To the Editor The systematic review by Hanna et al1 provides a comprehensive study of volatile organic compounds (VOCs). We have several concerns about the study’s methodology.
We noted the high heterogeneity among the studies analyzed (I2, ≥90%), but no further subgroup or sensitivity analysis was performed. After reviewing the Supplement and full text of the included studies, we believe that the meta-analysis might have been inappropriately performed. Judgements about the validity of pooled data should be made by considering the similarity of pooled tests and whether the results may be misleading.2 In this meta-analysis, we found that the VOCs tested as diagnostic biomarkers differed considerably among the included studies. The inclusion criteria of the meta-analysis did not limit the type of VOCs tested, the diagnostic standard was not defined, and the data of different diagnostic tests were inappropriately synthesized. For example, data on several cancer antigens for diagnosis of different types of cancer in a meta-analysis cannot be pooled, instead data on cancer antigen 125 for breast cancer diagnosis should be pooled.3 Similarly, only one type of VOC should be identified for diagnosis of a specific cancer in a meta-analysis. Moreover, receiver operating characteristic (ROC) curves generally can help determine the optimal cutoff value in common situations.4 However, owing to the inappropriate data pooling, the ROC curve could not be used to identify the optimal cutoff value for the diagnostic biomarkers. Therefore, the results might have little clinical significance and could not inform the clinical application of the test. The VOC test is far from a mature diagnostic test; its efficacy remains unclear. We hope that the authors will reconsider the data pooling methods to avoid misleading results.
Cao Y, Li C, Li L. Clinical Application of Volatile Organic Compound–Based Exhaled Breath Tests for Cancer Diagnosis. JAMA Oncol. Published online May 23, 20195(7):1068–1069. doi:10.1001/jamaoncol.2019.1143
Browse and subscribe to JAMA Network podcasts!
Customize your JAMA Network experience by selecting one or more topics from the list below.
Create a personal account or sign in to: