Neoplastic cellularity contributes to the analytic sensitivity of most present technologies for mutation detection, such that they underperform when stroma and inflammatory cells dilute a cancer specimen’s variant fraction. Thus, tumor purity assessment by light microscopy is used to determine sample adequacy before sequencing and to interpret the significance of negative results and mutant allele fraction afterwards. However, pathologist estimates of tumor purity are imprecise and have limited reproducibility. With the advent of massively parallel sequencing, large amounts of molecular data can be analyzed by computational purity algorithms. We retrospectively compared tumor purity of 3 computational algorithms with neoplastic cellularity using hematoxylin and eosin light microscopy to determine which was best for clinical evaluation of molecular profiling. Data were analyzed from 881 cancer patients from a clinical trial cohort, LCCC1108 (UNCseq), whose tumors had targeted massively parallel sequencing. Concordance among algorithms was poor, and the specimens analyzed had high rates of algorithm failure partially due to variable tumor purity. Computational tumor purity estimates did not add value beyond the pathologist’s estimate of neoplastic cellularity microscopy. To improve present methods, we propose a semiquantitative, clinically applicable strategy based on mutant allele fraction and copy number changes present within a given specimen, which when combined with the morphologic tumor purity estimate, guide the interpretation of next-generation sequencing results in cancer patients.
This is an open-access article distributed under the terms of the Creative Commons Attribution-Non Commercial-No Derivatives License 4.0 (CCBY-NC-ND), where it is permissible to download and share the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal. http://creativecommons.org/licenses/by-nc-nd/4.0/