In recent years, cryo-electron microscopy (cryo-EM) has become a practical and effective method of determining structures at previously unattainable resolutions due to advances in detection, automation, and data processing. However, sample preparation remains a major bottleneck in the cryo-EM workflow. Even after the arduous process of biochemical sample optimization, it often takes several iterations of grid vitrification and screening to determine the optimal grid freezing parameters that yield suitable ice thickness and particle distribution for data collection. Since a high-quality sample is imperative for high-resolution structure determination, grid optimization is a vital step. For researchers who rely on cryo-EM facilities for grid screening, each iteration of this optimization process may delay research progress by a matter of months. Therefore, a more strategic and efficient approach should be taken to ensure that the grid optimization process can be completed in as few iterations as possible. Here, we present an implementation of Design of Experiments (DOE) to expedite and strategize the grid optimization process. A Fractional Factorial Design (FFD) guides the determination of a limited set of experimental conditions which can model the full parameter space of interest. Grids are frozen with these conditions and screened for particle distribution and ice thickness. Quantitative scores are assigned to each of these grid characteristics based on a qualitative rubric. Input conditions and response scores are used to generate a least-squares regression model of the parameter space in JMP, which is used to determine the conditions which should, in theory, yield optimal grids. Upon testing this approach on apoferritin and L-glutamate dehydrogenase on both the Vitrobot Mark IV and the Leica GP2 plunge freezers, the resulting grid conditions reliably yielded grids with high-quality ice and particle distribution that were suitable for collecting large overnight datasets on a Krios. We conclude that a DOE-based approach is a cost-effective and time-saving tool for cryo-EM grid preparation.

Introduction European guidelines suggest that obesity should prompt screening to predict steatosis, steatohepatitis and fibrosis (EASL 2016). Bariatric surgery has been shown to have a favourable effect on non-alcoholic fatty liver disease (Lee et al. 2019), and screening patients for advanced fibrosis before surgery is suggested as a method of reducing complications and determining the appropriate surgical approach (Antipass et al. 2020). Regular attendance of a Hepatologist at our Tier 4 Bariatric multidisciplinary meeting (MDM) began in September 2022. Methods Patients discussed at the MDM between October and December 2022 were screened for significant fibrosis with a FIB-4 score. Patients with a FIB-4 score greater than or equal to 1.3, or greater than or equal to 2 if aged 65 or more, were invited to have a FibroScan. Patients with a median liver stiffness of 8 kPa or above were booked an outpatient appointment with Hepatology. Data was collected on the interval between the blood tests needed to calculate a FIB-4 score and when discussed in MDM. Results 126 patients were discussed at the MDM between October and December 2022. 25 patients (20%) had not had the blood tests required to calculate a FIB-4 score. When a FIB-4 could be calculated, 48 (48%) and 70 (69%) patients had the required blood tests in the last 1 and 2 years respectively. The median interval was 12 months, with an interquartile range of 2 to 32.5 months. 14 (14%) patients had elevated FIB-4 scores. 3 patients were already known to our hepatology service. 1 with cirrhosis secondary to hepatitis C virus and 2 with cirrhosis secondary to NAFLD. The remaining 11 were invited for a FibroScan. The majority of patients are still awaiting their FibroScan. of the 5 that have had a FibroScan, 1 has a median liver stiffness over 8 kPa and has been referred for hepatology input. Conclusions Over 40% of patients had no screening within the last two years. Given the significant risk of NAFLD in patients living with obesity, this is a missed opportunity to identify patients with potentially significant liver disease. Their presence in the MDM means they are receiving support in reducing weight, which will help with NAFLD, although we may be missing the chance to monitor for or prevent the complications of significant fibrosis. We cannot quantify the full impact of closer hepatology involvement in the MDM until more time has passed. The short period is unlikely to have significantly impacted behaviours. Until the remainder of FibroScans are performed, we do not know how successful we have been in finding cases of significant liver disease. We hope that after a longer period of data collection, we may see screening improve and have a better idea of the burden of liver disease within our MDM.

In the field of Structural Biology, Atom Probe Tomography (APT) is in the nascent stages of development wherein coarse-grained visuals of proteins have been captured. However, no aspect of protein structure has been determined. The characterization of organic samples or biomolecules through the technique is currently limited to the detection of a few dominant signatures. The problem of indecipherable characterization can inherently be traced back to multiple forms of technique-specific responses to organic samples and consequent triggers leading to organic-sample and sample-medium interactions. As a first counter to the characterization problem, by taking the experiment-specific uncertainties, and probable APT-centric organic sample-based variabilities into account, the current results of the study are presented. Through mass-spectrometric and computational analysis, specific individual amino acids (Sulfur-containing protein-bound amino acids) in proteins and aspects of protein structure (probable backbone fragments) have been identified and characterized. Additionally, to overcome technique-specific deterrents to the characterization of biomolecules in cryogenic sample medium, the development of a protein-labeling strategy tailored to APT is suggested.

Introduction We established an early diagnostic service with Cytosponge to triage patients for Barrett's surveillance and symptomatic acid reflux. It Aims to address COVID delays, identify disease, and establish a new sustainable service for earlier diagnosis. We now have the largest single site experience of Cytosponge in the UK. We report on the challenges establishing a new service and adopting new technology during uncertain times, and outcomes from the first 1000 procedures. Methods The service started November 2020. We trained 2 nurse endoscopists, 2 Band 6 nurses in Cytosponge delivery and follow up. There is administrative support and 2 consultant gastroenterologists lead the service. 2 patient cohorts:- Barrett's oesophagus on surveillance and patients with symptomatic acid reflux. Consultant led triage from waiting list, Barrett's surveillance and routine reflux referrals with screening telephone triage by nurse to confirm suitability. Cytosponge was delivered by nurses and a patient satisfaction survey completed. Cytology was analysed by Cyted and results relayed <4 weeks. Clinical decision making was pathway led.TFF3+ was used as marker for intestinal metaplasia, p53+ve and atypia for potential dysplasia. Patients were fast tracked to gastroscopy (atypia or p53+), routine gastroscopy (New TFF3+/persistent symptoms) ongoing Barrett's surveillance (according to risk factors) or discharged with advice and safety netting (Short segment Barrett's with -ve TFF3, uncomplicated reflux with -veTFF3) Patients were also referred to pH Manometry. Results 1000 patients agreed to Cytosponge. Adequate samples were reported in 91.2%. No adverse events. 12% reflux patients had +veTFF3, 5.6% confirmed new Barrett's. We identified 8.8% suspected dysplasia in Barrett's surveillance and in 0.9% reflux patients. Furthermore 11.5% have been discharged from surveillance and 72% with reflux to their GP. Those needing gastroscopy for persistent symptoms are reported elsewhere. The patient experience survey reports clear preference for Cytosponge over gastroscopy. Cytosponge has lower carbon footprint and prevented unnecessary gastroscopy in 670 patients. Challenges include ensuring financial stability and job security, encouraging new referral patterns and overcoming scepticism for change, improving success rates and adequacy of samples, securing sufficient administrative support and sufficient clinical lead time. There have been pathway changes as knowledge accrues. Conclusion We report on the largest single site series of Cytosponge and challenges establishing a new service. Our patients have welcomed Cytosponge and benefited from a speedier streamlined pathway and detection of early pathology.

Introduction During the COVID-19 pandemic gastroscopy was halted with large burden on recovery and risk of late diagnosis. We established an early diagnosis service using Cytosponge to triage patients to timely gastroscopy and management Methods 2 patient cohorts were used: 1. Barrett's oesophagus (BO) on endoscopic surveillance 2. Patients with symptoms of gastro-oesophageal reflux referred for routine gastroscopy. Exclusions were patients with dysplasia on last gastroscopy, fundoplication, pregnancy and patient preference. Triage of patients on the waiting list was from endoscopy referrals, Barrett's surveillance database and telephone triage. 2 research nurses and 2 clinical nurse specialists were trained in delivering Cytosponge. A patient satisfaction survey was completed. All cytology specimens were analysed by Cyted. Results were relayed to patients within 4 weeks by consultant led nurse-run teleclinic and letter. Clinical triage was according to the table below. TFF3+ was used as a marker for intestinal metaplasia (IM), P53+ve and atypia for potential dysplastic change. Results 470 patients agreed to Cytosponge over 14 months November 2020-January 2022. 22 cancelled-mostly COVID related. 34 failed to swallow (5.5% of Barrett's, 8.9% reflux). of those successfully swallowing the sponge 6% were inadequate samples in Barrett's and 9% reflux. No major adverse events occurred. Conclusion We report on the largest single site series of Cytosponge in non-specialist clinical practice in England and its pragmatic use in patients management and pandemic recovery. Significant benefits in the Barrett's cohort were timely identification of dysplasia and those longer requiring surveillance. Benefits in the reflux group include identification of new BO, avoiding unnecessary gastroscopy and early discharge. Overall reduced endoscopy resulted in reduced cost, lower carbon footprint and improved patient experience. Careful follow up and longer-term outcomes will provide confidence to continue this new technique in routine clinical practice.

Introduction Up to 6% patients with reflux may have underlying Barrett's Oesophagus (BO) but only 20% patients with BO are diagnosed. The BEST3 study group established the value of Cytosponge as a non-invasive tool that picks up 10x more BO than routine practice in a primary care setting but this has not been assessed in secondary care. We established an early diagnosis and dyspepsia service during the pandemic to aid early diagnosis, management and appropriate triage to gastroscopy. Methods All patients waiting for direct access gastroscopy or referred to the clinical service with reflux symptoms were assessed in a nurse-run clinic and offered Cytosponge as a first test. We excluded patients with known BO, cancer, previous fundoplication and strictures. Patients were enrolled in the DELTA study or the ongoing NHSE pilot study. Patients found to be trefoil factor 3 (TFF3) +ve, a marker for intestinal metaplasia (IM), had routine endoscopy. Patients with atypia, p53+ve or dysplasia had 2ww gastroscopy. Patients with other significant findings or symptoms were offered gastroscopy and investigation as appropriate. All patients were given lifestyle and management advice. Results over 14 months 235 patients had successful Cytosponge procedures with adequate sampling. 29 patients (12%) had positive or equivocal TFF3, 215 (88%) had negative tff3. There were no significant adverse effects. 3 (1.2%) patients had high risk biomarkers on Cytosponge: at gastroscopy 1 patient (p53 equivocal, atypia) had focal IM of the gastro-oesophageal junction but no BO/dysplasia. 1 Patient (p53+ atypia with inflammation) had a C4M5 BO biopsies showed regenerative atypia. 1 patient (focal p53+ squamous atypia) had a normal Lugol's assessment and ENT review and had helicobacter gastritis. All will have repeat gastroscopy within 1yr. 24/29 patients with TFF3+ have since had gastroscopy to date-13/24 (52%) had visible BO reported and will have ongoing surveillance as appropriate. This suggest around 6% of all patients had BO. 66% patients did not require gastroscopy, 22% have gone on to have diagnostic gastroscopy for ongoing symptoms,12% for positive TFF3. 154 (65.6%) of all patients with successful Cytosponge were discharged from follow up so far. Conclusions We report encouraging outcomes from the largest English site using Cytosponge in reflux patients in secondary care. Cytosponge detected pathology, aided decision making in a pandemic environment with limited resources and reduced unnecessary gastroscopy, giving environmental benefit. Our results support those reported by the BEST3 study. Longer term evaluation and cost-effectiveness on a National scale is in progress.

The apolipoprotein E e4 allele (APOE4) is a major risk factor, contributing to vascular pathologies in AD which include CAA, BBB dysfunction and reduced cerebral vascular integrity. The diminished capacity of APOE4 to transport docosahexaenoic acid (DHA), an essential fatty acid required for the structural and functional maintenance and vascular integrity of the brain, could also contribute to AD pathogenesis. However, it remains undetermined if changes in the brain DHA content of phospholipids (PL) and DHA transporters exist in relation to APOE4, AD diagnosis and CAA pathology. We performed liquid chromatography/mass spectrometry-based PL analysis of the cerebrovascular and parenchymal fractions from autopsied human brain of pathologically confirmed AD cases and controls. We performed an antibody-based examination of the major facilitator superfamily domain containing 2A (mfsd2a) protein in the cerebrovasculature from these subjects and brain homogenates from transgenic mice with human APOE (APOE-TR) and APOE-TR crossed with mice with five AD mutations (EFAD). We performed proteomic analyses of the cerebrovascular and parenchymal fractions of the brains from 50-week-old APOE-TR. In the cerebrovasculature and parenchyma, DHA containing PL species were lower in ε4 carriers with AD compared to control ε4 carriers. We observed an ε4-dependent decrease in mfsd2a level in the cerebrovasculature of ε4 carriers compared to non-carriers. Stratification of DHA containing PL by CAA showed that these PL levels were reduced in ε4 positive AD patients with severe CAA. In addition, brain levels of DHA containing PC species and mfsd2a were lower in APOE4-TR and E4FAD mice compared to other genotypes in each mouse model. Furthermore, protein levels of glucose transporter 1 (Glut1) and a metabolite transporter, monocarboxylate transporter 1 (MCT1), were lower in the cerebrovasculature of APOE4-TR relative to APOE3-TR and APOE2-TR mice. These findings demonstrate that brain DHA deficiencies in ε4 carriers may be due to reduced mfsd2a expression and partly associated with severe CAA. Thus, targeting this transport mechanism may improve the bioavailability of DHA into the brains of ε4 carriers who are at risk of developing AD.

Introduction Pandemic restrictions delayed endoscopic Barrett's surveillance programmes with a risk of late or missed dysplasia and cancer diagnoses. We established a new Cytosponge service and surveillance pathway to address this risk and aid triage of patients to timely gastroscopy alongside gastroscopic assessment for higher risk patients. Methods East and North Herts population is 600,000. 740 patients are on our Barrett's surveillance database. All patients due surveillance from November 2020 to January 2022 were offered Cytosponge instead of gastroscopy unless contraindicated or previous history of Dysplasia or oesophageal cancer. Contraindications included strictures, varices and fundoplication and these patients as well as those declining or failing Cytosponge were offered gastroscopy.TFF3 was used as a biomarker for intestinal metaplasia (IM). P53+ve and atypia were biomarkers for potential dysplasia. 33 patients were also included in the DELTA study. Results 230 patients with available data.152 patients successfully swallowed Cytosponge with adequate cytology. 78 patients had gastroscopy as a primary surveillance method. Of the Cytosponge group115/152 (76%) patients had positive biomarkers-97 (64%)TFF3+ve alone (suggesting non dysplastic Barrett's oesophagus (BO)) will have routine gastroscopy in <2y.18 (12%) had atypia and or p53+ve (suggesting dysplasia) and had 2ww gastroscopy. Confirmed dysplasia at endoscopy was found in 8/18 patients with atypia/p53+ve Cytosponge. 2 High grade dysplasia (HGD) 5 Low grade dysplasia (LGD) 1indefinite for dysplasia (IFD). 5% overall confirmed dysplasia. 37 (24%) were TFF3-ve and are described elsewhere. Of the Gastroscopy Group 9/78 (11.5%) were found to have dysplasia-2HGD 3LGD 4IFD In total 17/230 (7.4%) on surveillance had histologically confirmed dysplasia. Conclusion We were able to prevent delayed surveillance during the pandemic using Cytosponge in most patients and reduce unnecessary gastroscopy, Using a combination of Cytosponge in low risk and Gastroscopy in higher risk patients for Barrett's surveillance during the pandemic allowed identification of dysplasia. Further stratification of risk using Barrett's length, Male sex and age have been identified from the recently reported DELTA study1 which may further improve dysplasia detection by identifying those in the Cytosponge group who need more frequent surveillance. Comparison with non-pandemic years going forward will also be important to evaluate this strategy. Reference Use of Cytosponge biomarker panel to prioritise endoscopic Barrett's oesophagus surveillance: a cross-sectional study followed by a real-world prospective pilot Lancet Oncol Jan 11 2022.

Background The British Society of Gastroenterology has recommended the Edinburgh Dysphagia Score (EDS) to risk-stratify dysphagia referrals during the endoscopy COVID recovery phase. Aims External validation of the diagnostic accuracy of EDS and exploration of potential changes to improve its diagnostic performance. Methods A prospective multicentre study of consecutive patients referred with dysphagia on an urgent suspected upper gastrointestinal (UGI) cancer pathway between May 2020 and February 2021. The sensitivity and negative predictive value (NPV) of EDS were calculated. Variables associated with UGI cancer were identified by forward stepwise logistic regression and a modified Cancer Dysphagia Score (CDS) developed. Results 1301 patients were included from 19 endoscopy providers; 43% male; median age 62 (IQR 51-73) years. 91 (7%) UGI cancers were diagnosed, including 80 oesophageal, 10 gastric and one duodenal cancer. An EDS >= 3.5 had a sensitivity of 96.7 (95% CI 90.7-99.3)% and an NPV of 99.3 (97.8-99.8)%. Age, male sex, progressive dysphagia and unintentional weight loss >3 kg were positively associated and acid reflux and localisation to the neck were negatively associated with UGI cancer. Dysphagia duration <6 months utilised in EDS was replaced with progressive dysphagia in CDS. CDS >= 5.5 had a sensitivity of 97.8 (92.3-99.7)% and NPV of 99.5 (98.1-99.9)%. Area under receiver operating curve was 0.83 for CDS, compared to 0.81 for EDS. Conclusions In a national cohort, the EDS has high sensitivity and NPV as a triage tool for UGI cancer. The CDS offers even higher diagnostic accuracy. The EDS or CDS should be incorporated into the urgent suspected UGI cancer pathway.

Introduction Barrett's surveillance guidelines recommend patients with segments>=3cm should have endoscopic surveillance. Segments <3cm without intestinal metaplasia (IM) are not recommended for surveillance as they are low risk of progression to neoplasia. Patients with short segment BO and no IM may inadvertently end up on surveillance programmes due to several factors and contribute an unnecessary burden on resources, patients and a heavy environmental footprint. During the pandemic most surveillance procedures were stopped. To address this problem a clinical Cytosponge service was established as first line for Barrett's surveillance. BO is characterised by the presence of IM on histology-on Cytosponge cytology the surrogate biomarker for this is trefoil factor 3 (TFF3). Methods All patients on our Barrett's surveillance programme who were due surveillance between 30/11/2020 and 30/11/2021 were offered Cytosponge instead of gastroscopy. We excluded patients with previous dysplasia of any grade and those with fundoplication or strictures. Cytology results were obtained within 1/12 and patients who were negative or equivocal for TFF3 were identified. Patients identified with dysplasia/atypia/p53 and TFF3 positivity have been described elsewhere and not discussed here. Previous endoscopy reports were reviewed for Prague criteria and histology reports were reviewed for presence of IM with a view to discharging patients from ongoing surveillance according to Fig1 Decision tree. Results Over 1 year period 140 patients had successful Cytosponge procedures with adequate sampling. 25 (18%) had negative TFF3, 3 (2%) had equivocal TFF3. of these, 3 patients had maximal length of 3cm and 1 of 5cm (with lots of regenerative islands) and will continue on surveillance. 23/28 had segments with a maximal length <3cm. of these, 13/24 were IM negative at last endoscopy and were discharged from surveillance. 1 patient had an unknown segment length performed elsewhere with no previous IM but will have routine gastroscopy. 10/23 were IM positive at last endoscopy so will have routine surveillance (5 described as M2 5 as M1or less). 1 patient had an inlet patch only and was discharged. 1 patient was incidentally found to have eosinophilic oesophagitis but no BO. Conclusions Unexpected benefit of the Cytosponge service was help identifying patients not requiring ongoing surveillance.13 (9%) patients were identified by Cytosponge as suitable for immediate discharge from surveillance due to negative TFF3 no previous IM and <3cm length thus saving unnecessary gastroscopies , patient discomfort and reducing environmental impact.

To compare trends in publications of network meta-analyses (NMAs) in cardiovascular diseases (CVDs) in Asia-Pacific (APAC; China, Japan, Singapore, South Korea, Thailand) and Europe (United Kingdom [UK], Germany, France, Spain, Italy), with a focus on volume, collaborations and methods. Freely available NMAs assessing pharmacological or surgical interventions for CVD in terms of mortality or major adverse cardiovascular events, published in 2012 or later, by authors affiliated with institutions in the target countries were identified via MEDLINE and Embase. CVDs were grouped using the International Classification of Diseases, Tenth Edition (ICD-10). Across the 193 publications identified, heart diseases such as atrial fibrillation, aortic stenosis and heart failure (ICD-10 I30–I52) were the most common indications reported (38%). 63% of publications involved authors in APAC countries and 40% from Europe. Cumulative numbers of publications from APAC surpassed those from Europe from 2018 onwards. Authors were largely affiliated with institutions in China (50%), the UK (20%) and Italy (20%). 145 publications were by authors affiliated with institutions in a single country; 91% of publications from APAC and 44% from Europe, indicating different patterns of collaboration within these regions. In terms of reporting methodologies, 39% of included publications did not specify whether a frequentist or Bayesian framework was used (43% in APAC, 34% in Europe). Among those that reported, the Bayesian framework was more commonly used. Whilst there is a growing trend in NMA publication counts generally, the rate of increase in APAC was higher than Europe, particularly in the years following adoption of health technology assessment (HTA) procedures in APAC. The volume of publications not reporting the framework used was substantial, despite requirements for this in reporting guidelines e.g. PRISMA. Where reported, the Bayesian framework may have been favoured due to its advocacy by some HTA bodies.

Introduction Missed upper gastrointestinal (GI) cancers are associated with poorer survival outcomes. Missed cancer is defined as having had a negative upper GI endoscopy within three years of confirmatory diagnosis. The aims of this study are to quantify cases of missed upper GI cancer at endoscopy from a district general hospital and identify potential predisposing factors to improve future outcomes. Methods In this project, retrospective patient records were obtained from MDT meetings run between 2019 and 2020. For data comparison, similar records from previous cohorts were reviewed. The endoscopy records of these patients were extracted from the CIPTS endoscopy recording system and analysed to see if any patient had an endoscopy within the three years prior to diagnosis of their upper GI cancer. Previous endoscopies were scrutinised for the following factors: indication, visible abnormality, biopsy, endoscopist grade, sedation, presentation type, histology and follow up. Results Of the 280 patients in total, 20 (7.1%) had a negative endoscopy in three years prior to diagnosis. Subgroup analysis of 2019-20 patients showed identified 78 cases of upper GI cancer. We identified five formal cases of missed upper GI cancer in this subgroup. In the 2019-20 cohort the average age in the missed cancer group and non-missed cancer group were 83.8 and 73.2 years, respectively. Factors identified in missed cancer groups from earlier cohorts included emergency bleeding, visible abnormality not biopsied, administrative delay in rescope and lack of recognition of pathology. Conclusions The incidence of missed upper GI cancer diagnosis within this centre is similar to that of recent studies from other centres. Our study has highlighted that advanced age, accurate recognition and biopsy, and timely re-biopsy may be factors influencing missed upper GI cancers. This evidence has highlighted the importance of endoscopy training in the recognition and biopsy of pre neoplastic lesions. Furthermore, we intend to analyse data for the year 2020-21 to monitor the impact of COVID-19 on missed upper GI cancer diagnosis.

Introduction British Society of Gastroenterology has recommended to use the Edinburgh Dysphagia Score (EDS) (with a cut off ≥ 3.5) to risk stratify dysphagia referrals during the endoscopy COVID recovery phase. We have validated EDS performance in a national prospective evaluation of the UGI cancer two week wait pathway and developed a modified scoring system with improved diagnostic accuracy. Methods Data on patients referred with dysphagia on the 2WW pathway between May 2020 and February 2021 to 19 centres across the UK were collected at telephone triage and recorded on a standardised data collection tool. Sensitivity and Negative predictive values were calculated for EDS ≥3.5. Forward stepwise logistic regression model identified the factors associated with UGI cancer and their regression coefficients were used to develop a modified scoring system, called Cancer Dysphagia Score (CDS). Diagnostic accuracy was examined by comparing area under the receiver operating curves (AUROC). Results 1496 patients were studied: median age 62 (IQR 51-73), 58% female. Median EDS score was 4 (IQR 2.5-6); with 67% having an EDS ≥ 3.5. 64% were triaged to 2WW endoscopy; 18% to urgent (but not 2WW) endoscopy; 2.8% to urgent CT scan; 5.5% to routine OGD; and 4.4% to barium swallow. After excluding patients who declined investigation, results were available for 96%. 91 UGI cancers were diagnosed (prevalence 7.1%); with 3 (3%) UGI cancers diagnosed in patients with EDS<3.5 (one with EDS 3, two with EDS 1.5). EDS ≥ 3.5 had sensitivity of 96.7% and negative predictive value of 99% for UGI cancer. The factors associated with UGI cancer and hence selected to develop CDS included: age odds ratio 1.05 (95% CI 1.03-1.06); male 3.95 (2.36-6.58); progressive dysphagia 2.30 (1.39-3.79); unintentional weight loss 3.28 (2.02-5.31); acid reflux symptoms 0.47 (0.25-0.88) and dysphagia localised to the neck (0.26 (0.12-0.57). Dysphagia duration less than 6 months was not statistically significant (1.02 (0.45-2.22). AUROC was 0.83 for CDS as compared to 0.81 for EDS. Cancer dysphagia score and its receiver operating curve in comparison to Edinburgh dysphagia score is presented in figure 1. Conclusion A national prospective cohort confirms that EDS has high sensitivity and negative predictive value for UGI cancer, however, a modified cancer dysphagia score offers higher diagnostic accuracy. Our results suggest that CDS should be incorporated in the UGI cancer 2WW pathway for risk stratification of patients with dysphagia and further studies in primary care are needed.

Introduction British Society of Gastroenterology (BSG) guidance on endoscopy during recovery from the COVID pandemic (April 2020) recommended that two week wait (2WW) referrals are triaged, with patients risk stratified for endoscopy or other investigation. We have prospectively evaluated the 2WW upper gastrointestinal (UGI) cancer pathway and its outcomes following this guidance. Methods Data were collected at telephone triage by consultants and nurse endoscopists between May 2020 and February 2021 in 19 centres across the UK and recorded on a standardised data collection tool, which included recommendations on the timing of endoscopy based on the BSG recovery document. This project was supported by the BSG Clinical Services and Standards Committee. Results Data for 1793 UGI 2WW referrals were received: median age 63 (IQR 51-74), 58% female. Dysphagia and odynophagia were the commonest reasons (83%) for referral. Other symptoms included dyspepsia (55%), weight loss (32%), globus (3%), and anaemia (3%). 15.8% of 2WW referrals were downgraded at triage to routine endoscopy (6.6%) or no investigation at all (9.2%). 56% were triaged to 2WW endoscopy; 19.6% to urgent (non-2WW) endoscopy; 4.7% to urgent CT scan; and 3.8% to barium swallow. 6.3% had UGI cancer (5.2% oesophageal, 1.1% gastric) and 0.9% had cancer at other sites (6 colorectal, 2 lung, 2 breast, 2 hypopharyngeal, 1 pancreatic and 2 unknown primary). Endoscopy results were available for 1387 patients (97.5% of all endoscopy pathways). The prevalence of UGI cancer in different triage categories was: 2WW 9%; urgent (non-2WW) endoscopy 3%; urgent CT 8.5%; and routine endoscopy 1%. Triage based on the BSG recovery guidance was 97% sensitive and 19% specific for upper GI cancer (at 2WW or urgent endoscopy or CT scan), with a negative predictive value of 99% and a positive predictive value of 8%. A summary of all endoscopy findings from the service evaluation is presented in figure 1. Conclusions Triage based on the BSG recovery guidance was 97% sensitive with a negative predictive value of 99% in diagnosing UGI cancer at 2WW or urgent endoscopy or CT scan. 6.6% of 2WW referrals were safely investigated routinely and over 9% of 2WW referrals required no investigation at all following triage. These findings should guide reform of the upper GI 2WW pathway to reduce the burden on endoscopy during and after the COVID pandemic.