Article Text
PDF
Abstract
Introduction Haemoptysis can be a feature of lung cancer and patients are typically fast-tracked for evaluation with chest radiography, contrast-enhanced CT and fibreoptic bronchoscopy (FOB).
Objective We aim to explore whether FOB should be conducted as a component of the routine evaluation of non-massive haemoptysis, especially in the context of suspected lung cancer.
Methods MEDLINE, EMBASE and Cochrane Library were searched for studies comparing FOB with CT in the evaluation of non-massive haemoptysis while reporting at least one of the listed primary outcomes. Primary outcomes include sensitivity of diagnostic modality with respect to lung cancer. Secondary outcomes include detection of other aetiologies such as infection. Results were synthesised using a random effects meta-analysis. Sensitivity analysis was performed for patient age group and year of study. Risk of bias assessment was carried out with the Quality Assessment of Diagnostic Accuracy Studies-2 tool.
Results A total of 2273 citations were screened and 11 studies were included, comprising a total sample size of 2015 patients with 226 confirmed cases of lung cancer. A total of 1816 and 1734 patients received a CT scan and FOB, respectively. The pooled sensitivities for detection of lung cancer using CT scan and bronchoscopy were 98% (95% CI 93.0% to 99.0%) and 86% (95% CI 63.0% to 95.0%), respectively. The sensitivity of CT was higher than that of FOB for both primary and secondary outcomes.
Conclusion This study suggests that bronchoscopy does not offer significant additional diagnostic benefit in the evaluation of patients presenting with non-massive haemoptysis and a negative CT scan.
- Bronchoscopy
- Imaging/CT MRI etc
- Lung Cancer
Data availability statement
Data are available in a public, open access repository. Data are available on reasonable request. PubMed, EMBASE and the Cochrane Library were searched from 1983 to 31 December 2022. The complete search strategies are fully described in online supplemental digital content 1. The bibliographies of similar and related systematic reviewers also searched along with the reference sections of included studies. Only articles written in English were considered.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Current UK guidelines recommend consideration of fibre optic bronchoscopy in patients presenting with non-massive haemoptysis if they are at high risk of lung malignancy or if the haemoptysis persists despite a normal thoracic CT. However, as the sensitivity of modern CT scanners has increased, the additional diagnostic benefit of such an intervention remains uncertain.
WHAT THIS STUDY ADDS
This systemic review and meta-analysis indicates that thoracic CT scans exhibit a higher diagnostic efficacy for investigating malignant and non-malignant aetiologies of haemoptysis when compared with fibre optic bronchoscopy. Furthermore, fibre optic bronchoscopy had a higher false negative rate when compared with CT.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
This study highlights the limited role of fibreoptic bronchoscopy in the investigation of patients presenting with non-massive haemoptysis and with a normal thoracic CT scan. A normal scan may be sufficient to exclude thoracic malignancy in this cohort without the need for bronchoscopy.
Introduction
Lung cancer is the predominant cause of cancer-related mortality worldwide.1 2 Data from the National Cancer Registration and Analysis Service reported that in 2018, a total of 47 838 patients of the English population received lung cancer diagnosis while 35 137 events of associated mortality were recorded.3 The past decade has seen an increment of approximately 1% in the lung cancer age-standardised incidence rates with a reduction in the lung cancer age-standardised mortality rate by approximately 14%. Despite the promising findings with respect to mortality, lung cancer persists as the most common cause of cancer-related mortality in the UK, contributing to approximately 21% of all cancer-related mortality.3 This is likely to be associated with the poor prognostic outcomes linked with advanced disease as registry data conveyed a 1-year survival proportion of 83% and 17% for patients with stage 1 and stage 4 disease, respectively.4
As a result, early detection is a key focus in lung cancer management with a drive towards efficient and standardised lung cancer care. The National Optimal Lung Cancer Pathway in collaboration with the National Institute for Health and Care Excellence developed a set of recommendations to facilitate standardised care from the point of referral to multidisciplinary team discussion for treatment discussion and recommendations. Furthermore, several clinical trials have assessed the efficacy of low-dose CT screening in reducing mortality from lung cancer among high-risk populations.5–7 They have demonstrated promising results reporting a significant reduction in lung cancer mortality rates in the screened population. On this basis, the UK National Screening Committee (NSC) has recommended a national screening programme for people aged 55–74 identified as being at high risk of lung cancer, in an effort to facilitate detection at an earlier stage.8
Haemoptysis is a common clinical finding responsible for 6%–11% of all respiratory consultations. It is noted to exhibit a positive predictive value of 2.4%–17% and a prevalence of 2%–6% as an index presenting symptom for patients with lung cancer.9–11 Currently, lung cancer is diagnosed with the aid of a variety of modalities including chest X-rays, CT scans, positron-emission tomography CT) and fibreoptic bronchoscopy (FOB). British Thoracic Society (BTS) guidelines advice bronchoscopy for persistent haemoptysis despite a negative CT scan.12 Tissue sampling via processes such as endobronchial ultrasound or CT or ultrasound-guided percutaneous procedures also form an integral component as they enable ascertainment of pathological diagnosis, disease staging, tumour subtyping, molecular profiling and assessment of predictive markers.13 Advancement in multidetector CT has enabled high-resolution volumetric imaging for lung-specific applications at acceptable levels of radiation exposure.
FOB is another diagnostic modality that is highly used for pulmonary malignancies with sensitivity noted in the endobronchial lesions arising from the central airways.14 The diagnostic yield, however, reduces in the context of peripheral lesions. Furthermore, FOB is a generally a safe and well-tolerated procedure that can be performed in the outpatient setting. Facciolongo et al conducted a multicentre prospective study to evaluate the incidence of complications in 20 986 cases of bronchoscopy. The authors reported a complication prevalence of 1.08% with a total mortality of 0.02%.15 A randomised trial conducted by Laroche et al investigated the sequence of investigation modalities and reviewed the role of CT prior to FOB in the context of suspected lung cancer.16 CT prior to FOB was deemed to be a cost-effective direction with improved diagnostic yield from invasive procedures along with the potential to obviate the need for subsequent investigation.
Several studies have conveyed a superior sensitivity with CT scans for lung cancer in comparison to FOB.17–20 Given this finding along with the logistical and financial burden of facilitating FOB, there might reason to consider the actual necessity and yield of FOB in the diagnostic process of lung cancer in the context of a CT scan devoid of abnormal features. As a result, the aim of this study was to conduct a systematic review and meta-analysis (if appropriate) to compare the diagnostic yield of CT scans and FOB in the presentation of non-massive haemoptysis in addition to assessment of suspected lung cancer and to better define the role of FOB in future diagnostic pathways.
Methods
The Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines were adhered to in the conducting of this systematic review and meta-analysis. The systematic review protocol was published a priori on PROSPERO. The registration number is CRD42023432016.
Patient and public involvement
Patients were not involved in the design of this systematic review and meta-analysis as it is not a clinical trial of an investigational medicinal product.
Data sources and searches
PubMed, EMBASE and the Cochrane Library were searched from 1983 to 31 December 2022. The complete search strategies are fully described in the online supplemental file 1. The bibliographies of similar and related systematic reviewers also searched along with the reference sections of included studies. Only articles written in English were considered.
Supplemental material
Study selection
Randomised control trials, non-randomised control trials, case–control studies, cohort studies, retrospective studies and cross-sectional studies that included patients with haemoptysis that underwent evaluation with CT and FOB were eligible for this review. Studies that conveyed primary and/or secondary outcome variables in addition to the above were included. Four reviewers (SM, TW, AM and NK) independently screened the retrieved articles for eligibility following deduplication by title and abstract. Any conflicts were resolved by discussion with a third reviewer (GT).
Primary and secondary outcomes
The primary outcome was diagnostic test accuracy for lung cancer by the evaluated diagnostic modality. Secondary outcomes included detection of abnormalities, and potential explanatory aetiologies of haemoptysis such as infection or bronchiectasis.
Data extraction and quality assessment
Three reviewers (SM, TW and AM) independently extracted data from the included studies. Study characteristics collated included authors, year of study, country of study, study design, number and specific diagnostic modalities and presence of randomisation. Sample population characteristics included sample size, mean age (SD), gender composition, smoking history and associated smoking pack-years and described severity of haemoptysis according to the scale used by the respective studies. Primary outcomes collated included total number of confirmed lung cancer cases in each study, number of lung cancers detected as per diagnostic modality and number of false negatives (FN) from each diagnostic arm. Secondary outcome data included number of confirmed alternate aetiologies including infection and bronchiectasis according to diagnostic modality. Any discrepancies between two reviewers were arbitrated by a third reviewer and resolved by consensus. The Grading of Recommendations, Assessment, Development and Evaluation (GRADE) approach was applied to assess individual study methodological quality.10
Risk of bias in individual studies
Risk of bias was assessed by three reviewers (SM, TW and AM) using the Quality Assessment of Diagnostic Accuracy Studies-2 (QUADUS-2) tool for diagnostic test accuracy studies.11 Disparities between investigators were resolved following discussion with a third reviewer until consensus was reached.
Stata analysis
Sensitivity and specificity, diagnostic OR (DOR) and likelihood ratios with 95% CIs were calculated for each primary study from the contingency 2×2 tables of true positive, false positive), FN and true negative (TN) using a bivariate random-effects model estimation. The estimated sensitivity and specificity are presented as forest plots while a hierarchical modelling-based summary receiver-operating characteristic curve was synthesised for both diagnostic modalities (CT and FOB). Publication bias was assessed using Deeks’ funnel plot, with Deeks’ asymmetry test applied to calculate the p value and ascertain statistical significances. The χ2-based Q test was used to test for heterogeneity among studies while the I2 value was applied to assess the degree of interstudy variation. When I2>50%, significant heterogeneity was suspected. The meta-analysis was performed by using the ‘metadta’, ‘midas’ and ‘metandi’ module in the Stata SE V.17 software.21–23 Meta-regression using ‘metadta’ was performed to determine predefined sources of heterogeneity. The values of p<0.05 were considered statistically significant.
Results
Literature search
The search retrieved 2295 results out of which 173 duplicate records were manually removed. Following title and abstract review, 2065 were excluded. Of the remaining 57 full text articles, 43 were excluded for specified reasons illustrated in figure 1. The remaining 14 studies were included.17–20 24–33 Eleven of the included studies were used for meta-analysis.17–20 24 25 27 29–33 The overall quality of the studies using the GRADE approach was low.
PRISMA flow-chart.
Characteristics of the included studies
Detailed patient characteristics are presented in table 1. Included studies were composed of five prospective studies,19 27 30 31 33 five retrospective studies18 20 24–26 28 29 32 and one study17 exhibiting both a prospective and retrospective cohort. The year of included studies ranged from 1990 to 2021. The total patient population across all studies was 2438 with 2239 and 2250 patients receiving a CT and FOB, respectively. From studies where there was a distinction in terms of specific interventions received, 97 patients received a CT only while 108 patients received an FOB only and lastly, 2142 patients received both. The mean age range across studies was from 37.2 to 63 years old while the male gender prevalence ranged from 52.6% to 86.2%. A total of 8 studies conveyed a smoking history prevalence of 24%–93.8%. Eight studies reported the quantification of haemoptysis with some organising into ordinal scales usually in three independent categories reported.17 18 24 25 27 28 32 33
Study and patient group characteristics
Quality assessment and publication bias
Overall, the risk of bias of the studies according to the QUADAS-2 criteria was high (figure 2). The outcomes of the risk of bias assessment for each study are shown in figure 2. Overall, two studies demonstrated low,19 30 four studies demonstrated some concern17 27–29 31 33 and five studies demonstrated high risk of bias18 20 24–26 32 in the methodology of the studies, respectively. Deeks’ funnel plot asymmetry test conveyed a p value of 0.02 for studies evaluating CT diagnostic capabilities, implying evidence of significant publication bias (figure 3). There was no evidence of significant publication bias for studies evaluating and reporting FOB findings (figure 3).
Risk of bias evaluation – QUADUS2.
Deek’s asymmetry (A) profile for CT (B) profile for FOB.
Lung cancer diagnostic efficacy: CT imaging
The sensitivity and specificity of each individual study are shown in figure 4 and table 2. Pooled sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio (NLR) and DOR were 98.0% (95% CI 93.0% to 99.0%), 100% (95% CI 0% to 100%), 1.4×1012 (95% CI 1.4×1012 to 10.4×1012), 0.02 (95% CI 0.01 to 0.07) and 5.6×1013. The area under the curve (AUC) was 1.00 (95% CI 0.99 to 1.00) (figure 5).
Forest plot (A) CT detection of lung cancer, (B) FOB detection of lung cancer.
Aetiologies for haemoptysis not including malignancy
Prevalence and diagnostic sensitivity of lung cancer
Summary receiver operating characteristics (SROC) curve.
Lung cancer diagnostic efficacy: FOB
The sensitivity and specificity of each individual study are shown in figure 4. Pooled sensitivity, specificity, PLR, NLR and DOR were 86.0% (95% CI 63.0% to 95.0%), 100% (95% CI 99.0% to 100%), 1152.7 (95% CI 135.1 to 9837.2), 0.14 (95% CI 0.05 to 0.42) and 8091 (95% CI 816 to 80 187). The AUC was 1.00 (95% CI 0.99 to 1.00) (figure 5).
Comparison of the diagnostic performance of CT and FOB
The application of CT in the context of haemoptysis from suspected lung cancer was associated with a higher sensitivity and identical specificity and AUC. The overall pooled sensitivity and specificity for CT were 98% (95% CI 93.0% to 99.0%) and 100% (95% CI 0% to 100%) in comparison to FOB which was 86% (95% CI 63.0% to 95.0%) and 100% (95% CI 99.0% to 100%), respectively. The AUC for CT was 1.00 (95% CI 0.99 to 1.00) while FOB was also 1.00 (95% CI 0.99 to 1.00).
Comparison of the diagnostic performance of CT and FOB for secondary outcomes
The pooled analysis results for each of the diagnostic modalities are summarised in table 3. With respect to detection of infective aetiologies by CT, pooled sensitivity, specificity, PLR, NLR and DOR were 57.0% (95% CI 0.40% to 0.73%), 100% (95% CI 0.99% to 1.00%), 187 (95% CI 39.8 to 879), 0.43 (95% CI 0.29 to 0.64) and 435 (95% CI 77.0 to 2455). The area under curve (AUC) was 0.99 (95% CI 0.98 to 1.00). Regarding FOB, pooled sensitivity, specificity, PLR, NLR and DOR were 40.0% (95% CI 24.0% to 58.0%), 100% (95% CI 97.0% to 100%), 383 (95% CI 14.9 to 9893), 0.60 (95% CI 0.45 to 0.81) and 639 (95% CI 26.0 to 15 764). The AUC was 0.98 (95% CI 0.96 to 0.99).
With respect to CT detection of bronchiectasis, pooled sensitivity, specificity, PLR, NLR and DOR were 99.0% (95% CI 83.0% to 100%), 100% (95% CI 98.0% to 100%), 269 (95% CI 41.9 to 1724), 0.01 (95% CI 0.00 to 0.20) and 53 418 (95% CI 1171 to 2 435 966). The AUC was 1.00 (95% CI 0.99 to 1.00). Regarding FOB, pooled sensitivity, specificity, PLR, NLR and DOR were 7.00% (95% CI 0.01% to 0.30%), 100% (95% CI 98.0% to 100%), 31.0 (95% CI 2.40 to 402), 0.93 (95% CI 0.82 to 1.05) and 33.0 (95% CI 2.00 to 460). CT (43.5%) demonstrated superior sensitivity in diagnostic efficacy for non-malignant aetiologies compared with FOB (25.0%). With respect to infective aetiologies, CT was again associated with a higher detection of infective changes (55.6%) compared with FOB (32.1%). Lastly, for bronchiectasis, there was a 100% detection rate observed for the CT cohort in contrast to the FOB cohort (13.5%).
Heterogeneity and meta-regression analysis
The I2 did not reveal any evidence of substantial heterogeneity for sensitivity and specificity of CT (I2=0.00 (p=0.53) and I2=0.00 (p=1.00)), however, substantial heterogeneity was observed for FOB (I2=83.5 (p<0.01) and I2=33.1 (p=0.13)). Meta-regression was performed according to the year of the study and recorded median/mean age of the study patient cohort (figure 6).
Subgroup analysis (A) studies before 2010, (B) studies after 2010, (C) studies with population age under 55 years, (D) studies with population age above 55 years.
For studies conducted prior to 2010 (figure 6), CT and FOB demonstrated similar sensitivity (Relative Ratio 0.92 (95% CI 0.83 to 0.97)) and specificity (Relative Ratio 1.00 (95% CI 0.99 to 1.00)). There is more heterogeneity on the logit sensitivity (tau2=0.92) than on the logit specificity (tau2=0.00) and generalised I2 (tau2=0.00). For studies conducted after 2010 (figure 6), CT demonstrated a superior sensitivity (0.99 (95% CI 0.83 to 1.00)) compared with FOB (0.80 (95% CI 0.32 to 0.97)) while both groups demonstrated an identical specificity (Relative Ratio 1.00 (95% CI 1.00 to 1.00)). More heterogeneity was observed on the logit sensitivity (tau2=3.5) compared with the logit specificity (tau2=0.44) and generalised I2 (tau2=0.00). Meta-regression conducted for studies exhibiting a median/mean age below 55 (figure 6), demonstrated a greater sensitivity for CT (1.00 (95% CI 1.00 to 1.00)) compared with FOB (0.55 (95% CI 0.1 to 2.93)) while specificity was identical (Relative Ratio 1.00 (95% CI 0.99 to 1.00)). More heterogeneity was observed on the logit sensitivity (tau2=6.4) in comparison to logit specificity (tau2=0.45) and generalised I2 (Tau.sq.=0.00). For ages above 55 (figure 6), CT demonstrates a similar sensitivity to FOB (Relative Ratio 0.94 (95% CI 0.88 to 0.97)) and specificity (Relative Ratio 1.00 (95% CI 1.00 to 1.00)) with more heterogeneity observed on the logit sensitivity (tau2=1.25) in comparison to logit specificity (tau2=0.00) and generalised I2 (tau2=0.00).
Discussion
Our study compared the diagnostic value of CT and FOB for detecting lung cancer in the context of patients presenting with non-massive haemoptysis. There were a total of 247 patients with confirmed lung cancer that contributed to 10.1% of the pooled study population which is comparable to the incidence rate of lung cancer among all new cancer diagnosis in a year. The findings from this meta-analysis indicated that CT exhibits a higher diagnostic efficacy for investigating lung malignancies. Both diagnostic modalities exhibited an equal AUC despite demonstrating different sensitivity profiles towards detection of lung cancer. While AUC is an established measure of overall performance of a diagnostic test, an equal AUC in comparing two tests does not mean they have the same overall diagnostic performance as they may differ on sensitivity.34 As a result, the diagnostic test should generally be decided with a consideration towards an appropriate balance in sensitivity and specificity values. A total of 5 patients with confirmed lung cancer (2.02%) exhibited a negative CT in conjunction with positive FOB findings for malignancy while in comparison, 34 patients (13.8%) conveyed a negative FOB with positive CT findings. Furthermore, with the seven studies demonstrating a higher FN rate with FOB when compared with CT,17–20 29 30 32 they accounted for 145 (58.5%) and 841 (34.5%) of the study cancer and total population, respectively. In contrast, the two studies demonstrating a higher false rate with CT compared with FOB, contributed to 40 (16.1%) and 436 (17.9%) of the study cancer and total population, relatively.25 26 Of the five patients with positive bronchoscopy findings in the context of a negative CT, three patients were reported from study conducted in 1997 where it was likely that the imaging precision was not as sensitive as it is currently. Of the remaining two patients, one was diagnosed via cytology extracted from bronchial washings from an area of abnormality suggestive of squamous metaplasia which may have not been typically evident on CT. For the last patient, there was a paucity in details. As a result, it is currently difficult to specifically comment on selection factors towards choice of diagnostic modality.
Meta-regression demonstrated a greater diagnostic profile for CT in studies conducted after 2010 in comparison to FOB. This could be explained by the ongoing development and improvement of CT scanner technology across the last few decades. Meta-regression also evaluated the effect of age on the diagnostic efficacy of CT and FOB. While CT demonstrated a similar sensitivity to FOB in ages above 55, sensitivity was superior to FOB in populations under the age of 55. Explanations for this could include location, size and histology of the malignancies for the under 55 population. A population based study conducted by Shi et al conveyed that adenocarcinoma was the most prevalent histological subtype in the younger population.35 This finding could provide some explanation into our study’s age group meta-regression finding as lung adenocarcinoma is associated with peripheral lesion presentation for which FOB has well-documented limitations.36 Furthermore, older patient groups may expectedly present with larger lesions in comparison to younger patient groups37 and thus be better suited for FOB as lesion size is a well-documented predictor of diagnostic efficacy.38 Overall, CT demonstrated a reliable diagnostic profile irrespective of ages groups. Of the studies available, there was generally a paucity of information corresponding to lesion localisation. One study reported this information,27 however, due to substantial methodological heterogeneity, sensitivity meta-analysis could not be pursued for any meaningful outcomes.
CT scans have long played an integral role in haemoptysis evaluation and lung cancer management with transition of application from an adjunctive to definitive role in lung cancer diagnostics. During the 1980s, Colice et al described a moderate accuracy in the prediction of airway abnormalities when using a 10 mm collimation, however, the authors also mentioned inadequacies in definition of mucosal abnormalities as there was limited delineation between localised mucosal changes, endobronchial lesions and extrinsic compression which was a similar consensus to Naidich et al.29 39 An improved performance in accuracy was observed by Mayr et al following the application of a 5 mm collimation.40 Further refinement of the collimation resulted in better delineation of lesion extension and proximity to adjacent structures. Studies have evaluated the concomitant application of CT with FOB and conveyed a complimentary effect with FOB enabling an assessment of mucosal abnormalities and tissue sampling while CT was adept at compensating for FOB’s deficiencies in peripheral lesions.25 30 BTS guidance on lung cancer diagnostic approach recommends consideration of bronchoscopy for evaluation of haemoptysis in the context of a normal CT and high-risk features of lung carcinoma or persistent haemoptysis.12 Furthermore, the Danish Lung Cancer group (DLCG) conveyed recommendations of facilitating both CT and bronchoscopy in patients aged 40 and above, presenting with haemoptysis that exceeds a duration of 1 week in conjunction with a positive smoking history if the initial chest radiography is normal.41
Studies have evaluated the efficacy of this approach by attempting to discern additional lung cancer diagnosis from FOBs performed following CT evaluation. Nielsen et al reported that no additional lung cancers were detected following the implementation of FOB with CT compared with CT alone20 which was consistent with a study conducted by Bønløkke et al.24 It is also worth mentioning that the studies which endorsed combined application were conducted at a much earlier time period while in contrast, recent studies have actually seen an increasing advocation towards exclusively using CT as the diagnostic tool for haemoptysis when there exists an index suspicion of malignancies.17 24 28 The outcomes from the present systematic review and meta-analysis are consistent with the consensus of these recently published studies as there were considerably fewer FN malignancies in the CT group compared with the FOB group.
Bronchoscopy is an essential skill in respiratory medicine with both diagnostic and therapeutic application. From a diagnostic perspective, it permits visualisation and tissue sampling for histological and molecular assessment from central airways using a flexible bronchoscope. The diagnostic yield is, however, variable and significantly influenced by characteristics such as location with Rivera et al reporting a high sensitivity for endobronchial disease in comparison to peripheral lesions.14 42 Other predictors of diagnostic yield also include size of lesion (<2 cm) with multivariate logistic regression demonstrating a statistically significant association. To enhance the diagnostic yield of early cancers and precancerous lesions, bronchoscopic techniques such as autofluorescence bronchoscopy were introduced but with limited application to the proximal bronchial tree predominantly.43 There have been advances in bronchoscopy detection rates for lung cancer over the years.44 This along with other interventional technologies has expanded the sampling procedures that clinicians can use to diagnose and stage lung cancer, however, is not associated with a higher diagnostic yield when prior CT scans were negative. Furthermore, FOB introduces a degree of risk for complications such as bronchospasms, haemorrhage, infection, oropharyngeal trauma and pneumothorax along with risks from sedative agents. It is, however, a generally safe procedure with a very low incidence of mortality and has seen a progressively reducing frequency of adverse events since its initial decade which is likely attributed to improved monitoring, intervention and patient selection.45 46 Nonetheless, as an invasive diagnostic approach, it exposes the patient to a greater degree of risk in comparison to receiving a CT scan and patients are also likely to exhibit a greater tolerability for the latter. Lastly, from a logistical perspective, it is comparatively more straightforward to implement an infrastructure supporting CT scans as opposed to performing FOB.
Early detection of lung cancer has been recognised as an important opportunity in reduction of cancer-related mortality. Thus, the feasibility and efficacy of lung cancer screening has been gaining traction over the last decade with multiple trials conducted for evaluation.5 6 47 Of these, the two largest randomised trials, the US National Lung Screening Trial and the Dutch-Belgian Lung Cancer Screening Trial, have provided conclusive evidence that the intervention reduces lung cancer mortality.
In 2022, the UK NSC recommended targeted screening for lung cancer for people aged 55–74 years.8 The diagnostic accuracy profile of CTs from this meta-analysis provides support to this decision given the consistent sensitivity and specificity. The meta-regression performed also substantiates the reliability of CT irrespective of age groups in comparison to FOB.
The current systematic review and meta-analysis is not without limitations and like any systematic review, analysis is limited by each study’s own methodological discrepancies. First, the number of studies is less than desired and thus restricts deep exploration of potential sources of heterogeneity. Second, methodological quality of the studies included was limited given an absence of characteristics such as randomisation. Next, the QUADUS2 assessment detected a generally high degree of study bias throughout majority of the studies. Lastly, the Deeks’ asymmetry test indicated a high degree of publication bias in the studies included for the CT group. This could potentially suggest that studies with statistically insignificant results or negative outcome were not published. As a result, clinical outcomes derived from this meta-analysis may be tentative and inconclusive, thus caution should be applied for considerable judgement of the results.
Conclusion
This meta-analysis provides evidence to support the potential omission of FOB in the evaluation of patients presenting with haemoptysis in the context of suspected lung cancer, when the CT scan is either normal or reports other benign abnormalities. Future studies evaluating the role of advanced bronchoscopic techniques in diagnosing precancerous mucosal lesions in at-risk populations along with diagnostic modality efficacy based on lesion localisation should be considered.
Data availability statement
Data are available in a public, open access repository. Data are available on reasonable request. PubMed, EMBASE and the Cochrane Library were searched from 1983 to 31 December 2022. The complete search strategies are fully described in online supplemental digital content 1. The bibliographies of similar and related systematic reviewers also searched along with the reference sections of included studies. Only articles written in English were considered.
Ethics statements
Patient consent for publication
Ethics approval
Institutional ethical approval was received (23-GTR/0753).
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
SM, TW and AM are joint first authors.
Twitter @George_Tsaknis
Contributors GT was the study lead and is guarantor for the paper. SM, TW, AM, NK and GT developed the concept and study protocol. SM, TW, AM and NK collected data. SM, TW and AM were involved in data analysis. SM, TW, AM and GT drafted the manuscript. The corresponding author attests that all listed authors meet authorship criteria and that no others meeting the criteria have been omitted.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.