[Skip to Content]
[Skip to Content Landing]
Views 267
Citations 0
Comment & Response
October 2018

Regarding the Congruence Between 2 Circulating Tumor DNA Sequencing Assays

Author Affiliations
  • 1Jin-An Clinic, New Taipei City, Taiwan
JAMA Oncol. 2018;4(10):1429. doi:10.1001/jamaoncol.2018.2314

To the Editor Torga and Pienta1 reported low congruence between 2 circulating tumor DNA sequencing assays in 40 patients with metastatic prostate cancer. However, the mutant allele fraction (MAF) of reported genomic alterations, which may help clarify the cause of low congruence, is unmentioned in their report. Other than analytical factors, sampling error may affect the replicability of the circulating tumor DNA testing result when genomic alterations with low MAF are encountered. Five to 30 ng (ie, 1500-9000 genome equivalents) of cell-free DNA was used as input material in the Guardant360 assay.2 For example, for an alteration with an MAF of 0.1% and a cell-free DNA sample of 5 ng (ie, 1500 genome equivalents), the possibility that no single copy of the DNA molecule of the alteration is present in the sample is 22.3% (possibility that the DNA molecule of the alteration is not present in a given genome equivalent = 1 − 0.001 = 0.999; possibility that the DNA molecule of the alteration is not present in all 1500 genome equivalents = 0.9991500 ≈ 0.223). In other words, 22.3% of testing results of the alteration are negative simply due to chance. Therefore, the possibility of congruence between 2 assays for this alteration is as low as 65.4% (possibility of negative results in both assays ≈ 0.223 × 0.223 ≈ 0.050; possibility of positive results in both assays ≈ 0.777 × 0.777 ≈ 0.604; possibility of congruence ≈ 0.050 + 0.604 = 0.654). Furthermore, the congruence may be even lower if analytical factors are considered. Thus, MAF data may provide insights into the root cause of low congruence between these 2 assays.