Mismatch repair-deficient (MMRd) colorectal cancers (CRCs) have high mutation burdens, which make these tumours immunogenic and many respond to immune checkpoint inhibitors. The MMRd hypermutator phenotype may also promote intratumour heterogeneity (ITH) and cancer evolution. We applied multiregion sequencing and CD8 and programmed death ligand 1 (PD-L1) immunostaining to systematically investigate ITH and how genetic and immune landscapes coevolve. All cases had high truncal mutation burdens. Despite pervasive ITH, driver aberrations showed a clear hierarchy. Those in WNT/β-catenin, mitogen-activated protein kinase, and TGF-β receptor family genes were almost always truncal. Immune evasion (IE) drivers, such as inactivation of genes involved in antigen presentation or IFN-γ signalling, were predominantly subclonal and showed parallel evolution. These IE drivers have been implicated in immune checkpoint inhibitor resistance or sensitivity. Clonality assessments are therefore important for the development of predictive immunotherapy biomarkers in MMRd CRCs. Phylogenetic analysis identified three distinct patterns of IE driver evolution: pan-tumour evolution, subclonal evolution, and evolutionary stasis. These, but neither mutation burdens nor heterogeneity metrics, significantly correlated with T-cell densities, which were used as a surrogate marker of tumour immunogenicity. Furthermore, this revealed that genetic and T-cell infiltrates coevolve in MMRd CRCs. Low T-cell densities in the subgroup without any known IE drivers may indicate an, as yet unknown, IE mechanism. PD-L1 was expressed in the tumour microenvironment in most samples and correlated with T-cell densities. However, PD-L1 expression in cancer cells was independent of T-cell densities but strongly associated with loss of the intestinal homeobox transcription factor CDX2. This explains infrequent PD-L1 expression by cancer cells and may contribute to a higher recurrence risk of MMRd CRCs with impaired CDX2 expression. © 2023 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.

The gut microbiota is implicated in the pathogenesis of colorectal cancer (CRC). We aimed to map the CRC mucosal microbiota and metabolome and define the influence of the tumoral microbiota on oncological outcomes. A multicentre, prospective observational study was conducted of CRC patients undergoing primary surgical resection in the UK (n = 74) and Czech Republic (n = 61). Analysis was performed using metataxonomics, ultra-performance liquid chromatography-mass spectrometry (UPLC-MS), targeted bacterial qPCR and tumour exome sequencing. Hierarchical clustering accounting for clinical and oncological covariates was performed to identify clusters of bacteria and metabolites linked to CRC. Cox proportional hazards regression was used to ascertain clusters associated with disease-free survival over median follow-up of 50 months. Thirteen mucosal microbiota clusters were identified, of which five were significantly different between tumour and paired normal mucosa. Cluster 7, containing the pathobionts Fusobacterium nucleatum and Granulicatella adiacens, was strongly associated with CRC (PFDR = 0.0002). Additionally, tumoral dominance of cluster 7 independently predicted favourable disease-free survival (adjusted p = 0.031). Cluster 1, containing Faecalibacterium prausnitzii and Ruminococcus gnavus, was negatively associated with cancer (PFDR = 0.0009), and abundance was independently predictive of worse disease-free survival (adjusted p = 0.0009). UPLC-MS analysis revealed two major metabolic (Met) clusters. Met 1, composed of medium chain (MCFA), long-chain (LCFA) and very long-chain (VLCFA) fatty acid species, ceramides and lysophospholipids, was negatively associated with CRC (PFDR = 2.61 × 10−11); Met 2, composed of phosphatidylcholine species, nucleosides and amino acids, was strongly associated with CRC (PFDR = 1.30 × 10−12), but metabolite clusters were not associated with disease-free survival (p = 0.358). An association was identified between Met 1 and DNA mismatch-repair deficiency (p = 0.005). FBXW7 mutations were only found in cancers predominant in microbiota cluster 7. Networks of pathobionts in the tumour mucosal niche are associated with tumour mutation and metabolic subtypes and predict favourable outcome following CRC resection.

Immune-checkpoint inhibitors (ICIs) showed limited efficacy in mismatch repair proficient (MMRp) metastatic colorectal cancer (mCRC) to date. Many RAS/BRAF wildtype mCRCs respond to EGFR antibodies (cetuximab/panitumumab) and we previously showed that acquired resistance to these is characterised by an inflamed phenotype and PD-L1 and LAG3 upregulation. The single arm phase 2 iSCORE trial investigated whether RAS/BRAF wildtype MMRp mCRCs that had acquired resistance to prior chemotherapy and an EGFR antibody benefit from nivolumab (anti-PD1) and relatlimab (anti-LAG3) ICIs. Patients (pts) with RAS/BRAF wildtype MMRp mCRC that had responded to chemotherapy + EGFR antibody and subsequently progressed were recruited for treatment with nivolumab (480mg iv) and relatlimab (160mg iv) every 4 weeks. The primary endpoint was disease control rate at 6 months (DCR6) from treatment initiation. To detect an increase in DCR6 from 10% to 30%, with a two-sided 5% significance and power of 80%, 25 pts were needed. Secondary endpoints included duration of disease control, best objective response (ORR) during 6 months, progression free survival (PFS), overall survival (OS) and safety. Pre-treatment and on-treatment biopsies were obtained for biomarker analyses. 32 pts were registered. 25pts who received at least one dose of nivolumab/relatlimab were included in the primary endpoint analysis. 24% of pts had received ≥2 prior lines of systemic therapy. The median number of cycles administered was 2 (range: 1-12). 1pt remained on treatment (cycle 11) at the time of data cut-off (30/01/2023). Among 25 evaluable pts, best objective responses during 6 months by RECIST 1.1 were one CR, one PR, two SD and 20 PD; 1pt died without a follow-up scan. One PR and one SD were reported during treatment in two additional pts by iRECIST after pseudoprogression. Median duration of disease control for pts with clinical benefit (CR, PR, SD) by iRECIST was 9.0mo [range:1.8-11.1]. 4/7pts [57.1%] without liver metastases and 2/18pts [11.1%] with liver metastases achieved clinical benefit by iRECIST. The best ORR during 6 months was 8% [95% CI:1.0-26.0] by both RECIST 1.1 and iRECIST. DCR6 was 12% [95% CI:2.5-31.2] by RECIST 1.1 and 16% [95% CI:4.5-36.1] by iRECIST in all 25pts. At data cut-off, median PFS and OS were 1.6mo [95% CI:1.6-1.8] and 15.2mo [95% CI:6.4-18.5], respectively. Five grade 3 treatment related adverse events (TRAE) were reported across 5pts; no grade 4/5 TRAEs occurred. The most common TRAEs (any grade) were fatigue (24%) and acneiform rash (12%). In pre-treatment biopsies, the median PD-L1 combined positive score (CPS) was 0 [range:0-3] for pts without liver metastases and PD-L1 CPS was 0 for all four of these with clinical benefit. The median PD-L1 CPS was 1.5 [range:0-65] for pts with liver metastases, with PD-L1 CPS 6 and 12 in the 2pts with clinical benefit. The prespecified endpoint of 25% DCR6 was not met with nivolumab/relatlimab. However, the clinical benefit rate in pts without liver metastases and prolonged disease control in two pts with liver metastases are encouraging. Nivolumab/relatlimab was well tolerated. Biomarker analyses, including TMB and T-cell quantification, are ongoing and will be presented.

446 Background: ICONIC evaluated 4+4 cycles of FLOT-A (2-weekly standard FLOT with 10mg/kg of the anti-PDL1 antibody avelumab) for perioperative treatment of early-stage OGA. We report R0 resection rates, pathologic complete response rates (pCR), pathologic tumour regression grades (TRG) according to Mandard classification and progression free survival (PFS) data in the modified intention to treat population (mITT), and translational analyses. Methods: ICONIC is a single-arm phase II trial of FLOT-A in patients (pts) with ≥cT2-4 or N+ OGA. The pCR rate in surgical specimens was the primary endpoint. Response evaluation according to Mandard, R0 rates and PFS were among secondary endpoints. PD-L1 expression was assessed according to the combined positive score (CPS) with the 22C3 pharmDx kit. Results: The trial closed early after the pCR rate was 15% (95% CI: 49%-83%) once 34 pts in the mITT population (defined as pts who had 1-4 cycles FLOT-A and surgery) as the pre-specified aim to demonstrate a pCR rate of 25% in 40 pts was unlikely to be met. Baseline characteristics of the mITT population were: median age 64y, 79% male, 79% OG junction, 6% oesophageal, 15% stomach, 65% poorly differentiated, 79% T3/4, 59% N+. 91% of pts received all 4 pre-operative cycles and 97% of pts achieved R0 resections. PDL1 CPS ≥1 and CPS ≥5 were nevertheless associated with increased TRG3 and decreased TRG4/5 rates. 3/34 pts (9%) had hypermutated/MMRd tumours and after their exclusion, the association of higher CPS with better TRG3 and decreased TRG4/5 remained. With a median follow up duration of 15.8 months the 12-month PFS was 93.1% (95% CI: 75.1%-98.2%) which is promising compared to historic results with peri-operative FLOT. Conclusions: Although FLOT-A failed to increase pCR rates to 25%, there is a trend towards higher TRG3 and lower TRG4/5 in pts with PDL1 CPS≥1 and CPS≥5 and promising PFS indicate activity of immunotherapy in combination with FLOT chemotherapy. Translational and biomarker analyses by exome- and RNA-sequencing, and multiplex immune cell staining are ongoing and will be reported. Clinical trial information: NCT03399071 . [Table: see text]

Introduction: The evolution of metastatic cancers over time can be assessed in circulating tumor DNA (ctDNA) but sequencing and bioinformatics tools for ultra-deep ctDNA whole-exome sequencing (WES) analysis are lacking. Methods: We developed ctDNA WES that only requires 15ng DNA to achieve sequencing depths of 1000-2000x. Error correction with molecular barcodes and duplex DNA detection allowed calling of mutations ≥0.5% variant frequency (VF). This pilot study applied ctDNA WES to plasma from 20 EGA patients (pts) and standard WES to matched biopsies in order to assess ctDNA WES performance and whether clonal mutation burden (cMB), a critical immunotherapy biomarker and important for neoantigen vaccine designs, differed between ctDNA and biopsies. We furthermore established a mutation and copy number data analysis pipeline using Bayesian clustering to define subclones and track their evolution during therapy in 3pts. Results: The median age of pts was 71y, 95% had distant metastases and 5% locally advanced EGAs. At a median sequencing depth of ????x after de-duplication, VFs of mutations in pre-treatment ctDNA was low (<2% VF) in 8 pts, and intermediate (2-10%) to high (>10%) in 12. Whether cMB differed in biopsies vs ctDNA was assessed in 7 pts with high VFs in ctDNA and good cancer purity in matching biopsy WES. The median cMB was 82 in biopsies and 111 in ctDNA. The increase was driven by 3 cases with 30%, 35% and 59% higher cMB in ctDNA vs biopsies. In two pts, most mutations that only appeared clonal in ctDNA were subclonal rather than absent in the primary tumor, indicating the subclonal presence of the metastasis progenitor clone. Subclonal intermixing in primary tumors hence limits the accurate identification of mutations that are clonal in metastatic disease. Evolutionary dynamics analyses in 3 pts who had good responses to chemotherapy before progression showed a major clonal sweep in one, supporting monoclonal resistance, and evolution of small subclones in two, indicating polyclonal resistance. A MEK1 K57T mutation evolved at resistance in a HER2 amplified EGA treated with trastuzumab+chemotherapy, demonstrating the utility to identify mechanisms of acquired resistance. Conclusions: ctDNA WES can assess the genetics of entire metastatic cancer cell populations over time and deconvolute their evolutionary trajectories. 43% pts had higher cMB in ctDNA compared to biopsies. Liquid biopsy analyses by WES may be superior to tissue-based WES for mutation burden analysis and neoantigen vaccine designs. ctDNA WES identified distinct evolutionary modes of resistance. A larger cohort is being analyzed to define how these differ clinically and whether they can be predicted from pre-treatment ctDNA WES. Efforts to increase the ctDNA WES sensitivity are ongoing as 8 pts had low ctDNA VFs which increased false negative rates. Citation Format: Neil McCafferty, Caroline Fong, Louise J. Barber, Andrew Woolston, Dimitrios Kleftogiannis, Taqia Rana, Susan Cromarty, Shannon Kidd, Ruwaida Begum, Ian Chau, Naureen Starling, David Cunningham, Marco Gerlinger. Clonal mutation burden and evolutionary dynamics analysis in metastatic gastro-esophageal adenocarcinoma (GEA) by error corrected whole-exome circulating tumor DNA sequencing. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 5593.

Background: The utility of molecular residual disease (MRD) detection by circulating tumor (ct)DNA in early-stage (pT2+ or N+, M0) esophagogastric adenocarcinoma treated with peri-operative systemic therapy has not been assessed in prospective trials. Methods: This exploratory analysis of the phase 2 ICONIC trial (NCT03399071), assessed whether ctDNA can predict recurrence and determine the efficacy of 4xFLOT+avelumab (FLOT-A) before and after surgery. Exome sequencing of pre-treatment biopsies was successful in 24/26 patients (pts) (92.3%) who had received FLOT-A and had undergone surgery at the time of analysis. All pts had R0 resections. Tumor-informed ctDNA assays (SignateraTM) were designed for these 24 pts and 220 serial plasma samples were analyzed. Pathologic response was assessed using Mandard tumor regression grading (TRG1 complete, 2 excellent, 3 good, 4 poor and 5 no response). The median follow-up was 17.0m from surgery. Progression free survival (PFS) was calculated from surgery to radiological recurrence or death. Results: ctDNA was detected in 23/24 pts (95.8%) prior to treatment. PFS was significantly shorter for pts with a mean number of tumor DNA molecules per ml plasma in the middle & highest tertile (p=0.049, HR=8.33, 95% CI: 1.01-1083, Firth Correction for Cox regression). 6 pts remained ctDNA-positive post neoadjuvant chemotherapy (NAC). None of these pts had a TRG1/2, 3 (50.0%) had TRG3 and 3 (50%) TRG4/5 in the resection specimen. Of 18 ctDNA-negative pts 5 (27.8%) had TRG1/2, 10 (55.6%) TRG3 and 3 (16.7%) TRG4/5. Post-surgery and prior to adjuvant therapy, 6/24 pts were ctDNA-positive. ctDNA positivity at this time point was associated with significantly shorter median PFS (12.9m) compared to ctDNA-negative status (PFS not reached, p<0.0001, HR=27, 95%CI: 3.0-241). Nodal status and Mandard TRG are routinely used as clinical predictors. Only the former was significantly associated with poor median PFS (pN+: 13.5m, pN-: not reached, p=0.027, HR=11, 95%CI: 1.3-98). Of 6 pts who remained ctDNA positive post-operatively, none achieved ctDNA clearance despite adjuvant FLOT-A. The median lead-time from ctDNA positivity after surgery to recurrence was 11.4 months. Conclusions: Post-NAC/pre-surgical ctDNA positivity correlated with worse pathological response. Persistent ctDNA after NAC & surgery was a stronger predictor of recurrence than nodal status or TRG in the resection specimen. Post-operative adjuvant therapy failed to clear ctDNA in any of the pts who were ctDNA positive after surgery, indicating that administering more of the same treatment is ineffective. This provides an opportunity to test new adjuvant therapies in pts who remain ctDNA positive after surgery. Whether post-operative therapy can be omitted in pts who are ctDNA negative after surgery should be assessed in future trials. Citation Format: Marco Gerlinger, Anderley Gordon, Louise J. Barber, Georgios Laliotis, Avani Athauda, Benjamin Challoner, Andrew Woolston, Sonia Mansukhani, Matt Dunstan, Nikoletta Petrou, Komel Khabra, Retchel Lazaro-Alcausi, Richard Crux, Victoria Borja, Ruwaida Begum, Isma Rana, Charuta Palsuledesai, Meenakshi Malhotra, Minetta Liu, Adham Jurdi, Shruti Sharma, Sheela Rao, Sacheen Kumar, David Cunningham, Ian Chau, Naureen Starling, M Asif Chaudry. Circulating tumor DNA for recurrence prediction and efficacy analysis in the ICONIC trial of peri-operative FLOT and avelumab (PD-L1) in localized esophago-gastric adenocarcinoma. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 5591.

Cancer cells harbor molecular alterations at all levels of information processing. Genomic/ epigenomic and transcriptomic alterations are inter-related between genes, within and across cancer types and may affect clinical phenotypes. Despite the abundant prior studies of integrating cancer multi-omics data, none of them organizes these associations in a hierarchical structure and validates the discoveries in extensive external data. We infer this Integrated Hierarchical Association Structure (IHAS) from the complete data of The Cancer Genome Atlas (TCGA) and compile a compendium of cancer multi-omics associations. Intriguingly, diverse alterations on genomes/epigenomes from multiple cancer types impact transcriptions of 18 Gene Groups. Half of them are further reduced to three Meta Gene Groups enriched with (1) immune and inflammatory responses, (2) embryonic development and neurogenesis, (3) cell cycle process and DNA repair. Over 80% of the clinical/molecular phenotypes reported in TCGA are aligned with the combinatorial expressions of Meta Gene Groups, Gene Groups, and other IHAS subunits. Furthermore, IHAS derived from TCGA is validated in more than 300 external datasets including multi-omics measurements and cellular responses upon drug treatments and gene perturbations in tumors, cancer cell lines, and normal tissues. To sum up, IHAS stratifies patients in terms of molecular signatures of its subunits, selects targeted genes or drugs for precision cancer therapy, and demonstrates that associations between survival times and transcriptional biomarkers may vary with cancer types. These rich information is critical for diagnosis and treatments of cancers.

Mismatch repair deficient colorectal cancers have high mutation loads and many respond to immune checkpoint-inhibitors. We investigated how genetic and immune landscapes co-evolve in these tumors. All cases had high truncal mutation loads. Driver aberrations showed a clear hierarchy despite pervasive intratumor heterogeneity: Those in WNT/βCatenin, mitogen-activated protein kinase and TGFβ receptor family genes were almost always truncal. Immune evasion drivers were predominantly subclonal and showed parallel evolution. Pan-tumor evolution, subclonal evolution, and evolutionary stasis of genetic immune evasion drivers defined three MMRd CRC subtypes with distinct T-cell infiltrates. These immune evasion drivers have been implicated in checkpoint-inhibitor resistance. Clonality and subtype assessments are hence critical for predictive immunotherapy biomarker development. Cancer cell PD-L1 expression was conditional on loss of the intestinal homeobox transcription factor CDX2. This explains infrequent PD-L1 expression by cancer cells and likely contributes to the high recurrence risk of MMRd CRCs with impaired CDX2 expression.

Cancer cells harbor molecular alterations at all levels of information processing. Genomic/epigenomic and transcriptomic alterations are inter-related between genes, within and across cancer types and may affect clinical phenotypes. Despite the abundant prior studies of integrating cancer multi-omics data, none of them organizes these associations in a hierarchical structure and validates the discoveries in extensive external data. We infer this Integrated Hierarchical Association Structure (IHAS) from the complete data of The Cancer Genome Atlas (TCGA) and compile a compendium of cancer multi-omics associations. Intriguingly, diverse alterations on genomes/epigenomes from multiple cancer types impact transcriptions of 18 Gene Groups. Half of them are further reduced to three Meta Gene Groups enriched with (1) immune and inflammatory responses, (2) embryonic development and neurogenesis, (3) cell cycle process and DNA repair. Over 80% of the clinical/molecular phenotypes reported in TCGA are aligned with the combinatorial expressions of Meta Gene Groups, Gene Groups, and other IHAS subunits. Furthermore, IHAS derived from TCGA is validated in more than 300 external datasets including multi-omics measurements and cellular responses upon drug treatments and gene perturbations in tumors, cancer cell lines, and normal tissues. To sum up, IHAS stratifies patients in terms of molecular signatures of its subunits, selects targeted genes or drugs for precision cancer therapy, and demonstrates that associations between survival times and transcriptional biomarkers may vary with cancer types. These rich information is critical for diagnosis and treatments of cancers.

Anti-EGFR antibodies such as cetuximab are active against KRAS/NRAS wild-type colorectal cancers (CRCs), but acquired resistance invariably evolves. It is unknown which mutational mechanisms enable resistance evolution and whether adaptive mutagenesis (a transient cetuximab-induced increase in mutation generation) contributes in patients. Here, we investigate these questions in exome sequencing data from 42 baseline and progression biopsies from cetuximab-treated CRCs. Mutation loads did not increase from baseline to progression, and evidence for a contribution of adaptive mutagenesis was limited. However, the chemotherapy-induced mutational signature SBS17b was the main contributor of specific KRAS/NRAS and EGFR driver mutations that are enriched at acquired resistance. Detectable SBS17b activity before treatment predicted shorter progression-free survival and the evolution of these specific mutations during subsequent cetuximab treatment. This result suggests that chemotherapy mutagenesis can accelerate resistance evolution. Mutational signatures may be a new class of cancer evolution predictor.

Background Gastric and gastro-esophageal junction cancers (GCs) frequently recur after resection, but markers to predict recurrence risk are missing. T-cell infiltrates have been validated as prognostic markers in other cancer types, but not in GC because of methodological limitations of past studies. We aimed to define and validate the prognostic role of major T-cell subtypes in GC by objective computational quantification. Methods Surgically resected chemotherapy-naive GCs were split into discovery (n=327) and validation (n=147) cohorts. CD8 (cytotoxic), CD45RO (memory), and FOXP3 (regulatory) T-cell densities were measured through multicolor immunofluorescence and computational image analysis. Cancer-specific survival (CSS) was assessed. All statistical tests were two-sided. Results CD45RO-cell and FOXP3-cell densities statistically significantly predicted CSS in both cohorts. Stage, CD45RO-cell, and FOXP3-cell densities were independent predictors of CSS in multivariable analysis; mismatch repair (MMR) and Epstein-Barr virus (EBV) status were not statistically significant. Combining CD45RO-cell and FOXP3-cell densities into the Stomach Cancer Immune Score showed highly statistically significant (all P <=.002) CSS differences (0.9years median CSS to not reached). T-cell infiltrates were highest in EBV-positive GCs and similar in MMR-deficient and MMR-proficient GCs. Conclusion The validation of CD45RO-cell and FOXP3-cell densities as prognostic markers in GC may guide personalized follow-up or (neo)adjuvant treatment strategies. Only those 20% of GCs with the highest T-cell infiltrates showed particularly good CSS, suggesting that a small subgroup of GCs is highly immunogenic. The potential for T-cell densities to predict immunotherapy responses should be assessed. The association of high FOXP3-cell densities with longer CSS warrants studies into the biology of regulatory T cells in GC.

Epidermal growth factor receptor antibodies (EGFR-Abs) confer a survival benefit in patients with RAS wild-type metastatic colorectal cancer (mCRC), but resistance invariably occurs. Previous data showed that only a minority of cancer cells harboured known genetic resistance drivers when clinical resistance to single-agent EGFR-Abs had evolved, supporting the activity of non-genetic resistance mechanisms. Here, we used error-corrected ctDNA-sequencing (ctDNA-Seq) of 40 cancer genes to identify drivers of resistance and whether a genetic resistance-gap (a lack of detectable genetic resistance mechanisms in a large fraction of the cancer cell population) also occurs in RAS wild-type mCRCs treated with a combination of EGFR-Abs and chemotherapy. We detected one MAP2K1/MEK1 mutation and one ERBB2 amplification in 2/3 patients with primary resistance and KRAS, NRAS, MAP2K1/MEK1 mutations and ERBB2 aberrations in 6/7 patients with acquired resistance. In vitro testing identified MAP2K1/MEK1 P124S as a novel driver of EGFR-Ab resistance. Mutation subclonality analyses confirmed a genetic resistance-gap in mCRCs treated with EGFR-Abs and chemotherapy, with only 13.42% of cancer cells harboring identifiable resistance drivers. Our results support the utility of ctDNA-Seq to guide treatment allocation for patients with resistance and the importance of investigating further non-canonical EGFR-Ab resistance mechanisms, such as microenvironmentally-mediated resistance. The detection of MAP2K1 mutations could inform trials of MEK-inhibitors in these tumours.

Background: Image guided tissue biopsies are critically important in diagnosis and management of cancer patients. High yield samples are also vital for biomarker and resistance mechanism discovery through molecular/genomic analyses. Methods: All consecutive patients who underwent plugged image-guided biopsy at RM from June 2013 until September 2016 were included in the analysis. In second step, a second cohort of patients prospectively treated within two clinical trials (PROSPECT-C and R), were assessed for the DNA yield from biopsies required for complex genomic analysis. Results: A total of 522 plugged core biopsies were performed in 457 patients [52% men; median age 63 years (range 17-93)]. Histological diagnosis was achieved in 501/522 (96%) of performed biopsies. Age, gender, modality, metastatic site and seniority of the interventionist were not found to be significant factors associated with odds of failure on a logistic regression. Seventeen (3.3%) were admitted due to biopsy-related complications; 9, 3, 2, 1, 1,1 were admitted for grade I/II pain control, sepsis, vasovagal syncope, thrombosis, haematuria and deranged liver functions respectively; 2 patients with right upper quadrant pain after liver biopsy were found to have radiologically confirmed subcapsular haematoma requiring conservative treatment. One patient (0.2%) developed grade III haemorhage following biopsy of a gastric GIST tumour. Overall molecular analysis was successful in 89% (197/222 biopsies). Prospective validation in 82 biopsies revealed 92.06% and 87.3% success rate of DNA extraction and tumour content of >20% respectively. Conclusion: The probability of diagnostic success for complex molecular analysis is increased with plugged large co-axial needle biopsy technique, which also minimises complications and reduces hospital stay. High yield DNA acquisition allows genomic molecular characterisation for personalised medicine.

An amendment to this paper has been published and can be accessed via a link at the top of the paper.

Background The T cell bispecific antibody cibisatamab (CEA-TCB) binds Carcino-Embryonic Antigen (CEA) on cancer cells and CD3 on T cells, which triggers T cell killing of cancer cell lines expressing moderate to high levels of CEA at the cell surface. Patient derived colorectal cancer organoids (PDOs) may more accurately represent patient tumors than established cell lines which potentially enables more detailed insights into mechanisms of cibisatamab resistance and sensitivity. Methods We established PDOs from multidrug-resistant metastatic CRCs. CEA expression of PDOs was determined by FACS and sensitivity to cibisatamab immunotherapy was assessed by co-culture of PDOs and allogeneic CD8 T cells. Results PDOs could be categorized into 3 groups based on CEA cell-surface expression: CEAhi (n = 3), CEAlo (n = 1) and CEAmixed PDOs (n = 4), that stably maintained populations of CEAhi and CEAlo cells, which has not previously been described in CRC cell lines. CEAhi PDOs were sensitive whereas CEAlo PDOs showed resistance to cibisatamab. PDOs with mixed expression showed low sensitivity to cibisatamab, suggesting that CEAlo cells maintain cancer cell growth. Culture of FACS-sorted CEAhi and CEAlo cells from PDOs with mixed CEA expression demonstrated high plasticity of CEA expression, contributing to resistance acquisition through CEA antigen loss. RNA-sequencing revealed increased WNT/β-catenin pathway activity in CEAlo cells. Cell surface CEA expression was up-regulated by inhibitors of the WNT/β-catenin pathway. Conclusions Based on these preclinical findings, heterogeneity and plasticity of CEA expression appear to confer low cibisatamab sensitivity in PDOs, supporting further clinical evaluation of their predictive effect in CRC. Pharmacological inhibition of the WNT/β-catenin pathway may be a rational combination to sensitize CRCs to cibisatamab. Our novel PDO and T cell co-culture immunotherapy models enable pre-clinical discovery of candidate biomarkers and combination therapies that may inform and accelerate the development of immuno-oncology agents in the clinic.

Background Patient derived organoids (PDOs) can be established from colorectal cancers (CRCs) as in vitro models to interrogate cancer biology and its clinical relevance. We applied mass spectrometry (MS) immunopeptidomics to investigate neoantigen presentation and whether this can be augmented through interferon gamma (IFNγ) or MEK-inhibitor treatment. Methods Four microsatellite stable PDOs from chemotherapy refractory and one from a treatment naïve CRC were expanded to replicates with 100 million cells each, and HLA class I and class II peptide ligands were analyzed by MS. Results We identified an average of 9936 unique peptides per PDO which compares favorably against published immunopeptidomics studies, suggesting high sensitivity. Loss of heterozygosity of the HLA locus was associated with low peptide diversity in one PDO. Peptides from genes without detectable expression by RNA-sequencing were rarely identified by MS. Only 3 out of 612 non-silent mutations encoded for neoantigens that were detected by MS. In contrast, computational HLA binding prediction estimated that 304 mutations could generate neoantigens. One hundred ninety-six of these were located in expressed genes, still exceeding the number of MS-detected neoantigens 65-fold. Treatment of four PDOs with IFNγ upregulated HLA class I expression and qualitatively changed the immunopeptidome, with increased presentation of IFNγ-inducible genes. HLA class II presented peptides increased dramatically with IFNγ treatment. MEK-inhibitor treatment showed no consistent effect on HLA class I or II expression or the peptidome. Importantly, no additional HLA class I or II presented neoantigens became detectable with any treatment. Conclusions Only 3 out of 612 non-silent mutations encoded for neoantigens that were detectable by MS. Although MS has sensitivity limits and biases, and likely underestimated the true neoantigen burden, this established a lower bound of the percentage of non-silent mutations that encode for presented neoantigens, which may be as low as 0.5%. This could be a reason for the poor responses of non-hypermutated CRCs to immune checkpoint inhibitors. MEK-inhibitors recently failed to improve checkpoint-inhibitor efficacy in CRC and the observed lack of HLA upregulation or improved peptide presentation may explain this.

DNA somatic copy number aberrations (SCNAs) are key drivers in oesophagogastric adenocarcinoma (OGA). Whether minimally invasive SCNA analysis of circulating tumour (ct)DNA can predict treatment outcomes and reveal how SCNAs evolve during chemotherapy is unknown. We investigated this by low-coverage whole genome sequencing (lcWGS) of ctDNA from 30 patients with advanced OGA prior to first-line chemotherapy and on progression. SCNA profiles were detectable pretreatment in 23/30 (76.7%) patients. The presence of liver metastases, primary tumour in situ, or of oesophageal or junctional tumour location predicted for a high ctDNA fraction. A low ctDNA concentration associated with significantly longer overall survival. Neither chromosomal instability metrics nor ploidy correlated with chemotherapy outcome. Chromosome 2q and 8p gains before treatment were associated with chemotherapy responses. lcWGS identified all amplifications found by prior targeted tumour tissue sequencing in cases with detectable ctDNA as well as finding additional changes. SCNA profiles changed during chemotherapy, indicating that cancer cell populations evolved during treatment; however, no recurrent SCNA changes were acquired at progression. Tracking the evolution of OGA cancer cell populations in ctDNA is feasible during chemotherapy. The observation of genetic evolution warrants investigation in larger series and with higher resolution techniques to reveal potential genetic predictors of response and drivers of chemotherapy resistance. The presence of liver metastasis is a potential biomarker for the selection of patients with high ctDNA content for such studies.

535 Background: The bispecific antibody CEA-TCB binds Carcino-Embryonic Antigen (CEA) on cancer cells and CD3 on T cells. This triggers T cell killing of colorectal cancer cell lines expressing moderate to high levels of CEA at the cell surface (Bacac, Clin Cancer Res 2016). Patient derived organoids (PDOs) may more accurately represent patient tumors than established cell lines. Yet, determinants of CEA-TCB resistance have not been studied in PDOs. Methods: PDOs were established from biopsies of eight multidrug-resistant metastatic CRCs, GFP labelled and adapted to 2D culture. Allogenic CD8 T cells and CEA-TCB or a non-targeting control antibody were added and cancer cell killing and growth were monitored for 10 days. CEA expression of PDOs was determined by FACS. Results: CRC PDOs could be categorized into three groups based on CEA cell-surface expression: CEAhigh (n = 3), CEAlow (n = 2), and CEA heterogeneous PDOs (n = 3) that stably maintained populations of both CEAhigh and CEAlow cells, which has not previously been described in CRC cell lines. Heterogeneity of cell-surface CEA expression is common in CRC cells in patients, supporting that PDOs may better represent these tumors than established cell lines. CEAhigh cells were sensitive whereas CEAlow cells showed resistance to CEA-TCB. All PDOs with heterogeneous CEA expression were resistant to CEA-TCB, suggesting that CEA-negative cells maintain cancer cell growth. Culture of FACS sorted CEAhigh and CEAlow cells from PDOs with heterogeneous CEA expression demonstrated high plasticity of CEA expression which may contribute to rapid resistance acquisition through CEA antigen loss. Conclusions: These results suggest that cell-surface CEA expression is a major determinant of CEA-TCB sensitivity and resistance in PDOs. In addition, we identified heterogeneous CEA expression in several PDOs and demonstrated that this could confer CEA-TCB resistance in vitro. These PDO models are likely to provide insights into the mechanism of CEA loss and may inform therapeutic opportunities to counter CEA-TCB resistance. RNA-sequencing and functional experiments are ongoing to investigate this and will be presented.