Citations

INTRODUCTION: Reflux bile acids are believed to promote esophageal adenocarcinoma (EAC), but the role of systemic bile acids is unknown. This study aimed to assess associations between systemic bile acids and stages of Barrett's esophagus (BE) progression. METHODS: Subjects with and without BE were enrolled in this multicenter cross-sectional study. Targeted serum bile acid profiling was performed, and a subset of subjects completed a validated food frequency questionnaire. RNA sequencing was performed on BE or gastric cardia tissue to assess bile acid associations with gene expression. RESULTS: A total of 141 subjects were enrolled with serum bile acids profiled (49 non-BE; 92 BE: 44 no dysplasia, 25 indefinite/low grade dysplasia, 23 high-grade dysplasia/EAC). Lower Healthy Eating Index score, older age, higher body mass index, and no proton pump inhibitor use were associated with increased levels of multiple bile acids. Global bile acid pools were distinct between non-BE and stages of BE neoplasia ( P = 0.004). Increasing cholic acid was associated with high-grade dysplasia/EAC compared with non-BE, even after adjusting for EAC risk factors (adjusted odds ratio 2.03, 95% confidence interval 1.11–3.71) as was the combination of unconjugated primary bile acids (adjusted odds ratio 1.81, 95% confidence interval 1.04–3.13). High cholic acid levels were associated with tissue gene expression changes including increased DNA replication and reduced lymphocyte differentiation genes. DISCUSSION: Alterations in serum bile acids are independently associated with advanced neoplasia in BE and may contribute to neoplastic progression. Future studies should explore associated gut microbiome changes, proneoplastic effects of bile acids, and whether these bile acids, particularly cholic acid, represent potential biomarkers or viable therapeutic targets for advanced neoplasia in BE.

AbstractBackgroundAdaptive radiotherapy (ART) workflows have been increasingly adopted to achieve dose escalation and tissue sparing under shifting anatomic conditions, but the necessity of recontouring and the associated time burden hinders a real‐time or online ART workflow. In response to this challenge, approaches to auto‐segmentation involving deformable image registration, atlas‐based segmentation, and deep learning‐based segmentation (DLS) have been developed. Despite the particular promise shown by DLS methods, implementing these approaches in a clinical setting remains a challenge, namely due to the difficulty of curating a data set of sufficient size and quality so as to achieve generalizability in a trained model.PurposeTo address this challenge, we have developed an intentional deep overfit learning (IDOL) framework tailored to the auto‐segmentation task. However, certain limitations were identified, particularly the insufficiency of the personalized dataset to effectively overfit the model. In this study, we introduce a personalized hyperspace learning (PHL)‐IDOL segmentation framework capable of generating datasets that induce the model to overfit specific patient characteristics for medical image segmentation.MethodsThe PHL‐IDOL model is trained in two stages. In the first, a conventional, general model is trained with a diverse set of patient data (n = 100 patients) consisting of CT images and clinical contours. Following this, the general model is tuned with a data set consisting of two components: (a) selection of a subset of the patient data (m < n) using the similarity metrics (mean square error (MSE), peak signal‐to‐noise ratio (PSNR), structural similarity index (SSIM), and the universal quality image index (UQI) values); (b) adjust the CT and the clinical contours using a deformed vector generated from the reference patient and the selected patients using (a). After training, the general model, the continual model, the conventional IDOL model, and the proposed PHL‐IDOL model were evaluated using the volumetric dice similarity coefficient (VDSC) and the Hausdorff distance 95% (HD95%) computed for 18 structures in 20 test patients.ResultsImplementing the PHL‐IDOL framework resulted in improved segmentation performance for each patient. The Dice scores increased from 0.810.05 with the general model, 0.83 for the continual model, 0.83 for the conventional IDOL model to an average of 0.87 with the PHL‐IDOL model. Similarly, the Hausdorff distance decreased from 3.06 with the general model, 2.84 for the continual model, 2.79 for the conventional IDOL model and 2.36 for the PHL‐IDOL model. All the standard deviations were decreased by nearly half of the values comparing the general model and the PHL‐IDOL model.ConclusionThe PHL‐IDOL framework applied to the auto‐segmentation task achieves improved performance compared to the general DLS approach, demonstrating the promise of leveraging patient‐specific prior information in a task central to online ART workflows.

84 Background: Local therapies potentially offer a curative approach for patients with oligometastatic colorectal cancer (CRC). The effectiveness of systemic chemotherapy following definitive local treatment in this setting is not well-defined. Tumor-informed circulating tumor DNA (ctDNA) might guide management strategies after curative-intent local treatment. Methods: This is a single institution retrospective study of patients with CRC who underwent curative-intent local therapy to an isolated site of metastatic disease and had post-intervention tumor-informed ctDNA (Signatera) testing. The study was approved by the institutional review board. Clinical characteristics, including ctDNA results, were collected and evaluated. Kaplan-Meier method and log-rank test were used to calculate and compare disease-free survival (DFS). Factors associated with DFS were analyzed using multivariate Cox proportional hazards model. Results: 45 patients were included with median age of 59 years, 62% being male, 91% white, and 53% with stage IV disease at diagnosis. 37% of the patients had RAS mutations, no patient had BRAF mutation, and 2.2% patient had mismatch repair deficient disease. Isolated liver metastases were present in 80% of patients, lung metastases in 11% and other sites in 9%. 44 patients received chemotherapy prior to definitive treatment, with a median treatment duration of 16 weeks. ctDNA tests were done prior to local therapy in 23 patients, with 14 cases (61%) testing positive. Definitive treatment included resection (58%), ablation (13%), radiotherapy (4%), and multimodality (25%). Post-definitive treatment chemotherapy was administered to 24 (53%) patients. ctDNA positivity following definitive therapy was observed in 10 patients (22%), which was associated with a poorer prognosis, with a median DFS of 5.3 months compared to 21.3 months for the ctDNA-negative group (p<0.001). These findings were confirmed on multivariate analysis. In patients who had negative ctDNA after definitive local treatment, median DFS was 14.9 months for those who received post-intervention chemotherapy versus 21.3 months for those without post-intervention chemotherapy (p=0.835). Conclusions: Patients with negative ctDNA results following definitive local therapy for isolated metastatic CRC have better prognosis. Post-intervention chemotherapy in this group of patients was not associated with improved DFS. ctDNA guided omission of post-intervention chemotherapy following curative-intent local treatment of oligometastatic CRC warrants further study. [Table: see text]

50 Background: A+B is commonly used as first line systemic therapy in HCC patients with CP-B liver function, although in the IMbrave150 clinical trial, its efficacy was only assessed in patients with CP-A liver function. We evaluated outcomes in patients with CP-B liver function and compared to those with CP-A. Methods: We completed a multi-institution, retrospective review of patients with HCC from Mayo Clinic and University Hospitals Seidman Cancer Center, who had received A+B in the first line setting between March 2018 and November 2023. CP score was calculated at time of treatment, response was assessed, and progression free (PFS) and overall (OS) survival were determined using the Kaplan Meier method. Results: Among 322 patients with HCC, median age was 66.5 years, 78.6% were male, and 82.6% were white. Eighty-six patients (27%) had CP-B liver function, while 226 (70%) had CP-A and 10 (3%) had CP-C. Etiology of liver disease in patients with CP-B included hepatitis B in 7.0%, hepatitis C in 38.4%, alcohol use in 30.2%, and metabolic associated steatohepatitis in 20.9%. Median OS in patients with CP-B liver function was 6.4 months, while mOS was 21.6 months for those with CP-A ( p<0.0001). Further stratifying, mOS in those with CP-B7 was 9.1 months, CP-B8 was 5.5 months, and CP-B9 was 4.0 months. 12-month OS in patients with CP-B was 32.4%, compared to 69.3% in CP-A. Median PFS was 4.8 months in those with CP-B and 8.9 months in those with CP-A. Median PFS in those with CP-B7, -B8, and -B9 was 6.4 months, 2.7 months, and 2.7 months, respectively. Among those with CP-B, 1 patient had a complete response, 17 (19.8%) had a partial response, 29 (33.7%) had stable disease, and 37 (43.0%) had progressive disease. Adverse events in CP-B patients included hypertension in 14.0% (grade 3 in 1 patient), bleeding in 16.3% (grade 3 or greater in 11.6%), fatigue in 38.4%, and an immune-related adverse event in 17.4% (grade 3 in 5.8%). Conclusions: Patients with advanced HCC and CP-B7 liver function may still benefit from A+B. Importantly, no additional safety risks were identified in this subgroup. Those with CP-B8 and CP-B9 have significantly worse survival than those with CP-B7 and the likelihood of benefit from systemic therapy is minimal. It is important for providers to address the impact of baseline liver function on expected outcomes in patient with HCC.

Abstract Background Atezolizumab plus bevacizumab is the standard of care for advanced hepatocellular carcinoma (HCC) in the first-line setting, although was only evaluated in patients with Child-Pugh (CP) A liver function in the IMbrave150 trial. We sought to determine the outcomes of these patients based on CP score and ALBI grade in the US population. Methods This multicenter cohort study included patients with HCC who received atezolizumab with bevacizumab as first-line systemic therapy between March 2018 and November 2023. Overall survival (OS) was determined using the Kaplan-Meier method and multivariate analyses were performed using Cox proportional hazard regression method. Results Among 322 patients, 226, 86, and 10 patients had CP-A, CP-B, and CP-C liver function, respectively. Median age was 66.5 years, 78.6% were male, and 82.6% were White. Median OS (mOS) was 21.6 months for those with CP-A, 9.1 months for those with CP-B7, and 4.7 months for those with CP-B8-C12 (P &lt; .0001). Among patients with CP-A, those with ALBI grade 1 had an mOS of 34.9 months versus 14.2 months in those with grade 2. In multivariate analyses, CP score, ALBI grade, hepatitis B, performance status, and macrovascular invasion were significantly associated with survival. Conclusions CP score is an important prognostic tool for patients with HCC receiving atezolizumab plus bevacizumab, and this regimen remains a viable option for patients with CP-B7 with no additional safety concern, although the benefit is significantly less than those with CP-A. ALBI score has independent predictive value in patients with CP-A liver function.

Women with type 2 diabetes (T2D) have a higher risk of being diagnosed with breast cancer and have worse survival than non-diabetic women if they do develop breast cancer. However, more research is needed to elucidate the biological underpinnings of these relationships. Here, we found that forkhead box A1 (FOXA1), a forkhead family transcription factor, and metformin (1,1-dimethylbiguanide hydrochloride), a medication used to treat T2D, may impact hormone-receptor-positive (HR+) breast cancer (BC) tumor cell growth and metastasis. Indeed, fourteen diabetes-associated genes are highly expressed in only three HR+ breast cancer cell lines but not the other subtypes utilizing a 53,805 gene database obtained from NCBI GEO. Among the diabetes-related genes, FOXA1, MTA3, PAK4, FGFR3, and KIF22 were highly expressed in HR+ breast cancer from 4032 breast cancer patient tissue samples using the Breast Cancer Gene Expression Omnibus. Notably, elevated FOXA1 expression correlated with poorer overall survival in patients with estrogen-receptor-positive/progesterone-receptor-positive (ER+/PR+) breast cancer. Furthermore, experiments demonstrated that loss of the FOXA1 gene inhibited tumor proliferation and invasion in vitro using MCF-7 and T47D HR+ breast cancer cell lines. Metformin, an anti-diabetic medication, significantly suppressed tumor cell growth in MCF-7 cells. Additionally, either metformin treatment or FOXA1 gene deletion enhanced tamoxifen-induced tumor growth inhibition in HR+ breast cancer cell lines within an ex vivo three-dimensional (3D) organoid model. Therefore, the diabetes-related medicine metformin and FOXA1 gene inhibition might be a new treatment for patients with HR+ breast cancer when combined with tamoxifen, an endocrine therapy.

3579 Background: Fruquintinib (F) is a highly selective, oral tyrosine kinase inhibitor of all 3 VEGF receptors. Based on FRESCO (NCT02314819) and FRESCO-2 (NCT04322539) data, F was approved by the US FDA for previously treatedmetastatic colorectal cancer (mCRC). Efficacy and safety of F + best supportive care (BSC) vs placebo (P) + BSC were analyzed according to prior treatment (tx) sequence using FRESCO and FRESCO-2 data. Methods: In FRESCO and FRESCO-2, patients (pts) were randomized 2:1 to receive F 5 mg or matching P, PO daily x 21 days every 28 days. In FRESCO, eligible pts had received ≥2 prior chemotherapy regimens and were permitted to have received prior anti-VEGF and/or anti-EGFR therapies, whereas in FRESCO-2, in addition to standard chemotherapies, anti-VEGF, and anti-EGFR therapies if indicated, prior exposure to trifluridine/tipiracil (TAS) and/or regorafenib (rego) was required. FRESCO-2 pts must have also received a BRAF inhibitor and/or immunotherapy, if indicated and available in their country. Results: In FRESCO (F n=278 vs P n=138), no pts had received rego and 16/416 pts (4%) had received TAS. InFRESCO-2, prior therapies were balanced between F vs P arms; pts had received prior TAS or rego in these sequences: prior TAS only (n=240 vs n=121); prior rego only (n=40 vs n=18); prior rego then TAS (n=84 vs n=45); prior TAS then rego (n=95 vs n=44). Improvements in OS and PFS were seen with F vs P regardless of prior tx sequence (Table), with similar magnitudes of benefit for OS (HRs 0.53–0.77) and PFS (HRs 0.25–0.37). F safety profile was consistent regardless of prior tx sequence (Table). Conclusions: F is effective in mCRC with a manageable safety profile regardless of prior TAS or rego tx and sequence. Improvement with F was similar for OS and PFS despite the number and sequence of prior regimens. F provides a new tx option for pts with previously treated mCRC regardless of prior TAS or rego exposure. Clinical trial information: NCT04322539 . [Table: see text]

(1) Background: Local therapies offer a potentially curative approach for patients with oligometastatic colorectal cancer (CRC). An evidence-based consensus recommendation for systemic therapy following definitive locoregional therapy is lacking. Tumor-informed circulating tumor DNA (ctDNA) might provide information to help guide management in this setting. (2) Methods: A multi-institutional retrospective study was conducted, including patients with CRC that underwent curative-intent locoregional therapy to an isolated site of metastatic disease, followed by tumor-informed ctDNA assessment. The Kaplan–Meier method and log-rank tests were used to compare disease-free survival based on ctDNA results. ctDNA test performance was compared to carcinoembryonic antigen (CEA) test results using McNemar’s test. (3) Results: Our study cohort consisted of 87 patients treated with locoregional interventions who underwent ctDNA testing. The initial ctDNA test post-intervention was positive in 28 patients and negative in 59 patients. The median follow-up time was 14.0 months. Detectable ctDNA post-intervention was significantly associated with early disease recurrence, with a median disease-free survival (DFS) of 6.63 months compared to 21.30 months in ctDNA-negative patients (p < 0.001). ctDNA detected a numerically higher proportion of recurrences than CEA (p < 0.097). Post-intervention systemic therapy was not associated with improved DFS (p = 0.745). (4) Conclusions: ctDNA results are prognostically important in oligometastatic CRC, and further prospective studies are urgently needed to define its role in guiding clinical decisions.