Evidence-Based Health Program e Discipline of Evidence-Based Medicine, Escola Paulista de Medicina (EPM), Universidade Federal de São Paulo (UNIFESP) - São Paulo - Brazil.
We will conduct a systematic review accordantly to the Cochrane Handbook for Systematic Reviews of Interventions recommendations [Higgins 2011]. The protocol for this review was prospectively registered in PROSPERO database (CRD42018092367, available from: at https://www.crd.york.ac.uk/PROSPERO/display_record.php?RecordID = 92367). We developed this manuscript in accordance with the recommendations of PRISMA-P Statement [Moher 2015]
Criteria for including studies
Randomized clinical trials with parallel design, including those of cluster type.
Types of participants
Adults with malignancy-suspected peripheral pulmonary nodules, defined as pulmonary nodules between 8 mm and 30 mm, and characteristics such as spiculation, pleural retraction, notch sign, growing size and complex features. [Snoeckx 2017] It is important to mention that nodules with other characteristics but considered malignancy-suspected by the authors could be included as well.
Types of intervention:
Tomography-guided percutaneous biopsy.
Non-guided transbronchial biopsy.
Fluoroscopy-guided transbronchial biopsy.
Transbronchial biopsy guided by endobronchial ultrasound with radial probe
Transbronchial biopsy guided by electromagnetic navigation.
Studies comparing different sizes of bronchoscopes or studies comparing different nodule localization techniques will not be considered.
Diagnostic yield, measured as the proportion of biopsies in which it was possible to define the histological diagnosis of the pulmonary nodule.
Major adverse events including procedure complications such as pneumothorax (symptomatic and / or requiring drainage), hemothorax (symptomatic and / or requiring drainage), and death.
Need of approach by another technique.
Non-serious adverse events, including pain (frequency and intensity)
Time of procedure.
We will consider any time point reported in the included studies. However, we plan to pool only similar time points: short-term period (up to six months) or long term period (more than six months). Whenever a study reports outcomes at multiple time points within the same period (as listed above), we will use the last measurement.
We will consider studies regardless of whether they report an outcome of interest for this review. In this case, we will contact the authors to confirme if at least one of the outcomes planned in the systematic review was assessed (but not reported until the date).
Searching for studies
We will carry out comprehensive searches in the following electronic databases: CENTRAL (Cochrane Controlled Register of Trials, via Wiley), EMBASE (Excerpta Medica database, via Elsevier), and MEDLINE (Medical Literature Analysis and Retrieval System Online, via PubMed). All databases will be searched from inception to the date of search, with no restriction for language or publication status. The search strategy for MEDLINE is detailed in Table 1. For other electronic databases, we will develop search strategies based on the search strategy for MEDLINE by adjusting in accordance with syntax and controlled vocabulary, both specific for each database.
Table 1. Search strategy for MEDLINE
Medline via Pubmed
#1 "Multiple Pulmonary Nodules"[Mesh] OR (Multiple Pulmonary Nodules) OR (Multiple Pulmonary Nodule) OR (Pulmonary Nodule, Multiple) OR (Pulmonary Nodules, Multiple) OR (Lung Nodules) OR (Lung Nodule)
#2 "Solitary Pulmonary Nodule"[Mesh] OR (Solitary Pulmonary Nodule) OR (Pulmonary Nodule, Solitary) OR (Nodule, Solitary Pulmonary) OR (Solitary Pulmonary Nodules) OR (Nodules, Solitary Pulmonary) OR (Pulmonary Nodules, Solitary) OR (Pulmonary Coin Lesion) OR (Lesion, Pulmonary Coin) OR (Lesions, Pulmonary Coin) OR (Coin Lesions, Pulmonary) OR (Pulmonary Coin Lesions) OR (Coin Lesion, Pulmonary) OR (Pulmonary Nodules) OR (Pulmonary Nodule)
#3 (Pulmonary lesion*) OR (Lung lesion*)
#4 #1 OR #2 OR #3
#5 "Biopsy"[Mesh] OR Biopsy OR Biopsies
#6 #3 AND #4
#7 ((clinical[Title/Abstract] AND trial[Title/Abstract]) OR clinical trials as topic[MeSH Terms] OR clinical trial[Publication Type] OR random*[Title/Abstract] OR random allocation[MeSH Terms] OR therapeutic use[MeSH Subheading])
#8 #6 AND #7
We will scrutinize the reference lists of the included studies and of narrative reviews for further studies. We will contact experts in the area requesting information about ongoing or unpublished trials.
At first, we will select titles and abstracts of the retrieved records. At a second moment, we will assess the full‐texts of eligible studies and then classify them as included or excluded. The entire selection process will be conducted independently by two authors. Disagreements between them will be solved by a third author. The selection process will be performed through rayyan platform (http://rayyan.qcri.org) [Ouzzanni 2016]. The full process of selection will be detailed in a PRISMA flow.
Data extraction and management
Data will be extracted and placed together in a standard form. Data extraction will be performed independently by two authors. Disagreements will be solved by a third author. The following data will be considered: study design, follow up, number of study centers, setting, date of study (start-end), sample size, mean age and sex of participants, nodule characteristics, interventions, comparison, primary and secondary outcomes reported, time-points measures, funding sources and conflicts of interest.
Assessment of the risk of bias of included studies
We will evaluate the risk of bias by the use of Cochrane Risk of Bias (RoB) table [Higgins 2011]. Two authors will independently apply the seven domains of the RoB table: sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcomes assessors, incomplete outcome data, selective reporting of outcomes, and other potential sources of bias. For blinding of participants and personnel, blinding of outcomes assessors and incomplete outcome data, we will perform the assessment at the outcome level. We will present the reasons for each judgement.
Firstly we will assess the heterogeneity based on clinical and methodological diversity of the included studies. Clinical diversity will be assessed by comparing the PICO elements of included studies. Inconsistency (statistical heterogeneity) will be evaluated by Chi² test (p<0.1 as indicative of statistical heterogeneity) and its extension by I² test (I²>50 as indicative of significant inconsistency) [Higgins 2011]. Whenever possible, we will investigate the reasons for heterogeneity by additional analysis.
Measuring the treatment effect and data synthesis
When two or more trials assessing the same outcome exist and were homogeneous, we will perform quantitative synthesis. We will estimate treatment effects with a 95% confidence interval, calculating risk ratios (RR) for dichotomous outcomes and mean difference (MD) for continuous outcomes. We will summarize the data using random effects model meta-analysis in Review Manager 5.3 software [Higgins 2011]. When quantitative synthesis is not possible due to clinical and/or methodological diversity, we will present results by a narrative approach.
We plan to perform subgroup analysis for all primary outcomes considering the anatomic region of nodules: central or peripheral, since we expect different diagnostic yelds for each technique; bronchoscopic methods tend to have better results for central lesions and transthoracic approach tend to show better yelds in peripheral lesions. We also plan to carry out sensitivity analyses by excluding studies with unclear ou high risk of bias for at least one domain of RoB Table.
We plan to investigate publication bias by the use of funnel plots wether 10 or more studies are pooled into a meta-analysis [Higgins 2011].
Certainty of the body of the evidence
To evaluate the certainty of the body of the evidence, we will use the GRADE approach (Grading of Recommendations, Assessment, Development and Evaluations), which comprises five criteria: risk of bias, inconsistency, imprecision, indirectness and publication bias [Guyatt 2011]. We will develop a Summary of Findings (SoF) table in GRADEpro GDT considering all outcomes for each pair-wise comparison. We will report reasons to downgrade the evidence for each outcome.
Our group recognize the importance of documenting protocol amendments. Any relevant change in the protocol will be registered and described in the final review article.