Portal de Periódicos da CAPES

Comment on ‘Sensitive and specific multi-cancer detection and localization using methylation signatures in cell-free DNA’ by M. C. Liu et al.

2020; Elsevier BV; Volume: 31; Issue: 9 Linguagem: Inglês

10.1016/j.annonc.2020.04.013

1569-8041

William C. Taylor,

Genetic factors in colorectal cancer

Liu et al.1Liu M.C. Oxnard G.R. Klein E.A. Swanton C. Seiden MV, on behalf of the CCGA ConsortiumSensitive and specific multi-cancer detection and localization using methylation signatures in cell-free DNA.Ann Oncol. 2020; 31: 745-759Abstract Full Text Full Text PDF PubMed Scopus (145) Google Scholar reported the results of a study of targeted methylation analysis of circulating cell-free DNA as part of a project to develop a screening test for multiple cancers, supported by a company with the laudable mission 'to detect cancer early, when it can be cured.'2GRAIL, Inc.https://grail.com/Date accessed: April 8, 2020Google Scholar Unfortunately, methodological problems and misleading characterization of results detract from the usefulness of the authors' report. Although the authors correctly emphasized the risks and harms of false-positive results from screening, they characterized their study's false-positive results in a misleading manner. When asymptomatic patients receive positive results from a cancer screening test, they and the clinicians caring for them need to know the chance that their test results are wrong. Because the authors used false-positive rate to mean the proportion of false-positive results among persons thought to be cancer-free, they described their results as showing a false-positive rate of 0.7%. As shown in Table 1, constructed according to the authors' assumptions (sensitivity 55%, specificity 99.3%, pre-test probability 1.3%), the fraction of positive tests shown to be falsely positive is 49% (691/1406), not the low rate of 0.7% noted by the authors. This high rate at which test results are falsely positive would be expected to cause considerable 'subsequent morbidity and psychological, physical and financial costs' that the authors rightly stated should be minimized in screening.Table 1A 2 × 2 table constructed according to the assumptions of Liu et al.1Liu M.C. Oxnard G.R. Klein E.A. Swanton C. Seiden MV, on behalf of the CCGA ConsortiumSensitive and specific multi-cancer detection and localization using methylation signatures in cell-free DNA.Ann Oncol. 2020; 31: 745-759Abstract Full Text Full Text PDF PubMed Scopus (145) Google ScholarCancerNo cancerTotalTest positive7156911406Test negative58598 00998 594Total130098 700100 000 Open table in a new tab In addition, the authors inappropriately selected the annual incidence of all invasive cancers to estimate pre-test probability in their calculation of predictive value. The correct input should have been the prevalence of cancer among asymptomatic persons. If the appropriate pre-test probability based on prevalence is lower than the incorrectly chosen annual incidence, the percentage of positive test results that are false positives would be correspondingly higher than 49%. Finally, the authors miscalculated sensitivity as the rate of test positivity among persons known to have cancer. For a screening test, the relevant group consists of asymptomatic persons, not those with cancer. As shown in their Figure 4, the authors found test positivity less often at progressively earlier stages of cancer. Earlier testing among these same persons when they were asymptomatic, presumably with at least some of these cancers at earlier stages, might therefore have resulted in fewer positive test results—and thus lower sensitivity, contributing additionally to the proportion of positive test results that would be false positives. Although the benefits of screening seem obvious, multiple commentators over the past half century have identified many challenges in providing screening that does more good than harm.3Wilson J.M.G. Jungner G. Principles and Practice of Screening for Disease. World Health Organization, Geneva1968Google Scholar, 4Taylor W.C. Delbanco T.L. Looking for early cancer.Ann Intern Med. 1980; 93: 773-775Crossref PubMed Scopus (7) Google Scholar, 5Sackett D.L. Screening in family practice: prevention, levels of evidence and the pitfalls of common sense.J Fam Pract. 1987; 24: 233-234PubMed Google Scholar Along with others looking for opportunities to improve human health presented by recent advances in molecular biology, genetics and computational medicine, those reading the report from Liu et al.1Liu M.C. Oxnard G.R. Klein E.A. Swanton C. Seiden MV, on behalf of the CCGA ConsortiumSensitive and specific multi-cancer detection and localization using methylation signatures in cell-free DNA.Ann Oncol. 2020; 31: 745-759Abstract Full Text Full Text PDF PubMed Scopus (145) Google Scholar would be well served to recall what was written about screening for the World Health Organization in 1968:"In theory…, screening is an admirable method of combating disease, since it should help detect it in its early stages and enable it to be treated adequately before it obtains a firm hold on the community. In practice, there are snags."3Wilson J.M.G. Jungner G. Principles and Practice of Screening for Disease. World Health Organization, Geneva1968Google Scholar There are no funders to report for this submission.