Comprehensive profiling and quantitation of oncogenic mutations in non small-cell lung carcinoma using single molecule amplification and re-sequencing technology

Abstract

Activating and resistance mutations in the tyrosine kinase domain of several oncogenes are frequently associated with non-small cell lung carcinoma (NSCLC). In this study we assessed the frequency, type and abundance of EGFR, KRAS, BRAF, TP53 and ALK mutations in tumour specimens from 184 patients with early and late stage disease using single molecule amplification and re-sequencing technology (SMART). Based on modelling of EGFR mutations, the detection sensitivity of the SMART assay was at least 0.1%. Benchmarking EGFR mutation detection against the gold standard ARMS-PCR assay, SMART assay had a sensitivity and specificity of 98.7% and 99.0%. Amongst the 184 samples, EGFR mutations were the most prevalent (59.9%), followed by KRAS (16.9%), TP53 (12.7%), EML4-ALK fusions (6.3%) and BRAF (4.2%) mutations. The abundance and types of mutations in tumour specimens were extremely heterogeneous, involving either monoclonal (51.6%) or polyclonal (12.6%) mutation events. At the clinical level, although the spectrum of tumour mutation(s) was unique to each patient, the overall patterns in early or advanced stage disease were relatively similar. Based on these findings, we propose that personalized profiling and quantitation of clinically significant oncogenic mutations will allow better classification of patients according to tumour characteristics and provide clinicians with important ancillary information for treatment decision-making.

Keywords: allele-specific amplification refractory mutation system; non small-cell lung carcinoma; oncogenic mutations; single molecule amplification and re-sequencing technology.

Figure 1. Study design

Figure 2. Sensitivity of SMART assay for detection of common EGFR variants

A single nucleotide substitution L858R, an exon 19 deletion (746-750) and the exon 20 insertion (V774insH) were modelled at levels of 0.01%, 0.05%, 0.1%, 0.5%, 1%, 5%, 10% and 50%. By Pearson correlation, there was a linear relationship between actual (Y-axis) and theoretical log₁₀ values (X-axis).

Figure 3. Levels of oncogenic mutations in matching tumour and non-tumour specimens

A total of 25 patients were studied with either activating EGFR or ALK mutations identified in the original tumour specimens. All non-tumour tissues were negative but low levels of new mutations were identified in four cases. Closed symbols represent the level of activating mutation in the tumour tissue whereas open symbols represent the level of activating or new mutations in the non-tumour tissue.

Figure 4. Spectrum of oncogenic mutations

Pie charts show the frequency of mutations detected and the six different patterns of mutations (groups A-F) in the 184 NSCLC tumour specimens.

Figure 5. Nature and abundance of oncogenic mutations according to disease stage

Gene mutations are coloured coded (legend).

Figure 6. Molecular analysis of EML4-ALK fusion variants

Panel 1. Sanger sequencing confirmed the fusion site between EML4 and ALK predicted from the paired end sequence reads generated by SMART assay. Panel 2. Diagrammatical representation of each fusion variant, showing the fusion site with respect to hg19 genome reference co-ordinates, EML4 and ALK exon (E) fusion positions and their tissue abundance. Panel 3. Level of EML4-ALK fusion protein detected by IHC in FFPE tissue, indicated by brown staining. Panel 4. In situ EML4-ALK DNA fusions are indicated by co-localization of orange and green FISH signals.