3533 Background: Patients with metastatic CRC are typically treated with fluorouracil-based chemotherapy in combination with leucovorin and irinotecan (FOLFIRI) or oxaliplatin (FOLFOX). However, patients who progress after 2 lines of therapy have limited treatment options, with poor outcomes (response rates of ≤6%). Ozekibart (INBRX-109) is a next-generation, tetravalent death receptor 5 agonist that has demonstrated robust efficacy as a single agent in chondrosarcoma. Additionally, ozekibart in combination with FOLFIRI showed early signs of antitumor activity in a small cohort of patients with CRC in the ongoing INBRX-109 phase 1 study (NCT03715933) (Berz D, et al. J Clin Oncol. 2025. Abstract 129). Based on these findings, the CRC cohort in this study was expanded to include additional patients who had received 2 or 3 prior lines of systemic therapy. We present initial findings from this cohort. Methods: Ozekibart + FOLFIRI was examined in patients with locally advanced or metastatic, unresectable CRC in the expansion cohort C4d of the open-label, INBRX-109 phase 1 study. Eligible patients were aged 18 to < 85 years, had received 2 or 3 prior lines of systemic therapy, and had no chronic or acute liver disease. Previous irinotecan-containing regimens were allowed but not as an immediate prior line of therapy. Patients received ozekibart 3 mg/kg every 4 weeks + FOLFIRI every 2 weeks. Safety and clinical response were primary endpoints. Results: As of the data cutoff (October 15, 2025), 44 patients had received ozekibart + FOLFIRI in this cohort. Median age was 54.5 years (range, 29-77 years), and > 70% were in the fourth-line treatment setting; 81.8% had previously received an irinotecan-based regimen. Overall, 28 patients (63.6%) remain on treatment; 11 patients had disease progression, 1 patient discontinued due to adverse events (AEs), and 4 patients discontinued for other reasons. Among evaluable patients for response at data cutoff (n = 26), ozekibart + FOLFIRI led to an ORR of 23% (all partial responses) and a disease control rate of 92%. Consistent with the safety profile of FOLFIRI, the most common AEs (in > 25% of patients) were anemia (any grade, 36.4%; grade ≥3, 9.1%), diarrhea (34.1%; 9.1%), nausea (31.8%; 0%), fatigue (27.3%; 2.3%), and alopecia (27.3%; 0%). Increased alanine aminotransferase was the only grade ≥3 hepatotoxicity event and occurred in 1 patient. Conclusions: Ozekibart + FOLFIRI demonstrated encouraging preliminary efficacy in a heavily pretreated patient population with metastatic, unresectable CRC. The safety profile of ozekibart + FOLFIRI was manageable, with most AEs being low grade. Our findings support further evaluation of ozekibart + FOLFIRI in advanced CRC. Updated data will be provided at the time of presentation. Clinical trial information: NCT03715933 .

Pancreatic ductal adenocarcinoma (PDAC) is a highly lethal malignancy, with rising incidence and a 5-year survival rate of 13%. Late presentation, early metastasis, and intrinsic resistance constrain the efficacy of cytotoxic chemotherapy, which remains the backbone of PDAC treatment, with only modest survival gains and resistance nearly universal. Although KRAS mutations dominate tumour biology (~90% of cases), PDAC is a heterogeneous disease with distinct molecular subtypes that confer differential therapeutic vulnerabilities. Advances in comprehensive molecular profiling have catalysed a paradigm shift toward precision oncology in PDAC. In KRAS-mutant PDAC, mutation-specific inhibitors have established proof-of-concept, particularly in KRAS G12C disease, while next-generation approaches including KRAS G12D inhibitors, RAS-“ON” inhibitors, proteolysis-targeting chimeras (PROTACs), and KRAS-targeted vaccine strategies are expanding the therapeutic landscape. Combination strategies targeting upstream and downstream effectors of the RAS–MAPK pathway are also being explored to enhance the depth and durability of response. In parallel, KRAS-wild-type PDAC has emerged as a molecularly distinct subgroup enriched for rare but actionable alternative oncogenic fusion drivers including NRG1, NTRK, RET, ALK, and FGFR. Additional molecularly directed strategies targeting HER2 alterations, BRAF mutations, EGFR-dependent signalling, and tumour-selectively exposed surface antigens such as CLDN18.2 are under investigation across PDAC irrespective of KRAS mutation status. Synthetic lethal approaches, including targeting the PRMT5/CDKN2A/MTAP axis, represent a further emerging therapeutic strategy. Germline homologous recombination repair defects, particularly involving BRCA1/2 and PALB2, further define clinically important subsets with sensitivity to platinum chemotherapy and PARP inhibition. This review summarises current and emerging targeted and molecularly directed therapeutic strategies in PDAC, emphasising the importance of molecular stratification and recent advances shaping precision oncology in this historically treatment-refractory disease.

384 Background: Advanced oesophago-gastric cancers have showcased varied responses to immunotherapy based on biomarkers such as tumour mutational burden, microsatellite instability, and PD-L1 expression. However, these markers may not fully capture tumour heterogeneity and evolutionary dynamics. Hence, understanding the intricate relationship between immune selection, specifically immune dN/dS,1 and tumour evolution is vital for devising personalized therapeutic strategies. Methods: In our phase 2 open label, multicentre, randomised PLATFORM trial (NCT02678182), patients received maintenance durvalumab post first-line chemotherapy. A survival advantage was not seen with maintenance durvalumab compared to surveillance and exploratory analysis of PD-L1 expression was not associated with improved survival outcomes.2 However, a subset of patients treated with durvalumab experienced durable responses and some experienced incremental radiological responses. A pilot substudy on a selected cohort (n=24) of these patients was designed to explore the potential of immunogenomics evolutionary-based metrics, notably dN/dS, as predictive biomarkers. DNA and RNA from FFPE baseline tumour samples underwent analysis, examining genomic factors including tumour mutational burden, indel signatures, PDL1 status, and immune evasion mechanisms. Furthermore, the evolutionary metric, immune dN/dS, which has implications in tumour immunoediting and its response to immunotherapies, was scrutinized. Results: A dominant indel mutational signature emerged from the analysis. The majority of patients displayed a significantly higher number of indel variants, compared to non-synonymous and synonymous single-nucleotide variants. A recurring observation was dN/dS values consistently above 1, indicating positive selection for non-synonymous somatic variants. Additionally, hypermutation in driver genes, notably TP53 and XIRP2, was detected. One patient, with a unique dN/dS < 1, showcased survival beyond four years without liver metastases. Immune dN/dS > 1 and no evidence of liver metastases displayed the best prognosis after treatment in contrast with those that have immune dN/dS < 1 with or without liver metastases. Conclusions: Immune dN/dS emerges as a promising tool to decode tumour evolutionary mechanisms, providing profound insights into the dynamics of tumour evolution in advanced HER2 negative oesophago-gastric cancer patients receiving durvalumab immunotherapy. This pilot study underscores the potential of this novel immunogenomic approach in improving patient stratification and tailoring therapeutic strategies in cancer immunotherapy. 1. Zapata et al, Nature Genetics, 2023. 2. Fong et al, J Clin Oncol, 2021. Clinical trial information: NCT02678182 .

Immunotherapy has revolutionised cancer treatment over the past decade. Long-term durable responses can be achieved in some cancer patient populations that were previously facing terminal disease. In this chapter, we summarise current phase 3 clinical trial evidence for the use of immunotherapy in gastrointestinal cancers (oesophageal squamous cell carcinoma, oesophago-gastric adenocarcinoma, pancreatic cancer, biliary cancer, hepatocellular carcinoma, colorectal cancer, and squamous cell cancer of the anus). We discuss meaningful biomarkers used in clinical trials to select patients most likely to benefit from immunotherapy, such as mismatch-repair deficiency (MMRd)/microsatellite instability (MSI) and programmed-death-ligand-1 (PD-L1) immunohistochemistry (IHC) expression. Clinical questions are arising regarding the role of immunotherapy in the adjuvant/perioperative setting, optimal timing of surgery in patients who respond to immunotherapy, and toxicities specific to patients with gastrointestinal malignancies. We outline the current landscape and future horizon of immunotherapy in gastrointestinal cancers, such as strategies to increase effectiveness of checkpoint blockade through combinations with other checkpoint inhibitors, cytotoxic chemotherapy, targeted agents, radiotherapy, CAR-T therapy, and cancer vaccines.

419 Background: A predictive biomarker for resistance to trastuzumab in HER2 positive oesophago-gastric cancer would refine patient selection. Our preclinical studies suggest that patients with HIGH baseline plasma miR-148a-3p levels will experience shorter Overall Survival (OS), Progression Free Survival (PFS), and worse Progression Free Rates (PFR) than those with LOW plasma miR-148a-3p levels. Methods: This sub-study is a prospective biomarker analysis of baseline plasma samples for HER2 positive advanced oesophago-gastric cancer patients registered within Arm B1 of the Phase 2 open label, multicentre, randomised PLATFORM trial (NCT02678182). All HER2 positive patients assessable for clinical response and miR signatures were included. Copies of miR-148a-3p and miR-16 per ml of plasma were quantified using Digital Droplet Polymerase Chain Reaction (ddPCR). OS as primary endpoint for miR-148a-3p LOW (≤median) versus HIGH (>median) was analysed using Kaplan-Meier curves and Cox model regression. Secondary endpoints were PFS and PFR. OS and PFS are from start of 1st line therapy; PFR are from start of maintenance trastuzumab. Sensitivity analysis normalised miR-148a-3p to miR-16. Results: Of 63 patients with analysable lab samples and available survival data, normalisation was possible for 41 patients. Median follow-up was 38 months. There was no statistically significant relationship between OS and miR-148-3p copies/μl of ddPCR reaction LOW versus HIGH (n= 62, Hazard Ratio (HR) 0.98, p=0.933), PFS (n=62, HR 1.08, p=0.759) or PFR (n=31, Odds Radio (OR) 0.67, p=0.577). Normalised miR-148a-3p (NmiR-148a-3p) LOW versus HIGH demonstrated a statistically significant difference in PFR at 3 months (n=23, OR=0.11, p=0.027) but no difference in OS or PFS. A model adjusting for primary tumour site, metastatic disease and number of sites demonstrated a statistically significant difference in PFR at 3 months (aOR=0.03, p=0.029). Conclusions: Patients with HIGH NmiR-148-3p have 0.03 times the odds of being progression-free at 3 months than patients with LOW NmiR-148a-3p in this population, when adjusted for key factors. Limitations include small sample size and normalisation. Clinical trial information: NCT02678182 . [Table: see text]

BackgroundThis study aimed to identify microRNAs (miRs) as circulating biomarkers of resistance to first-line trastuzumab-based therapy in advanced HER2-positive oesophago-gastric cancer patients.MethodsA high-throughput 1015 Exiqon miRCURY LNA™ microRNA inhibitor library screen was performed in trastuzumab-treated HER2-positive NCI-N87 and HER2-negative FLO-1 oesophago-gastric cancer cell lines. NanoString nCounter® miR analysis was performed in NCI-N87, FLO-1, and MAGIC trial (ISRCTN93793971) formalin-fixed paraffin-embedded (FFPE) oesophago-gastric cancer patient samples. MiR-148a-3p copies in plasma samples were quantified using digital droplet polymerase chain reaction (ddPCR) from HER2-positive oesophago-gastric cancer patients treated with standard-of-care trastuzumab-based therapy within the FOrMAT (NCT02112357) and PLATFORM (NCT02678182) clinical trials. The primary endpoints were overall survival (OS) for plasma miR-148a-3p HIGH (>median) versus LOW (≤median). The secondary endpoints were progression-free survival (PFS) and 3-month progression-free rates (PFRs) miR-148a-3p HIGH versus LOW. PLATFORM sensitivity analysis normalised miR-148a-3p (NmiR-148a-3p).ResultsThe inhibition of miR-148a-3p reduced NCI-N87 relative cell viability (<0.6) and expression was high (>242) in NCI-N87 and HER2-positive MAGIC trial patients (n=5). Normalised-miR-148a-3p (NmiR-148a-3p) LOW versus HIGH demonstrated a statistically significant difference in 3-month PFRs (n=23; OR, 0.11 [0.02–0.78]; p=0.027; aOR, 0.03 [0.001–0.71], p=0.029) but no difference in OS or PFS. There was no statistically significant relationship between miR-148-3p LOW versus HIGH for OS (PLATFORM, n=62; hazard ratio [HR], 0.98 [0.57–1.66]; p=0.933; FOrMAT, n=8; HR, 0.54 [0.13–2.31]; p=0.322), PFS (n=62; HR, 1.08 [0.65–1.81]; p=0.759; FOrMAT, n=8; HR, 1.26 [0.31–5.07]; p=0.714), or PFRs (PLATFORM, n=31; odds ratio [OR], 0.67 [0.2–2.8]; p=0.577).ConclusionNormalised miR-148a-3p may be a relevant biomarker for trastuzumab-based therapy in advanced HER2-positive oesophago-gastric cancer patients.

Supplementary methods and Figures S1-S17

Survival in mismatch-repair proficient (MMRp) metastatic colorectal cancer (mCRC) remains poor and chemotherapy is the mainstay of treatment. Immunotherapy has demonstrated durable responses and a favourable side-effect profile in various cancer types and multiple clinical trials have been conducted in MMRp mCRC. In this review we summarise emerging trial data which demonstrate promising immunotherapy combinations in MMRp mCRC. We outline barriers to success, evaluate emerging biomarkers and discuss potential strategies to increase the effectiveness of immunotherapy in MMRp mCRC.

Introduction HER2 positive gastric cancer is a distinct subgroup overexpressing the HER2 receptor. For a decade, first-line Trastuzumab was the only licensed HER2-directed therapy for HER2 positive advanced gastric cancer following results from the ToGA trial in 2010 demonstrating a survival benefit when added to chemotherapy. Within the last year, significant advances have taken place in the field of HER2-directed gastric cancer therapy. Areas covered This review discusses pivotal clinical trial results and summarizes current clinical trials of HER2-directed therapy in gastric cancer. Evidence for HER2-directed antibodies, immunotherapy, immune stimulating antibody conjugates, antibody-drug conjugates (including DESTINY trial results), and tyrosine kinase inhibitors are placed into clinical context. Key challenges including resistance mechanisms and drug toxicities are outlined. Search terms 'HER2' and 'gastric cancer' were entered into ClinicalTrials.gov, PubMed, and Google. Only English-language studies were included. Expert Opinion Clinical management of HER2 positive gastric cancer patients is likely to change significantly over the next 5 years. Checkpoint inhibition is likely to be used alongside HER2-directed therapy and chemotherapy first-line in advanced disease. Trastuzumab deruxtecan is likely to be offered second-line and beyond. The sheer number of clinical trials of HER2-directed therapy in gastric cancer are testament to progress and potential.

Abstract Trastuzumab in combination with chemotherapy represents the standard of care in HER2-positive advanced gastro-esphageal cancer (GOC), but development of resistance limits response. MicroRNAs (miRs) modulate key pathways in GOC. Identification of miRs responsible for resistance to HER2 inhibition may help stratify patients, predict response and define novel strategies to restore drug sensitivity. A high-throughput LNA™ miR-inhibitor screen in HER2-positive NCI-N87 and HER2-negative FLO-1 GOC cell lines was performed in order to identify potential miRs involved in trastuzumab sensitivity/resistance. Cells were treated with a combination of cisplatin, 5FU, and trastuzumab. MiRs were considered significant if LNAs caused >40% reduction in cell viability 72 hours post treatment relative to controls with a t-test p-value of <0.001 across 3 biological replicates. Putative hits with higher baseline miR levels on nCounter® NCI-N87 cell line analysis were validated in two HER2-positive cell lines by calculating the difference (Δ) in cell viability between cells treated with trastuzumab+chemotherapy versus ‘no drug' control. MiR levels in translational samples from ten HER2-positive GOC patients recruited to the FOrMAT trial (NCT02112357) were analysed by ddPCR. MiR levels were categorized as ‘low' or ‘high' relative to median expression and correlated with clinical response. Kaplan-Meier curves for PFS and OS were generated. In NCI-N87 and FLO-1 cell lines, 59 and 37 LNAs respectively caused >40% reduction in cell viability compared to controls (p<0.001). Eighteen miRs were significant exclusively in the NCI-N87 cell line screen, suggesting potential involvement in the HER2-signaling pathway. nCounter® analysis showed 6 of these were elevated at baseline in NCI-N87: miR-1260, miR-148-3p, miR-28-3p, miR-331-3p, miR-7-5p and miR-365. Inhibition of miR-148a-3p increased sensitivity to trastuzumab + 5FU+cisplatin in both the NCI-N87 cell line (Δ 67% versus Δ 55% in the control) and OACP4C cell line (Δ 50% versus Δ 47% in the control). Patients who responded to trastuzumab + chemotherapy in FOrMAT had a trend towards lower average miR-148a-3p plasma levels than those who failed to respond. Median PFS for plasma miR-148a-3p ‘low' subgroup was 146 days versus 154 days in ‘high' subgroup. Median OS for plasma miR-148a-3p ‘low' subgroup was 357 days versus 226 days for miR-148a-3p ‘high' subgroup. Our screen identified a panel of miRs associated with GOC resistance to HER2 inhibitors in combination with chemotherapy. Inhibition of these miRs affects GOC cell viability and restores sensitivity. MiR-148a-3p emerged as a potential biomarker of resistance and further evaluation is underway in the PLATFORM trial (NCT02678182). Citation Format: Hazel Lote, Andrea Lampis, George Vlachogiannis, Jens Hahne, Sing-Yu Moorcraft, Michael Davidson, Carolin Fong, Ruwaida Begum, Matteo Fassan, Sheela Rao, David Watkins, Naureen Starling, Ian Chau, David Cunningham, Nicola Valeri. MicroRNAs as biomarkers of resistance to HER2 inhibition in combination with chemotherapy in gastro-esophageal cancer [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 258.

Abstract Purpose: Anti-EGFR mAbs are effective in the treatment of metastatic colorectal cancer (mCRC) patients. RAS status and tumor location (sidedness) are predictive markers of patients' response to anti-EGFR mAbs. Recently, low miR-31-3p expression levels have been correlated with clinical benefit from the anti-EGFR mAb cetuximab. Here, we aimed to validate the predictive power of miR-31-3p in a prospective cohort of chemorefractory mCRC patients treated with single-agent anti-EGFR mAbs. Experimental Design: miR-31-3p was tested by in situ hybridization (ISH) in 91 pretreatment core biopsies from metastatic deposits of 45 patients with mCRC. Sequential tissue biopsies obtained before treatment, at the time of partial response, and at disease progression were tested to monitor changes in miR-31-3p expression overtreatment. miR-31-3p expression, sidedness, and RAS status in pretreatment cell-free DNA were combined in multivariable regression models to assess the predictive value of each variable alone or in combination. Results: Patients with low miR-31-3p expression in pretreatment biopsies showed better overall response rate, as well as better progression-free survival and overall survival, compared to those with high miR-31-3p expression. The prognostic effect of miR-31-3p was independent from age, gender, and sidedness. No significant changes in the expression of miR-31-3p were observed when sequential tissue biopsies were tested in long-term or poor responders to anti-EGFR mAbs. miR-31-3p scores were similar when pretreatment biopsies were compared with treatment-naïve archival tissues (often primary colorectal cancer). Conclusions: Our study validates the role of miR-31-3p as potential predictive biomarker of selection for anti-EGFR mAbs.

ObjectiveRegorafenib demonstrated efficacy in patients with metastatic colorectal cancer (mCRC). Lack of predictive biomarkers, potential toxicities and cost-effectiveness concerns highlight the unmet need for better patient selection.DesignPatients with RAS mutant mCRC with biopsiable metastases were enrolled in this phase II trial. Dynamic contrast-enhanced (DCE) MRI was acquired pretreatment and at day 15 post-treatment. Median values of volume transfer constant (Ktrans), enhancing fraction (EF) and their product KEF (summarised median values of Ktrans× EF) were generated. Circulating tumour (ct) DNA was collected monthly until progressive disease and tested for clonal RAS mutations by digital-droplet PCR. Tumour vasculature (CD-31) was scored by immunohistochemistry on 70 sequential tissue biopsies.ResultsTwenty-seven patients with paired DCE-MRI scans were analysed. Median KEF decrease was 58.2%. Of the 23 patients with outcome data, >70% drop in KEF (6/23) was associated with higher disease control rate (p=0.048) measured by RECIST V. 1.1 at 2 months, improved progression-free survival (PFS) (HR 0.16 (95% CI 0.04 to 0.72), p=0.02), 4-month PFS (66.7% vs 23.5%) and overall survival (OS) (HR 0.08 (95% CI 0.01 to 0.63), p=0.02). KEF drop correlated with CD-31 reduction in sequential tissue biopsies (p=0.04). RAS mutant clones decay in ctDNA after 8 weeks of treatment was associated with better PFS (HR 0.21 (95% CI 0.06 to 0.71), p=0.01) and OS (HR 0.28 (95% CI 0.07–1.04), p=0.06).ConclusionsCombining DCE-MRI and ctDNA predicts duration of anti-angiogenic response to regorafenib and may improve patient management with potential health/economic implications.

Patient-derived organoids (PDOs) have recently emerged as robust preclinical models; however, their potential to predict clinical outcomes in patients has remained unclear. We report on a living biobank of PDOs from metastatic, heavily pretreated colorectal and gastroesophageal cancer patients recruited in phase 1/2 clinical trials. Phenotypic and genotypic profiling of PDOs showed a high degree of similarity to the original patient tumors. Molecular profiling of tumor organoids was matched to drug-screening results, suggesting that PDOs could complement existing approaches in defining cancer vulnerabilities and improving treatment responses. We compared responses to anticancer agents ex vivo in organoids and PDO-based orthotopic mouse tumor xenograft models with the responses of the patients in clinical trials. Our data suggest that PDOs can recapitulate patient responses in the clinic and could be implemented in personalized medicine programs.

Human epidermal growth factor receptor 2 (HER2)-inhibition is an important therapeutic strategy in HER2-amplified gastro-oesophageal cancer (GOC). A significant proportion of GOC patients display HER2 amplification, yet HER2 inhibition in these patients has not displayed the success seen in HER2 amplified breast cancer. Much of the current evidence surrounding HER2 has been obtained from studies in breast cancer, and we are only recently beginning to improve our understanding of HER2-amplified GOC. Whilst there are numerous licensed HER2 inhibitors in breast cancer, trastuzumab remains the only licensed HER2 inhibitor for HER2-amplified GOC. Clinical trials investigating lapatinib, trastuzumab emtansine, pertuzumab and MM-111 in GOC have demonstrated disappointing results and have not yet changed the treatment paradigm. Trastuzumab deruxtecan may hold promise and is currently being investigated in phase II trials. HER2 amplified GOC differs from breast cancer due to inherent differences in the HER2 amino-truncation and mutation rate, loss of HER2 expression, alterations in HER2 signalling pathways and differences in insulin-like growth factor-1 receptor and MET expression. Epigenetic alterations involving different microRNA profiles in GOC as compared to breast cancer and intrinsic differences in the immune environment are likely to play a role. The key to effective treatment of HER2 amplified GOC lies in understanding these mechanisms and tailoring HER2 inhibition for GOC patients in order to improve clinical outcomes.

There are limited data on circulating, cell-free, tumour (ct)DNA analysis in locally advanced rectal cancer (LARC). Digital droplet (dd)PCR was used to investigate KRAS/BRAF mutations in ctDNA from baseline blood samples of 97 LARC patients who were treated with CAPOX followed by chemoradiotherapy, surgery and adjuvant CAPOX ± cetuximab in a randomised phase II trial. KRAS mutation in G12D, G12V or G13D was detected in the ctDNA of 43% and 35% of patients with tumours that were mutant and wild-type for these hotspot mutations, respectively, according to standard PCR-based analyses on tissue. The detection rate in the ctDNA of 10 patients with less common mutations was 50%. In 26 cases ctDNA analysis revealed KRAS mutations that were not previously found in tissue. Twenty-two of these (84.6%) were detected following repeat tissue testing by ddPCR. Overall, the ctDNA detection rate in the KRAS mutant population was 66%. Detection of KRAS mutation in ctDNA failed to predict prognosis or refine patient selection for cetuximab. While this study confirms the feasibility of ctDNA analysis in LARC and the high sensitivity of ddPCR, larger series are needed to better address the role of ctDNA as a prognostic or predictive tool in this setting.

Abstract Sequential profiling of plasma cell-free DNA (cfDNA) holds immense promise for early detection of patient progression. However, how to exploit the predictive power of cfDNA as a liquid biopsy in the clinic remains unclear. RAS pathway aberrations can be tracked in cfDNA to monitor resistance to anti-EGFR monoclonal antibodies in patients with metastatic colorectal cancer. In this prospective phase II clinical trial of single-agent cetuximab in RAS wild-type patients, we combine genomic profiling of serial cfDNA and matched sequential tissue biopsies with imaging and mathematical modeling of cancer evolution. We show that a significant proportion of patients defined as RAS wild-type based on diagnostic tissue analysis harbor aberrations in the RAS pathway in pretreatment cfDNA and, in fact, do not benefit from EGFR inhibition. We demonstrate that primary and acquired resistance to cetuximab are often of polyclonal nature, and these dynamics can be observed in tissue and plasma. Furthermore, evolutionary modeling combined with frequent serial sampling of cfDNA allows prediction of the expected time to treatment failure in individual patients. This study demonstrates how integrating frequently sampled longitudinal liquid biopsies with a mathematical framework of tumor evolution allows individualized quantitative forecasting of progression, providing novel opportunities for adaptive personalized therapies. Significance: Liquid biopsies capture spatial and temporal heterogeneity underpinning resistance to anti-EGFR monoclonal antibodies in colorectal cancer. Dense serial sampling is needed to predict the time to treatment failure and generate a window of opportunity for intervention. Cancer Discov; 8(10); 1270–85. ©2018 AACR. See related commentary by Siravegna and Corcoran, p. 1213. This article is highlighted in the In This Issue feature, p. 1195