NEW YORK (GenomeWeb) – Two new studies resulting from the ongoing Tracking Cancer Evolution through Therapy (TRACERx) trial in the UK illustrate important features of the heterogeneity and clonal evolution of non-small cell lung cancer and begin to build evidence for the use of liquid biopsy for recurrence monitoring and predicting response to adjuvant treatment.
Led by Charles Swanton, a professor at the Francis Crick Institute, TRACERx aims to systematically characterize the genomic paths lung cancers take in growing, spreading, and becoming resistant to therapy in about 900 NSCLC patients.
In one of the new reports, published in the New England Journal of Medicine today, Swanton and colleagues conducted genomic analyses of tumor tissue from 100 patients, comparing the genetic features of different regions of these tumors to elucidate their clonal evolution and heterogeneity.
The authors wrote that chromosomal instability appeared to not only reflect tumor evolution but also predicted worse outcomes for patients in the study. Those with a high proportion of unstable chromosomes were more than four times more likely than those without to have their cancer return, or die from their disease, within two years.
According to Swanton and colleagues, the study provides a newly comprehensive analysis of NSCLC evolution, and demonstrates that intratumor heterogeneity and branched evolution are almost universal, at least across the TRACERx study cohort.
This heterogeneity is characterized by early clonal genome doubling followed by extensive subclonal diversification. "Certain targetable drivers, including those in EGFR, MET, and BRAF, were almost exclusively clonal and early, explaining the robust and uniform responses often seen across multiple sites of disease when targeting these alterations. However, over 75 percent of tumors harbored a subclonal driver alteration, in genes such as PIK3CA, NF1, KRAS, TP53, and NOTCH family members," the authors wrote.
Additionally, the team found that 62 clonal driver mutations detected in the study appeared clonal in a single tumor region, but were absent or subclonal in other regions, confirming the limitations of single tumor sampling strategies.
The multi-region tissue sequencing approach also appeared to offer promising predictions of disease relapse and poor outcomes, by allowing assessment of dynamic chromosomal instability and revealing the differences in chromosomal karyotypes between NSCLC subclones
"Elevated [somatic copy number alteration] heterogeneity was associated with shorter [survival], suggesting patients with low stage tumors harboring high levels of SCNA heterogeneity may represent a high-risk group who might benefit from close monitoring and early therapeutic intervention during follow up," the authors wrote.
In the other analysis, published in Nature, Swanton and colleagues analyzed blood samples from the same 100 patients, using a liquid biopsy test developed by Natera.
The team used their multi-region exome sequencing results, obtained from patients' tumor tissue samples, to design bespoke multiplex-PCR assays for each individual patient, and then used those to search for mutations in circulating cell-free DNA.
Samples from four patients failed, but in the remaining 96, about half had at least two single nucleotide variants in their pre-operative blood samples. A single SNV was detected in another 12 cases.
On average, 94 percent of the clonal alterations targeted by the bespoke panels showed up in patients' blood. Subclonal SNVs appeared in 68 percent of patients whose panels included them.
The investigators were also interested in why some patients had detectable ctDNA while others did not. According to the authors, predictors of ctDNA detection included non-adenocarcinoma histology, necrosis, increased proliferative indices, and lymphovascular invasion. Tumor volume also appeared to play a role in whether a patient had detectable ctDNA alterations, and at what allele frequency.
Expanding upon earlier results shared at last year's annual meeting of the American Association for Cancer Research, the team also confirmed that ctDNA can predict treatment outcomes, such as resistance to post-operative chemotherapy, and can track the evolution of different cancer clones and subclones during relapse.
Blinded to relapse status, investigators tested 24 patients before and after receiving surgery and found that they could identify most of those who later had their disease recur — in some cases up to a year before clinical imaging confirmed the relapse.
On average, there was a 70-day lead time for ctDNA evidence of relapse before imaging-based diagnosis. Among 13 relapse cases, four showed lead times of more than six months. In two cases, ctDNA provided advance notice more than 150 days before imaging.
The team also compared ctDNA levels immediately before and after adjuvant chemotherapy in a subset of patients after surgery.
According to the authors, adjuvant platinum-based chemotherapy improves cure rates in only 5 percent of NSCLC patients, while 20 percent experience acute toxicities. Therefore, the field greatly needs ways to better target the use of these drugs.
Several TRACERx patients had detectable ctDNA in their plasma that increased despite adjuvant chemotherapy, with disease then recurring within one year.
In contrast, one patient showed a complete clearance of ctDNA mutations 51 days following completion of adjuvant chemo — progressing from 20 ctDNA mutations detected after surgery to 13 before beginning adjuvant chemo, to zero, and remaining relapse free at the last timepoint investigators analyzed over 650 days later.
The authors concluded that although the data suggest that the sensitivity of the approach may be constrained by tumor size, and is expensive — about $1,750 per patient for the analysis of five samples — the results suggest that liquid biopsy could aid, and potentially simplify, clinical evaluation of cancer progression and direction of appropriate treatment.
In the case of adjuvant chemo, results of the study also suggest that ctDNA analysis could make feasible new avenues of drug development in NSCLC, using ctDNA to identify residual disease, define adjuvant treatment response, and target emerging subclones prior to clinical recurrence, the authors wrote.
"We are pleased with the groundbreaking results of this collaboration with the TRACERx consortium," Natera CEO and founder Matthew Rabinowitz said in a statement. "We are working towards bringing this technology to market later this year, initially to serve researchers in academia and the pharmaceutical industry, and ultimately to save lives in the clinic by enabling earlier diagnosis and personalized treatment of disease."
This application of liquid biopsy technologies to recurrence monitoring and prediction is gaining attention in other cancer types as well. For example, encouraging data were published last year by collaborators at Johns Hopkins and the Institute of Medical Research in Melbourne, Australia, showing that ctDNA can be used to identify colorectal cancer patients who are more likely to recur after surgical tumor resection, and can serve as a marker of response or resistance to adjuvant therapy much earlier than current clinical standards.