The impact of tumor profiling approaches and genomic data strategies for cancer precision medicine

Abstract

Background: The diversity of clinical tumor profiling approaches (small panels to whole exomes with matched or unmatched germline analysis) may engender uncertainty about their benefits and liabilities, particularly in light of reported germline false positives in tumor-only profiling and use of global mutational and/or neoantigen data. The goal of this study was to determine the impact of genomic analysis strategies on error rates and data interpretation across contexts and ancestries.

Methods: We modeled common tumor profiling modalities-large (n = 300 genes), medium (n = 48 genes), and small (n = 15 genes) panels-using clinical whole exomes (WES) from 157 patients with lung or colon adenocarcinoma. We created a tumor-only analysis algorithm to assess germline false positive rates, the impact of patient ancestry on tumor-only results, and neoantigen detection.

Results: After optimizing a germline filtering strategy, the germline false positive rate with tumor-only large panel sequencing was 14 % (144/1012 variants). For patients whose tumor-only results underwent molecular pathologist review (n = 91), 50/54 (93 %) false positives were correctly interpreted as uncertain variants. Increased germline false positives were observed in tumor-only sequencing of non-European compared with European ancestry patients (p < 0.001; Fisher's exact) when basic germline filtering approaches were used; however, the ExAC database (60,706 germline exomes) mitigated this disparity (p = 0.53). Matched and unmatched large panel mutational load correlated with WES mutational load (r(2) = 0.99 and 0.93, respectively; p < 0.001). Neoantigen load also correlated (r(2) = 0.80; p < 0.001), though WES identified a broader spectrum of neoantigens. Small panels did not predict mutational or neoantigen load.

Conclusions: Large tumor-only targeted panels are sufficient for most somatic variant identification and mutational load prediction if paired with expanded germline analysis strategies and molecular pathologist review. Paired germline sequencing reduced overall false positive mutation calls and WES provided the most neoantigens. Without patient-matched germline data, large germline databases are needed to minimize false positive mutation calling and mitigate ethnic disparities.

Keywords: Disparities; Genomics; Immuno-oncology; Neoantigens; Panel testing; Precision medicine.

Fig. 1

Germline false positives in tumor-only clinical sequencing. Sensitivity and positive predictive value (PPV) curves for multiple germline filtering strategies identifies optimal approaches for unmatched large targeted panel testing (a) and whole-exome sequencing (b). For 91 patients, germline exome data were used to identify false positives post-filtering. Subsequent molecular pathologist review of variants was performed on individual cases to further classify putative germline variants. With molecular pathology review, 50/54 false positive variants were correctly classified as unknown (“tier 4”), with the remaining variants classified as having uncertain (n = 3; “tier 3”) or potential (n = 1; “tier 2”) clinical utility (c, d). Please see “Methods” for detailed descriptions of the four-tier classification schema

Fig. 2

Ancestry and germline false positives using different analysis strategies. a The use of dbSNP as the primary germline filtration strategy results in a significant increase in false positives among non-white patients (p < 0.001). b A similar increase was observed with the use of 1000 Genomes (p < 0.001). c With larger germline databases such as ExAC, this disparity is mitigated

Fig. 3

Mutational load predictions with different panel tests. Comparison of mutational load predictions using WES or either matched (a) or unmatched (b) large panel tests (n = 300 genes) demonstrates both can reliably predict the mutational load. The linear regression line is shown in black with 95 % confidence bands shaded in grey. The identity line (dashed) is shown for comparison. With medium sized panels (n = 48 genes), this ability decreases in both the matched and unmatched setting and is not possible with small (n = 15) gene panels

Fig. 4

Neoantigen predictions in panels. a The proportion of neoantigens called in large panel targeted sequencing data demonstrates an inability to identify as a broad spectrum of neoantigens compared to WES. b Nonetheless, there is a linear relationship between large panel neoantigens recovered from exome and germline-matched large panel data. This linear relationship no longer holds when considering neoantigen data from medium (c) and small (d) targeted panels