Background Lung cancer screening (LCS) with low-dose CT (LDCT) reduces lung cancer mortality. National implementation is now recommended in England, with government support announced. Following successful UK pilot studies between 2016–2019, NHS England (NHSE) launched the Targeted Lung Health Check (TLHC) programme. This abstract summarises lung cancer detection data from the combined NHSE TLHCs. Methods In 2019, face-to-face or telephone lung health checks (LHCs) commenced for ever-smokers aged 55–74 in underserved populations with high lung cancer mortality. Current smokers are offered Very Brief Advice and cessation referral. Those at higher risk of lung cancer (PLCOm2012 ≥1.51% or LLPv2 ≥2.5%) are offered LDCT reported by quality-assured thoracic radiologists with lung MDT expertise using adapted British Thoracic Society nodule guidelines. The TLHC programme includes several phases of deployment, now incorporating 43 live sites. Site level data are presented for all phase 1 (original), phase 2 (onboarded) sites, and phase 3 (expansion) sites. Results By March 2023, 892,404 people had been invited for a LHC across all TLHC 'live' sites. 377,791 attended a LHC (uptake rate 42.1%) of whom 176,572 were identified as high risk and offered LDCT screening. After DNAs/exclusions, 156,032 (88.4% of referred) participants underwent LDCT screening (LHC to CT conversion rate 41.3%). 2,056 participants have been diagnosed with lung cancer, equating to a prevalence of 1.3%. 75.1% were diagnosed at an early-stage (I-II). Further cancer diagnoses are expected from nodule surveillance. As a proportion of baseline LDCT, surveillance scans had been performed at 3mo (11.3%), and 12mo (5.8%) respectively. In parallel, rapid registration data for lung cancer showed greatest improvement in stage of diagnosis in the most deprived quintile. Conclusions Through the work of a multi-disciplinary team, the NHSE TLHC programme, working across multiple project sites has successfully delivered LCS to high-risk participants across England. This has resulted in increased early stage diagnosis, especially for people experiencing higher deprivation. The programme continues to expand in a phased manner, and will play a key role informing full national LCS roll-out by 2030.

Objectives Targeted lung health checks (TLHC) with low dose computed tomography (LDCT) have been shown to identify an increased proportion of lung cancers at an earlier stage in the UK. The reporting of index LDCT without access to prior imaging affects many programs due to reporting capacity and the need for outsourcing. We assessed the impact of historical imaging on reported nodules through our local TLHC. Methods All patients with TLHC index LDCT scan between 5/8/22 and 11/5/23 were included. Historical imaging within the Trust was identified via CRIS codes and if present was reviewed at the Specialist Review Meeting (SRM). Results Of 4,422 index LDCT scans performed, nodules were reported in 674(15.2%). Of these 125(20%) patients had prior CT scans with a median interval time of 3.9[1.6–6.8] years. Nodules were present in 75(60%) of the prior CT images. In 67/75(89%) cases the nodules had improved or were stable. In 7/75(11%) patients the nodules had increased in size. A progressive peri-cystic lesion was noted in one case (unmeasurable), four cases showed nodules <5mm, and two cases were >5mm. All were unreported. The median growth in diameter was 3[2.25–4.5]mm. A direct referral to the 2WW pathway was made in one patient. The prior CT scans were dedicated thoracic imaging in 45 cases, while other CT imaging included CT coronary angiograms, CT colons and CT thorax-abdomen-pelvis. We found only 24/67(32%) of nodules were reported on the prior imaging. In those with dedicated prior thoracic imaging 19/45(42%) were reported while in non-dedicated thoracic CT scans only 4/29(14%) were reported (p<0.01). Conclusions Our data shows that 1 in 5 patients with reported nodules at index LDCT had relevant prior imaging. Of these the reported nodules were present in nearly two thirds. Most were stable and this avoided an additional 134 surveillance LDCT scans. This indicates that reporters of index LDCT should routinely have access to prior CT imaging.

Introduction and Objectives SARS-CoV-2 has reemphasised the importance of respiratory virus infections (RVIs). UNIVERSAL aims to characterise the impact of a broad range of RVIs. We hypothesise that analysis of a prospective cohort of hospitalised adults with RVI will allow determination of clinical and biological profiles associated with virus type and disease severity. Methods Excluding co-infection, we recruited 313 hospitalised, symptomatic adults, with PCR evidence of RVI. Clinical characteristics were collected prospectively. We compared admission total/differential white cell count (WCC), CRP and supplementary O2-requirement. Multiple logistic regression (MLR) was utilised to calculate the OR and 95%CI associated with oxygen requirement for all viruses. Adjustments included: Age>65, smoking status, Charlson co-morbidity score (CCS), COPD, asthma, hypertension, diabetes, obesity, congestive heart failure, pneumococcal/covid/influenza-vaccination status, white cell differentials and CRP. Results Table-1 displays patient characteristics by virus group. Overall, 62.9% of participants required supplemental-O2 on admission. There was no significant difference in admission-O2 requirement between virus groups (p=0.651). Lymphocyte count <1.0*109/L(p=0.018) and CRP>55 mg/L(p<0.001) were associated with admission-O2 requirement SARS-CoV-2 positive patients (median age-72) were significantly older than Rhino/enterovirus (RhV/EV) (median age-64, p=0.006), and influenza positive patients (median age-64, p=0.001).Furthermore, SARS-CoV-2 patients (median CCS-10) were significantly more comorbid than the RhV/EV (median CCS-4, p=0.043) and influenza patients(median CCS-4, p=0.001). For all viruses, median WCC was 10.3*109/L. WCC differential median values included: neutrophils (8.0*109/L), lymphocyte count (1.0*109/L), eosinophil count (0.0*109/L) Median CRP was 55 mg/L and there was no significant CRP difference between the viruses(p=0.389). SARS-CoV-2 WCC (median-8.8*109/L) was significantly lower when compared to RhV/EV (median-11.7*109/L, p=0.003) and influenza (median-8.9*109/L, p=0.041).Neutrophils were significantly lower in SARS-CoV-2 (median-7.2) than in RhV/EV(median-8.9*109/L, p=0.030).Additionally, lymphocytes were significantly lower in SARS-CoV-2(median-0.9*109/L) compared to RhV/EV(median-1.3 p=0.002). Eosinophil count in RhV/EV (median-0.1 IQR:0.0–0.3) was significantly higher than in influenza (median-0.0*109/L, p<0.001) and SARS-CoV-2(median-0.0*109/L p=0.043). MLR revealed lymphocytes <1.0*109/L (OR 1.838 CI:1.05–3.26) and CRP >55 mg/L (OR-2.58 CI:1.45–4.67) were independent risk factors for admission supplemental-O2 requirement. Conclusion Common RVIs including RhV cause considerable morbidity. Lymphopenia and CRP>55 mg/L were associated with more severe disease in adults admitted with RVI. More precise characterisation of biomarkers associated with severity is an essential prerequisite to novel therapeutics development. Please refer to page A286 for declarations of interest related to this abstract.

Introduction Targeted lung health checks (TLHC) with low dose CT thorax (LDCT) detect interstitial lung abnormalities (ILAs) and opens the opportunity to initiate targeted therapies. Aims To assess the distribution of reported ILAs through the pilot program Methods Patients undergoing a LDCT through the local TLHC had their scans reported within the pre-determined ILAs protocol. Available clinical data was then reviewed. Those felt to have clinically significant ILAs underwent questionnaire evaluation. Results Between August 2022 and May 2023, 33/4422 (0.7%) patients were found to have a reportable ILA. This is lower than found in previous studies.1 The median age of patients was 69 [65–74] years with 22/33 (67%) being male. Extent of lung involvement was reported as none/ILA without reticulation, <5%, 5–10% and >10% in 8, 10, 12 and 3 patients respectively. In 18/33 (55%) reported ILAs, radiological features were reported as uncharacterised. Usual interstitial pneumonia (UIP) pattern was seen in 3 patients, 4 had smoking-related ILD (SR-ILD), 2 nonspecific interstitial pneumonia (NSIP), 3 post-inflammatory/COVID-19 pneumonitis, 1 sarcoidosis, and 1 had hypersensitivity pneumonitis. In 17(52%) patients, ILAs were radiographically evident on prior scans and mild progression was noted in 5/17 (29%) of these. A total of 8 patients were contacted post LDCT and 5 had a subsequent outpatient appointment. All patients reported exertional breathlessness and the median MRC dyspnoea scale was 1[1–3]. In half of patients a cough was reported. Methotrexate usage was note in one patient, and one patient had a reported autoimmune condition. In nine patients with full pulmonary function, the median TLco was 51[43–65%] predicted. Conclusion The TLHC provides an opportunity to identify ILA early in patients who have a low symptom burden. Progression, course and timing of intervention need to be fully characterized in this population to optimize management in patients passing through future lung cancer screening programs. Reference Hewitt RJ, Bartlett EC, Ganatra R, et al. Lung cancer screening provides an opportunity for early diagnosis and treatment of interstitial lung disease. Thorax 2022;77:1149–1151.

Introduction Aortic valve calcium (AVC) scores on low dose CT (LDCT) as part of the targeted lung health check (TLHC) opens an opportunity to identify cardiovascular disease in an asymptomatic high-risk group. Objective To describe the prevalence of reported AVC on LDCT. Methods Index LDCT scans from 4,422 participants in the local pilot Targeted Lung Health Check (TLHC) programme between August 2022 and May 2023 were assessed for reported moderate to severe AVC scores. Results In total, 76/4422 (1.7%) scans showed evidence of moderate/severe AVC on CT with 36/76 (46%) being severe. Echocardiograms were available for review in 41/77 (53%) patients (including 28/36 severe cases) and confirmed aortic stenosis (AS) in 32/41(78%) cases. Of these seven patients had severe AS. In those with severe AVC, 19/28 had a program-initiated echocardiogram. Of these, 5/19(26%) had severe AS (new or progressive) and were referred for intervention. AVC and echo AS severity were concordant in 14/41 (34%) patients with available echocardiograms. In all discordant cases, CT calcification was of higher severity and those with severe AVC were the most discordant. In seven patients with severe calcification of the AV, no AS was detected on echocardiography. The incidence of moderate/severe AVC was greater in patients with moderate/severe coronary artery calcification (CAC) (3.2% vs 0.9% no/mild CAC) p <0.01). Severity of CAC correlated with degree of AV calcification (r2=0.24). No association was seen between the degree of emphysema and severity of AVC (1.7% vs 2% p=0.28). Conclusions In our local population the incidence of moderate to severe AVC was 1.7%. In patients with severe AVC, 1 in 4 patients undergoing an echocardiogram were subsequently referred for cardiology assessment, however there was poor agreement between AVC severity and echocardiography results.

IntroductionThe approach to central airways obstruction in lung cancer varies. Having previously established benchmarking in disease prevalence, management and outcomes in 2014, we re-audited our local approach following the introduction of a dedicated interventional service.MethodNew lung cancer diagnoses in 2019 were assessed for evidence of central airways disease. Comparisons were made with our previous audit from 2014. Simple descriptive statistics were used. To identify factors associated with death, all data were combined (2014/2015 and 2019/2020). Logistic regression was used to determine the effect of age, gender, tumour type and degree of obstruction to identify independent factors associated with 60-day mortality.ResultsNo differences were seen in the clinical characteristics of people with CAO presenting in 2014/2015 and 2019. Deaths were high, with 45% of people dying within 90 days of a diagnostic CT. There was no difference in the proportion dying in 2014/2015 and 2019. Over the two time periods the proportion with CAO<50% increased (12/30 [40%] in 2014/2015 versus 18/30 [60%] in 2020; p=0.07). The proportion of people eligible for intervention and receiving intervention was unchanged between 2014/2015 and 2019 (5/16 [31%] in 2014/2015 and 3/9 [33%] in 2019). However the time to intervention improved (60.5 [interquartile range {IQR} 29.5–120.0] days in 2014/2015 and 7.0 [IQR 6.0–7.0] days in 2019; p=0.03). Although limited by the number of outcomes, after adjusting for age, gender and degree of obstruction, the odds of death was lower among people with squamous cell carcinoma compared to those with other tumour types (squamous cell cancers 8/36 [22%] vs small cell lung cancer 5/21 [24%] versus other cancers 30/74 [40%]; adjusted odds ratio [aOR] 0.40 [95%CI 0.16–1.02], aOR 0.46 [0.15–1.40]; p=0.09).ConclusionsThe impact of the interventional service did not lead to an increase in intervention, however the time to procedure improved significantly. Understanding decision making and identifying those that benefit the greatest from intervention needs further work.

Aim: To compare the adequacy of different diagnostic procedures for tissue subtyping of NSCLC and analysis of epidermal growth receptor (EGFR) mutation, anaplastic lymphoma kinase (ALK) gene fusion and programmed death ligand (PD-L1). Method: Patients with a final diagnosis of NSCLC undergoing either a fibreoptic bronchoscopy (FOB), enodbronchial ultrasound (EBUS) tranbronchial needle aspiration (TBNA), or a physicina ultrasound-guided biopsy (including lung, chest wall, pleural and neck lymph node) were included. The adequacy of diagsnotic procedures to determine cellular subtype and suitability for molecular testing were assessed. Results: 492 subjects with NSCLC were identified.289/492 (59%) were male and the median age was 71.6 years. Adenocarcinoma and squamous cell carcinoma accounted for 278/492 and 184/492 cases respectively. Other subtypes were seen in 7/492 cases and 23/492 were not otherwsie specified (NOS). Overall biopsies subtyped NSCLC in between 87-100% of cases., while neck node being the largest group unable to subtyped in 12.7% cases with 8.3% classified as NSCLC-NOS. However, no statistically significant difference was seen across sampling techniques (p=0.323). Therpeutic target testing for EGFR, ALK and PD-L1 was reliable ranging from 73-100% with all Biopsy techniques. No difference was seen across groups in adequacy for EGFR (p=0.09), ALK (p=0.052) and PD-L1 (p sy=0.215) testing. Conclusion: We found no significant difference in biopsies providing suitable tissue for subtyping in NSCLC, and adequate tissue for therapeutic target testing. Standard biopsy techniques perform equally well in informing on treatment

Introduction and Objectives The relationship between symptoms, pleural effusion size and the diaphragm is unclear. We conducted a pilot study to understand the role of diaphragm shape and movement in patients with unilateral pleural effusions. Method We prospectively recruited patients with unilateral pleural effusions. Routine investigations were collected. Study-specific thoracic ultrasounds (TUS) were performed at baseline, post intervention, and at day 7. A seven-day visual analogue score (VAS) diary was completed for breathlessness, starting at baseline, immediately post aspiration and then daily thereafter. Results Of the 45 patients recruited, 17/45(38%) were female. The median [interquartile] age was 71[66–77] years. The most common reported symptom was breathlessness in 43/45(96%). At baseline, the medial effusion depth was 100[80–126]mm over 4[3–5] rib spaces. Procedures were performed in 40/45(89%), including 32 therapeutic-interventions and 8 diagnostic aspirations. A median of 1,000 [481–1,500]mls of pleural fluid was aspirated. Malignancy was diagnosed in 20/45(44%) patients. A diaphragm abnormality (abnormal shape, movement or both) was seen in 22/45(49%) with a flattened diaphragm in 7/45(16%), an inverted diaphragm in 2/45(4%), paradoxical movement in 13/45(29%) and no movement in 8/45(36%). A malignant diagnosis was found in 14/22(64%) of those with a diaphragm abnormality at baseline, compared to 6/23(35%) with normal diaphragm (p<0.05). Of those undergoing a therapeutic intervention diaphragm abnormalities persisted in 4/21(19%) with improvement in 15/21(71%) (two were unreported). Diaphragm shape improved in all patients, however two patients had a persistent paradoxically moving diaphragm and two had no movement. In 27 patients undergoing therapeutic intervention and completing follow up, 19/27(70%) had a diaphragm abnormality at baseline, 4/27(15%) post intervention and 11/27(41%) at day 7. VAS scores at baseline, post aspiration and day 7 were 44[27–53.5]mm, 25[13–44]mm and 36[13.5–58.5]mm in those with a diaphragm abnormality compared with 46.5[34.25–72.5]mm, 34.5[18.5–54.75]mm and 22.5[14.25–32.25]mm in those with an normal diaphragm. In those with an abnormal diaphragm at day 7, the change in VAS was -4[-11.5–1] in the abnormal diaphragm group and -23[-31- -10.25] in the normal diaphragm group (p<0.05). Conclusion A diaphragm abnormality was common, demonstrated reversibility, but recurrence by day 7 was associated with loss of therapeutic benefit.

Introduction Central airways obstruction (CAO) carries significant morbidity and mortality. Previous local auditing demonstrated a small number of lung cancer patients with CAO underwent interventional bronchoscopy. Following service development we prospectively re-audited. Methods We prospectively recorded cases referred to our interventional service between October 2018 and February 2020. Details on referrals of diagnoses, treatments performed and outcomes were collected. Comparisons were made with our baseline audit data from 2014. Results During the prospective audit period, 32 patients were recorded. The median[interquartile range] age was 68[59.5–73.3] years with 20/32(62.5%) being female. In 22/32(69%) cases, referral was based on index presentation. Of these, 69% were formally reported in CT scan. Symptomatic disease was noted in 23/32(72%), and 24/32(75%) had performance status of ≤2. Lung cancer was the primary diagnosis in 26/32(81%) of cases, with NSCLC-adenocarcinoma being the commonest. Rigid bronchoscopy was performed in 19/32(59%) of patients referred, and time from CT to procedure was 11[6.5–22.5] days. Debulking was the commonest intervention, while stents were placed in 3/19(16%) cases. In those patients undergoing interventional bronchoscopy, 12/19(63%) were alive at 3 months. Rigid bronchoscopy was not performed in 13/32(41%) patients. Intervention was deemed to carry too high a risk/benefit ratio in 3/13(23%) patients and 3/13(23%) patients with small cell lung cancer were referred for systemic therapy over interventional bronchoscopy. In comparison with the previous audit cohort – gender distribution, age and cancer represented were similar. More patients with CAO underwent intervention (p<0.001). Non-significant trends in improvement in time to rigid bronchoscopy and survival at 3 months were noted. Reporting of CAO on CT was unchanged. Conclusion Establishing a dedicated interventional service may impact on number of interventions, and potentially reduce time to procedure in a patient group that has a poor prognosis. However, improving recognition, standardising reporting and referral is required, as the true burden of disease remains underestimated in routine practice. Despite this, understanding reasons for not intervening need further study, as even with a pro-active team, of those referred, half of patients still did not undergo intervention.

Background EBUS has revolutionised lung cancer diagnostics. As with mediastinal nodal disease, centrally located lung tumours can be sampled using curvilinear EBUS. We evaluated our local five year experience. Method We retrospectively evaluated patients who underwent EBUS targeting the primary tumour. Sampling was performed using either a 22 gauge FNA or core biopsy needle (EchoTip®, Cook). Samples were placed directly into formalin and processed as cell blocks. Results Between January 2015 and July 2020, 87/1352 EBUS procedures performed sampled primary lesions in 84 patients. The median[interquartile] age was 68[60–73]years and 45/84(54%) were male. The commonest site was the right upper lobe (n=28/84; 33%), with a median lesion size of 30[24–35]mm. FNA needles were used in 79/84(94%) while in five cases core needles were used. Total representative sampling was 80/87(92%), with a false negative rate of 11%. In three patients who underwent repeat EBUS, one was representative but further tissue was required for molecular testing and the remaining two were diagnostic, but underwent sampling for recurrence >12 months after index presentation. Of those negative and non-representative samples – final diagnoses were defined by a clinco-radiological diagnosis in four cases, resections and CT guided biopsies in two cases each, and one patient underwent a repeat EBUS under general anaesthesia. Of diagnostic samples, lung cancer was diagnosed in 66/84(76%) cases, of which NSCLC-squamous was the commonest (n=27/66; 41%). In patients undergoing molecular/therapeutic target testing: EFGR was successful in 28/28 (100%), ALK-FISH in 25/26 (96%) patient tested and PDL-1 was tested in 39 patients. One repeat EBUS was required for molecular tissue. Complications included minor bleeding in 4/87(5%), with only one requiring cold saline instillation, desaturation in 1/87(1%), and tachycardia in 1/87(1%). Only one procedure was abandoned, due to patient tachyarrhythmia. Delayed complications occurred in one patient who was hospitalized ≤seven days with a pneumonia. Conclusion EBUS sampling of centrally located lung tumour provides a similar diagnostic yield to lymph node sampling, provides suitable material for molecular/therapeutic target testing and has a low complication rate.

Background Recurrent pleural intervention may complicate the pleural space by inducing pleural inflammation with subsequent septation formation. We evaluated our five year experience in the incidence of pleural septations in patients undergoing pleural interventions. Method We retrospectively identified patients who underwent thoracic ultrasound (TUS) in our pleural service from our reporting database between August 2015 and February 2021. Categorical reporting of the presence of septations was used, reporting septations as either present or absent. Repeated TUS, types of pleural interventions, and time between these interventions were analysed. Results Of the 2737 index TUS performed, we recorded whether septations were present or not in 2684 (98%) patients. Of these, septations were present in 715/2684 (26.6%: 95% CI 25–28.3%) cases. In 297 patients with >1 TUS reports, 187 underwent an intervention at the index visit. At baseline, septations were present in 39/187 (20.9%) of these patients. Of the remaining 148/187 (79.1%) patients without septations on index scan, 24/148 (16.2%; 95% CI 10.7–23.2) reported the formation of new septations at the second TUS visit at a median [interquartile range] time interval of 21 [9–63] days. No association was seen between the type of intervention and development of septations [chest drain 14.2% (1/7), diagnostic aspiration only 22.2% (6/27), therapeutic aspiration 15.3%( 17/111) p=0.68]. No difference was observed in those patients with serial scans, not undergoing intervention, with new septations reported in 15/75 (20%) (p=0.48), while a shorter time interval between scans reporting conversion to septations (median [IQR] 7 [2.25–57] days (p=0.04)) was noted. Conclusion Overall, in this large cohort of patients seen through our pleural service, septations were present in a quarter of baseline thoracic ultrasound examinations. Septations formed with or without intervention in around 1 in 5 patients. Understanding this further has significant implications for diagnostic and management pathways.

Background and Objectives Ways to assess and track progress of new EBUS operators and trainees is desirable to ensure training goals and procedural competence are achieved and maintained. While important, relying on the diagnostic yield or on question-based assessments alone is not sufficient. Methods This study examined the longitudinal change in times taken between needle passes (needle pass time; NPT) during EBUS lymph node sampling as a metric to monitor progress. The EBUS database of a tertiary hospital that employs 1–2 lung cancer fellow per year was accessed to extract data on the first 50 EBUS procedures for three trainees were collected. The NPT was derived using PACS images that are stored to document every needle pass during an EBUS procedure and an average NPT per procedure was calculated. Results Between the three trainees 157 procedures were carried out within the study period with 302 LN stations sampled. Station 7 was the most commonly sampled (36.9%). The mean NPT (n=204 stations) was 2:49±0:49 mins. The mean lymph node short axis diameter (n=210) was 15.5±8.7 mm. There was a negative correlation between node size and time per pass (r -0.146, p=0.045). The change in average NPT and time between passes during the study period for the trainees is plotted in figure 1 showing a consistent decrease in average times between passes during the first 50 procedures. A point of 'convergence' around the 30th procedure with less variation between procedures was noted (red vertical lines) for the three trainees. On multivariate regression, NPT was significantly associated with procedure order and type of station sampled but not lymph node diameter. Conclusion NPT and time between stations are easy metrics that can potentially help ensure EBUS trainees are advancing in a given training programme.

Background The UK lockdown introduced on 23rd March 2020 to flatten the curve of the COVID-19 pandemic was associated with a decrease, or cessation, of most non-COVID-19 NHS services, affecting other patient groups requiring time-critical access to NHS services. Objectives We assessed the impact of COVID-19 on our pleural service. Methods All patients referred to the pleural service during the official period of lockdown 23/3/20–11/5/20 (7 weeks) were identified alongside those referred in the preceding and following 7 weeks. Patient demographics, number and type of referrals (2 week wait, in-patient or out-patient), length of time to see the patient, procedures performed and ultimate diagnosis from the referral were collected. We defined cancer based on new diagnosis or management of malignant effusions and suspected pleural infection based on investigation for pH<7.2. Results During this 21 week period 401 patients were referred to the pleural service. The mean±SD age was 67.4± 15.4 years and 285/401 (71%) were male. Referrals dropped by 23% during lockdown returning to normal in the 7 weeks post lockdown. From baseline cancer diagnoses increased by 24% during lockdown and 53% in the 7 weeks following lockdown (Graph 1). This increase in cancer diagnosis was secondary to in-patient referrals (pre, during and post lockdown: 6/23(26%), 14/22(63%) and 15/25(60%) respectively (p=0.02). Suspected pleural infection referrals reduced by 21% during the lockdown. Time from referral to review increased from 1[1–2]days to 3[1–5]days during the lockdown, reducing to 2[0–4]days post lockdown (p=0.002). The subsequent delays driven by mandatory swabs to exclude SARS-CoV-2 infection prior to review. Conclusion Due to the national lockdown a reduction in referrals to our pleural service was observed, particularly affecting reviews and suspected pleural infection, but interestingly not cancer diagnoses. However, consequently we have seen an increase in diagnosis in malignancy and pleural infection driven through our in-patient cohort rather than 2 week wait referrals, indicating a probable delay in seeking medical attention or higher threshold for referral during the lockdown. COVID swabbing has impacted our timeliness to review urgent outpatients. The effect of COVID-19 lockdown will continue, and the true impact is yet to be determined.

Background: To determine if tuberculosis (TB) screening improves patient outcomes, we conducted two systematic reviews to investigate the effect of TB screening on diagnosis, treatment outcomes, deaths (clinical review assessing 23 outcome indicators); and patient costs (economic review). Methods: Pubmed, EMBASE, Scopus and the Cochrane Library were searched between 1/1/1980-13/4/2020 (clinical review) and 1/1/2010-14/8/2020 (economic review). As studies were heterogeneous, data synthesis was narrative. Findings: Clinical review: of 27,270 articles, 18 (n=3 trials) were eligible. Nine involved general populations. Compared to passive case finding (PCF), studies showed lower smear grade (n=2/3) and time to diagnosis (n=2/3); higher pre-treatment losses to follow-up (screened 23% and 29% vs PCF 15% and 14%; n=2/2); and similar treatment success (range 68-81%; n=4) and case fatality (range 3-11%; n=5) in the screened group. Nine reported on risk groups. Compared to PCF, studies showed lower smear positivity among those culture-confirmed (n=3/4) and time to diagnosis (n=2/2); and similar (range 80-90%; n=2/2) treatment success in the screened group. Case fatality was lower in n=2/3 observational studies; both reported on established screening programmes. A neonatal trial and post-hoc analysis of a household contacts trial found screening was associated with lower all-cause mortality. Economic review: From 2841 articles, six observational studies were eligible. Total costs (n=6) and catastrophic cost prevalence (n=4; range screened 9-45% vs PCF 12-61%) was lower among those screened. Interpretation: We found very limited patient outcome data. Collecting and reporting this data must be prioritised to inform policy and practice. Funding: WHO and EDCTP.