Table 3. Studies of the mutation status in cfDNA samples from NSCLC patients.
| Study | Patients included | Stage | Technology | Sample | Sensitivity (%) | Specificity (%) | Conclusions | Refs. |
|---|---|---|---|---|---|---|---|---|
| Rosell et al. [2009] | 164 | I–IV | ME-PCR | Serum | 59 | NA | Ninety-seven of 164 patients carried mutations (del19 and L858R) in 64 and 33, respectively | (119) |
| Taniguchi et al. [2011] | 44 | III–IV | BEAMing | Plasma | 73 | NA | From 44 patients that were confirmed to have activating mutations in their primary tumor, 32 mutations were detected in the plasma DNA | (125) |
| Goto et al. [2012] | 194 | III–IV | ARMS | Serum | 43 | 100 | Fewer patients were EGFR mutation positive when assessed using pretreatment cfDNA (23.7%) vs. tumor tissue-derived DNA (61.5%) | (132) |
| Huang et al. [2012] | 822 | III–IV | DHPLC | Plasma | 70 | 80 | Concordance of EGFR mutations between tissue and plasma samples was 77%. In patients with two or more lines of EGFR-TKI therapy, EGFR mutation status in plasma samples was a predictor for PFS | (133) |
| Zhao et al. [2013] | 111 | I–IV | ME-PCR | Plasma | 36 | 95.5 | The sensitivity was 10% in early-stage patients and 56% in advanced-stage patients. Patients with poorly differentiated tumors showed the highest sensitivity (77.8%) | (134) |
| Kim et al. [2013] | 60 | III–IV | PNA/LNA PCR | Plasma | 17 | 100 | There was no statistically significant difference in clinical parameters between patients with EGFR mutations in plasma and those without EGFR mutations | (121) |
| Kim et al. [2013] | 57 | III–IV | PNA/LNA PCR | Serum | 73 | 91 | The status of EGFR and KRAS mutation in serum was not prognostic in patients with advanced NSCLC | (135) |
| Wang et al. [2014] | 134 | III–IV | ARMS | Plasma | 22 | 97 | Patients with high TGF-β1 levels showed shorter OS and worse response to EGFR-TKIs than patients with low TGF-β1 levels | (136) |
| Couraud et al. [2014] | 107 | I–IV | Ion Torrent PGM | Plasma | 58 | 87 | In tumor DNA, 50 mutations (36 EGFR, 5 ERBB2, 4 KRAS, 3 BRAF, and 2 PIK3CA) were identified, of which 26 were detected in cfDNA | (127) |
| Weber et al. [2014] | 199 | II–IV | Cobas® EGFR blood test | Plasma | 61 | 96 | Patients with activating EGFR mutations in plasma DNA had a longer PFS than patients with wild-type status | (137) |
| Karachaliou et al. [2015] | 97 | III–IV | PNA/LNA PCR | Plasma | 78 | NA | Median OS was shorter in patients with the L858R mutation in cfDNA than in those with the exon 19 deletion | (138) |
| Thress et al. [2015] | 38 | IV | Cobas®/Therascreen®/DDPCR/BEAMing | Plasma | 78–100 | 93-100 | For the T790M mutation, the sensitivity and specificity were 73% and 67%, and 81% and 58% with the Cobas®Mutation Test and the BEAMing dPCR, respectively | (139) |
| Wei et al. [2016] | 50 | IV | DDPCR | Plasma | 76 | 88 | Concurrent mutant T790M DNA detection of lung cancer patients at baseline achieved 82% concordance with matched tissue analysis | (140) |
| Newman et al. [2016] | 142 | III–IV | CAPP-Seq | Plasma | 92 | 100 | In silico elimination of highly stereotypical background artifacts with a molecular barcoding strategy improves the sensitivity of cancer profiling by CAPP-Seq | (131) |
| Zheng et al. [2016] | 25 | III–IV | DDPCR | Plasma | 81 | 100 | In patients receiving 2nd line or later TKI treatment, the T790M ctDNA positive group had a significantly shorter OS than the negative group | (141) |
cfDNA, circulating free DNA; ctDNA, circulating tumor DNA; TKI, tyrosine-kinase inhibitor; PFS, progression-free survival; OS, overall survival; ME-PCR, mutant-enriched PCR; BEAMing, beads, emulsification, amplification, and magnetics; DHPLC, denaturing high performance liquid chromatography; ARMS, amplification-refractory mutation system; PNA/LNA PCR, peptide nucleic acid/locked nucleic acid PCR; DDPCR, droplet digital PCR; CAPP-Seq, cancer personalized profiling by deep sequencing.