Positron emission tomography-computed tomography (PET-CT) in suspected malignant pleural effusion. An updated systematic review and meta-analysis

The role of PET and integrated PET-CT in the diagnostic workup of suspected malignant pleural effusions is unknown. Earlier systematic reviews (published 2014 and 2015) both included pleural pathology without effusion, and reached contradictory conclusions. Five studies have been published since the latest review. This systematic review and meta-analysis aims to summarise the evidence of PET and integrated PET-CT in predicting pleural malignancy in patients suspected of having malignant pleural effusions. A meta-analysis based on a systematic literature search in Cochrane Library, Medline, EMBASE and Clinicaltrials.gov was performed. Diagnostic studies evaluating the performance of PET or PET-CT in patients with suspected malignant pleural effusion, using pleural fluid cytology or histopathology as the reference test, and presenting sufficient data for constructing a 2x2 table were included. The quality of the studies was assessed by the Quality Assessment of Diagnostic Accuracy Studies-2 score. Subgroup analyses on image modality, interpretation method and known malignancy status pre index-test application were planned. Seven studies with low risk of bias were included. The pooled ability to separate benign from malignant effusions varied with image modality, interpretation method and known malignancy status pre index-test application. In studies using PET-CT, visual/qualitative image analysis was superior to semi-quantitative with positive (LR + ) and negative likelihood ratio (LR-) of 9.9 (4.5-15.3) respectively 0.1 (0.1-0.2). There was considerable heterogeneity among studies. In conclusion, visual/qualitative image analysis of integrated PET-CT seems to add relevant information in the work-up of suspected malignant pleural effusions with LR + and LR- close to rigorous pre-set cut-offs of > 10 and < 0.1. However, the quality of evidence was low due to inter-study heterogeneity, and inability to assess meta-bias. Clinical Trial Registration: The protocol was uploaded to the PROSPERO database (CRD42020213319) on the 13th of October 2020.