key: cord-1028580-u1h9ckye
authors: Subesinghe, Manil; Bhuva, Shaheel; Dunn, Joel T; Hammers, Alexander; Cook, Gary J; Barrington, Sally F; Fischer, Barbara M
title: A case–control evaluation of pulmonary and extrapulmonary findings of incidental asymptomatic COVID-19 infection on FDG PET-CT
date: 2022-02-01
journal: Br J Radiol
DOI: 10.1259/bjr.20211079
sha: b679d361a7f4b42da83d8784f02d0d63db47916b
doc_id: 1028580
cord_uid: u1h9ckye

OBJECTIVES: To describe the findings of incidental asymptomatic COVID-19 infection on FDG PET-CT using a case–control design. METHODS: Incidental pulmonary findings suspicious of asymptomatic COVID-19 infection on FDG PET-CT were classified as a confirmed (positive RT-PCR test) or suspected case (no/negative RT-PCR test). Control cases were identified using a 4:1 control:case ratio. Pulmonary findings were re-categorised by two reporters using the BSTI classification. SUV metrics in ground glass opacification (GGO)/consolidation (where present), background lung, intrathoracic nodes, liver, spleen and bone marrow were measured. RESULTS: 7/9 confirmed and 11/15 suspected cases (COVID-19 group) were re-categorised as BSTI 1 (classic/probable COVID-19) or BSTI 2 (indeterminate COVID-19); 0/96 control cases were categorised as BSTI 1. Agreement between two reporters using the BSTI classification was almost perfect (weighted κ = 0.94). SUV(max) GGO/consolidation (5.1 vs 2.2; p < 0.0001) and target-to-background ratio, normalised to liver SUV(mean) (2.4 vs 1.0; p < 0.0001) were higher in the BSTI 1 & 2 group vs BSTI 3 (non-COVID-19) cases. SUV(max) GGO/consolidation discriminated between the BSTI 1 & 2 group vs BSTI 3 (non-COVID-19) cases with high accuracy (AUC = 0.93). SUV metrics were higher (p < 0.05) in the COVID-19 group vs control cases in the lungs, intrathoracic nodes and spleen. CONCLUSION: Asymptomatic COVID-19 infection on FDG PET-CT is characterised by bilateral areas of FDG avid (intensity > x2 liver SUV(mean)) GGO/consolidation and can be identified with high interobserver agreement using the BSTI classification. There is generalised background inflammation within the lungs, intrathoracic nodes and spleen. ADVANCES IN KNOWLEDGE: Incidental asymptomatic COVID-19 infection on FDG PET-CT, characterised by bilateral areas of ground glass opacification and consolidation, can be identified with high reproducibility using the BSTI classification. The intensity of associated FDG uptake (>x2 liver SUV(mean)) provides high discriminative ability in differentiating such cases from pulmonary findings in a non-COVID-19 pattern. Asymptomatic COVID-19 infection causes a generalised background inflammation within the mid-lower zones of the lungs, hilar and central mediastinal nodal stations, and spleen on FDG PET-CT.

The coronavirus disease 2019 (COVID-19) pandemic has created the biggest global health crisis in generations. The spread of infection has been difficult to control due to asymptomatic infection, estimated to account for over 50% of all transmissions. 1 Nasopharyngeal swab reverse transcriptase polymerase chain reaction (RT-PCR) is considered the gold standard for diagnosing COVID-19 infection. 2, 3 The sensitivity of RT-PCR in symptomatic patients ranges between 82 and 97% 4 but detection rates are lower in asymptomatic individuals. 5, 6 During the early stages of the pandemic in 2020, numerous studies described computed tomography (CT) features characteristic of COVID-19 infection with some suggesting sufficient diagnostic accuracy of CT in the absence of RT-PCR testing 7, 8 ; significant selection bias and several confounding factors have since undermined such conclusions. 9 International guidelines and a recent umbrella review recommend CT as a problem-solving tool to identify complications of COVID-19 infection or when an alternative diagnosis is suspected in symptomatic individuals. 10, 11 The British Society of Thoracic Imaging (BSTI) published discriminated between the BSTI 1 & 2 group vs BSTI 3 (non-COVID-19) cases with high accuracy (AUC = 0.93). SUV metrics were higher (p < 0.05) in the COVID-19 group vs control cases in the lungs, intrathoracic nodes and spleen. Conclusion: Asymptomatic COVID-19 infection on FDG PET-CT is characterised by bilateral areas of FDG avid (intensity > x2 liver SUV mean ) GGO/consolidation and can be identified with high interobserver agreement using the BSTI classification. There is generalised background inflammation within the lungs, intrathoracic nodes and spleen. Advances in knowledge: Incidental asymptomatic COVID-19 infection on FDG PET-CT, characterised by bilateral areas of ground glass opacification and consolidation, can be identified with high reproducibility using the BSTI classification. The intensity of associated FDG uptake (>x2 liver SUV mean ) provides high discriminative ability in differentiating such cases from pulmonary findings in a non-COVID-19 pattern. Asymptomatic COVID-19 infection causes a generalised background inflammation within the mid-lower zones of the lungs, hilar and central mediastinal nodal stations, and spleen on FDG PET-CT. (Continued) criteria for the diagnosis of COVID-19 infection on CT, based on the presence and distribution of ground glass opacification (GGO), consolidation and varied patterns of organising pneumonia (OP). 12 Asymptomatic individuals can have normal lungs on CT or alternatively demonstrate radiological features compatible with COVID-19 infection. [13] [14] [15] Several case reports/series of asymptomatic COVID-19 infection on 2-deoxy-2-[ 18 F]fluoro-D-glucose (FDG) positron emission tomography (PET)-CT report metabolically active findings mainly confined to the lungs and mediastinal lymph nodes [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] ; most studies have been purely descriptive, however. A few studies reporting increased FDG uptake in extrathoracic nodes, spleen, and bone marrow, suggest that FDG PET-CT can demonstrate the immune response to viral infections. 25 PET-CT in patients with suspected asymptomatic COVID-19 infection scanned during the 'first wave' of UK pandemic, are significantly different to those in a control group scanned prior to the pandemic, matched for age, gender and scan indication. We will also determine the ability of FDG uptake in conjunction with pulmonary findings categorised using the BSTI classification to discriminate between COVID-19 and non-COVID-19 infection, whilst assessing the interobserver agreement between two reporters using the BSTI classification.

Institutional review board approval was obtained for this retrospective non-interventional observational case-control study. Inclusion criteria were:

• FDG PET-CT examination performed between 23/03/2020 and 29/05/2020 during the 'first wave' of the UK pandemic. • Absence of new continuous cough or high temperature, i.e., asymptomatic. • Expedited (via email notification) FDG PET-CT report due to incidental pulmonary findings suspicious for asymptomatic COVID-19 infection.

Referring clinicians either opted for confirmation of COVID-19 infection via RT-PCR hospital testing or recommendation for a period of self-isolation for 14 days as per UK government guidance due to a lack of community testing at the time. 30 Patients with a positive RT-PCR test within 28 days of scanning were classified as a confirmed case. Patients with pulmonary findings suspicious of COVID-19 infection on FDG PET-CT but no or negative RT-PCR test within 28 days of scanning were classified as a suspected case. Information on RT-PCR testing and clinical follow-up was obtained from institutional electronic databases. Consecutive control cases matched for age, gender and scan indication without exclusion criteria were identified from spring 2019 ± 3 months, using a 4:1 control: case ratio.

FDG PET-CT imaging review FDG PET-CT examinations were performed using methodology aligned to EANM guidance and described previously. 31 All examinations were anonymised (including date of examination) and analysed using Hybrid Viewer (Hermes Medical Solutions, Sweden). Independent blinded review of pulmonary findings was undertaken 6 months after the 'first wave' by board certified radiologists (S.B) and consultant radiologist (M.S) with 1 and 10 years of PET-CT reporting experience, respectively, and each with 12 years of diagnostic CT (including thoracic CT) reporting experience. Pulmonary findings were categorised using the BSTI classification 12 ; classic/probable COVID-19 (BSTI 1), indeterminate COVID-19 (BSTI 2), non-COVID-19 (BSTI 3), and normal (BSTI 4). The normal category (BSTI 4) included findings considered within the spectrum of normality for PET-CT, e.g., gravity-dependent GGO and basal linear atelectasis. Following independent review, examinations with disagreement in BSTI classification had consensus reads. For examinations with clinically significant pulmonary parenchymal findings, i.e., BSTI 1-3, the highest maximum standardised uptake value (SUV max ) in an area of GGO/consolidation was documented, enabling targetto-background ratio (TBR) calculation, normalised to the mean standardised uptake value (SUV mean ) in the liver.

SUV metrics were derived from normal lung and from extrapulmonary sites (intrathoracic nodes, liver, spleen and bone marrow) by a consultant nuclear medicine physician (B.M.F) with 15 years of PET-CT reporting experience. Freehand regions of interest (ROIs) following the contours of the lungs but excluding subpleural regions and avoiding major vessels or parenchymal abnormalities were drawn in the upper (level of suprasternal notch), mid (1 cm below the carina) and lower zones (2.5 cm above the right hepatic dome) of both lungs to calculate SUV mean of background lung. ROIs were drawn around the major intrathoracic nodal stations 32 to calculate nodal SUV max ; SUV max was only measured if lymph nodes were visible on CT. Spherical volumes of interest (VOIs) were placed in the right lobe of the liver (6 cm diameter), spleen (3 cm diameter) and L4 vertebral body (2 cm diameter) as a representation of marrow uptake, to calculate SUV mean in these VOIs.

Interobserver agreement using the BSTI classification was assessed using the weighted κ method. 33 Non-parametric tests were used to assess for group-wise (Kruskal-Wallis) and pairwise (Mann Whitney U) differences. The Benjamini-Hochberg method to estimate the false discovery rate (FDR) was used to correct for multiple comparisons; an FDR < 0.05 was considered significant. Receiver operating characteristic (ROC) curves were generated with an area under the curve (AUC) calculated for each ROC 34 with the best thresholds for group discrimination defined using Youden's method. 35 All analyses were performed in R v. 4.0.0 with the base and stats packages while ROC analyses were performed using the pROC package. More detailed information can be found in the Supplementary Material 1 (range 285-392 MBq), mean uptake time: 63 ± 4.5 min (range 56-78 min)).

BSTI classification 7/9 confirmed and 11/15 suspected cases, were categorised as BSTI 1 or 2 (Figures 1-3) , 0/96 control cases were categorised as BSTI 1, and only two control cases categorised as BSTI 2 ( Figure 4 ). 2/9 confirmed cases (Figures 5 and 6 ) and 4/15 suspected cases were categorised as BSTI 3, whilst the remaining control cases were categorised as BSTI 3 (18/96) or BSTI 4 (76/96) (Figure 7 ). The BSTI classification had a sensitivity of 75% and specificity of 97.9% for the detection of COVID-19 infection on the CT component of FDG PET-CT, assuming that only BSTI 1 and 2 appearances represent COVID-19 infection.

There was almost perfect agreement (weighted κ = 0.94) between the two reporters using the BSTI classification across all four categories with an overall agreement of 94% (113/120). Excluding BSTI 4, which had 100% agreement (76/76), there remained almost perfect agreement for the BSTI 1-3 categories (weighted κ = 0.83) with an overall agreement of 84% (37/44); cases with disagreement only differed by one category between reporters ( Table 2) .

There were highly significant group-wise differences (p < 0.0001) across both the COVID-19 and BSTI classifications. Pairwise comparisons across the COVID-19 classification revealed no difference in SUV max (p = 0.056) or TBR (p = 0.066) GGO/ consolidation between confirmed vs suspected cases after correction for multiple comparisons. SUV max GGO/consolidation was, however, significantly higher in the COVID-19 group (confirmed and suspected cases) vs control cases (4.7 vs 2.1; p < 0.0001) as was TBR (2.2 vs 1.0; p < 0.0001), (Tables 3 and 4 , Figures 8 and 9 ).

Pairwise comparisons across the BSTI classification revealed no differences in SUV max GGO/consolidation (p = 0.088) or TBR (p = 0.064) between BSTI 1 and 2 cases. SUV max GGO/consolidation Figures 8 and 9 ). SUV max GGO/consolidation ROC analysis ROC curves indicated excellent discrimination using SUV max GGO/consolidation with an AUC of 0.93 (0.84-1.00) for differentiating between the BSTI 1 & 2 group and BSTI 3 cases and 0.87 (0.75-0.99) between the COVID-19 group and control cases (Figure 10 ). Using a SUV max 3.15 cut-off, discrimination between the BSTI 1 & 2 group and BSTI 3 cases was achievable with a sensitivity of 0.90 and specificity of 0.88, whilst a SUV max 3.45 cut-off enabled discrimination between the COVID-19 group and control cases with a sensitivity of 0.75 and specificity of 0.95.

There were significantly higher SUV metrics (p < 0.05) in the COVID-19 group vs control cases in 9/15 regions; 3/6 pulmonary regions (right mid zone, right lower zone, left lower zone), 5/6 nodal regions (bilateral hilar, bilateral paratracheal and subcarinal mediastinal nodal stations) and in the spleen. There was no significant difference in SUV mean in the liver or bone marrow (Supplementary Table 2 , Supplementary Figure 1) .

3.3% (24/732) of FDG PET-CT examinations performed during spring 2020 had incidental pulmonary findings suspicious of asymptomatic COVID-19 infection, which is within the quoted incidence range (2.1-16.2%) from a systematic review of 11 studies. 36 Our incidence is lower than reported in a study from a similar sized London institution (9.4%), 19 but this may be related to potential false-positive observations secondary to unilateral rather bilateral pulmonary findings, i.e., indeterminate for COVID-19 (BSTI 2) coupled with most of their cases with thoracic findings on PET-CT, either negative (4/15) or without RT-PCR confirmation (10/15).

Blinded consensus pulmonary analysis performed 6 months later, with a greater experience of reporting COVID-19 infection on FDG PET-CT, categorised 18/24 of the confirmed and suspected cases as either BSTI 1 (classic/probable COVID-19) or BSTI 2 (indeterminate for COVID-19), whilst none of the control cases from 2019 were categorised as BSTI 1. This confirms that the pattern of FDG avid pulmonary parenchymal changes observed in spring 2020 (Figures 1-3) was a novel phenomenon not experienced before, and also demonstrates the high specificity (97.9%) achieved through using the BSTI classification.

Several studies report an increased incidence of pulmonary findings suspicious for COVID-19 infection during the 'first wave' compared to control cases [20] [21] [22] 24 similar to ours, except that patterns compatible with COVID-19 interstitial pneumonia were observed in their control cohorts; this is likely due to the presence of COVID-19 mimics on FDG PET-CT, e.g., influenza pneumonia or OP related to connective tissue disease or drug toxicity. However, most studies did not use a standardised CT grading system for categorising pulmonary changes, likely reducing specificity. Maurea et al 20 using the COVID-19 Reporting and Data System (CORADS) 37 reported 14/335 (4%) control cases with 'abnormal PET-CT findings suspicious for COVID-19 infection' . However, 9/14 (64%) were classified as CO-RADS 2 (CT abnormalities consistent infection other than or CO-RADS 3 (uncertain CT findings for suggesting an overestimation of PET-CT findings suspicious for COVID-19 infection in their control population.

Our study, the first to formally assess agreement between two reporters using the BSTI classification on FDG PET-CT, demonstrated almost perfect agreement (weighted κ = 0.94), even when BSTI 4 cases were excluded from the analysis (weighted κ = 0.83). Inui et al, 38 compared different CT grading systems for COVID-19 infection, and showed that all had reasonable diagnostic performance (0.80-0.84), albeit with lower interobserver agreement (Cohen κ = 0.61-0.63) than ours, but that CO-RADS and BSTI outperformed the other two classifications. Our higher interobserver agreement may be augmented by amalgamation of the 'classic' and 'probable' COVID-19 categories to represent BSTI 1 as per the published COVID-19 CT reporting proforma 12 rather than interpretate them as two separate categories. The inadequacies of the low-dose non-breath hold CT component of a PET-CT examination, requiring a more pragmatic approach to assessing the lungs, i.e., forgoing subtleties, also likely contributed to more consistent and reproducible observations. FDG uptake in areas of GGO/consolidation was significantly higher in the COVID-19 group vs control cases (SUV max 4.7 vs 2.1) and BSTI 1 & 2 group vs BSTI 3 cases (SUV max 5.1 vs 2.2); similar values have been reported in an early systematic review of incidental COVID-19 infection on FDG PET-CT (mean SUV max 4.9), 39 and Italian multicentre study (mean SUV max 4.1). 24 TBR analysis demonstrated that the intensity of FDG uptake in GGO/ consolidation was >x2 liver SUV mean in the COVID-19 and BSTI 1 & 2 groups, and lower and comparable to liver SUV mean for control and BSTI 3 cases. These findings confirm that a distinctive feature of COVID-19 infection is the association of high FDG uptake with areas of GGO/consolidation, related to multinucleated giant cells and focal clusters of lymphomonocytic infiltration in the context of diffuse alveolar damage, demonstrable even in early COVID-19 infection. 40, 41 The discriminative ability of SUV max in areas of GGO/consolidation to differentiate between the BSTI 1 & 2 group and BSTI 3 cases was high; using a SUV max 3.15 cut-off, discrimination between BSTI 1 & 2 group and BSTI 3 cases was achievable with high sensitivity and specificity. From a clinical viewpoint, pulmonary findings compatible with classic/probable COVID-19 (BSTI 1), e.g., bibasal peripheral GGO/consolidation with 'reversehalo' or perilobular opacity, i.e., OP pattern, or indeterminate COVID-19 (BSTI 2), e.g., unilateral, non-peripheral GGO/ consolidation, are likely to have higher levels of associated FDG uptake (SUV max >3.15) in comparison with GGO/consolidation in a non-COVID-19 (BSTI 3) pattern.

Our study reaffirms that the low-dose non-breath hold CT component of the study can enable diagnosis despite not being of 'diagnostic quality' and is not solely for the purposes of attenuation correction and localisation. The absence of a full inspiratory effort and breathing artefact during scanning can limit accuracy, however. Difficulty in detection/characterisation of smaller lesions particularly towards the lung bases limits sensitivity, whilst an increased incidence of dependent GGO alongside areas of basal atelectasis, can be potentially misinterpreted as significant pathology, reducing specificity. Unrelated pulmonary pathologies, e.g., other viral pneumonias, OP secondary to connective tissue disease or drug toxicity or active pulmonary fibrosis can have similar appearances to COVID-19 infection, and will reduce specificity, although in our study this was only encountered in 2/96 control cases (Figure 4 ).

We found significantly higher SUV metrics in the COVID-19 group vs control cases in the mid-lower lung zones, both hilar and central mediastinal nodal stations, and spleen, suggesting the presence of generalised background inflammation. Lower zone predominant background pulmonary inflammation correlates with the tendency of COVID-19 infection to present with bilateral abnormalities affecting both lower lobes. 42,43 FDG avid intrathoracic and supraclavicular nodes with COVID-19 infection have been reported in several studies with varied frequency, with or without CT enlargement, 20, [22] [23] [24] [25] whilst only one study has reported increased splenic uptake (5/13 patients) and increased bone marrow uptake 25 ; extrapulmonary abnormalities involving the salivary glands 19 and gastro-intestinal tract 18 were not routinely assessed for during our study. The presence of generalised systemic inflammation has been confirmed in small cohorts of patients recovering from COVID-19 infection (lungs, mediastinal nodes, spleen, liver, large vessels), 44, 45 as well as in patients with post-COVID syndrome 46 in conjunction with findings of brain hypometabolism. 47, 48 The major limitation to our study, common to many, is the absence of RT-PCR testing for all FDG PET-CT examinations during spring 2020, due to a lack of testing capacity. 30 Patients with COVID-19 infection but without pulmonary findings suspicious of infection will have been missed using our methodology, which will undoubtedly affect the sensitivity and specificity estimate of the BSTI classification on FDG PET-CT. In addition, this limitation also brings into question the classification of suspected cases which had either no (12/15) or a negative (3/15) RT-PCR test despite the presence of pulmonary findings suspicious of COVID-19 infection on FDG PET-CT (Supplementary Table 1 ). However, SUV max GGO/consolidation and TBR analysis demonstrated that confirmed and suspected cases were similar to each other but were individually as well as in combination (COVID-19 group), distinct from control cases, supporting our methodology of combining these cases for analysis (Tables 3 and 4 , Figures 8  and 9 ). In addition, the sensitivity of the gold standard nasopharyngeal RT-PCR in symptomatic individuals is not 100% 4 and is lower in asymptomatic individuals, i.e., higher false-negative rates. 5, 6 Sensitivity can be improved through repeat RT-PCR testing 49 or with bronchoalveolar lavage; a study of 46 patients reported 18 patients (39%) had a positive bronchoalveolar lavage RT-PCR despite two preceding negative nasopharyngeal RT-PCR tests with importantly 13 of these 18 patients (72%) with two preceding negative nasopharyngeal RT-PCR tests having CT findings compatible with COVID-19 infection. 50 This confirms the imperfection of single/multiple nasopharyngeal RT-PCR tests and that pulmonary changes compatible with COVID-19 infection in the context of a negative RT-PCR test(s) cannot be readily dismissed.

Asymptomatic COVID-19 infection on FDG PET-CT is characterised by bilateral areas of GGO/consolidation that are associated with increased FDG uptake (>x2 liver SUV mean ) and which can be identified with high reproducibility using the BSTI classification. These changes occur on the background of generalised inflammation within the mid-lower zones of the lungs, hilar and central mediastinal nodal stations, and spleen. This analysis will enable better preparedness for identification of asymptomatic COVID-19 infection on FDG PET-CT, prompting early confirmation RT-PCR testing, and minimising the risk of undetected infection to both the individual and society as a whole.

AVAILABILITY OF DATA AND MATERIALS: Anonymised data that supports these findings are available from the corresponding author upon reasonable request.

SARScov-2 transmission from people without covid-19 symptoms

Comparison of saliva and nasopharyngeal swab nucleic acid amplification testing for detection of sars-cov-2: a systematic review and meta-analysis

Diagnostic performance of different sampling approaches for sars-cov-2 rt-pcr testing: a systematic review and meta-analysis

Clinical sensitivity of severe acute respiratory syndrome coronavirus 2 nucleic acid amplification tests for diagnosing coronavirus disease 2019

Interpreting diagnostic tests for sars-cov-2

At what times during infection is sars-cov-2 detectable and no longer detectable using rt-pcr-based tests? a systematic review of individual participant data

Sensitivity of chest ct for covid-19: comparison to rt-pcr

Correlation of chest ct and rt-pcr testing for coronavirus disease 2019 (covid-19) in china: a report of 1014 cases

Chest ct and coronavirus disease (covid-19): a critical review of the literature to date

A british society of thoracic imaging statement: considerations in designing local imaging diagnostic algorithms for the covid-19 pandemic

The accuracy of chest CT in the diagnosis of COVID-19: An umbrella review

Covid-19 BSTI reporting templates

Radiological findings from 81 patients with covid-19 pneumonia in wuhan, china: a descriptive study

CT features of sarscov-2 pneumonia according to clinical presentation: a retrospective analysis of 120 consecutive patients from wuhan city

Chest CT Findings in Cases from the Cruise Ship Diamond Princess with Coronavirus Disease (COVID-19)

18F-fdg pet/ct findings of covid-19: a series of four highly suspected cases

FDG pet/ct of covid-19

Identification and resolution of asymptomatic covid-19 pneumonitis and colitis: serial assessment of fluorodeoxyglucose positron emission tomography-computed tomography for evaluation of lung cancer

COVID-19 in the act: incidental 18f-fdg pet/ct findings in asymptomatic patients and those with symptoms not primarily correlated with covid-19 during the united kingdom coronavirus lockdown

FDG-pet/ct imaging during the covid-19 emergency: a southern italian perspective

Incidental findings suggestive of covid-19 in asymptomatic cancer patients undergoing 18f-fdg pet/ct in a low prevalence region

Increased incidence of interstitial pneumonia detected on [18f]-fdg-pet/ct in asymptomatic cancer patients during covid-19 pandemic in lombardy: a casualty or covid-19 infection?

Assessment of extra-parenchymal lung involvement in asymptomatic cancer patients with covid-19 pneumonia detected on 18f-fdg pet-ct studies

Prevalence of interstitial pneumonia suggestive of covid-19 at 18f-fdg pet/ct in oncological asymptomatic patients in a high prevalence country during pandemic period: a national multi-centric retrospective study

COVID-19 pneumonia: relationship between inflammation assessed by whole-body fdg pet/ct and short-term clinical outcome

The potential added value of fdg pet/ct for covid-19 pneumonia

Nuclear medicine in sars-cov-2 pandemia: 18f-fdgpet/ct to visualize covid-19

Mortality in patients with cancer and coronavirus disease 2019: a systematic review and pooled analysis of 52 studies

COVID-19 prevalence and mortality in patients with cancer and the effect of primary tumour subtype and patient demographics: a prospective cohort study

Covid-19: lack of capacity led to halting of community testing in march, admits deputy chief medical officer

The effect of a novel bayesian penalised likelihood pet reconstruction algorithm on the assessment of malignancy risk in solitary pulmonary nodules according to the british thoracic society guidelines

International association for the study of lung cancer (iaslc) lymph node map: radiologic review with ct illustration

The measurement of observer agreement for categorical data

ROCing along: evaluation and interpretation of receiver operating characteristic curves

Index for rating diagnostic tests

Role of 2-[18f]fdg as a radiopharmaceutical for pet/ct in patients with covid-19: a systematic review

CO-rads: a categorical ct assessment scheme for patients suspected of having covid-19-definition and evaluation

Comparison of chest ct grading systems in coronavirus disease 2019 (covid-19) pneumonia

Coronavirus disease 2019 (covid-19) in molecular imaging: a systematic review of incidental detection of sars-cov-2 pneumonia on pet studies

Pulmonary pathology of early-phase 2019 novel coronavirus (covid-19) pneumonia in two patients with lung cancer

Pulmonary post-mortem findings in a series of covid-19 cases from northern italy: a two-centre descriptive study

Chest ct imaging signature of coronavirus disease 2019 infection: in pursuit of the scientific evidence

Chest ct in covid-19: what the radiologist needs to know

Inflammatory response in lungs and extrapulmonary sites detected by [18f] fluorodeoxyglucose pet/ct in convalescing covid-19 patients tested negative for coronavirus

Vasculitis changes in covid-19 survivors with persistent symptoms: an [18f]fdg-pet/ct study

Long covid hallmarks on [18f]fdg-pet/ct: a casecontrol study

18F-fdg brain pet hypometabolism in patients with long covid

Brain metabolic correlates of persistent olfactory dysfunction after sars-cov2 infection

Clinical performance of sars-cov-2 molecular tests

COVID-19 diagnosis in case of two negative nasopharyngeal swabs: association between chest ct and bronchoalveolar lavage results

The authors of this manuscript declare no relationships with any companies, whose products or services may be related to the subject matter of the article.

Anonymised data that supports these findings are available from the corresponding author upon reasonable request.