Análisis del estado mutacional de los genes RAS en cáncer colorrectal metastásico: correlación entre tejido y biopsia líquida. l

UNIVERSITAT POLITÈCNICA DE VALÈNCIA DEPARTAMENTO DE BIOTECNOLOGÍA Mutational analysis of RAS genes in metastatic colorectal cancer: concordance between tumour tissue and liquid biopsy TRABAJO FIN DE MÁSTER EN BIOTECNOLOGÍA BIOMÉDICA ALUMNO: BORJA LAFUENTE GUTIÉRREZ TUTORA: ELOISA JANTUS LEWINTRE DIRECTORA (Co-tutor externo): SILVIA CALABUIG FARIÑAS Curso Académico: 2014/2016 VALENCIA, JULIO DE 2016 DATOS DEL ALUMNO/A - Dades de l'alumne/a - Student’s identification TÍTULO DEL TRABAJO FIN DE MÁSTER - Títol del Treball Fi de Màster - Title of the Final Master’s degree thesis Considero que el TFM está finalizado y se puede aceptar para la defensa del mismo SI Considere que el TFM està finalitzat i es pot aceptar la defensa d’aquest I consider that the TFM is completed and can accept for the defense Evaluación de la calidad del TFM y la labor del estudiante. Avaluació de la qualitat del TFM i la labor del estudiant. Evaluation of the quality of the Final Master’s degree thesis and the work of the student DEPARTAMENTO DE BIOTECNOLOGÍA INFORME TUTOR TRABAJO FIN DE MÁSTER (para cualquiera de las modalidades) Informe del tutor (per a qualsevol de les modalitats del treball fin de master) TUTOR’s REPORT (for any type of final Master’s degree thesis) CURSO - Curs - Academic year 2015-2016 Máster (Màster –Master's degree): Biotecnología Biomédica Apellidos (Cognoms - Surname(s)): Lafuente Gutiérrez Nombre (Nom – Name): Teléfono (Telèfon - Phone nr.): 662275327 DNI (ID card nr.): 21007055M Correo electr. (Correu electr. – email): borjalafuente.biomedi[email protected]m Borja Mutational analysis of RAS genes in metastatic colorectal cancer: concordance between tumor tissue and liquid biopsy Tutor/a (Tutor): Cotutor/a (Co-Tutor): Silvia Calabuig Fariñas Eloísa Jantus Lewintre El trabajo presentado tiene un claro objetivo traslacional, valorando la posibilidad de usar muestras de plasma para analizar mutaciones en genes RAS presentes en el tumor, específicamente a través de una técnica de muy alta sensibilidad como lo es la PCR-digital. En este caso, el estudio ha mostrado una alta concordancia entre las técnicas comparadas, y por lo tanto su posible implementación en la práctica clínica. El estudiante ha participado activamente en el trabajo experimental y también en el análisis y discusión de datos, habiendo adquirido las competencias requeridas para un grado de master. FECHA (Data) Date: 08 de Julio de 2016 Universitat Politècnica de València Departamento de Biotecnología Edificis 3J bajo. Camí de Vera, s/n, 46022 València Tel. +34 96 387 74 20 • Fax +34 96 387 74 29 [email protected] DEPARTAMENTO DE BIOTECNOLOGÍA ESCUELA TÉCNICA SUPERIOR DE INGENIERÍA AGRONÓMICA Y DEL MEDIO NATURAL AUTORIZACIÓN DEL COTUTOR PARA PRESENTACIÓN DEL TRABAJO FIN DE MÁSTER DE BIOTECNOLOGÍA BIOMÉDICA DEPARTAMENTO DE BIOTECNOLOGÍA.E.T.S.I.A.M.N Camino de Vera, s/nº.46022VALENCIA●Tel.+34963877420●Fax+34963877429 LA COTUTORA: D.ª SILVIA CALABUIG FARIÑAS Adscrita al organismo FUNDACIÓN PARA LA INVESTIGACIÓN DEL HOSPITAL GENERAL UNIVERSITARIO DE VALENCIA (FIHGUV) donde se ha realizado el Trabajo Fin de Máster titulado: Mutational analysis of RAS genes in metastatic colorectal cancer: concordance between tumor tissue and liquid biopsy Del que es AUTOR: D. BORJA LAFUENTE GUTIÉRREZ AUTORIZA la presentación del TRABAJO FIN DE MÁSTER para su defensa. Valencia, Julio 2016 INDEX 1. INTRODUCTION.........................................................................................................1 1.1. The concept of cancer........................................................................................1 1.2. Colorectal cancer................................................................................................2 1.2.1. Epidemiology.................................................................................................3 1.2.2. Diagnosis and staging....................................................................................3 1.3. CRC classification: from histology to molecular biology.................................4 1.3.1. Individual somatic mutations.........................................................................5 1.3.2. Detection of RAS mutations in CRC...............................................................6 1.3.3. Direct Sequencing: principles of pyrosequencing..........................................7 1.4. Treatment of CRC.................................................................................................9 1.5. Liquid biopsy......................................................................................................11 1.5.1. Digital PCR: BEAMing system.......................................................................13 2. OBJECTIVES..............................................................................................................15 3. MATERIALS AND METHODS....................................................................................16 3.1. Study design and patients................................................................................16 3.2. Biological samples: FFPE and blood.................................................................16 3.3. DNA extraction..................................................................................................16 3.4. Pyrosequencing..................................................................................................17 3.5. BEAMing determination of RAS mutations…………………………...........................18 3.6. Statistical analysis..............................................................................................20 4. RESULTS AND DISCUSSION......................................................................................21 4.1. Patient characteristics.......................................................................................21 4.2. RAS mutational analysis in tissue and ctDNA..................................................22 4.2.1. RAS analysis in FFPE samples.................................................................22 4.2.2. RAS analysis in ctDNA.............................................................................24 4.3. Concordance between plasma and tissue RAS mutational status.................28 4.4. Association of RAS mutational status with clinicopathological characteristics....................................................................................................33 4.5. Survival analysis according to RAS mutational status.....................................36 5. CONCLUSIONS.........................................................................................................39 6. REFERENCES.............................................................................................................40 7. APPENDICES.............................................................................................................46 1 1. INTRODUCTION 1.1. The Concept of Cancer Cancer is a group of related diseases, involving two common phenomena: uncontrolled proliferation and the potential to invade surrounding tissues or even spread to other organs or tissues (metastasize). The process of carcinogenesis involves dynamic changes in the genome of normal cells, eventually leading to the transformation into tumour cells (Hanahan & Weinberg 2000). This allows tumour cells to escape from homeostatic mechanisms that control proliferation. Hanahan and Weinberg proposed that there are ten essential characteristics, known as the “hallmarks of cancer”, for the development of cancer disease (Hanahan & Weinberg 2011) (Figure 1). Tumour cells present genomic instability, resulting in the accumulation of genomic mutations. The different steps involved in tumour progression are a succession of clonal expansions produced by the accumulation of mutations that generate selectively advantageous neoplastic cells. Figure 1: The ten hallmarks of cancer, modified from Hanahan & Weinberg 2011. In addition to cancer cells, tumours possess another dimension of complexity: they contain a repertoire of recruited cells that are active players in the tumour “microenvironment”, one of the main hallmark traits. In the microenvironment (Figure 2), epithelial neoplastic cells and the tumour-associated stroma form two well differentiated compartments, with distinct cell types and molecules that modulate tumour growth and invasiveness. 2 Figure 2: Influence of microenvironment in primary tumour and metastasis (Korkaya et al. 2011). 1.2. Colorectal Cancer Colorectal cancer (CRC) is a form of cancer of the intestinal gland cells that starts in the colon or the rectum. Both environmental and genetic factors play key roles in its etiology. Genetic susceptibility ranges from well-defined inherited syndromes (familial adenomatous polyposis, Lynch syndrome, etc.), to less defined familial aggregations. CRC arises after accumulation of acquired genetic and epigenetic changes that transform normal glandular epithelial cells into invasive adenocarcinomas. The classic CRC progression model (depicted in Figure 3) proposed by Vogelstein and colleagues describes the transformation of normal colonic epithelium into carcinomas (Vogelstein et al. 1988). A number of key pathways are involved in the oncogenesis of colon cancer, with the clinicopathological features of specific subgroups being driven by underlying molecular changes. 9 Gly12Ser (G12S) GGT  AGT Gly12Ser (G12S) GGT  AGT Gly12Ala (G12A) GGT  GCT Gly12Ala (G12A) GGT  GCT Gly12Arg (G12R) GGT  CCT Gly12Arg (G12R) GGT  CGT Codon 13 Codon 13 Gly13Asp (G13D) GCG  GAC Gly13Ser (G13S) GGT  AGT Codon 61 (CAA)1 Gly13Cys (G13C) GGT  TGT Gln61His (Q61H) TTG  GTG Gly13Arg (G13R) GGT  CGT Gln61Leu (Q61L) TTG  TAG Gly13Asp (G13D) GGT  GAT Gln61Arg (Q61R) TTG  TCG Gly13Val (G13V) GGT  GTT Gln61His (Q61H) TTG  ATG Gly13Ala (G13A) GGT  GCT Gln61Glu (Q61E) TTG  TTC Codon 59 Codon 59 Ala59Thr (A59T) GCT  ACT Ala59Thr (A59T) GCA  ACA Ala59Pro (A59P) GCT  CCT Ala59Ser (A59S) GCA  TCA Ala59Gly (A59G) GCT  GGT Ala59Gly (A59G) GCA  GGA Ala59Asp (A59D) GCT  GAT Ala59Glu (A59E) GCA  GAA Ala59Val (A59V) GCT  GTT Ala59Leu (A59L) GCA  GTA Codon 61 Codon 117 Gln61Lys (Q61K) CAA  AAA Lys117Glu (K117E) AAA  GAA Gln61Arg (Q61R) CAA  CGA Lys117Gln (K117Q) AAA  CAA Gln61Leu (Q61L) CAA  CTA Lys117Asn (K117N) AAA  AAC Gln61His (Q61H) CAA  CAT Lys117Asn (K117N) AAA  AAT Gln61His (Q61H) CAA  CAC Codon 146 Codon 117 Ala146Thr (A146T) GCA  ACA Lys117Asn (K117N) AAG  AAC Ala146Pro (A146P) GCA  CCA Lys117Asn (K117N) AAG  AAT Ala146Val (A146V) GCA  GTA Codon 146 Ala146Gly (A146G) GCA  GGA Ala146Thr (A146T) GCC  ACC Ala146Pro (A146P) GCC  CCC Ala146Ser (A146S) GCC  TCC Ala146Val (A146V) GCC  GTC Ala146Gly (A146G) GCC  GGC 1KRAS codon 61 is assayed in the reverse direction by Pyrosequencing. Mutations covered by OncoBEAM RAS CRC kit are highlighted in yellow. 1.4. Treatment of CRC The management of CRC depends mainly on clinicopathological characteristics of the patients, tumour stage and also on the molecular alterations found in tumour cells. The SEOM guidelines (Aranda et al. 2015) recommend, for most patients with good performance status (PS) and no significant comorbidities, the combination of 5-FU/leucovorin (5-FU/LV) with either oxaliplatin (FOLFOX) or irinotecan (FOLFIRI) as backbone for first-line 10 treatment. Other option includes capecitabine, an oral fluoropyrimidine with similar efficacy in first-line treatment of mCRC (Van Cutsem et al. 2004). On the other hand, first-line targeted therapies include the anti-vascular endothelial growth factor (VEGF) agent bevacizumab and the anti-EGFR drugs cetuximab and panitumumab. An apparent lack of response to targeted therapy with anti-EGFR monoclonal antibodies in around 40-50% of the patients with KRAS exon 2 wild-type (WT) tumours observed in some clinical trials boosted the search for additional predictive biomarkers. The effect of mutations in other members of the EGFR signaling pathway like BRAF, PIK3CA and NRAS was analysed in multiple studies (Karapetis et al. 2014; Sartore-Bianchi et al. 2009). It was found that there are around 5% of CRC patients who have mutations in KRAS exons 3 or 4, and a further 5% with mutations in NRAS exons 2, 3 or 4. These RAS mutations previously suggested were tested by Sanger sequencing: KRAS exon 3 (codons 59 and 61) and exon 4 (codons 117 and 146), and NRAS exon 2 (codons 12 and 13), exon 3 (codons 59 and 61), exon 4 (codons 117 and 146) (see Figure 8). Figure 8: Frequency of RAS mutations beyond KRAS codon 2 in CRC (Hecht et al. 2015). Retrospective analyses of several phase III trials indicated that all RAS mutations were regarded as a negative predictive factor of anti-EGFR therapy: one example was the PRIME trial (Douillard et al. 2013). These mutations were associated with inferior progression-free survival (PFS) and overall survival (OS) with panitumumab-FOLFOX4 treatment, like KRAS exon 2 mutations. The evidence supported the addition of all these mutations to the routinely tested KRAS mutation analysis, in order to further improve the selection of patients for antiEGFR therapy. As for the effect of mutations beyond anti-EGFR treatment efficacy, mutant KRAS and BRAF have been independently associated with worse overall survival in metastatic CRC patients. In the largest study, patients with CRC that harbored a KRAS mutation had a worse overall survival (OS) but similar progression-free survival (PFS) compared to patients with tumours bearing WT KRAS (Richman et al. 2009), as seen in Figure 9. 11 Figure 9: Prognostic impact of KRAS and BRAF mutations in A) progression-free survival (PFS) and B) overall survival (OS), compared to non-mutated patients, in any treatment arm. Modified from Richman et al. 2009. 1.5.Liquid biopsy Since the advent of targeted therapies, increased survival periods and improved quality of life are achieved for patients whose cancers harbor specific molecular alterations. However, targeted therapies have brought new challenges: high costs, potential morbidity of the necessary biopsies, lack of effective drugs against most genomic aberrations, technical limitations and regulatory obstacles. In addition, almost all tumours develop resistance mechanisms through tumour heterogeneity, clonal evolution and selection. Therefore, new methodologies are needed, that allow us to overcome these difficulties. Liquid biopsies appear to be a reliable alternative to conventional biopsies. They can provide both precise molecular data useful for improving the clinical management of mCRC cancer patients, and a less invasive way of monitoring tumour behavior. Multiple studies have shown that it is possible to reconstruct tumour genomes from plasma DNA (Thierry 2016; Goto et al. 2016). Traces of tumour DNA (circulating tumour DNA, ctDNA) can be found in the cell-free fraction of blood, together with DNA fragments from normal cells (cfDNA). After first description of fragments of DNA existing in the blood (Mandel & Metais 1948), higher levels of so called circulating free DNA (cfDNA) were identified in cancer patients compared to healthy controls, suggesting that this correlated with malignancy and tumour stage (Leon et al. 1977). To date, two main mechanisms for releasing circulating tumour DNA (ctDNA), “passive” and “active”, have been postulated. The passive mechanism involves the release of nucleic acids directly from apoptotic and necrotic tumour cells into the bloodstream or indirectly by necrotic tumour cells engulfed by macrophages (Diehl et al. 2005). This was further supported by measuring the size distribution of DNA fragments (Jahr et al. 2001; Heitzer et al. 2013). In contrast, fragments of ctDNA can also be “actively” secreted into the circulation, perhaps in association with a protein complex to act as an intercellular messenger of sorts (Peters & Pretorius 2012). 12 Considering that releasing of DNA into the bloodstream is not an exclusive process for primary or metastatic sites, ctDNA can provide a better overall representation of the malignant disease as a whole (Kuo et al. 2014). In cancer patients, ctDNA levels can vary according to tumour burden and stage, anatomical proximity to vasculature, and biological features like apoptotic rate and metastatic potential. The clinical applications of ctDNA (see Figure 10) can be divided into three main categories: a) early diagnosis and prognosis; b) profiling and molecular characterization of tumour genomic alterations, and c) monitoring treatment response and detection of resistance mechanisms. Figure 10: Uses of liquid biopsies in clinical oncology (Siravegna & Bardelli 2014). First, it has been proposed that monitoring tumour-specific changes may be a useful tool for early cancer detection and/or prognosis. A recent example of diagnostic approach involved blood-based CRC screening test using the SEPT9 biomarker that specifically detects a majority of CRCs of all stages and colorectal locations (Church et al. 2014). The simple measurement of the quantity of cfDNA in plasma by quantitative PCR is positively correlated with tumour burden in mCRC (Schmitt et al. 2012; Spindler et al. 2012). Second, as a prognostic biomarker, several studies have demonstrated that circulating-free DNA levels (cfDNA) or the number of circulating tumour cells (CTCs) are positively correlated with patient outcome in mCRC. In surgically resected CRC patients, the detection of ctDNA after surgery was related to an increased relapse rate (Diehl et al. 2008). Furthermore, it was shown that high concentrations of cfDNA and KRAS mutation were clear indicators of poor outcome for advanced CRC patients (Spindler et al. 2012). 13 Finally, one of the most widespread applications of liquid biopsy is monitoring response to therapy, particularly for those therapies with known resistance mechanisms. Several studies have reported that anti-EGFR resistant clones are present in the circulation months before progression was clinically obvious (Misale et al. 2014; Mohan et al. 2014). In summary, the determination of molecular alterations enables for the selection of adequate targeted therapies for each patient, and allows the clinician to make rapid therapeutic decisions if resistant clones are detected in circulation. Although liquid biopsy is very useful and advantageous compared to tumour tissue biopsies, the detection of scarce ctDNA mixed with relatively abundant WT cfDNA requires innovative ultra-sensitive techniques, such as digital PCR. 1.5.1. Digital PCR: BEAMing system Generally, there are two approaches for the analysis of ctDNA. A targeted approach: analysis of a small set of frequently occurring driver mutations with implications for therapy decisions, such as mutations in KRAS or EGFR. The second involves an untargeted approach without knowledge of any specific changes present in the primary tumour (Heitzer et al. 2015). Given the small proportion of ctDNA present in the total cfDNA samples, it is important to select the correct methods for its analysis; several highly sensitive techniques have been developed for the latter, ranging from real time PCR-based to more complex digital-PCR based technologies BEAMing (Beads-Emulsion-Amplification-Magnetics, by Sysmex Inostics) is a targeted, quantitative digital PCR technology that employs bead-based amplification in water-in-oil emulsions, and allele-specific hybridization followed by flow cytometry, for the detection of small amounts of mutated DNA released by tumours into the blood circulation. BEAMing is highly sensitive, able to detect mutant ctDNA in very low proportion (as low as 0.01% of total DNA fragments; see Figure 11) in a backgroud of normal (WT) DNA. Figure 11: Methodologies for detecting circulating tumour DNA (ctDNA). Pyrophosphorolysisactivated polymerization (PAP); tagged-amplicon deep sequencing (TAM-Seq) (Díaz Jr et al. 2014). 14 Ideally, after preamplification, a single DNA fragment is captured in a single magnetic bead, and each bead falls into one hydrophilic droplet of the water-oil emulsion. The amplification PCR takes place with TaqMan probes designed to cover mutational hotspots in exons 2, 3 and 4 of both genes (KRAS and NRAS). The mutant fraction cannot be interpreted as the fraction of cancer cells that harbor a particular mutation. While ctDNA quantification by quantitative PCR is positively correlated with tumour burden, mutant fraction may represent a combination of clones produced by different tumour lesions. In the context of CRC, RAS WT tumours are often sensitive to EGFR blockade with cetuximab or panitumumab, but almost all patients develop resistance within a few months (Karapetis et al. 2008). Liquid biopsy can substitute serial tumour tissue sampling, and may provide a global and evolving picture of the disease. Several studies have reported that anti-EGFR resistant clones are present in the circulation months before progression was clinically obvious (Misale et al. 2014; Mohan et al. 2014). Figure 12 shows the evolution of mutant cfDNA fraction in a patient; the increase in concentration predicted the emergence of secondary resistance to cetuximab treatment (Misale et al. 2014). Figure 12: Quantitative analysis of KRAS Q61H mutation in plasma by BEAMing (Misale et al. 2014). Proven useful in cancer research (Tabernero et al. 2015), BEAMing is now being validated for diagnostic determination of mutations in KRAS and NRAS in plasma of mCRC patients. Our laboratory had a chance to collaborate in this enterprise, and in this study we will determine whether blood-based RAS mutation testing is an appropriate surrogate for tissue-based RAS testing to assess eligibility of mCRC patients for anti-EGFR therapy by comparing the degree of concordance of plasma and tissue-based RAS testing in metastatic CRC patients. 15 2. OBJECTIVES Tumour tissue is currently used for RAS testing in mCRC patients, but the detection of circulating tumour DNA (ctDNA) is being actively investigated as a new method for the detection of actionable mutations in plasma samples. Therefore, the main objective of this work is to evaluate the concordance of RAS mutational status by comparison of results from ctDNA and tissue-based testing in a cohort of mCRC patients. The specific aims of this study are the following: a) To analyse concordances and discrepancies in RAS (KRAS and NRAS) mutational status between blood samples and tissue samples in a small cohort of newly diagnosed mCRC patients (n=30). b) To study correlations between the mutations detected in ctDNA and some relevant clinicopathological features. c) To integrate the results and determine whether RAS ctDNA testing is a feasible alternative for tumour tissue-based RAS testing. 16 3. MATERIALS AND METHODS 3.1. Study design and patients This is a retrospective analysis in 30 therapy-naïve patients with histologically or cytologically documented metastatic colorectal cancer. Patients having a history of another malignancy or having received any previous treatment (chemotherapy or targeted therapy) were excluded. We enriched our cohort with patients with detected mutations in RAS genes (KRAS and NRAS), so that we could perform the concordance study. The study was conducted in accordance with the Declaration of Helsinki, and the institutional ethical review board approved the protocol. 3.2. Biological samples: FFPE and blood First of all, patients must sign an informed consent document, by which they are given all the information regarding the samples that they will donate, how they will be taken and the relevance that they may have on treatment and/or prognosis. A total of 30 formalin-fixed, paraffin-embedded (FFPE) CRC samples obtained by colonoscopy or surgical resection were provided by the Pathology Service at the Consorcio Hospital General Universitario de Valencia. A pathology report was available for all the samples, enabling their characterisation. For mutational analysis of tumour tissue, the specimens were examined by a pathologist, and those having >5% of tumour cells were selected. Three to five 5 µM thick tissue sections were used for DNA isolation. Blood samples were obtained at the time of diagnosis of metastatic disease, previous to any systemic treatment. 10 mL of blood were collected from each patient in K2 EDTA BD Vacutainer® tubes, and processed to obtain plasma within an hour. Briefly, a series of centrifugations are performed, at increasing speeds, in order to isolate and clean-up the plasma. The isolated plasma is stored in Sarstedt™ CryoPure tubes, at -80 °C, until cfDNA extraction. 3.3. DNA extraction FFPE: DNA was isolated from FFPE tumour tissue sections. First paraffin was removed by incubating the samples with mineral oil at 95°C, followed by the addition of xylene. After centrifugation, the xylene supernatant was removed. Then ethanol 100% was added to clean residual xylene, followed by another centrifugation step. After that, the samples were allowed to air-dry for 20-30 minutes. The Cobas DNA Sample Preparation Kit (Roche®) was used for the extraction of DNA from deparaffinized samples, following manufacturer’s recommendations. Briefly, DNA Tissue Lysis 17 Buffer and Proteinase K were added. After two incubations (56 °C and 95 °C) and brief cooling, DNA Paraffin Binding Buffer and isopropanol were added, with respective incubation periods, and the content of each tube was transferred to a filtered tube. Then a series of centrifugations with washing steps in between were performed, and finally the DNA Elution Buffer was added in order to collect DNA in the definitive tube. Plasma: DNA from plasma samples was obtained by the commercial QIAamp® "Circulating Nucleic Acid" kit, following manufacturer's instructions. This protocol is based on affinity columns for retaining nucleic acids. FFPE samples DNA quantification was performed in the NanoDrop® 2000C system (Thermo Fisher Scientific). For plasma cfDNA, quantification was performed using Qubit® (Life Technologies) Fluorometer: A Working solution was prepared, diluting Qubit® reagent 1:200 in Qubit® Buffer. Sample DNA was diluted 1:200 in Working solution, and standard samples were diluted 1:20. The tubes were incubated at room temperature before readings. 3.4. Pyrosequencing All tissue biopsies were analysed for RAS genotyping assessment using two CE-IVD marked commercial kits: “Therascreen® KRAS Pyro Kit” and “Therascreen® RAS Extension Pyro Kit” (both from QIAGEN, Hilden, Germany) according to the producer protocols (“Therascreen KRAS Pyro Kit Handbook”, version 1, July 2011, and “Therascreen RAS Extension Pyro Kit Handbook”, version 1, October 2014). From each sample, 10 ng DNA were amplified for determining mutations status in: KRAS 12-13, NRAS 12-13, NRAS 61, KRAS 59-61, KRAS 117, KRAS 146, NRAS 58-59, NRAS 117 and NRAS 146. Pyrosequencing was performed using 10 μL of each PCR product with PyroMark Gold Q96 reagents (QIAGEN), Streptavidin Sepharose (GE Healthcare Bio-Science AB, Uppsala, Sweden), in the PyroMark Q24 instrument (QIAGEN). The results were analysed using PyroMark Q24 2.0.7 software (QIAGEN). The protocol is based on template DNA immobilization on Sepharose (beaded form of agarose), PCR amplification and sequencing. The reverse amplification primer (RP) is biotinylated, and so the sequencing is performed on the forward strand (except for KRAS codons 59/61, as seen in Figure 13). 18 Figure 13: Amplification primers (black arrows) and sequencing primers (white arrows) for KRAS codons 12-13 (A) and 59-61 (B). The gray circles represent biotinylated primers. A single nucleotide is incorporated in each sequencing step. If it is incorporated into the DNA strand, a pulse of light is generated (see Figure 14) and the intensity is registered (if more than one nucleotide of the same type is incorporated, the intensity is higher). Finally, the nucleotides are degraded and another nucleotide is incorporated, starting a new cycle. Figure 14: Biochemical basis for the generation of light by DNA pyrosequencing. ATP, adenosine triphosphate; ADP, adenosine diphosphate; dNDP, deoxy-nucleotidyl diphosphate; dNMP, deoxy-nucleotidyl monophosphate; PPi, pyrophosphate. (Petrosino et al. 2009). 3.5. BEAMing determination of RAS mutations RAS mutational analysis on ctDNA was done with BEAMing digital-PCR (OncoBEAM™ RAS CRC Kit Sysmex® Inostics), a technique based on emulsion PCR that allows detection of one mutant allele in 10000 WT alleles. As shown in Figure 15, DNA isolation and pre-amplification reactions were performed on the pre-PCR laboratory, whereas from emulsion-PCR step until the final flow-cytometry analysis the experimental work was done on the post-PCR laboratory (physically separated area), in order to avoid cross-contamination.A workflow of the different steps in the technique is shown in Figure 15: 25 KRAS 13 3 10.0% KRAS 61 2 6.7% KRAS 117 1 3.3% Total mutations identified in NRAS NRAS 12 3 10.0% NRAS 13 1 3.3% NRAS 61 6 20.0% Mutations identified in one codon KRAS 12 9 30.0% KRAS 13 2 6.6% KRAS 61 1 3.3% NRAS 13 1 3.3% NRAS 61 1 3.3% Mutations identified in more than one codon KRAS 12 + KRAS 117 1 3.3% KRAS 12 + NRAS 61 2 6.6% KRAS 12 + NRAS 12 + NRAS 61 1 3.3% KRAS 13 + NRAS 12 + NRAS 61 1 3.3% KRAS 12 + KRAS 61 + NRAS 12 + NRAS 61 1 3.3% As expected, KRAS codon 12 was the most frequently mutation found in our cohort. In these cases, no information on the specific base substitution was available, because BEAMing is only able to distinguish between the WT form and all the MUT forms for each codon. Mutant fractions obtained in the MUT cases ranged from 0.004% (case 11), which is extremely low, to 25.045% in case 6. Notably, case 11 was one of the three cases with especially high cfDNA yield (25.8 ng/µL); this probably allowed us to detect mutations present at extremely low concentration in cfDNA. In Figure 17 some examples of high (Fig 17 A) and low (Fig 17 B and C) mutant fraction are shown. The cutoff for KRAS codon 12 mutations was established at 40 mutant beads. As seen in Figure 17, case 15 (Fig 17 B) was just over the cutoff value (56 mutated beads), whereas case 3 (Fig 17 C) was just below it (22 mutated beads). In fact, case 3 was a discordant case, in which tissue sample was informed as MUT by pyrosequencing (KRAS Gly12Val mutation), although it showed a relatively low mutant fraction (10%). 26 Figure 17: Examples of BEAMing results. A) Case 8, KRAS codon 12 MUT. B) Case 15, KRAS codon 12 MUT. C) Case 3, KRAS codon 12 WT. 27 Table 8 contains a summary of all the results obtained for RAS mutational analysis in ctDNA by BEAMing technology: Table 8: Summary of the results for RAS plasma analysis by BEAMing (n=30). Case OncoBEAM RAS [cfDNA] (ng/µL) Mutant beads fraction (%) (same order) 1 WT 0.370 2 KRAS 13 0.554 49.442 3 WT 1.640 4 KRAS 12 1.760 28.042 5 WT 0.782 6 KRAS 12 NRAS 12 NRAS 61 1.980 25.045 0.022 0.007 7 KRAS 12 0.174 2.124 8 KRAS 12 0.876 19.032 9 KRAS 12 0.296 0.791 10 KRAS 12 NRAS 61 1.880 23.254 0.013 11 KRAS 13 NRAS 12 NRAS 61 25.800 10.926 0.013 0.004 12 WT 0.262 13 KRAS 13 54.600 5.329 14 WT 0.200 15 KRAS 12 KRAS 117 0.446 0.072 0.264 16 KRAS 12 0.300 13.056 17 KRAS 12 0.666 10.515 18 WT 0.168 19 KRAS 12 0.294 6.575 20 WT 0.362 21 KRAS 12b KRAS 61 NRAS 12 NRAS 61 3.820 0.013 0.016 0.018 0.021 22 Invalida 0.224 23 KRAS 12 0.230 9.843 24 WT 6.640 28 25 NRAS 61 0.526 0.262 26 KRAS 12 0.504 10.498 27 KRAS 12 NRAS 61 10.500 8.061 0.006 28 NRAS 13 87.600 29.580 29 WT 0.326 30 KRAS 61 0.226 0.122 a Sample from case 22 reported invalid resultsdue to low DNA amount. b KRAS 12 mutation was conditioned because the beads were too dispersed. Of interest, we found 6 MUT cases displaying coexistent RAS mutations in ctDNA (Table 8; cases: 6, 10, 11, 15, 21, and 27). This was not entirely unexpected due to the high sensitivity of BEAMing, but the biological and therapeutic relevance of these mutated subclones that are present in very low proportion needs to be further investigated. May be, is exactly in this point where the advantages of the high-sensitivity blood-based RAS testing platforms will have more relevance in the clinical practice, since it is very well known that in metastatic patients, a single tumor tissue biopsy may not represent the evolving RAS mutational status of the disease. In patients whose tumours show extensive heterogeneity, a comprehensive surveying of RAS status by testing multiple tissue samples would be desirable; however, this is impractical and unfeasible. Instead, serial blood sampling and BEAMing analysis of RAS status could be the solution. Finally, it is worth pointing out that up until now, almost all studies that included BEAMing analysis had been performed centralized in the Sysmex facilities in Hamburg. The Molecular Oncology Laboratory at FIHGUV is one of the first laboratories using the BEAMing technology out of the central laboratory of Sysmex, , as part of a pilot project conducted on 8 Spanish centers. 4.3. Concordance between plasma and tissue RAS mutational status Determinations of RAS mutational status in ctDNA by BEAMing and in tissue by pyrosequencing have been summarized in Table 9. Table 9: Summary of RAS mutational status as determined in tumour tissue by pyrosequencing, and in plasma ctDNA by BEAMing. Case RAS tissue RAS ctDNA Case RAS tissue RAS ctDNA 1 WT WT 16 KRAS G12V KRAS G12A KRAS 12 2 KRAS G13D KRAS 13 17 KRAS G12D KRAS 12 3 KRAS G12V WT 18 WT WT 4 KRAS G12V KRAS 12 19 KRAS G12V KRAS 12 5 KRAS G12D WT 20 WT WT 6 KRAS G12V KRAS 12 21 WT KRAS 12b 29 NRAS 12 NRAS 61 KRAS 61 NRAS 12 NRAS 61 7 KRAS G12C KRAS 12 22 KRAS G13D NAa 8 KRAS G12S KRAS 12 23 KRAS G12V KRAS 12 9 KRAS G12D KRAS 12 24 WT WT 10 KRAS G12V KRAS 12 NRAS 61 25 NRAS Q61R NRAS 61 11 KRAS G13D KRAS 13 NRAS 12 NRAS 61 26 KRAS G12V KRAS 12 12 WT WT 27 KRAS G12A KRAS 12 NRAS 61 13 KRAS G13D KRAS 13 28 NRAS G13R NRAS 13 14 WT WT 29 WT WT 15 KRAS G12C KRAS 12 KRAS 117 30 KRAS Q61H KRAS 61 Of the 30 paired tissue/plasma samples analysed, there was one case that had to be excluded from the analysis due to low concentration of DNA in plasma, which resulted in invalid RAS genotyping by BEAMing. Of the 29 remaining cases, 21 were RAS MUT in FFPE (72.4%) whereas in ctDNA we found 20 RAS MUT cases (69.0%). RAS mutation positive samples were called above mutant allelic fraction thresholds of 0.02% in plasma and 5% in tissue. The agreement between BEAMing system in plasma samples and the determination by pyrosequencing in tissue samples was estimated by calculating the raw agreement and performing the concordance test. Overall agreement was obtained, along with positive and negative agreement (Table 10) Table 10: Concordance of plasma and tissue for RAS mutational status. Tissue RAS status Plasma RAS status Positive (Mut) Negative (WT) Total Positive (Mut) 19 1 20 Negative (WT) 2 7 9 Total 21 8 29 Overall agreement = 0.896  89.6 % Positive agreement = 0.905  90.5 % Negative agreement = 0.875  87.5 % 30 These results show a good concordance between BEAMing and pyrosequencing, with an overall agreement (OA) of 89.6%. Although there are few studies in the literature, concordance between liquid biopsy, using highly sensitive methods such as digital PCR and tumour genotyping have usually resulted in concordance. One example is a report using datasets with 76 paired tissue-blood samples from two clinical trials (OPUS and CRYSTAL), where OA between tissue (RAS DNA sequencing of FFPE tumour samples) and plasma (BEAMing RAS 33 Mutation Panel) was 93.4% (Jones et al. 2015). In other tumours, like breast and lung cancer, Higgins et al found a 100% OA in the retrospective assessment of PI3KCA mutations in breast cancer samples (BEAMing ctDNA vs. sequencing tumour tissue); in the prospective analysis, however, only a 72.5% of concordance was achieved (Higgins et al. 2012). Finally, Karlovich and colleagues found an OA of 67% when assessing T790M mutation by cobas® Tumour test and BEAMing plasma ctDNA on non-small cell lung cancer samples (Karlovich et al. 2016). Although highly concordant, we had 3 discordant cases, which are summarized in Table 11: Table 11: Discordant cases summary. Case Tumour site Plasma result Tissue result #3 Recto-sigmoid junction WT KRAS exon 2 G12V #5 Transverse colon WT KRAS exon 2 G12D #21 Rectum KRAS 12, 61 + NRAS 12, 61 WT Among the two patients in which RAS mutation was identified in tissue but not in plasma, we investigated possible causes. First, we re-examined FFPE samples by pyrosequencing, confirming previous results. Second, preanalytical procedures for plasma processing could also explain the lack of concordance. Since this is a retrospective study, characteristics of the collection tube, time from collection until processing samples and plasma storage conditions might have affected the yield and/or quality of ctDNA obtained. In both cases (#3 and #5), blood samples were collected in standard K2-EDTA tubes without any nucleic-acid stabilizing agent. Apart from this, samples were stored at -80 °C for more than 5years. Third, there is also a possible correlation between tumour burden, number and location of metastatic disease lesions and the amount of plasma DNA recovered. Interestingly, case #3 had a unique metastatic lesion (low tumour burden), even though the amount of plasma obtained was in the median yield. On the contrary, case #5 was a patient with high tumour burden at the time of diagnosis (primary tumour and several liver metastasis, the greater measuring more than 90 mm). Consequently, we would expect higher ctDNA levels, but in this case the yield was half of that obtained in case #3. 31 Standardized methods are still needed in order to minimize their impact on mutation detection rates (El Messaoudi et al. 2013). Some important factors are:  Processing the blood in 4-6 hours after drawing is essential, because half-life of ctDNA in circulation is between 16 minutes and 2 hours (Diehl et al. 2008).  Use of Cell-Free DNA™ Blood Collection Tubes (Streck, Omaha, NE) is advantageous over other collection tubes: they contain a formaldehyde-free preservative that prevents white blood cells from breaking, avoiding WT DNA contamination up to 72 hours post-collection (Xue et al. 2009; Sherwood et al. 2016). The other discordant case was RAS MUT in plasma but no mutation was detected on FFPE sample. Of interest, in this case (#21) RAS genotyping was performed on a FFPE sample corresponding to a liver metastasis, since primary tumour tissue was not available. Figure 18 shows the plots obtained for all four mutations (A-D): Figure 18: Plots for all four mutations detected in ctDNA of case 21, with number of mutated beads and mutant fraction. Differences in RAS mutation status between plasma and tissue may be attributed to intra or inter-tumour molecular heterogeneity. For instance, RAS WT in tissue but MUT in plasma, such as in case 21, may arise in patients having heterogeneous distribution of RAS mutant clones in 32 the primary tumour and/or in the metastases, which are not represented in the analysed tissue sample but are detected systemically by the plasma test. Interestingly, patient #21 was treated with a combination of FOLFOX with anti-EGFR agent panitumumab for 7 cycles. No response was observed, only achieving disease stabilization. The treatment was stopped and four months later, the patient progressed and a second line treatment was initiated. So, these apparently discordant results between plasma and tissue make sense when we look closely at the clinical evolution of the patient. Probably, if we had followed the patient with serial blood sampling we could have detected an increase in the mutant fraction during treatment. Although not a discordance, there were 5 cases in which BEAMing detected the mutation observed by pyrosequencing and additional mutations present in low proportion in ctDNA. The relevance of these results is still unknown, and they should be checked; analysing tissue samples using BEAMing technology could reveal if those mutations were present in the tissue in such low proportion that pyrosequencing could not detect them. This is being performed right now on our laboratory. Finally, case #22 did not yield any results because of low DNA concentration and subsequent failure of emulsion PCR. The concentration of cfDNA after nucleic acid extraction was 0.224 , which is just over the value we estimate to be sufficient to detect at least one mutation in ctDNA. Therefore, the best course of action would be to repeat cfDNA extraction from the same plasma sample (if still possible), increasing the amount of input plasma: there is evidence that increasing the amount of input plasma can improve ctDNA mutation detection if the processing of the plasma is optimal (Sherwood et al. 2016). Clinical validation of BEAMing has arrived first to tumour tissue analysis. Over 1200 patients from different clinical trials had tumour tissue samples tested using the BEAMing platform in a retrospective manner:  In OPUS trial sample set (Bokemeyer et al. 2015), mutation frequency was broadly concordant with those reported in similar studies using pyrosequencing (Douillard et al. 2013; Schwartzberg et al. 2014; Stintzing et al. 2012).  In the CRYSTAL study, RAS mutations were found in 14.7% of evaluable patients.  Finally, results from the subgroup analysis of CALGB/SWOG 80405 have not been published yet. Preliminary results show that new RAS mutations were identified in 15.3% of analysed patients. In conclusion, liquid biopsy could hold the key to earlier detection and treatment of relapsed disease, and ultimately improve the outcome of a patient. As blood serial sampling is much less invasive and safe than metastasis or tumour biopsies, ctDNA analysis for RAS mutational 33 status assessment represents a potential surrogate for solid biopsies. Therefore, sensitive techniques such as BEAMing system able to detect minimal quantities of ctDNA carrying actionable mutations are desirable and will become an essential tool for molecular oncology diagnostics. Moreover, BEAMing technique is versatile, able to analyse DNA from both plasma and tissue samples: an invaluable resource for molecular oncology researchers and oncologists. 4.4. Association of RAS mutational status with clinico-pathological characteristics We analysed association between clinicopathological variables and RAS mutational status as determined by both techniques. RAS status was dichotomized as follows: "RAS WT" (meaning no detection of mutated DNA) and "Any RAS MUT" (meaning that at least one KRAS or NRAS mutation was detected). First, results from tumour tissue pyrosequencing are shown in Table 12; no statistically significant correlation was found between RAS mutational status in tumour tissue and these characteristics: Table 12: Correlation analysis between RAS mutational status (tumour tissue) and clinicopathological variables of interest in CRC using Mann-Whitney U test. Characteristics All (n=30) All RAS WT (n=8; 26.7%) Any RAS mutation (n=22; 73.3%) pvalue Age at diagnosis (years) Median [range] 68 [43-83] 67 [43-71] 69 [47-83] *0.270 Sex Man 20 (66.7%) 6 (20%) 14 (46.7%) †0.682 Woman 10 (33.3%) 2 (6.7%) 8 (26.7%) CEA (ng/mL) at diagnosisa Median [range] 31 [2.77059.5] 19.9 [6.51000] 36.1 [2.77059.5] *1.000 Primary tumour localization Ascending colon 6 (20%) 1 (3.3%) 5 (16.7%) †0.650 Transverse colon 3 (10%) 0 (0%) 3 (10%) Descending colon 8 (26.7%) 2 (6.7%) 6 (20%) Rectum 13 (43.3%) 5 (16.7%) 8 (26.7%) Primary tumour histologyb ADC 24 (85.7%) 7 (25%) 17 (60.7%) †1.000 MUC 4 (14.3%) 1 (3.6%) 3 (10.7%) Grade of differentiation Well diff. 9 (45%) 2 (10%) 7 (35%) †1.000 Moderately diff. 10 (50%) 3 (15%) 7 (35%) Poorly diff. 1 (5%) 0 (0%) 1 (5%) Tumour size and invasiveness (T) T2 2 (11.8%) 1 (5.9%) 1 (5.9%) †1.000 T3 8 (47.1%) 3 (17.6%) 5 (29.4%) T4 7 (41.2%) 3 (17.6%) 4 (23.5%) Regional lymph N0 6 (33.3%) 2 (11.1%) 4 (22.2%) †0.350 34 nodes involvement (N) N1 6 (33.3%) 4 (22.2%) 2 (11.1%) N2 6 (33.3%) 1 (5.6%) 5 (27.8%) Lymphatic invasion (L) L0 9 (64.3%) 2 (14.3%) 7 (50%) †0.580 L1 5 (35.7%) 2 (14.3%) 3 (21.4%) Vascular invasion (V) V0 13 (86.7%) 4 (26.7%) 9 (60%) †1.000 V1 2 (13.3%) 1 (6.7%) 1 (6.7%) Perineural invasion (Pn) Pn0 12 (85.7%) 4 (28.6%) 8 (57.1%) †1.000 Pn1 2 (14.3%) 0 (0%) 2 (14.3%) Liver metastasis No 7 (23.3%) 3 (10%) 4 (13.3%) †0.345 Yes 23 (76.7%) 5 (16.7%) 18 (60%) Organs affected by metastasis 1 organ 14 (46.7%) 6 (75%) 6 (31.6%) †0.101 >1 organ 16 (53.3%) 2 (25%) 13 (68.4%) *Kruskal-Wallis test; †Mann-Whitney U test. a CEA levels only available for 17 patients. b ADC = Adenocarcinoma; MUC = Mucinous adenocarcinoma. Second, the results obtained from RAS mutational status by ctDNA analysis are shown in Table 13: Table 13: Correlation analysis between RAS mutational status and clinicopathological variables of interest in CRC using Mann-Whitney U test. Characteristics All (n=29) All RAS WT (n=9; 31%) Any RAS mutation (n=20; 69%) p-value Age at diagnosis (years) Median [range] 68 [43-83] 68 [43-81] 69 [47-83] *0.850 Sex Man 20 (69%) 6 (20.7%) 14 (48.3%) †1.000 Woman 9 (31%) 3 (10.3%) 6 (20.7%) CEA (ng/mL) at diagnosisa Median [range] 36.1 [2.77059.5] 17.25 [2.7-1000] 41.45 [2.77059.5] *0.302 Primary tumour localization Ascending colon 6 (20.7%) 1 (3.4%) 5 (17.2%) †0.933 Transverse colon 3 (10.3%) 1 (3.4%) 2 (6.9%) Descending colon 7 (24.1%) 2 (6.9%) 5 (17.2%) Rectum 13 (44.8%) 5 (17.2%) 8 (27.6%) Primary tumour histologyb ADC 24 (88.9%) 8 (29.6%) 16 (59.3%) †1.000 MUC 4 (11.1%) 1 (3.7%) 2 (7.4%) Grade of differentiation Well diff. 9 (47.4%) 3 (15.8%) 6 (31.6%) †1.000 Moderately diff. 9 (47.4%) 3 (15.8%) 6 (31.6%) Poorly diff. 1 (5.3%) 0 (0%) 1 (5.3%) Tumour size and invasiveness (T) T2 2 (12.5%) 0 (0%) 2 (12.5%) †0.621 T3 8 (50%) 4 (25%) 4 (25%) T4 6 (37.5%) 3 (18.8%) 3 (18.8%) 41 pp.1659–1665. Jones, F.S. et al., 2015. Blood-based detection of RAS mutations to guide anti-EGFR therapy in CRC patients: Concordance of results from circulating tumor DNA and tissue-based RAS testing. European Journal of Cancer, (Special issue). Karapetis, C.S. et al., 2008. K-ras mutations and benefit from cetuximab in advanced colorectal cancer. The New England Journal of Medicine, 359(17), pp.1757–1765. Karapetis, C.S. et al., 2014. PIK3CA, BRAF, and PTEN status and benefit from cetuximab in the treatment of advanced colorectal cancer-results from NCIC CTG/AGITG CO.17. Clinical Cancer Research, 20(3), pp.744–753. Karlovich, C. et al., 2016. Assessment of EGFR Mutation Status in Matched Plasma and Tumor Tissue of NSCLC Patients from a Phase I Study of Rociletinib (CO-1686). Clinical Cancer Research, 22(10), pp.2386–2395. Korkaya, H., Liu, S. & Wicha, M.S., 2011. Breast cancer stem cells , cytokine networks , and the tumor microenvironment. The Journal of Clinical Investigation, 121(10), pp.3804–3809. Kotoula, V. et al., 2009. Targeted KRAS mutation assessment on patient tumor histologic material in real time diagnostics. PLoS ONE, 4(11), p.e7746. Kuo, Y.-B. et al., 2014. Comparison of KRAS mutation analysis of primary tumors and matched circulating cell-free DNA in plasmas of patients with colorectal cancer. Clinica chimica acta; international journal of clinical chemistry, 433, pp.284–289. Lauring, J. & Park, B.H., 2011. BEAMing sheds light on drug resistance. Clinical Cancer Research, 17(24), pp.7508–7510. Leon, S.A. et al., 1977. Free DNA in the serum of cancer patients and the effect of therapy. Cancer Research, 37(3), pp.646–650. Mandel, P. & Metais, P., 1948. Les acides nucléiques du plasma sanguin chez l’homme. Comptes rendus des seances de la Societe de biologie et de ses filiales, 142(3-4), pp.241– 243. Marisa, L. et al., 2013. Gene expression classification of colon cancer into molecular subtypes: characterization, validation, and prognostic value. PLoS Medicine, 10(5), p.e1001453. McBride, D.J. et al., 2010. Use of cancer-specific genomic rearrangements to quantify disease burden in plasma from patients with solid tumors. Genes, Chromosomes & Cancer, 49(11), pp.1062–1069. El Messaoudi, S. et al., 2013. Circulating cell free DNA: Preanalytical considerations. Clinica Chimica Acta; International Journal of Clinical Chemistry, 424, pp.222–230. Misale, S. et al., 2014. Emergence of KRAS mutations and acquired resistance to anti EGFR therapy in colorectal cancer. Nature, 486(7404), pp.532–536. Modest, D.P. et al., 2011. Clinical characterization of patients with metastatic colorectal cancer depending on the KRAS status. Anti-cancer Drugs, 22(9), pp.913–918. Mohan, S. et al., 2014. Changes in colorectal carcinoma genomes under anti-EGFR therapy identified by whole-genome plasma DNA sequencing. PLoS Genetics, 10(3), p.e1004271. 42 Ogino, S. et al., 2009. KRAS mutation in stage III colon cancer and clinical outcome following intergroup trial CALGB 89803. Clinical Cancer Research, 15(23), pp.7322–7329. Ogino, S. et al., 2005. Sensitive sequencing method for KRAS mutation detection by Pyrosequencing. The Journal of Molecular Diagnostics, 7(3), pp.413–421. Patwardhan, M.B. et al., 2006. Cancer care quality measures: diagnosis and treatment of colorectal cancer. Evid Rep Technol Assess, (138), pp.1–116. Perez-Villamil, B. et al., 2012. Colon cancer molecular subtypes identified by expression profiling and associated to stroma, mucinous type and different clinical behavior. BioMed Central Cancer, 12, p.260. Peters, D.L. & Pretorius, P.J., 2012. Continuous adaptation through genetic communication - a putative role for cell-free DNA. Expert Opinion on Biological Therapy, 12 Suppl 1, pp.S127–32. Petrosino, J.F. et al., 2009. Metagenomic pyrosequencing and microbial identification. Clinical Chemistry, 55(5), pp.856–866. Richman, S.D. et al., 2009. KRAS and BRAF mutations in advanced colorectal cancer are associated with poor prognosis but do not preclude benefit from oxaliplatin or irinotecan: results from the MRC FOCUS trial. Journal of Clinical Oncology, 27(35), pp.5931–5937. Roepman, P. et al., 2014. Colorectal cancer intrinsic subtypes predict chemotherapy benefit, deficient mismatch repair and epithelial-to-mesenchymal transition. International Journal of Cancer, 134(3), pp.552–562. Rustgi, A.K., 1994. Hereditary gastrointestinal polyposis and nonpolyposis syndromes. The New England Journal of Medicine, 331(25), pp.1694–1702. Sadanandam, A. et al., 2013. A colorectal cancer classification system that associates cellular phenotype and responses to therapy. Nature Medicine, 19(5), pp.619–625. Sahin, I.H. et al., 2013. Rare though not mutually exclusive: A report of three cases of concomitant kras and braf mutation and a review of the literature. Journal of Cancer, 4(4), pp.320–322. Sartore-Bianchi, A. et al., 2009. PIK3CA mutations in colorectal cancer are associated with clinical resistance to EGFR-targeted monoclonal antibodies. Cancer Research, 69(5), pp.1851–1857. Schlicker, A. et al., 2012. Subtypes of primary colorectal tumors correlate with response to targeted treatment in colorectal cell lines. BioMed Central Medical Genomics, 5, p.66. Schmitt, M.W. et al., 2012. Detection of ultra-rare mutations by next-generation sequencing. Proceedings of the National Academy of Sciences of the United States of America, 109(36), pp.14508–14513. Schwartzberg, L.S. et al., 2014. PEAK: a randomized, multicenter phase II study of panitumumab plus modified fluorouracil, leucovorin, and oxaliplatin (mFOLFOX6) or bevacizumab plus mFOLFOX6 in patients with previously untreated, unresectable, wildtype KRAS exon 2 metastatic colorectal. Journal of Clinical Oncology, 32(21), pp.2240– 2247. 43 Sherwood, J.L. et al., 2016. Optimised pre-analytical methods improve KRAS mutation detection in circulating tumour DNA (ctDNA) from patients with non-small cell lung cancer (NSCLC). PLoS ONE, 11(2), pp.1–14. Siravegna, G. & Bardelli, A., 2014. Genotyping cell-free tumor DNA in the blood to detect residual disease and drug resistance. Genome Biology, 15(8), p.449. Sociedad Española de Oncología Médica, 2016. Las Cifras del Cáncer en España 2016. Sociedad Española de Oncología Médica, pp.1–24. Sorich, M.J. et al., 2015. Extended RAS mutations and anti-EGFR monoclonal antibody survival benefit in metastatic colorectal cancer: a meta-analysis of randomized controlled trials. Annals of Oncology, 26(1), pp.13–21. De Sousa E Melo, F. et al., 2013. Poor-prognosis colon cancer is defined by a molecularly distinct subtype and develops from serrated precursor lesions. Nature Medicine, 19(5), pp.614–618. Spindler, K.L. et al., 2012. Quantitative cell-free DNA, KRAS, and BRAF mutations in plasma from patients with metastatic colorectal cancer during treatment with cetuximab and irinotecan. Clinical Cancer Research, 18(4), pp.1177–1185. Spindler, K.-L.G. et al., 2015. Clinical utility of KRAS status in circulating plasma DNA compared to archival tumour tissue from patients with metastatic colorectal cancer treated with anti-epidermal growth factor receptor therapy. European Journal of Cancer, 51(17), pp.2678–2685. Stintzing, S. et al., 2012. FOLFIRI plus cetuximab versus FOLFIRI plus bevacizumab as first-line treatment for patients with metastatic colorectal cancer-subgroup analysis of patients with KRAS: Mutated tumours in the randomised German AIO study KRK-0306. Annals of Oncology, 23(7), pp.1693–1699. Stintzing, S. et al., 2009. The treatment of colorectal carcinoma with monoclonal antibodies: the importance of KRAS mutation analysis and EGFR status. Deutsches Ärzteblatt international, 106(12), pp.202–206. Sundström, M. et al., 2010. KRAS analysis in colorectal carcinoma: analytical aspects of Pyrosequencing and allele-specific PCR in clinical practice. BioMed Central Cancer, 10, p.660. Tabernero, J. et al., 2015. Analysis of circulating DNA and protein biomarkers to predict the clinical activity of regorafenib and assess prognosis in patients with metastatic colorectal cancer: a retrospective, exploratory analysis of the CORRECT trial. The Lancet Oncology, 16(8), pp.937–948. The Cancer Genome Atlas Network, 2012. Comprehensive molecular characterization of human colon and rectal cancer. Nature, 487(7407), pp.330–337. Thierry, A.R., 2016. A Targeted Q-PCR-Based Method for Point Mutation Testing by Analyzing Circulating DNA for Cancer Management Care. Methods in Molecular Biology, 1392, pp.1–16. Tsiatis, A.C. et al., 2010. Comparison of Sanger Sequencing, Pyrosequencing, and Melting Curve Analysis for the Detection of KRAS Mutations. The Journal of Molecular Diagnostics, 12(4), pp.425–432. 44 Vogelstein, B. et al., 1988. Genetic alterations during colorectal-tumor development. The New England Journal of Medicine, 319(9), pp.525–532. Xue, X. et al., 2009. Optimizing the yield and utility of circulating cell-free DNA from plasma and serum. Clinica chimica acta; international journal of clinical chemistry, 404(2), pp.100–104. Yaeger, R. et al., 2015. RAS mutations affect pattern of metastatic spread and increase propensity for brain metastasis in colorectal cancer. Cancer, 121(8), pp.1195–1203. 45 7. APPENDICES 7.1. Communications derived from this study VI Meeting of Young Researchers (RTICC). Salamanca, Spain 2016. Mutational analysis of RAS genes in metastatic colorectal cancer: Concordance between tissue and liquid biopsy. Borja Lafuente Gutiérrez, Alejandro Herreros Pomares, Silvia Calabuig-Fariñas, Marais Mosqueda, Eva Escorihuela, Sandra Gallach, María José Safont, Eloisa Jantus-Lewintre, Carlos Camps. BACKGROUND: Mutations in proto-oncogenes RAS (KRAS and NRAS) are routinely tested in metastatic colorectal cancer (mCRC). Tumors with activating mutations in RAS genes do not respond to anti-EGFR targeted therapy (cetuximab and panitumumab). In consequence, these drugs are restricted to patients with RAS wild-type tumors. Eventually tumors develop resistance by selection of RAS mutant subclones. Thus, serial sampling for mutational analysis is desirable, but it entails multiple biopsies (invasive and expensive). Liquid biopsy systems, like BEAMing, allow for continuous mutational analysis in a non-invasive, ultrasensitive manner. OBJECTIVE: We aim to compare our standard technique for RAS assessment in tumor tissue, Pyrosequencing, with a "liquid biopsy" approach based on ctDNA analysis: BEAMing system. METHODS: We used TheraScreen® KRAS Pyro Kit and RAS Extension Pyro Kit for pyrosequencing of tumor tissue DNA, and Sysmex® OncoBEAM CRC RAS kit for BEAMing ctDNA analysis. Concordance was determined by the number of cases reported as “mutant” or “wildtype/no mutation detected” in each system. RESULTS: We tested RAS mutational status in tumor tissue and plasma samples of 30 patients. We found an overall agreement of 89.6% (Table 1), with three discordant cases (Table 2). One case could not be analyzed by BEAMing because of low DNA input: Table 3: Concordance table of plasma and tissue results for RAS mutational status. Tissue RAS status Plasma RAS status Positive (Mut) Negative (WT) Total Positive (Mut) 19 1 20 Negative (WT) 2 7 9 Total 21 8 29 46 Table 4: Summary of discordant cases, type of discordance, plasma and tissue results. Case Discordant type Plasma result Tissue result #3 False Negative WT KRAS G12V #5 False Negative WT KRAS G12D #21 False Positive KRAS 12, 61 + NRAS 12, 61 WT The false negatives were investigated (Table 2), pointing towards ctDNA degradation over time and preanalytical suboptimal processing as the main factors involved. In case 21, the presence of RAS mutant ctDNA in plasma before anti-EGFR treatment could have predicted early disease progression, which occurred just four months after treatment. Probably, RAS mutant subclones began to proliferate and sustain the tumor, generating the secondary resistance to anti-EGFR treatments. CONCLUSIONS: Our results indicate that BEAMing analysis of RAS mutations in ctDNA has a high concordance rate when compared to pyrosequencing tumor tissue. Therefore, BEAMing is an optimal technique for molecular diagnosis of RAS mutational status of mCRC patients. In addition, it is very useful in disease monitoring as a non-invasive, specific method for detecting low prevalence RAS mutations in plasma, allowing earlier interventions to modify and improve treatments. This project was supported by [RD12/0036/0025] from RTICC.