News

Monitoring lung cancer evolution with circulating tumour cell DNA

8 May 2017
Monitoring lung cancer evolution with circulating tumour cell DNA

by ecancer reporter Clare Sansom

Cancers of the lung are almost invariably fatal unless they are detected early; lung cancer is the leading cause of cancer death worldwide and survival times have improved very little in forty years.

About 85% of all lung cancers are classified as non-small cell lung carcinoma (NSCLC), which is invariably fatal once it has spread.
However, a proportion of patients gain a modest but significant benefit if treated with adjuvant chemotherapy following surgery.

Early identification of patients who are likely to relapse and whose tumours will be tractable to chemotherapy will be useful for maximising the benefit of this therapy while limiting unnecessary treatment.

The presence of DNA from cancer cells in blood plasma, known as circulating tumour DNA (ctDNA), is known to correlate with cancer progression after surgery in patients with breast and colorectal cancer, and it can be detected before any symptoms of relapse occur.

Furthermore, genetic profiling of this DNA – which carries the same mutations as the tumour region from which it derives – can be useful for selecting therapies and predicting prognosis.

Researchers led by Christopher Abbosh of University College London Cancer Institute, London, UK have now carried out a detailed genetic profile of ctDNA following surgery in patients diagnosed with NSCLC.

This study was carried out in the context of the ambitious TRACERx (TRAcking non-small cell lung Cancer Evolution through therapy [Rx]) study, which is led by Charles Swanton from the Francis Crick Institute in London and aims to follow the evolution of a large NSCLC tumour cohort from diagnosis to death.

This study involves next-generation sequencing of the exomes (protein coding regions of the genomes) from multiple regions of each patient’s tumour to determine clonal and subclonal mutations.

Abbosh and his co-workers obtained pre-operative plasma samples from 100 patients recruited into this study and then extracted, amplified and sequenced cell-free DNA in these samples.

A sample was classed as circulating tumour DNA positive if two or more single nucleotide variants (SNVs) were detected in cell-free DNA from the pre-operative plasma.

It was not possible to analyse the plasma obtained from four of the patients; 46 of the samples from the remaining 96 patients were classed as ctDNA positive, and a further 12 contained DNA with a single SNV.

The proportion of patient samples that contained ctDNA was strongly dependent on tumour subtype: a total of 30/31 (97%) of samples from patients diagnosed with lung squamous cell carcinoma (LUSC) were ctDNA positive, compared to only 11/58 (19%) from patients diagnosed with lung adenocarcinoma (LUAD).

Release of ctDNA into plasma has been associated with necrosis, and as expected the squamous cell carcinomas were significantly more necrotic than the adenocarcinomas.

From multivariate analysis, other strong predictors of ctDNA release included invasion into the lympho-vascular system and a high concentration of the proliferation marker antigen Ki-67.

Clonal SNVs – that is, nucleotide variants found throughout the primary tumour – were present in all 46 of the ctDNA positive patients, but subclonal SNVs were present in only 27 of the cases.

The variant allele fraction (VAF) in the plasma of the ctDNA positive patients correlated with the size of the tumour.

The researchers then profiled ctDNA from plasma obtained from a subset of 24 of the patients following surgery and the patients followed up in order to detect signals of drug resistance and relapse.

At least two SNVs were detected in plasma samples from 13 of 14 patients with clinically confirmed relapse either before or at the time of the relapse.

The median time between ctDNA detection and disease progression that was detectable on a CT scan ranged from 10 to 346 days, with a mean of 70 days; in four cases the relapse appeared to be driven by a single tumour subclone.

Data from some patients suggested that it would be possible to determine which tumours would respond to chemotherapy through sequencing clonal and subclonal ctDNA.

Tumour DNA from one patient who died during the analysis was sequenced through the PEACE (Posthumous Evaluation of Advanced Cancer Environment) post-mortem study, showing that all that patient’s metastases evolved from one subclone within the primary tumour.

In conclusion, Abbosh and his colleagues show that ctDNA sequencing can map the progress of tumour evolution in real time and suggest that these results can be used to select chemotherapy options and even drive drug development.

It is likely that these findings will be extensible beyond NSCLC.

Reference
Abbosh, C., Birkbak, N.J., Wilson, G.A., the TRACERx consortium, the PEACE consortium and many others (2017). Phylogenetic ctDNA analysis depicts early stage lung cancer evolution. Nature, published online ahead of print 26 April 2017.  doi:10.1038/nature22364