Imperial College London

Professor Anand Devaraj

Faculty of MedicineNational Heart & Lung Institute

Professor of Practice (Thoracic Radiology)
 
 
 
//

Contact

 

anand.devaraj Website

 
 
//

Location

 

South BlockRoyal Brompton Campus

//

Summary

 

Publications

Publication Type
Year
to

198 results found

Bartlett E, Belsey J, Derbyshire J, Morris K, Chen M, Addis J, Martins M, Ridge CA, Desai SR, Mirsadraee S, Padley S, Whiteside S, Vaghani P, Morjaria JB, Kemp SV, Devaraj Aet al., 2021, Implications of incidental findings from lung screening for primary care: data from a UK pilot, npj Primary Care Respiratory Medicine, Vol: 31, ISSN: 2055-1010

Regional lung cancer screening (LCS) is underway in England, involving a “lung health check” (LHC) and low-dose CT scan for those at high risk of cancer. Incidental findings from LHCs or CTs are usually referred to primary care. We describe the proportion of participants referred from the West London LCS pilot to primary care, the indications for referral, the number of general practitioner (GP) attendances and consequent changes to patient management, and provide an estimated cost-burden analysis for primary care. A small proportion (163/1542, 10.6%) of LHC attendees were referred to primary care, primarily for suspected undiagnosed chronic obstructive pulmonary disease (55/163, 33.7%) or for QRISK® (63/163, 38.7%) assessment. Ninety one of 159 (57.2%) participants consenting to follow-up attended GP appointments; costs incurred by primary care were estimated at £5.69/LHC participant. Patient management changes occurred in only 36/159 (22.6%) referred participants. LHCs result in a small increase to primary care workload provided a strict referral protocol is adhered to. Changes to patient management arising from incidental findings referrals are infrequent.

Journal article

Bartlett EC, Silva M, Callister ME, Devaraj Aet al., 2021, False-Negative Results in Lung Cancer Screening- Evidence and Controversies, JOURNAL OF THORACIC ONCOLOGY, Vol: 16, Pages: 912-921, ISSN: 1556-0864

Journal article

Kishore AK, Devaraj A, Vail A, Ward K, Thomas PG, Sen D, Procter A, Win M, James N, Roffe C, Meisel A, Woodhead M, Smith CJet al., 2021, Use of Pulmonary Computed Tomography for Evaluating Suspected Stroke-Associated Pneumonia., J Stroke Cerebrovasc Dis, Vol: 30

OBJECTIVES: Accurate and timely diagnosis of pneumonia complicating stroke remains challenging and the diagnostic accuracy of chest X-ray (CXR) in the setting of stroke-associated pneumonia (SAP) is uncertain. The overall objective of this study was to evaluate the use of pulmonary computed tomography (CT) in diagnosis of suspected SAP. MATERIALS AND METHODS: Patients with acute ischemic stroke (IS) or intracerebral hemorrhage (ICH) were recruited within 24h of clinically suspected SAP and underwent non-contrast pulmonary CT within 48h of antibiotic initiation. CXR and pulmonary CT were reported by two radiologists. Pulmonary CT was used as the reference standard for final diagnosis of SAP. Sensitivity, specificity, positive and negative predictive values (PPV and NPV), and diagnostic odds ratio (OR) for CXR were calculated. RESULTS: 40 patients (36 IS, 4 ICH) with a median age of 78y (range 44y-90y) and a median National Institute of Health Stroke Scale score of 13 (range 3-31) were included. All patients had at least one CXR and 35/40 patients (88%) underwent pulmonary CT. Changes consistent with pneumonia were present in 15/40 CXRs (38%) and 12/35 pulmonary CTs (34%). 9/35 pulmonary CTs (26%) were reported normal. CXR had a sensitivity of 58.3%, specificity of 73.9%, PPV of 53.8 %, NPV of 77.2 %, diagnostic OR of 3.7 (95% CI 0.7 - 22) and an accuracy of 68.5% (95% CI 50.7% -83.1%). DISCUSSION: CXR has limited diagnostic accuracy in SAP. The majority of patients started on antibiotics had no evidence of pneumonia on pulmonary CT with potential implications for antibiotic stewardship. CONCLUSIONS: Pulmonary CT could be applied as a reference standard for evaluation of clinical and biomarker diagnostic SAP algorithms in multi-center studies.

Journal article

Mirsadraee S, Gorog DA, Mahon CF, Rawal B, Semple TR, Nicol ED, Arachchillage DRJ, Devaraj A, Price S, Desai SR, Ridge CA, Singh S, Padley SPGet al., 2021, Prevalence of Thrombotic Complications in ICU-Treated Patients With Coronavirus Disease 2019 Detected With Systematic CT Scanning, CRITICAL CARE MEDICINE, Vol: 49, Pages: 804-815, ISSN: 0090-3493

Journal article

Wells AU, Devaraj A, Desai SR, 2021, Interstitial Lung Disease after COVID-19 Infection: A Catalog of Uncertainties COMMENT, RADIOLOGY, Vol: 299, Pages: E216-E218, ISSN: 0033-8419

Journal article

Behr J, Brown KK, Walsh SLF, Devaraj A, Song JW, Wuyts WA, Valenzuela C, Goeldner RG, Stowasser S, Schlenker-Herceg R, Wells AUet al., 2021, Does HRCT pattern influence the effect of nintedanib in patients with progressive fibrosing interstitial lung diseases?, Publisher: GEORG THIEME VERLAG KG, Pages: S31-S32, ISSN: 0934-8387

Conference paper

Koschel D, Flaherty KR, Wells AU, Cottin V, Devaraj A, Inoue Y, Richeldi L, Walsh SLF, Kolb M, Moua T, Stowasser S, Goeldner RG, Schlenker-Herceg R, Brown KKet al., 2021, Effects of nintedanib on progression of ILD in patients with fibrosing ILDs and a progressive phenotype: further analyses of the INBUILD trial, Publisher: GEORG THIEME VERLAG KG, Pages: S30-S31, ISSN: 0934-8387

Conference paper

Horst C, Dickson J, Tisi S, Hall H, Verghese P, Mullin A, Farrelly L, Levermore C, Gyertson K, Clarke C, Allen B, Hamilton S, Hartman A, Nair A, Devaraj A, Hackshaw A, Janes Set al., 2021, The SUMMIT Study: Pulmonary Nodule and Incidental Findings in the First 10,000 Participants of a Population-Based Low-Dose CT Screening Study, Publisher: ELSEVIER SCIENCE INC, Pages: S473-S474, ISSN: 1556-0864

Conference paper

Invernizzi R, Wu BG, Barnett J, Ghai P, Kingston S, Hewitt RJ, Feary J, Li Y, Chua F, Wu Z, Wells AU, Renzoni EA, Nicholson AG, Rice A, Devaraj A, Segal LN, Byrne AJ, Maher TM, Lloyd CM, Molyneaux PLet al., 2021, The respiratory microbiome in chronic hypersensitivity pneumonitis is distinct from that of idiopathic pulmonary fibrosis, American Journal of Respiratory and Critical Care Medicine, Vol: 203, Pages: 339-347, ISSN: 1073-449X

RATIONALE: Chronic hypersensitivity pneumonitis (CHP) is a condition that arises following repeated exposure and sensitisation to inhaled antigens. The lung microbiome is increasingly implicated in respiratory disease but to date, no study has investigated the composition of microbial communities in the lower airways in CHP. OBJECTIVE: To characterise and compare the airway microbiome in subjects with CHP, idiopathic pulmonary fibrosis (IPF) and controls. METHODS: We prospectively recruited individuals diagnosed with CHP (n=110), IPF (n=45) and controls (n=28). Subjects underwent bronchoalveolar lavage and bacterial DNA was isolated, quantified by qPCR and the 16S rRNA gene was sequenced to characterise the bacterial communities in the lower airways. MAIN MEASUREMENTS AND RESULTS: Distinct differences in the microbial profiles were evident in the lower airways of subjects with CHP and IPF. At the phylum level, the prevailing microbiota of both IPF and CHP subjects included Firmicutes, Bacteroidetes, Proteobacteria and Actinobacteria. However, in IPF, Firmicutes dominated while the percentage of reads assigned to Proteobacteria in the same group was significantly lower compared to CHP subjects. At the genus level, Staphylococcus was increased in CHP and Actinomyces and Veillonella in IPF. The lower airway bacterial burden in CHP subjects was higher than controls but lower than those with IPF. In contrast to IPF, there was no association between bacterial burden and survival in CHP. CONCLUSIONS: The microbial profile of the lower airways in subjects with CHP is distinct from that of IPF and, notably, bacterial burden in individuals with CHP fails to predict survival.

Journal article

Barnett J, Pulzato I, Javed M, Lee YJ, Choraria A, Kemp S, Rice A, Jordan S, Shah PL, Nicholson AG, Padley S, Devaraj Aet al., 2021, Radiological-pathological correlation of negative CT biopsy results enables high negative predictive value for thoracic malignancy, CLINICAL RADIOLOGY, Vol: 76, ISSN: 0009-9260

Journal article

Dafydd DAP, O'Mahony M, Jhanji S, Devaraj A, Allum W, Nicol D, Blunt DM, Riddell AMet al., 2021, The role of CT chest in screening for asymptomatic COVID-19 infection in self-isolating patients prior to elective oncological surgery: findings from a UK Cancer Hub, BRITISH JOURNAL OF RADIOLOGY, Vol: 94, ISSN: 0007-1285

Journal article

George PM, Hida T, Putman RK, Hino T, Desai SR, Devaraj A, Kumar S, Mackintosh JA, Gudnason V, Hatabu H, Gudmundsson G, Hunninghake GM, George PM, Hida T, Putman RK, Desai SR, Devaraj A, Kumar S, Mackintosh JA, Gudnason V, Hatabu H, Gudmundsson G, Hunninghake GMet al., 2020, Hiatus hernia and interstitial lung abnormalities, EUROPEAN RESPIRATORY JOURNAL, Vol: 56, ISSN: 0903-1936

Journal article

Arachchillage DJ, Desai SR, Devaraj A, Ridge CA, Padley SPG, Patel BV, all authors of Pulmonary Angiopathy in Severe COVID-19 Physiologic, Imaging and Hematologic Observationset al., 2020, Reply to: Sanfilippo et al Caviedes et al., American Journal of Respiratory and Critical Care Medicine, Vol: 203, Pages: 261-263, ISSN: 1073-449X

Journal article

George PM, Barratt SL, Condliffe R, Desai SR, Devaraj A, Forrest I, Gibbons MA, Hart N, Jenkins RG, McAuley DF, Patel BV, Thwaite E, Spencer LGet al., 2020, Respiratory follow-up of patients with COVID-19 pneumonia, Thorax, Vol: 75, Pages: 1009-1016, ISSN: 0040-6376

The COVID-19 pandemic has led to an unprecedented surge in hospitalised patients with viral pneumonia. The most severely affected patients are older men, individuals of black and Asian minority ethnicity and those with comorbidities. COVID-19 is also associated with an increased risk of hypercoagulability and venous thromboembolism. The overwhelming majority of patients admitted to hospital have respiratory failure and while most are managed on general wards, a sizeable proportion require intensive care support. The long-term complications of COVID-19 pneumonia are starting to emerge but data from previous coronavirus outbreaks such as severe acute respiratory syndrome (SARS) and Middle East respiratory syndrome (MERS) suggest that some patients will experience long-term respiratory complications of the infection. With the pattern of thoracic imaging abnormalities and growing clinical experience, it is envisaged that interstitial lung disease and pulmonary vascular disease are likely to be the most important respiratory complications. There is a need for a unified pathway for the respiratory follow-up of patients with COVID-19 balancing the delivery of high-quality clinical care with stretched National Health Service (NHS) resources. In this guidance document, we provide a suggested structure for the respiratory follow-up of patients with clinicoradiological confirmation of COVID-19 pneumonia. We define two separate algorithms integrating disease severity, likelihood of long-term respiratory complications and functional capacity on discharge. To mitigate NHS pressures, virtual solutions have been embedded within the pathway as has safety netting of patients whose clinical trajectory deviates from the pathway. For all patients, we suggest a holistic package of care to address breathlessness, anxiety, oxygen requirement, palliative care and rehabilitation.

Journal article

Ridge CA, Desai SR, Jeyin N, Mahon C, Lother DL, Mirsadraee S, Semple T, Price S, Bleakley C, Arachchillage DJ, Shaw E, Patel BV, Padley SPG, Devaraj Aet al., 2020, Dual-energy CT pulmonary angiography (DECTPA) quantifies vasculopathy in severe COVID-19 pneumonia, Radiology: Cardiothoracic Imaging, Vol: 2, ISSN: 2638-6135

BackgroundThe role of dual energy computed tomographic pulmonary angiography (DECTPA) in revealing vasculopathy in coronavirus disease 2019 (COVID-19) has not been fully explored.PurposeTo evaluate the relationship between DECTPA and disease duration, right ventricular dysfunction (RVD), lung compliance, D-dimer and obstruction index in COVID-19 pneumonia.Materials and MethodsThis institutional review board approved this retrospective study, and waived the informed consent requirement. Between March-May 2020, 27 consecutive ventilated patients with severe COVID-19 pneumonia underwent DECTPA to diagnose pulmonary thrombus (PT); 11 underwent surveillance DECTPA 14 ±11.6 days later. Qualitative and quantitative analysis of perfused blood volume (PBV) maps recorded: i) perfusion defect ‘pattern’ (wedge-shaped, mottled or amorphous), ii) presence of PT and CT obstruction index (CTOI) and iii) PBV relative to pulmonary artery enhancement (PBV/PAenh); PBV/PAenh was also compared with seven healthy volunteers and correlated with D-Dimer and CTOI.ResultsAmorphous (n=21), mottled (n=4), and wedge-shaped (n=2) perfusion defects were observed (M=20; mean age=56 ±8.7 years). Mean extent of perfusion defects=36.1%±17.2. Acute PT was present in 11/27(40.7%) patients. Only wedge-shaped defects corresponded with PT (2/27, 7.4%). Mean CTOI was 2.6±5.4 out of 40. PBV/PAenh (18.2 ±4.2%) was lower than in healthy volunteers (27 ±13.9%, p = 0.002). PBV/PAenh correlated with disease duration (β = 0.13, p = 0.04), and inversely correlated with RVD (β = -7.2, p = 0.001), persisting after controlling for confounders. There were no linkages between PBV/PAenh and D-dimer or CTOI.ConclusionPerfusion defects and decreased PBV/PAenh are prevalent in severe COVID-19 pneumonia. PBV/PAenh correlates with disease duration and inversely correlates with RVD. PBV/PAenh may be an important marker of vasculopathy in severe COVID-19 pneumonia

Journal article

Bartlett EC, Kemp S, Ridge CA, Desai SR, Mirsadraee S, Morjaria JB, Shah PL, Popat S, Nicholson AG, Rice AJ, Jordan S, Begum S, Mani A, Derbyshire J, Morris K, Chen M, Peacock C, Addis J, Martins M, Kaye SB, Padley SPG, Devaraj A, McDonald F, Robertus JL, Lim E, Barnett J, Finch J, Dalal P, Yousaf N, Jamali A, Ivashniova N, Phillips C, Newsom-Davies T, Lee R, Vaghani P, Whiteside S, Vaughan-Smith Set al., 2020, Baseline Results of the West London lung cancer screening pilot study - Impact of mobile scanners and dual risk model utilisation, LUNG CANCER, Vol: 148, Pages: 12-19, ISSN: 0169-5002

Journal article

Horst C, Dickson JL, Tisi S, Ruparel M, Nair A, Devaraj A, Janes SMet al., 2020, Delivering low-dose CT screening for lung cancer: a pragmatic approach, THORAX, Vol: 75, Pages: 831-832, ISSN: 0040-6376

Journal article

Ruparel M, Quaife SL, Dickson JL, Horst C, Tisi S, Hall H, Taylor M, Ahmed A, Shaw P, Burke S, Soo M-J, Nair A, Devaraj A, Sennett K, Duffy SW, Navani N, Bhowmik A, Baldwin DR, Janes SMet al., 2020, Lung Screen Uptake Trial: results from a single lung cancer screening round, THORAX, Vol: 75, Pages: 908-912, ISSN: 0040-6376

Journal article

George P, Hida T, Putman R, Desai S, Devaraj A, Kumar S, MacKintosh J, Gudnason V, Hatabu H, Gudmundsson G, Hunninghake Get al., 2020, Hiatus hernia and interstitial lung abnormalities, Publisher: EUROPEAN RESPIRATORY SOC JOURNALS LTD, ISSN: 0903-1936

Conference paper

Feary J, Parfrey H, Burge S, Nicholson AG, Devaraj A, Cullinan Pet al., 2020, Interstitial Lung Disease (ILD) in aluminium welders, Publisher: EUROPEAN RESPIRATORY SOC JOURNALS LTD, ISSN: 0903-1936

Conference paper

Flaherty KR, Wells AU, Cottin V, Devaraj A, Inoue Y, Richeldi L, Walsh SLF, Kolb M, Koschel D, Moua T, Stowasser S, Goeldner R-G, Schlenker-Herceg R, Brown KKet al., 2020, Effects of nintedanib on progression of ILD in patients with fibrosing ILDs and a progressive phenotype: further analyses of the INBUILD trial, Publisher: EUROPEAN RESPIRATORY SOC JOURNALS LTD, ISSN: 0903-1936

Conference paper

Nicholson AG, Osborn M, Devaraj A, Wells AUet al., 2020, COVID-19 related lung pathology: old patterns in new clothing?, Histopathology, Vol: 77, Pages: 169-172, ISSN: 0309-0167

Journal article

Nair A, Rodrigues JCL, Hare SS, Edey A, Devaraj A, Jacob J, Johnstone A, McStay R, Denton E, Robinson Get al., 2020, A British Society of Thoracic Imaging statement: considerations in designing local imaging diagnostic algorithms for the COVID-19 pandemic. A reply, CLINICAL RADIOLOGY, Vol: 75, Pages: 637-637, ISSN: 0009-9260

Journal article

Patel BV, Arachchillage DJ, Ridge CA, Bianchi P, Doyle JF, Garfield B, Ledot S, Morgan C, Passariello M, Price S, Singh S, Thakuria L, Trenfield S, Trimlett R, Weaver C, Wort SJ, Xu T, Padley SPG, Devaraj A, Desai SRet al., 2020, Pulmonary angiopathy in severe COVID-19: physiologic, imaging and hematologic observations, American Journal of Respiratory and Critical Care Medicine, Vol: 202, Pages: 690-699, ISSN: 1073-449X

Rationale: Clinical and epidemiologic data in coronavirus disease 2019 (Covid-19) have accrued rapidly since the outbreak but few address the underlying pathophysiology. Objectives: To ascertain the physiologic, hematologic and imaging basis of lung injury in severe Covid-19 pneumonia. Methods: Clinical, physiologic and laboratory data were collated. Radiologic (computed tomography pulmonary angiography [CTPA, n=39] and dual-energy CT [DECT, n=20]) studies were evaluated: observers quantified CT patterns (including the extent of abnormal lung and the presence/extent of dilated peripheral vessels) and perfusion defects on DECT. Coagulation status was assessed using thromboelastography (TEG). Measurements and Results: In 39 consecutive patients (M:F 32:7; mean age, 53±10 years [range 29-79 years]; black and ethnic minority, n=25 [64%]), there was a significant vascular perfusion abnormality and increased physiologic dead-space (dynamic compliance, 33.7±14.7 mls/cmH2O; Murray Lung Injury Score, 3.14±0.53; mean ventilatory ratios, 2.6±0.8) with evidence of hypercoagulability and fibrinolytic ‘shutdown’. The mean CT extent (±SD) of normally-aerated lung, ground-glass opacification and dense parenchymal opacification were 23.5±16.7%, 36.3±24.7% and 42.7±27.1%, respectively. Dilated peripheral vessels were present in 21/33 (63.6%) patients with at least two assessable lobes (including 10/21 [47.6%] with no evidence of acute pulmonary emboli). Perfusion defects on DECT (assessable in 18/20 [90%]), were present in all patients (wedge-shaped, n=3; mottled, n= 9; mixed pattern, n=6). Conclusions: Physiologic, hematologic and imaging data show not only the presence of a hypercoagulable phenotype in severe Covid-19 pneumonia but also markedly impaired pulmonary perfusion likely caused by pulmonary angiopathy and thrombosis.

Journal article

Ruparel M, Quaife SL, Dickson JL, Horst C, Tisi S, Hall H, Taylor MN, Ahmed A, Shaw PJ, Burke S, Soo M-J, Nair A, Devaraj A, Sennett K, Hurst JR, Duffy SW, Navani N, Bhowmik A, Baldwin DR, Janes SMet al., 2020, Prevalence, Symptom Burden, and Underdiagnosis of Chronic Obstructive Pulmonary Disease in a Lung Cancer Screening Cohort, ANNALS OF THE AMERICAN THORACIC SOCIETY, Vol: 17, Pages: 869-878, ISSN: 1546-3222

Journal article

AbdulJabbar K, Raza SEA, Rosenthal R, Jamal-Hanjani M, Veeriah S, Akarca A, Lund T, Moore DA, Salgado R, Al Bakir M, Zapata L, Hiley CT, Officer L, Sereno M, Smith CR, Loi S, Hackshaw A, Marafioti T, Quezada SA, McGranahan N, Le Quesne J, TRACERx Consortium, Swanton C, Yuan Yet al., 2020, Geospatial immune variability illuminates differential evolution of lung adenocarcinoma., Nat Med, Vol: 26, Pages: 1054-1062

Remarkable progress in molecular analyses has improved our understanding of the evolution of cancer cells toward immune escape1-5. However, the spatial configurations of immune and stromal cells, which may shed light on the evolution of immune escape across tumor geographical locations, remain unaddressed. We integrated multiregion exome and RNA-sequencing (RNA-seq) data with spatial histology mapped by deep learning in 100 patients with non-small cell lung cancer from the TRACERx cohort6. Cancer subclones derived from immune cold regions were more closely related in mutation space, diversifying more recently than subclones from immune hot regions. In TRACERx and in an independent multisample cohort of 970 patients with lung adenocarcinoma, tumors with more than one immune cold region had a higher risk of relapse, independently of tumor size, stage and number of samples per patient. In lung adenocarcinoma, but not lung squamous cell carcinoma, geometrical irregularity and complexity of the cancer-stromal cell interface significantly increased in tumor regions without disruption of antigen presentation. Decreased lymphocyte accumulation in adjacent stroma was observed in tumors with low clonal neoantigen burden. Collectively, immune geospatial variability elucidates tumor ecological constraints that may shape the emergence of immune-evading subclones and aggressive clinical phenotypes.

Journal article

Biswas D, Birkbak NJ, Rosenthal R, Hiley CT, Lim EL, Papp K, Boeing S, Krzystanek M, Djureinovic D, La Fleur L, Greco M, Döme B, Fillinger J, Brunnström H, Wu Y, Moore DA, Skrzypski M, Abbosh C, Litchfield K, Al Bakir M, Watkins TBK, Veeriah S, Wilson GA, Jamal-Hanjani M, Moldvay J, Botling J, Chinnaiyan AM, Micke P, Hackshaw A, Bartek J, Csabai I, Szallasi Z, Herrero J, McGranahan N, Swanton C, TRACERx Consortiumet al., 2020, Publisher Correction: A clonal expression biomarker associates with lung cancer mortality., Nat Med, Vol: 26

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

Journal article

Joshi K, de Massy MR, Ismail M, Reading JL, Uddin I, Woolston A, Hatipoglu E, Oakes T, Rosenthal R, Peacock T, Ronel T, Noursadeghi M, Turati V, Furness AJS, Georgiou A, Wong YNS, Ben Aissa A, Sunderland MW, Jamal-Hanjani M, Veeriah S, Birkbak NJ, Wilson GA, Hiley CT, Ghorani E, Guerra-Assunção JA, Herrero J, Enver T, Hadrup SR, Hackshaw A, Peggs KS, McGranahan N, Swanton C, TRACERx consortium, Quezada SA, Chain Bet al., 2020, Publisher Correction: Spatial heterogeneity of the T cell receptor repertoire reflects the mutational landscape in lung cancer., Nat Med, Vol: 26

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

Journal article

Chemi F, Rothwell DG, McGranahan N, Gulati S, Abbosh C, Pearce SP, Zhou C, Wilson GA, Jamal-Hanjani M, Birkbak N, Pierce J, Kim CS, Ferdous S, Burt DJ, Slane-Tan D, Gomes F, Moore D, Shah R, Al Bakir M, Hiley C, Veeriah S, Summers Y, Crosbie P, Ward S, Mesquita B, Dynowski M, Biswas D, Tugwood J, Blackhall F, Miller C, Hackshaw A, Brady G, Swanton C, Dive C, TRACERx Consortiumet al., 2020, Publisher Correction: Pulmonary venous circulating tumor cell dissemination before tumor resection and disease relapse., Nat Med, Vol: 26

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

Journal article

Folgoc LL, Baltatzis V, Alansary A, Desai S, Devaraj A, Ellis S, Manzanera OEM, Kanavati F, Nair A, Schnabel J, Glocker Bet al., 2020, Bayesian sampling bias correction: training with the right loss function, Publisher: arXiv

We derive a family of loss functions to train models in the presence ofsampling bias. Examples are when the prevalence of a pathology differs from itssampling rate in the training dataset, or when a machine learning practionerrebalances their training dataset. Sampling bias causes large discrepanciesbetween model performance in the lab and in more realistic settings. It isomnipresent in medical imaging applications, yet is often overlooked attraining time or addressed on an ad-hoc basis. Our approach is based onBayesian risk minimization. For arbitrary likelihood models we derive theassociated bias corrected loss for training, exhibiting a direct connection toinformation gain. The approach integrates seamlessly in the current paradigm of(deep) learning using stochastic backpropagation and naturally with Bayesianmodels. We illustrate the methodology on case studies of lung nodule malignancygrading.

Working paper

This data is extracted from the Web of Science and reproduced under a licence from Thomson Reuters. You may not copy or re-distribute this data in whole or in part without the written consent of the Science business of Thomson Reuters.

Request URL: http://wlsprd.imperial.ac.uk:80/respub/WEB-INF/jsp/search-html.jsp Request URI: /respub/WEB-INF/jsp/search-html.jsp Query String: respub-action=search.html&id=00528873&limit=30&person=true