PURPOSE: Somatic mutations derived from the expansion of clonal populations of blood cells (clonal hematopoiesis of indeterminate potential, or CHIP) may be detected in sequencing of cell-free DNA (cfDNA) samples. We evaluated the potential implications of CHIP in targeted sequencing of plasma samples using matched peripheral blood mononuclear cells (PBMCs) from patients with lung cancer to identify potential CHIP-associated mutations. MATERIALS AND METHODS: A total of 332 plasma and corresponding PBMC samples were collected predose, cycle 1 day 1 (C1D1), from the randomized, phase III study (OAK) comparing atezolizumab versus docetaxel in previously treated patients with non–small-cell lung cancer (NSCLC). The samples were analyzed with the AVENIO ctDNA Surveillance Kit (for research use only; not for use in diagnostic procedures), a 198-kb next-generation sequencing panel targeting cancer-related genes. CHIP variants were assessed by analyzing both plasma and PBMC sequencing data. RESULTS: A range of zero to eight CHIP variants (median = one) was detected per cfDNA sample. Most of these variants were not in the Database of Single Nucleotide Polymorphisms (dbSNP). The number of CHIP variants was positively associated with age, and TP53 was the most frequently mutated gene. Furthermore, the allele frequency was less variable over time for CHIP variants than for tumor-derived variants. CONCLUSION: CHIP-derived mutations are present in late-stage NSCLC. However, not all plasma samples had CHIP mutations detected with targeted panel sequencing. Paired PBMC sequencing analysis may be needed to remove CHIP variants for comprehensive genomic profiling using plasma samples to identify true somatic mutations.

PURPOSE: Identification of predictors for overall survival (OS) allows timely detection of clinical efficacy signals and therefore facilitates treatment decisions. We assessed the association between circulating tumor DNA (ctDNA) metrics and the primary end point of OS in a subset of previously treated patients with locally advanced or metastatic non–small-cell lung cancer, who underwent atezolizumab or docetaxel treatment in the open-label randomized phase III OAK trial. MATERIALS AND METHODS: Plasma from 94 patients at baseline and at subsequent cycles of therapy every 3 weeks was analyzed retrospectively for ctDNA. ctDNA was measured by allele frequency and mutant molecules per milliliter (MMPM). Concordance between various per-sample metrics and clinical outcome were assessed using C index. RESULTS: Of all the ctDNA metrics tested, the association of median MMPM at 6 weeks with OS in patients treated with atezolizumab or docetaxel had a C index > 0.7. The OS hazard ratios relative to high ctDNA above median MMPM within each arm were 0.28 (95% CI, 0.11 to 0.75) for atezolizumab and 0.19 (95% CI, 0.08 to 0.48) for docetaxel. For patients who had ctDNA median MMPM levels of < 4.79, the median survival time was more than 17 months in docetaxel-treated patients and the median survival time was not reached in the atezolizumab-treated patients. CONCLUSION: ctDNA MMPM levels measured at 6 weeks post-treatment are associated with OS in advanced non–small-cell lung cancer. Our results suggest that ctDNA has the potential for a noninvasive early liquid biopsy predictor for OS that warrants further studies to demonstrate its utility in clinical development.

Molecular biomarkers hold promise for personalization of cancer treatment. However, a typical tumor biopsy may be difficult to acquire and may not capture genetic variations within or across tumors. Given these limitations, tumor genotyping using next-generation sequencing of plasma-derived circulating tumor (ct)-DNA has the potential to transform non-small cell lung cancer (NSCLC) management. Importantly, mutations detected in biopsied tissue must also be detected in plasma-derived ctDNA at different disease stages. Using the AVENIO ctDNA Surveillance kit (research use only), mutations in ctDNA from NSCLC subjects were compared with those identified in matched tumor tissue samples, retrospectively. Plasma and tissue samples were collected from 141 treatment-naïve NSCLC subjects (stage I, n = 48; stage II, n = 37; stage III, n = 33; stage IV, n = 23). In plasma samples, the median numbers of variants per subject were 4, 6, 8, and 9 in those with stage I, II, III, and IV disease, respectively. The corresponding values in tissue samples were 5, 5, 6, and 4. The overall tissue-plasma concordance of stage II through IV was 62.2% by AVENIO software call. On multivariate analysis, concordance was positively and significantly associated with tumor size and cancer stage. Next-generation sequencing-based analyses with the AVENIO ctDNA Surveillance kit could be an alternative approach to detecting genetic variations in plasma-derived ctDNA isolated from NSCLC subjects.

Purpose: We assessed plasma circulating tumor DNA (ctDNA) level as a prognostic marker for progression-free survival (PFS) following first-line metastatic colorectal cancer (mCRC) therapy. Experimental Design: STEAM was a randomized, phase II trial investigating efficacy of bevacizumab (BEV) plus 5-fluorouracil/leucovorin/oxaliplatin (FOLFOX) and 5-fluorouracil/leucovorin/irinotecan (FOLFIRI), administered concurrently or sequentially, versus FOLFOX-BEV in first-line mCRC. Evaluation of biomarkers associated with treatment outcomes was an exploratory endpoint. Patients in the biomarker evaluable population (BEP) had 1 tissue sample, 1 pre-induction plasma sample, and 1 post-induction plasma sample collected ≤60 days of induction from last drug date. Results: Among the 280 patients enrolled in STEAM, 183 had sequenced and evaluable tumor tissue, 118 had matched pre-induction plasma, and 54 (BEP) had ctDNA-evaluable sequencing data for pre- and post-induction plasma. The most common somatic variants in tumor tissue and pre-induction plasma were TP53, APC, and KRAS. Patients with lower-than-median versus higher-than-median post-induction mean allele fraction (mAF) levels had longer median PFS (17.7 vs. 7.5 months, hazard ratio = 0.33, 95% confidence interval, 0.17-0.63). Higher levels of post-induction mAF and post-induction mean mutant molecules per milliliter (mMMPM), and changes in ctDNA (stratified by a 10-fold or 100-fold reduction in mAF between pre- and post-induction plasma), were associated with shorter PFS. Post-induction mAF and mMMPM generally correlated with each other (r = 0.987, P < 0.0001). Conclusions: ctDNA quantification in post-induction plasma may serve as a prognostic biomarker for mCRC post-treatment outcomes.