanic Black patients had worse cause-specific survival than non-Hispanic White patients. Among patients with wild-type KRAS, Hispanic patients had worse survival compared with non-Hispanic White patients. These findings highlight the need for further research on racial and ethnic differences in KRAS-related outcomes.

The development of effective antibody therapeutics has been hampered by a lack of methods to measure drug delivery and activity within tumors at single-cell resolution. Here we introduce single-cell spatial pharmacobiology (SSP), an experimental and analytical framework that integrates in situ imaging of a systemically infused, fluorescently labeled therapeutic antibody with high-plex spatial proteomics to quantify antibody distribution, target engagement and tumor microenvironment (TME) architecture. We applied SSP to tumor tissues from participants with head and neck squamous cell carcinoma and pancreatic ductal adenocarcinoma who received the antibody panitumumab-IRDye800 in phase 1 trials. SSP identified pronounced spatial heterogeneity in single-cell drug delivery and target engagement, shaped by conserved stromal barriers, including periostin-rich extracellular matrix assemblies and fibroblast-activation-protein-positive cancer-associated fibroblast neighborhoods, which were associated with reduced antibody delivery in both tumor types. SSP measures drug-target-TME interactions in human tumors and can support studies of resistance mechanisms, dosing strategies and discovery of spatial biomarkers for precision oncology.

Reliable computational pathology depends on preprocessing methods that identify informative tissue regions while excluding artifacts and low-utility regions from whole-slide images (WSI). However, existing preprocessing pipelines often retain such regions or discard diagnostically relevant tissue, thereby limiting downstream model performance, reliability, and robustness across heterogeneous cohorts. Here, we systematically evaluate how these regions affect downstream AI model performance across multiple clinically relevant applications and introduce MUFASA, a generalizable, information utility-aware preprocessing framework for H E-stained WSI that excludes artifacts and low-utility regions while preserving biologically meaningful tissue. MUFASA integrates slide-level artifact masking, stain-aware tile filtering, reconstruction-based utility stratification of tiles, and targeted recovery of tissue tiles that are over-filtered by earlier phases. Across tumor diagnosis, tumor subtyping, biomarker status prediction, and survival prognostication tasks in diverse cancer cohorts, MUFASA consistently improves downstream model performance relative to widely used preprocessing baselines. These gains are accompanied by reduced artifact-associated attribution in model heatmaps, indicating improved alignment between retained tissue and model attention. Our findings establish WSI preprocessing as a critical determinant of downstream model performance and validity, revealing that even accurate predictions can conceal important failure modes stemming from anatomically implausible reasoning driven by retained artifact-containing and low information-utility tiles.

350 Background: Esophageal (EAC), gastroesophageal junction (GEJ) or gastric (GAC) adenocarcinoma share similar etiologies and treatment approaches. However, it is not clear if they possess similar or different prognostic and biological characteristics. We aimed to examine overall survival (OS) in associations of the common genomic alterations with these three anatomically closely linked adenocarcinomas. Methods: Eligible patients had recurrent or de novo metastatic GAC, GEJAC or EAC whose tumors had undergone next generation sequencing (NGS) performed from November 2017 to December 2023. We used Cox regression modeling to examine the association between GAC, GEJAC, EAC and OS, adjusting for demographics, performance status, Charlson comorbidity index, receipt of chemotherapy, and HER2 overexpression or amplification (HER2 positive), p53 (mutp53 ) , KRAS (mutKRAS), CDKN2A , PIK3CA co-mutations and MYC amplification. Results: Of 875 total eligible patients, 173 had EAC, 276 had GEJAC and 426 had GAC. GEJAC had substantially better OS than EAC (HR = 0.68, [95% CI, 0.54-0.86]), and modestly better OS than GAC (HR = 0.85, [95% CI, 0.67-1.09]). HER2 positivity was associated with substantially better OS among EAC (HR = 0.62; [95% CI, 0.40-0.96]) and GEJAC (HR = 0.59; [95% CI, 0.38-0.87]) but not among GAC (HR = 0.98; [95% CI, 0.78-1.23]) patients. In addition, p53 gain-of-function versus non-gain-of-function mutations were associated with substantially worse OS among GAC (HR = 1.41; [95% CI, 0.98-2.01]) but not among EAC or GEJAC. MutKRAS was associated with substantially worse OS among EAC (HR = 1.81; [95% CI, 1.10-2.99]) but not among GEJAC or GAC. Surprisingly, MYC amplification was associated with dramatically better OS among EAC (HR = 0.19; [95% CI, 0.08-0.42]) but substantially worse OS among GEJAC (HR = 1.98, 95% CI, 1.14-3.40]) and GAC (HR = 1.75; [95% CI, 1.10-2.80]). Conclusions: GEJAC, EAC and GAC possess substantially different OS that appears differentially associated with HER2 status, mutp53, mutKRAS and MYC amplification. These results suggest distinct prognostic and biological characteristics of these three anatomically closely linked adenocarcinomas that could have important implications in clinical practice and on further investigations on their biological and topological mechanisms.

Chemotherapy plus epidermal growth factor receptor (EGFR) inhibitors, such as cetuximab, is standard therapy for KRAS wild-type (KRASwt) colorectal cancer (CRC); however, responses are infrequent. Magrolimab is a monoclonal antibody targeting CD47, an antiphagocytic signal overexpressed in solid tumors (STs). This open-label, multicenter phase 1b/2 study (NCT02953782) aimed to determine the recommended phase 2 dose (RP2D) and evaluate the safety, tolerability, and efficacy of magrolimab + cetuximab in patients with advanced CRC or other STs. A total of 78 patients were enrolled at eight study sites in the USA. In phase 1b, patients with advanced STs received weekly maintenance doses of magrolimab at 10–45 mg/kg and cetuximab at 200–250 mg/m2 following 3 + 3 dose-escalation. In phase 2, patients with anti–EGFR-refractory CRC received magrolimab + cetuximab at RP2Ds. Primary endpoints were dose-limiting toxicities, adverse events, and objective response rate (ORR; phase 2). The maximum tolerated dose was not reached in phase 1b. Two RP2Ds were explored in phase 2: magrolimab at 30 or 45 mg/kg plus cetuximab at 250 mg/m2. Most common treatment-related adverse events (TRAEs) were dermatitis acneiform (35.9

e16457 Background: Patients with inoperable stage II-III pancreatic cancer commonly undergo sequential stereotactic body radiation therapy (SBRT) and chemotherapy. However, there is variability in clinical outcomes, and the predictors of high-risk patients with rapid tumor progression (within 3 months) and poor overall survival (OS) despite treatment are not well characterized. We developed a radiomic (imaging-derived features) signature (RS) that predicts rapid progression. We then investigated clinical characteristics, including pathologic features and treatments received, and RS in a combined analysis to develop and validate the highest-performing model to predict OS from the time of SBRT initiation. We identified the most predictive feature from the combined model that best predicts OS. Methods: In our retrospective study, we examined a cohort of 124 stage II-III pancreatic cancer patients who underwent sequential SBRT and systemic chemotherapy and had pre-treatment pancreatic protocol computed tomography (CT) imaging. We examined 10 clinical features and extracted 900 radiomic features from each segmented tumor per patient using PyRadiomics. Dividing our cohort into training and test sets (60:40), we built a prediction model for rapid tumor progression using radiomic data on the training set (n = 74), applying a LASSO-based algorithm for feature selection and 5-fold cross-validation for parameter optimization. We validated the model performance on the held-out test set (n = 50) and generated the RS for predicting rapid progression. We examined the performance of clinical features and RS in predicting OS in univariate and multivariate Cox regression models. Results: Analysis on our cohort (57 men, 67 women; mean age 67 ± 11 years) generated a 43-feature RS that predicted rapid tumor progression (AUC 0.83, 95% CI: 0.70–0.94) in the test set. High RS was a significant predictor of mortality with hazard ratio (HR) 2.22 (95% CI: 1.32–3.73, p = 0.003) within the first year and 2.85 (95% CI: 1.35–6.03, p = 0.006) thereafter. Non-intensive chemotherapy increased only early mortality risk (HR 1.95, 95% CI: 1.08–3.53, p = 0.03), while older age was significant in later years (HR 1.80, 95% CI: 1.02–3.15, p = 0.04). Conclusions: CT-derived RS accurately predicted rapid tumor progression in stage II-III pancreatic cancer patients undergoing SBRT sequentially with chemotherapy. High RS was the strongest prognostic indicator for increased mortality risk, suggesting its utility for guiding treatment or selection for clinical trials.

BACKGROUND:Advanced esophageal (EAC), gastroesophageal junction (GEJAC) and gastric (GAC) adenocarcinomas with HER2 amplification or overexpression (HER2+) are routinely treated with trastuzumab. However, it remains unclear if HER2+ is associated with superior overall survival (OS). METHODS:The cohort included recurrent or de novo metastatic GAC, GEJAC and EAC from Kaiser Permanente Northern California. We used Cox regression modelling to examine association between HER2+ and OS, adjusting for demographics, performance status, CCI, receipt of chemotherapy and p53 (mutp53), KRAS (mutKRAS), CDKN2A, PIK3CA co-mutations and MYC amplification. RESULTS:Of 875 total eligible patients, 173 had EAC, 276 had GEJAC and 426 had GAC. HER2+ was associated with better OS among the full cohort (HR = 0.74, 95% CI [0.60-0.93]), among EAC (HR = 0.62; [95% CI, 0.40-0.96]) and GEJAC (HR = 0.59; [95% CI, 0.38-0.87]), but not among GAC (HR = 0.89; [95% CI, 0.59-1.35]) patients. GEJAC had better OS than EAC (HR = 0.68, [95% CI, 0.54-0.86]). Trastuzumab treatment was associated with better OS (HR = 0.40, 95% CI [0.21-0.77]). In addition, HER2+ was associated with better OS across the molecular subgroups except that of KRAS mutation (mutKRAS). Our data also show that GEJAC, EAC and GAC were differentially associated with mutp53, mutKRAS and MYC amplification. CONCLUSION:HER2+ and treatment with trastuzumab in HER2+ patients were associated with superior OS in upper gastrointestinal adenocarcinomas across molecular subgroups except that of mutKRAS. These results reaffirm the importance of anti-HER2 treatment in HER2+ patients and provide insight on the prognostic and biological divergence among these anatomically linked upper gastrointestinal adenocarcinomas.

PURPOSE:To examine potential overall survival (OS) differences between males and females with advanced gastric (GAC), gastroesophageal junction (GEJAC) and esophageal (EAC) adenocarcinoma. PATIENTS AND METHODS:The study included patients from Kaiser Permanente Northern California with de novo metastatic or relapsed EAC, GEJAC and GAC. We used Cox regression modeling to examine association of sex with OS adjusting for demographics, performance status, Charlson comorbidity index, histology (Lauren's classification), receipt of chemotherapy, and HER2 amplification or overexpression, p53, KRAS, CDKN2A, PIK3CA co-mutations and MYC amplification. RESULTS:Of 875 total eligible patients, 426 had GAC, of whom 224 were male and 202 were female. Among patients with GAC, males had better OS than females (HR = 0.73; [95% CI, 0.59-0.92]), and this OS difference was preserved across the molecular subgroups except mutKRAS. Intriguingly, among GAC patients with a p53 mutation, males versus females had better OS if tumor carried a non-gain-of-function mutation (non-GOF, HR = 0.59; [95% CI, 0.40-0.85]) but worse OS if tumor carried gain-of-function mutation (GOF, HR = 1.80; [95% CI, 0.83-3.99]). Sex was not associated with OS among patients with GEJAC (HR = 1.14); (95% [CI, 0.77-1.67]) or EAC (HR = 1.0; [95% CI, 0.57-1.74]). These results remained similar when separate analyses were performed among patients who received and among patients who did not receive chemotherapy. CONCLUSIONS:Males had better OS than females among patients with advanced GAC. In addition, among GAC patients with a mutp53, sex and OS association was inversely driven by the presence of GOF versus non-GOF. Our data reveal a previously unappreciated sex disparity in survival outcomes among patients with advanced GAC. If confirmed, this finding could have important implications for clinical practice and for further understanding the biology of GAC.

4019 Background: Inhibitors of Programmed Cell Death Protein-1 (PD-1) have demonstrated remarkable activity in dMMR/MSI-H cancers, leading to the first tissue agnostic FDA approval for an oncologic indication. We report herein results of long-term follow up of KEYNOTE-016, the first study to demonstrate pan-tumor activity of the PD-1 inhibitor pembro in dMMR/MSI-H solid tumors. Methods: KEYNOTE-016 was a multi-center open-label phase 2 study evaluating pembro in patients with advanced colorectal cancer (CRC) (Cohort A) or non-CRC solid tumors (Cohort C) that were dMMR and had progressed after ≥1 prior line of therapy (or ≥2 prior lines for CRC). Eligible patients were age ≥ 18, and had measurable disease per RECIST 1.1. Patients with active CNS metastases, who were on immunosuppressive therapy, had autoimmune disease, or were previously treated with immune checkpoint inhibitors were excluded. Patients received pembro IV every 2 weeks until progression, intolerance, withdrawal of consent or up to a maximum of 2 years. Results: Between 9/2013 - 9/2017, 88 patients (Cohort A: 41; Cohort C: 47) enrolled at 7 sites and received ≥1 dose of pembro. Tumor types enrolled on Cohort C included endometrial (N = 15), pancreatic (N = 9), small intestinal (N = 5), gastroesophageal (N = 5), biliary (N = 4), ampullary (N = 4), and other (N = 5). Median follow up time was 49.7 mos for all patients and 99.8 mos for alive patients. Objective response rate (ORR) was 58% with 23 partial (PR) and 28 complete responses (CR). 16 patients experienced a best response of stable disease (SD) for a disease control rate of 76%. Median PFS and OS were 34.9 mos (95% CI: 14.8-NR) and 80.8 mos (95% CI: 33.2-NR) respectively. The 3-, 5-, and 10-year OS rates were 55.1%, 53.7% and 47.4% respectively. Outcomes were similar between Cohorts A and C (see Table). Conclusions: In summary, long term follow up of KEYNOTE-016 confirms high rates of durable remission from pembro in patients with dMMR/MSI-H solid tumors, with several patients remaining alive and in remission at 10+ years follow up. Responses were seen across tumor types. Clinical trial information: NCT01876511 . Results by cohort. Cohort ACRCN=41 Cohort Cnon-CRC N=47 ORR, % 56.1 59.6 PR, N (%) 11 (27) 12 (25) CR, N (%) 12* (29) 16 (34) SD, N (%) 10 (24) 6 (13) PD, N (%) 5 (12) 9 (19) NE, N (%) 3 (7) 4 (9) PFS, median months (95% CI) 38.8 (8.1-NR) 20.5 (14.3-NR) OS, median months (95% CI) 80.8 (33.2-NR) 86.4 (21.8-NR) Follow up time, median months 51.2 35.9 3-year OS rate (%) 60.5 50.3 5-year OS rate (%) 57.5 50.3 10-year OS rate (%) 47.3 47.2 *Includes 3 patients with unconfirmed CR.

Effective targeting of somatic cancer mutations to enhance the efficacy of cancer immunotherapy requires an individualized approach. Autogene cevumeran is a uridine messenger RNA lipoplex-based individualized neoantigen-specific immunotherapy designed from tumor-specific somatic mutation data obtained from tumor tissue of each individual patient to stimulate T cell responses against up to 20 neoantigens. This ongoing phase 1 study evaluated autogene cevumeran as monotherapy (n = 30) and in combination with atezolizumab (n = 183) in pretreated patients with advanced solid tumors. The primary objective was safety and tolerability; exploratory objectives included evaluation of pharmacokinetics, pharmacodynamics, preliminary antitumor activity and immunogenicity. Non-prespecified interim analysis showed that autogene cevumeran was well tolerated and elicited poly-epitopic neoantigen-specific responses, encompassing CD4+ and/or CD8+ T cells, in 71 NCT03289962 . In this phase 1 trial, patients with locally advanced or metastatic solid tumors were treated with the individualized mRNA neoantigen-specific immunotherapy (iNeST) autogene cevumeran alone or in combination with the anti-PD-L1 agent atezolizumab, showing long-lasting neoantigen-specific immune responses and preliminary clinical activity, supporting further development of this therapeutic approach.

676 Background: Data is lacking to guide the care of vulnerable older adults (OA) with newly diagnosed metastatic pancreatic adenocarcinoma (mPDAC). EA2186 trial demonstrated poor outcomes among vulnerable OA with mPDAC treated with dose-reduced chemotherapy. To understand the factors driving treatment outcomes in this patient population, we analyzed the correlation between baseline geriatric and quality of life (QOL) assessments and treatment outcomes. Methods: Vulnerable OA ≥70 yo with mPDAC, ECOG PS 0-2 were enrolled. Vulnerability was defined by screening geriatric assessment (GA) demonstrating mild abnormalities in function, comorbidities, cognition, or age≥80y. Pts were randomized to Arm A: Gemcitabine (1000mg/m2) + Nab-Paclitaxel (125mg/m2) q14 days or Arm B: 5-Fluorouracil (2400mg/m2 46hr) + Leucovorin (400mg/m2) + Liposomal Irinotecan (50mg/m2) q14 days. GA and QOL evaluations were completed at baseline and 3 time points. Secondary endpoints of the study included evaluating the correlation between baseline GA, QOL and treatment outcomes. Regression models were used to evaluate the associations between baseline GA and QOL factors, survival and grade 3 or higher toxicity, with 80% power to detect doubling in grade 3 toxicity for GA/QOL measures. Results: 176 pts (88 per arm) enrolled with median age 77 (range 70-90), 24% ECOG-0, 64% ECOG-1 and 12% ECOG-2. Pts were deemed vulnerable by cognition (46%), age (36%) or comorbidities (31.4%), with 35% meeting vulnerability criteria in ≥2 domains. No significant difference was seen in median OS (4.7 vs. 4.4 months; p=0.72) or ≥grade 3 toxicity rate (45.6% vs. 58.7%; p=0.10) between arms A and B, respectively. Strong correlation was found between OS and baseline instrumental activities of daily living score (HR 0.84; p=0.02), nutritional scores (HR 0.82; p<0.0001), depression scores (HR 1.07; p=0.02), and scores of all QOL measures (HR 0.98; p<0.0001). No correlation was found between OS and comorbidity, cognition, and Activities of Daily Living scores. After adjustment for age and PS, only baseline WBC level (OR=0.35; p=0.0054), and depression scores (OR=1.20 per score unit; p=0.021) and to a lesser extent FACT-G score (OR=0.98 per score unit; p=0.061) were found to correlate with rates of ≥grade 3 toxicity. Conclusions: Baseline GA and QOL factors among vulnerable older adults with mPDAC correlate strongly with survival and treatment tolerance. Supportive care to address these factors may favorably affect outcomes in this patient population. Clinical trial information: NCT04233866 .

BACKGROUND AND PURPOSE:As patients with rectal cancer with clinical complete response (cCR) after neoadjuvant therapy may be safely spared Total Mesorectal Excision (TME), strategies to maximize cCR are needed. MATERIALS AND METHODS:We conducted a single-arm phase II study to determine whether dose-escalated short-course radiotherapy (25 Gy/5 fractions + 5 Gy/1 fraction boost) followed by eight cycles of FOLFOXIRI increased cCR rates among adult patients with > T2N0M0 or low T2N0 rectal cancer. RESULTS:Between 2020 and 2023, we enrolled 37 patients, of whom 27 (73 %) had at least one high-risk feature (cT4, extramural vascular invasion [EMVI], N2, threatened circumferential resection margin, positive lateral node). At primary endpoint assessment, nine (24 %) patients had cCR on both endoscopy and MRI, and pursued organ preservation (OP). Fourteen (38 %) patients had cCR only on endoscopy, nine of whom pursued OP. Of the 18 patients who pursued OP, nine had local regrowth at two years from radiotherapy start, with two-year TME-free survival of 26 %. Baseline factors significantly associated with not achieving OP included age < 50 years and T4 disease. At mid-treatment restaging, patients who achieved OP were significantly less likely to have persistent node positivity, EMVI, and endoscopically visible tumor. Grade 3+ adverse events at least possibly attributed to chemotherapy and radiotherapy occured in 51% and 43% of patients, respectively. CONCLUSION:Short-course radiotherapy with a boost followed by FOLFIXIRI results in OP in one-quarter of patients with high-risk rectal cancer, with poorer response among younger patients and T4 disease. Mid-treatment response may help guide timely decision-making regarding treatment.

Background The inflamed immune phenotype (IIP), defined by enrichment of tumor-infiltrating lymphocytes (TILs) within intratumoral areas, is a promising tumor-agnostic biomarker of response to immune checkpoint inhibitor (ICI) therapy. However, it is challenging to define the IIP in an objective and reproducible manner during manual histopathologic examination. Here, we investigate artificial intelligence (AI)-based immune phenotypes capable of predicting ICI clinical outcomes in multiple solid tumor types.Methods Lunit SCOPE IO is a deep learning model which determines the immune phenotype of the tumor microenvironment based on TIL analysis. We evaluated the correlation between the IIP and ICI treatment outcomes in terms of objective response rates (ORR), progression-free survival (PFS), and overall survival (OS) in a cohort of 1,806 ICI-treated patients representing over 27 solid tumor types retrospectively collected from multiple institutions.Results We observed an overall IIP prevalence of 35.2% and significantly more favorable ORRs (26.3% vs 15.8%), PFS (median 5.3 vs 3.1 months, HR 0.68, 95% CI 0.61 to 0.76), and OS (median 25.3 vs 13.6 months, HR 0.66, 95% CI 0.57 to 0.75) after ICI therapy in IIP compared with non-IIP patients, respectively (p<0.001 for all comparisons). On subgroup analysis, the IIP was generally prognostic of favorable PFS across major patient subgroups, with the exception of the microsatellite unstable/mismatch repair deficient subgroup.Conclusion The AI-based IIP may represent a practical, affordable, clinically actionable, and tumor-agnostic biomarker prognostic of ICI therapy response across diverse tumor types.

AbstractWhile high circulating tumor DNA (ctDNA) levels are associated with poor survival for multiple cancers, variant-specific differences in the association of ctDNA levels and survival have not been examined. Here we investigate KRAS ctDNA (ctKRAS) variant-specific associations with overall and progression-free survival (OS/PFS) in first-line metastatic pancreatic ductal adenocarcinoma (mPDAC) for patients receiving chemoimmunotherapy (“PRINCE”, NCT03214250), and an independent cohort receiving standard of care (SOC) chemotherapy. For PRINCE, higher baseline plasma levels are associated with worse OS for ctKRAS G12D (log-rank p = 0.0010) but not G12V (p = 0.7101), even with adjustment for clinical covariates. Early, on-therapy clearance of G12D (p = 0.0002), but not G12V (p = 0.4058), strongly associates with OS for PRINCE. Similar results are obtained for the SOC cohort, and for PFS in both cohorts. These results suggest ctKRAS G12D but not G12V as a promising prognostic biomarker for mPDAC and that G12D clearance could also serve as an early biomarker of response.

OBJECTIVE:To examine the optimal method of assessing response to neoadjuvant therapy (NAT) in patients with operable pancreatic ductal adenocarcinoma (PDAC). BACKGROUND:PDAC response to NAT is measured with biochemical, radiographic, and pathologic parameters, which can often be discordant with each other. METHODS:Patients with PDAC undergoing resection after NAT at a single institution were retrospectively analyzed. Tumor response was assessed using pre/post-NAT carbohydrate antigen 19-9 (CA19-9) levels, radiographic decrease in tumor diameter, and pathologic tumor regression grade. The association of these factors with overall survival (OS) was compared using Kaplan-Meier, Cox regression, and recursive partitioning analysis, a machine learning technique that can validate prediction models for complex hierarchical relationships. RESULTS:From 2011 to 2022, 225 patients underwent pancreatectomy after NAT (Folfirinox, 70%; gemcitabine + nab-paclitaxel, 19%; radiation, 18%). Almost half required vascular resection (portal vein, 39%; celiac axis 8%). Improved OS was observed after CA19-9 decrease >50% (32 vs 24 months, P = 0.0028), but not after major pathologic (tumor regression grade: 0-1, P = 0.067) or radiographic response (tumor diameter decrease >30%, P = 0.89). However, recursive partitioning analysis identified that the coexistence of biochemical and major pathologic response (achieved in 9% of patients) was associated with the longest OS (40 months, P = 0.0086). This optimal dual response combination was more commonly observed after neoadjuvant radiotherapy was used after systemic chemotherapy (45% vs 11%, P < 0.001). CONCLUSIONS:CA19-9 response to NAT alone is not enough to identify long-term postresection PDAC survivors. The coexistence of CA19-9 and major pathologic response was predictive of the most optimal survival outcome.

Introduction:Learning health networks (LHNs) improve clinical outcomes by applying core tenets of continuous quality improvements (QI) to reach community-defined outcomes, data-sharing, and empowered interdisciplinary teams including patients and caregivers. LHNs provide an ideal environment for the rapid adoption of evidence-based guidelines and translation of research and best practices at scale. When an LHN is established, it is critical to understand the needs of all stakeholders. To accomplish this, we used ethnographic methods to develop personas of different stakeholders within The Canopy Cancer Collective, the first oncology LHN. Methods:We partnered with a firm experienced in qualitative research and human-centered design to conduct interviews with stakeholders of The Canopy Cancer Collective, a newly developed pancreatic cancer LHN. Together with the firm, we developed a personas model approach to represent the wide range of diverse perspectives among the representative stakeholders, which included care team members, patients, and caregivers. Results:Thirty-one stakeholders from all facets of pancreatic cancer care were interviewed, including 20 care team members, 8 patients, and 3 caregivers. Interview transcripts were analyzed to construct 10 personas felt to represent the broad spectrum of stakeholders within The Cancer Canopy Collective. These personas were used as a foundation for the design and development of The Cancer Canopy Cancer Collective key drivers and aims. Conclusions:As LHNs continue to facilitate comprehensive approaches to patient-centered care, interdisciplinary teams who understand each other's needs can improve Network unity and cohesion. We present the first model utilizing personas for LHNs, demonstrating this framework holds significant promise for further study. If validated, such an approach could be used as a dynamic foundation for understanding individual stakeholder needs in similar LHN ecosystems in the future.

74 Background: We identified computed tomography (CT)-derived radiomic features predictive of tumor progression within three months, then examined their ability to prognosticate overall survival (OS) along with clinical features in pancreatic cancer. We evaluated these features in patients with unresected pancreatic cancer who underwent stereotactic body radiation therapy (SBRT) in sequence with chemotherapy, but not surgery. Methods: In this retrospective study, we examined a cohort of 101 patients with stage II-III pancreatic cancer who underwent SBRT with sequential chemotherapy at a single institution (Stanford Health Care) between 1999-2020. From their pre-SBRT contrast-enhanced CT images with segmented tumors, delineating regions-of-interest, we extracted 900 radiomic (quantitative pixel-level imaging characteristic) features. In the first phase, we identified radiomic features that predicted rapid tumor progression within three months following SBRT. We divided the dataset into a training set (n = 53) for model development and a test set (n = 48) for evaluation. Using logistic regression with the Least Absolute Shrinkage and Selection Operator algorithm for feature selection and classification, we built a binary prediction model on the training set to identify patients at risk of progression within three months of SBRT. To fine-tune parameters, we performed five-fold cross-validation (CV) on the training set, repeating each set of parameters five times. We assessed model performance on the test set using the area under the curve (AUC). We selected the model with the best AUC, generating the predictive radiomic feature set. In the second phase, we conducted univariate and multivariate Cox regression analyses to assess the relationship between OS and individual clinical variables (age, sex, stage, vessel involvement, tumor location, performance status, body mass index, biological equivalent dose of radiation) and the radiomic feature set as high versus low risk. Results: Our cohort consisted of 48 men (mean age, 70 years ± 11 [SD]) and 53 women (mean age, 67 years ± 13 [SD]). From the first phase, 32 textural features comprised the radiomic feature set that best predicted rapid tumor progression, with mean AUCs of 0.852 (CV, n=53) and 0.814 (test, n=48). In the univariate Cox model, only the radiomic feature set was predictive of OS (hazard ratio, HR, 1.724, p=0.011). In the multivariate Cox model, radiomic features and age were significant predictors of OS, with HR of 1.819 (p=0.007) and 1.024 (p=0.024), respectively. Conclusions: CT-derived radiomic features predict rapid tumor progression following SBRT, confer nearly a twofold increase in mortality risk, and, along with patient age, enhance the identification of patients with stage II-III pancreatic cancer with poor OS. [Table: see text]