E-Poster Presentation 33rd Lorne Cancer Conference 2021

Diagnostic performance of blood based liquid biopsies for screening and early detection of cancers – a narrative review with a systematic approach (#203)

William WW Wang 1 , Stephen SW Wong 2 , Maarten MI IJzerman 1 2
  1. Cancer Health Services Research, University of Melbourne Centre for Cancer Research, Melbourne School of Population and Global Health, Parkville, Parkville, victoria , Australia
  2. Peter MacCallum Cancer Centre , melbourne , victoria

Background: Several targeted and pan-cancer blood-based liquid biopsy tests are being developed to facilitate the early detection of cancer, with the aim to improve outcomes. Here we review the clinical diagnostic performance novel, minimally invasive and relative low-cost blood tests - for early cancer detection.

Methods: A systematic search and review of studies available in the literature was performed according to the PRISMA criteria to assess the analytical and clinical validity of current liquid biopsy assays in the setting of early cancer diagnosis or screening. Data regarding test sensitivity, specificity and predictive values were extracted.

Results: Twenty-two publications were retrieved from a selection of 180 papers. Publications with data related to the analytical and clinical validity of liquid biopsy assays indicated moderate to good sensitivity range (35% - 100%, mean of 79.2%) and specificity range (60.7 % - 100%, mean of 88.7%) to diagnose cancer from healthy controls. Two distinct classes of tests were observed, (1) liquid biopsies for targeted screening of single cancer types using circulating tumour DNA (ctDNA) or miRNA panels and (2) pan-cancer tests such as those that analyse DNA methylation or multi-analyte approach achieved high specificity but lower sensitivity. The positive predictive values are reported in only two studies at 17% and 5.9% highlights high false negative rates in current liquid biopsy testing in population screening.

Conclusion: Current early detection liquid biopsy assays are capable of identifying between cancer and non-cancer patients. But current diagnostic performance data does not support the clinical utility for any assays, in a real-world population with low cancer prevalence. Further improvement in the performance of assays is required, together with larger prospective trials in real-world patient populations, as well as a further analysis of how these test could be integrated into the current clinical practice.