Liquid Biopsy-Directed Treatment Improves Advanced NSCLC Overall Survival

Author: By Lynda Williams, Senior medwireNews Reporter 

medwireNews: A study of advanced non-small-cell lung cancer (NSCLC) patients has demonstrated a significant overall survival (OS) benefit for patients who are matched to a targeted therapy by circulating tumour (ct)DNA or tissue sampling. 

Presenting the research at the 2021 ASCO Annual Meeting, Justin Jee, from Memorial Sloan Kettering Cancer Center in New York, USA, explained that while it is faster to perform next-generation sequencing for potential treatment targets on plasma than tumour biopsy samples, not every patient has ctDNA. 

To explore the relationship between ctDNA, prognosis and survival further, the researchers analysed plasma and tissue samples from 1002 patients with recurrent or metastatic NSCLC treated at the Memorial Sloan Kettering Cancer Center. 

As expected, plasma ctDNA analysis took significantly less time than tumour sample analysis (median 10 vs 31 days), which Justin Jee attributed to delays in scheduling biopsies and processing samples. 

In all, 348 patients were matched with targeted therapy, 220 of whom were matched using ctDNA and 128 by tissue analysis. 

After correcting for immortality bias, OS from time of metastatic diagnosis to death or last follow-up was significantly better in patients who were matched to a targeted therapy by ctDNA or by tissue sample alone than those who were not matched to a targeted therapy, with hazard ratios (HRs) for death of 0.63 and 0.46, respectively. 

Justin Jee noted that the target gene proportions were comparable in the ctDNA- and tissue-only patient groups and that the OS benefit with targeted therapy was “driven largely by patients with EGFR and ALK mutations who were treatment naïve.” 

The presenter highlighted an “important cohort” of 197 patients who did not undergo tissue biopsy because of poor performance status or other issues and were matched to a targeted treatment only by ctDNA. Again patients derived a significant OS benefit with versus without use of a targeted therapy (HR=0.49). 

Turning to the hypothesis that ctDNA alterations may influence survival, Justin Jee said that the dataset “vigorously confirmed” the theory that increasing numbers of mutations and copy number changes in plasma are associated with worse OS. Patients with one or more ctDNA alterations had significantly poorer OS than those without (HR=2.26) and the risk of death increased further in patients with three or more alterations (HR=2.75). And this increased risk of death with ctDNA alterations was true for both treatment-naïve and previously treated patients (HRs=2.15 and 2.50, respectively). 

The investigator reminded delegates that among treatment-matched patients there had been a trend towards better OS with therapy matched by tissue only than matching by ctDNA. 

Hypothesising that ctDNA alteration numbers may explain this discrepancy, they assessed OS in patients with at least one ctDNA alteration in their plasma, and receipt of targeted therapy versus no targeted therapy was associated with a significant HR for death of 0.47, “basically equal to [the] hazard ratio of tissue-matched targeted therapy.” 

The presenter noted that around 15% of patients in the study cohort had mutations or copy number changes detected in their plasma that were not present in tissue samples. After filtering out clonal haematopoietic and germline mutations, 11 of the 92 mutations in the plasma sample remained and repeat tissue analysis detected the mutations in approximately half these cases. 

Justin Jee noted that these particular patients with alterations detected by ctDNA but not tumour tissue had a “much worse prognosis” than those without ctDNA-only alterations and those with only tissue ctDNA alterations, with HRs for death of 2.3 and 1.7, respectively. 

“We really think that [the] ctDNA-only mutations aren’t artefactual and are probably just related to heterogeneity and extent of disease”, he commented. 

The presenter said that ctDNA alteration burden was lower in patients who had received prior treatment, patients with adenocarcinoma and older patients but was higher in patients with extrapulmonary disease. 

In a multivariate model taking into consideration smoking status, tumour histology, treatment history, sex, age, disease extent and use of targeted therapy, the strongest predictor of OS was the presence of ctDNA alterations (HR=2.10), and therefore ctDNA “is really a poor prognostic sign” that is not modifiable, he commented. But use of targeted treatments was associated with a significant OS benefit in this analysis (HR=0.61). 

“We think that studies that use ctDNA matching as a means of linking patients to targeted therapy should account for the lower baseline survival of those patients”, Justin Jee remarked.  

An additional analysis of patients in the community oncology setting is now underway that “will provide further real-world validation in a very different clinical setting from what we’ve studied here”, he concluded. 

Reference  

Jee J, Lebow ES, Murciano-Goroff YR, et al. Overall survival with circulating tumor DNA-guided therapy in advanced non-small cell lung cancer. J Clin Oncol 2021;39(suppl 15; abstr 9009). DOI: 10.1200/JCO.2021.39.15_suppl.9009