Abstract DDR1 (Discoidin domain receptor 1) is a collagen-activated receptor that facilitates tumor growth by orchestrating extracellular matrix (ECM) remodeling, pro-survival signaling, epithelial-mesenchymal transition (EMT), metastasis, and immune exclusion. Interactions between collagen and DDR1 generate a tightly organized matrix that restricts T-cell infiltration and promotes CAF-mediated desmoplasia, leading to immune-cold tumor microenvironments. Inhibition of DDR1 remodels the extracellular matrix, suppresses AKT/MAPK/NF-κB signaling pathways, eliminates cancer stemness, and disrupts tumor–stroma interactions, thereby enhancing immune infiltration and significantly improving the efficacy of PD-1/PD-L1 blockade. Collectively, these features establish DDR1 as an ideal therapeutic target in ECM-abundant malignancies, including PDAC, TNBC, and NSCLC. Employing a non-kinase scaffold-hopping approach combined with fragment-based drug discovery, we developed a selective series of Type II, DFG-out state-specific DDR1 inhibitors (CVD series). Lead optimization generated sub-10 nM compounds with high potency and excellent selectivity. In collaboration with structural biologists, we solved DDR1-CVD co-crystal structures and elucidated their binding mechanism using synchrotron X-ray crystallography. Our inhibitors act as Type II, DFG-out state-specific binders, occupying a previously unexploited pocket in DDR1. This poster describes the discovery, structural validation, and optimization of a novel, lead-stage selective DDR1 inhibitor. Citation Format: Jungwook Chin, Sun Jun Park, Su-Jeong Lee, Jong-Hyun Park, Jun Young Hong. Discovery of highly selective, orally available small-molecule DDR1 inhibitors that reprogram the collagen-DDR1 tumor microenvironment [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 2 (Late-Breaking, Clinical Trial, and Invited Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(8_Suppl):Abstract nr LB351.

Abstract Patient-derived tumor organoids have emerged as a promising translational model for oncology drug development. However, its utility has been limited by restricted access to patient-derived materials and insufficient donor-related data such as clinical records. Here, we report the establishment of the Samsung Organoids, a GxP-based, standardized patient-derived organoid designed to support data-driven cancer drug development. All organoids were generated and processed by GxP-qualified personnel under rigorously controlled and fully documented procedures, ensuring reproducibility, traceability, and regulatory readiness. Each organoid line is linked to clinical information of the corresponding donor and multi-omics data both original tumor tissues and derived organoids. Comparative analyses demonstrated high concordance of mutational profiles and gene expression patterns between tumor tissues and matched organoids, confirming preservation of the patient-specific characteristics. To examine the responsiveness of organoids to various drugs including small molecules and ADCs, we performed high-throughput drug screening combined with high-content imaging. Drug responsiveness data revealed substantial inter-patient heterogeneity, enabling classification of organoids into distinct groups. Further, we identified particular gene expression signatures by cross-group comparison, demonstrating explanation of different drug sensitivity between organoids. Conclusively, by integrating patient clinical information with genomic/genetic alterations, transcriptomic signatures, and functional drug response data, the Samsung Organoids provides translational insights into determinants of therapeutic sensitivity and resistance, helping successful cancer drug development such as drug candidate selection, biomarker discovery, and preclinical decision-making. Citation Format: Minhyung Lee, Sekyu Oh, Seongju Jeong, Sung Min Ha, Chihah Moon, Nayoun Choi, Yoonhyeok Kwon, Seahee Kim, Sangmyung Lee, Brian Hosung Min. Samsung Organoids: A GxP-based drug screening leveraging patient-derived tumor organoids platform accompanied with clinical records and multi-omics data [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 2 (Late-Breaking, Clinical Trial, and Invited Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(8_Suppl):Abstract nr LB397.

Abstract Background: Postoperative driver gene-specific circulating tumor DNA (ctDNA) monitoring stratifies recurrence risk in resectable stage III driver-mutant NSCLC (NCT06443684). Whether driver-specific ctDNA MRD can evaluate neoadjuvant response and predict recurrence in stage III driver-mutant NSCLC remains to be defined. Methods: From 2022 to 2025, 203 resectable stage III NSCLC patients with tissue-confirmed driver mutations were prospectively enrolled at 14 centers in China (NCT06443684). Peripheral blood was collected before neoadjuvant therapy, preoperatively, 3 days and 1 month after surgery, and every 3 months until investigator-confirmed recurrence. ctDNA was analyzed using a 10-gene driver panel (LC10). This analysis focused on peri-neoadjuvant ctDNA dynamics and clinical outcomes among 82 patients who received neoadjuvant therapy followed by surgery. Results: Among the 82 neoadjuvant-treated patients, regimens included chemo-immunotherapy (57.3%), TKI (3rd-EGFR TKI/2nd or 3rd-ALK TKI, 39.0%), and chemotherapy alone (3.7%). Post-neoadjuvant ctDNA positivity was associated with a significantly lower MPR rate versus ctDNA negativity (0% vs 42.3%, p=0.004), whereas baseline ctDNA status was not associated with MPR (16.7% vs 21.2%). Post-neoadjuvant ctDNA molecular residual level (hGE/mL) positively correlated with residual viable tumor cells (R=0.332, p=0.009). Post-neoadjuvant ctDNA-positive patients had significantly shorter event-free survival (median EFS 11.4 vs 23.8 months; HR=8.82; p=0.009). For non-MPR patients, post-neoadjuvant negative ctDNA status also indicated favorable prognosis (median EFS 23.8 vs 4.7 months; HR=5.80; p=0.054). Compared with other regimens, neoadjuvant TKI was associated with a lower post-treatment ctDNA residual rate (0% vs 15.3%; p=0.05). Patients with persistent ctDNA positivity (baseline→post-neoadjuvant) had the shortest EFS (median 4.7 months), followed by negative→positive (23.8 months), while positive to negative and persistent negative ctDNA indicated disease free status(log-rank p<0.001). Conclusions: A driver-specific ctDNA MRD strategy provides a practical tool for neoadjuvant efficacy assessment and postoperative recurrence risk stratification for stage III driver-mutant NSCLC patients. Citation Format: Jian Hu, Xiao Teng, Ziming Li, Yu Qi, Feng Li, Qixun Chen, Changbin Zhu, Xing Li, Shun Lu. Driver-specific MRD strategy associated with pathological response and predicted recurrence in stage III driver-mutant NSCLC receiving neoadjuvant treatment [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 2 (Late-Breaking, Clinical Trial, and Invited Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(8_Suppl):Abstract nr LB120.

Abstract Introduction: V-domain Ig Suppressor of T-cell Activation (VISTA) is an immune checkpoint regulator found on tumor, myeloid, and other immune cells. Its presence has been shown to enhance tumor growth, create an immunosuppressive microenvironment and may potentially contribute to resistance to anti-CTLA-4 and anti-PD-1/PD-L1 therapies. Therefore, VISTA represents a promising therapeutic target. HMBD-002, a non-depleting, high-affinity IgG4 monoclonal antibody against VISTA, has demonstrated significant inhibition of tumor growth in preclinical studies, both as a monotherapy and in combination with pembrolizumab. HMBD-002 is intended to increase T-cell activity and reprogram the suppressive tumor microenvironment to a proinflammatory antitumor phenotype. Cancer types including triple-negative breast cancer (TNBC) and non-small cell lung cancer (NSCLC) exhibit high expression levels of VISTA in the TME providing a rational basis for exploring these solid tumor indications. Methods: Phase 1, first-in-human, open-label, study evaluating multiple repeat doses of intravenously (IV) administered HMBD-002, with or without pembrolizumab, in participants with advanced solid tumors (i.e., locally advanced and unresectable, or metastatic). A standard 3+3 design was utilized. Dose escalation was initiated with single-agent HMBD-002. After completion of the first 4 cohorts as monotherapy HMBD-002 dose escalation, HMBD-002 was dose escalated in combination with fixed dose pembrolizumab (200 mg IV Q3W). HMBD-002 was administered as a single weekly dose, with or without pembrolizumab, in 21-day treatment cycles. Primary endpoints were safety and tolerability, maximum tolerated dose (MTD) and recommended phase 2 dose (RP2D). Secondary endpoints included pharmacokinetics and preliminary antitumor activity. Results: A total of 48 subjects were enrolled, 28 receiving HMBD-002 as monotherapy at doses ranging from 20mg to 1400mg. Dose escalation in combination with pembrolizumab (20 subjects) ranged from 180mg to 1400mg. 18 subjects (37.5%) experienced a treatment related AE with 3 subjects having a grade 3 or greater TRAE. The most common TRAEs were fatigue, rash and nausea. 1 dose limiting toxicity, immune related hepatitis, was seen at 360mg monotherapy and resolved with corticosteroids after drug was held. No cytokine release syndrome (CRS) was observed. No MTD was declared. PK data indicates a RP2D of 720mg QW. Best observed response was stable disease (SD) in 5/28 (18%) monotherapy subjects and 5/20 (25%) combination therapy subjects. A subset of subjects experienced SD for 6 cycles or greater, including subjects with NSCLC and TNBC. Conclusions: HMBD-002 demonstrated preliminary safety and tolerability in this phase 1 study. A RP2D of 720mg QW was determined. Further investigation in tumor specific phase 2 cohorts is warranted and planned for 2026. Citation Format: Jordi Ahnert Rodon, Joshua J. Gruber, Melinda L. Telli, Andrew Hendifar, Joseph W. Kim, Sharonlin Bhardwaj, Padmanee Sharma, Dipti Thakkar, Jerome Boyd-Kirkup, Eugene Kennedy. Results of phase 1 first-in-human clinical trial of HMBD-002, an IgG4 monoclonal antibody targeting VISTA, in advanced solid tumors [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 2 (Late-Breaking, Clinical Trial, and Invited Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(8_Suppl):Abstract nr CT126.

Abstract Cancer arises within a complex cellular milieu, where the extent and composition of tumor-infiltrating immune cells strongly influence tumor progression and patient prognosis. Emerging research highlights the therapeutic potential of fasting mimicking diets (FMDs) in delaying cancer onset, providing cellular protection and regulating immunity. My project investigates immune profiles, before and after the transfer of donor-derived immune cells previously exposed to fasting, either independently or combined with therapeutic drugs. We aim to identify immune populations, enhanced by this combined treatment, and to recognize their impact on tumor advancement and immune activation. Using the syngeneic 4T1, triple negative breast cancer mouse model, we evaluated how fasting and chemotherapy affect tumor growth and immune competence. Flow cytometry and immunohistochemistry were used to characterize immune cells, recruited at tumors, spleens and bone marrows of treated hosts. A subpopulation of preconditioned splenocytes and tumor-infiltrating lymphocytes, recognized through immunophenotyping, were transplanted into tumor-bearing recipients to assess tumor control, immune priming and cytotoxicity. We observed reduced tumor volume, delayed progression and increased survival rate in the transplants exposed to FMD and doxorubicin. The treatment preserved the size, morphology, and cellularity of healthy, primary and secondary lymphoid organs, while expanded beneficial, antitumor immune signatures. Integrating FMD with standard therapeutic strategies could enhance antitumor immunity and broaden the spectrum of malignancies that respond effectively. Exploring the underlying mechanisms may reveal immune pathways that can be therapeutically leveraged to eradicate tumors, in a field that currently remains unexplored. Citation Format: Evangelia Pavlou, Olga Blazevits, Giulia Salvadori, Sara Martone, Valter Longo. Integrating a fasting mimicking diet to augment antitumor immunity following adoptive cell transfer in breast cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 2 (Late-Breaking, Clinical Trial, and Invited Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(8_Suppl):Abstract nr LB146.

Abstract Accurate identification of somatic mutations is crucial to early diagnosis of cancers and optimal personalized treatments for cancer patients. However, current somatic mutation detection mostly relies on inference from standard reference genome (such as GRCh38) based read alignments, leading to some reference related false positives or false negatives. Genome-in-a-Bottle (GIAB) consortium has publicly released a wide variety of sequencing data, including multiple short- and long-read technologies, Hi-C, and optical mapping for a new broadly-consented tumor-normal paired samples derived from the same individual (HG008), aiming for development of high-quality somatic mutation benchmarks to advance the development of sequencing technologies and analytical methods. Leveraging the latest advancements in long-read technologies and assembly algorithms, we have used the above data to generate and subsequently curate near-complete chromosomal-scale haplotype-resolved assemblies for both normal tissue and tumor cell lines, thus enabling the direct comparison of the tumor haplotype genome to its corresponding normal haplotype genome for accurate somatic mutation detection. Here we report that we implemented an integrated analysis workflow for accurate detection of haplotype-specific somatic structural variations in paired tumor-normal samples by adopting genetic marker identification, haplotype assembly-to-haplotype assembly mapping (SyRI, svim-asm, PAV) and read-alignment approaches (Severus, savana, Sniffles2 etc). We first identified high-quality haplotype-specific genetic markers so that each of the chromosomes from two sets of haplotype assemblies in the tumor sample could be matched with their corresponding haplotype chromosomes in the normal sample. These genetic markers also allowed us to explore and identify genome-wide recombination events in this cancer sample, such as a complex series of inverted duplications on chr19 that is attached to chr22 on one end and to the opposite haplotype of chr19 on the other end. We then generated two comprehensive somatic SV sets by incorporating multiple lines of evidence from matched tumor-normal haplotypes and long/short read mappings (including PacBio HiFi, ONT, and Illumina), when two haplotype-resolved assemblies from normal sample were used as reference. We curated each of the selected SVs using IGV and compared them with GIAB’s curated draft somatic SV benchmark (v0.4, GRCh38-based). While most somatic SVs were identical between the two callsets, our assembly-based approach was complementary to the mapping-based benchmark in certain repetitive and complex regions. Finally, we assigned each of the somatic SVs to its correct haplotype and generated final haplotype-specific somatic structural variation callset. At present, we confirm 32 insertions and 62 deletions (greater than 50 bps), and 10 translocation and inversion breakpoints that do not have clear GRCh38 coordinates, mostly in centromeric satellite regions. While we continue to improve the haplotype-based somatic SVs, these analyses are already providing more insights into the ongoing refinement of high-quality somatic SV benchmarks for this tumor-normal pair. Citation Format: Chunlin Xiao, Justin Wagner, Jennifer McDaniel, Francoise Thibaud-Nissen, Justin Zook. Marker-based tumor-normal haplotypes matching reveals genome-wide recombination and haplotype specific somatic structural variation discovery using GIAB de novo pancreatic tumor-normal assemblies [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 2 (Late-Breaking, Clinical Trial, and Invited Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(8_Suppl):Abstract nr LB104.

Abstract Glioblastomas are incurable primary brain tumors that depend on neural-like cellular processes, tumor microtubes (TM), to invade the brain. TMs also interconnect single tumor cells to a communicating multicellular network that resists current therapies. In this study, we developed a combined, comprehensive in vitro/in vivo anti-TM drug screening approach, including machine learning-based analysis tools. Two protein kinase C (PKC) modulators robustly inhibited TM formation and pacemaker tumor cell-driven, TM-mediated glioblastoma cell network communication. As TM-unconnected tumor cells exhibited increased sensitivity to cytotoxic therapy, the PKC activator TPPB was combined with radiotherapy, and long-term intravital two-photon microscopy paired with spatially resolved multiomics revealed anti-TM and antitumor effects. TPPB treatment also decreased the expression of tweety family member 1 (TTYH1), a key driver of invasive TMs. Our study establishes a novel screening pipeline for anti-TM drug development, identifies a TM master regulator pathway, and supports the approach of TM targeting for efficient brain tumor therapies. Citation Format: Daniel Azorín. Screening for tumor microtube-targeting drugs identifies PKC modulators as multipotent inhibitors of glioblastoma progression [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 2 (Late-Breaking, Clinical Trial, and Invited Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(8_Suppl):Abstract nr NG03.

Abstract Introduction: Lung adenocarcinoma (LuAd) is thought to arise from the type 2 alveolar epithelial (AT2) cell, the stem cell of the alveolus of the lung. However, we recently found in mice that driving oncogenic KRAS in the non-stem cell of the lung alveolus, the type 1 alveolar epithelial (AT1) cell, drove it to take on an AT2 phenotype and go on to form indolent tumors that modeled lepidic LuAd in humans. Given that it is perhaps the most stable epigenetic change, CpG methylation may be the best method to determine the lineage of human tumors. Using whole genome methylation sequencing, we validated this approach in a genetic mouse model with defined AT1 and AT2 origin tumors and then applied it to human non-small cell lung cancer (NSCLC) resection specimens. Methods: Fluorescence assisted cell sorting was used to purify AT1, AT2, and tumor cells from genetically modified mice and from human NSCLC resection specimens. DNA was converted for methylation sequencing using an enzymatic approach. Libraries were prepared and then sequenced to a target depth of approximately 5X for mouse specimens and 30X for human specimens. Results: AT1-derived tumor cell populations in mice had a significantly higher degree of AT1-specific methylation than AT2-derived tumors. Twelve LuAd, three squamous cell lung cancer (LuSC), four normal AT1 populations, and four normal AT2 populations were collected from specimens from the Stanford University Hospital operating rooms. In the LuAd specimens, EGFR was the most commonly mutated driver with KRAS, KDR, P53, BRAF, BRCA2, MET, RB1, ARID1A, KEAP1, NF1, and STK11 also represented. Driver mutations and abnormal copy number were verified in all LuAd populations, and purity was determined to be up to ∼99%. Comparison of the methylation profile of the tumors to an existing methylation atlas that included profiles of alveolar, bronchial, and upper airway epithelial cells suggested a basal cell origin for LuSc and an alveolar origin for LuAd. One LuSC and three LuAd had such extensive methylation derangement that they lost a methylation signature of any particular tissue origin. We then defined 1808 differentially methylated regions in AT1 versus AT2 cells. Of the remaining 9 LuAd that retained their lung methylation identity, 8 appeared to be of AT1 cell origin and 1 of AT2 cell origin. Higher AT1 methylation score correlated with lower clinical grade, more lepidic and acinar histology, and more ground glass appearance on CT scan. Conclusion: Whole genome methylation sequencing strongly suggests that human LuAd is derived from alveolar epithelial cells, with some tumors being derived from AT1 cells, the non-stem cell of the alveolus. The presence of non-stem cell methylation correlated with favorable clinical indices. Of note, our analysis includes only resectable LuAd, potentially biasing towards indolent tumors more likely to be derived from AT1 cells. Citation Format: Nicholas Juul, Diego Almanza, Maximilian Diehn, Tushar Desai. Whole genome methylation sequencingstrongly suggests a novel, non-stem cell origin for some lung adenocarcinomas [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 2 (Late-Breaking, Clinical Trial, and Invited Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(8_Suppl):Abstract nr LB314.

Abstract Background: PARP inhibitors (PARPi) selectively kill tumor cells harboring genetic mutations in critical DNA repair genes (e.g., BRCA1/2). Approved nonselective PARPi have demonstrated robust antitumor activity, but are associated with significant hematologic toxicities that limit dose intensity and clinical benefit. Drugs that selectively inhibit PARP1, but spare PARP2, may improve the risk-benefit profile by retaining antitumor activity while avoiding PARP2-related toxicities. Furthermore, current PARPi have variable brain penetrance, limiting their utility for targeting CNS tumors or brain metastases. EIK1004 (IMP1707) is a potent, CNS-penetrant, PARP1-selective inhibitor that demonstrates tumor growth inhibition in preclinical brain metastasis models. Methods: Study EIK1004-001 (IMP1707-101) is a FIH, global, multi-center, Phase 1/2 study evaluating the safety and potential antitumor activity of EIK1004 monotherapy in patients with advanced ovarian, breast, prostate, or pancreatic cancer, with or without brain metastases (NCT06907043). The study consists of Part 1 (Dose Escalation; using the Bayesian optimal interval [BOIN] design) and Part 2 (Dose Optimization). All participants must have a deleterious or suspected deleterious mutation in select homologous recombination repair genes and received no more than one prior line of PARPi treatment. Eligible participants must be ≥ 18 years, have histologically or cytologically confirmed tumors and received appropriate prior antitumor therapies, and have evaluable disease (eg, at least 1 measurable lesion by RECIST 1.1 [or RANO-BM for brain metastases] and/or serum tumor markers). Primary endpoints include safety and tolerability including identifying the maximum tolerated dose or maximum achievable dose, and recommended dose(s) for expansion (RDE). Secondary endpoints include evaluation of EIK1004 pharmacokinetics and preliminary antitumor activity including overall response, duration of response, and progression-free survival. Part 2 will open once the RDE is determined from Part 1. This study opened on 23-Jan-2025 and is actively enrolling participants. Citation Format: Timothy A. Yap, Jian Zhang, Yanhua Xu, Sanum Chaudry, Kevin H. Eng, Yawei Zhang, Viola J. Chen, Gerald Falchook. A first-in-human (FIH), Phase 1/2, dose-escalation and dose-optimization study of central nervous system (CNS)-penetrant, PARP1-selective inhibitor EIK1004 in patients with advanced solid tumors with or without brain metastases [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 2 (Late-Breaking, Clinical Trial, and Invited Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(8_Suppl):Abstract nr CT287.

A defining feature of epithelial ovarian cancer, irrespective of histologic subtype, is its predominant spread through transcoelomic metastasis, where tumor cells disseminate into the peritoneal fluid, resist anoikis, and form multicellular aggregates that invade the peritoneum. This tumor progression represents the main driver of mortality for ovarian cancer patients. Identification of the earliest adaptations necessary for metastasizing ovarian cancer cells to survive matrix detachment could help develop strategies to prevent the initiation of transcoelomic metastasis. In this study, we identified a conserved detachment-sensitive gene signature activated shortly after matrix-detachment across multiple ascites-derived ovarian cancer cell lines. Within this signature, RHOV, an atypical and fast-cycling Rho GTPase, emerged as a top transcript that was confirmed to be highly induced in patient-ascites derived cells. Loss of RHOV impaired anoikis resistance, multicellular aggregate compaction, migration, and invasion in vitro, and it completely abolished metastasis in vivo. Mechanistically, RHOV enhanced c-Jun signaling and cytoskeletal remodeling to support pro-metastatic signaling. Rescue experiments showed that both GTP-binding and membrane localization were required for the pro-metastatic function of RHOV. Together, these findings define RHOV as a unique detachment-sensitive Rho GTPase and establish RHOV as a critical and necessary mediator of early adaptations that prime ovarian cancer cells for peritoneal metastatic progression. This work provides key insights into the molecular vulnerabilities of disseminating tumor cells, establishes the targeting of early molecular adaptations following matrix detachment as a potential therapeutic strategy for metastatic disease, and uncovers functions of an understudied member of the Rho GTPase family.