Esmo 2019 – Foundation eyes liquid biopsy rule change

New data could change liquid biopsy guidelines and boost sales, though regulatory approval would not be a bad idea.

The European Society for Medical Oncology suggests use of cancer blood assays in lung cancer only when traditional surgical biopsy does not yield enough tissue for testing. There is a chance this guidance might change – at least for some non-small cell lung cancer patients – following data on FoundationOne Liquid presented today at the 2019 Esmo conference.

The BFast trial used FoundationOne Liquid, the liquid biopsy developed by Roche’s subsidiary Foundation Medicine, to detect patients with Alk fusions who were then treated with Roche’s Alecensa. This allowed a higher objective response rate than in Alecensa’s first-line NSCLC approval trial, suggesting that the blood test is at least as good as tissue testing at finding the right patients for the drug. 


Alecensa was approved by the FDA for first-line treatment of Alk-positive NSCLC in 2017 on the strength of the Alex trial, in which it trounced Pfizer’s Xalkori in an Alk-positive population. This study used Roche’s Ventana Alk (D5F3) CDx Assay, which works via tissue staining, to identify suitable patients. Alk-positive patients are one of a few genetically defined groups in which Merck’s lung cancer juggernaut Keytruda is not recommended first line. 

The phase II/III BFast trial is an open-label study looking at various group including Ret mutants and patients with high tumour mutation burden. The data presented this morning at Esmo concern the Alk-positive cohort.

FoundationOne Liquid was used to screen 2,219 patients, picking out 119 with Alk mutations in circulating tumour DNA. 87 of these were treated with 600mg Alecensa twice daily. The idea was to prove that using the liquid biopsy has a positive effect on patient outcomes.

At a median follow-up of 12.6 months confirmed objective response rate was 87% – numerically higher than the 79% seen with Roche’s drug in the Alex trial. More than three quarters of patients were still responding after a year, also better than in Alex. 

In 35 patients with asymptomatic baseline CNS metastases ORR was 91%, and 12-month PFS 78%. In exploratory analyses of patients with measurable CNS lesions in the Alex trial ORR was 81%.

Comparing two separate trials is imperfect, but the data might be good enough to persuade Esmo to put FoundationOne Liquid on an equal footing with tissue biopsy. The BFast researchers concluded that the data “validate the clinical utility of blood-based next-generation sequencing as an additional method to inform clinical decision-making in Alk-positive NSCLC".


The central question is whether the higher response rate in BFast versus Alex was to do with more accurate identification of Alk mutants.

Around a quarter of NSCLC patients either have tumours that are unsuitable for biopsy, or undergo a biopsy that does not yield enough tissue for testing. If FoundationOne Liquid is as accurate as tissue testing but can be used in more patients, and those patients are then assigned targeted therapy, that would theoretically allow greater response rates. 

If Esmo is convinced to change guidelines to back FoundationOne Liquid’s use to pick out Alk-positive NSCLC patients the test will likely receive a sales fillip – though only a small one, since Alk is only one of the more than 70 genes FoundationOne Liquid assesses. 

What might provide the group with a bigger sales bump is regulatory approval. Roche almost always seeks formal clearance or approval of its diagnostics, rather than seeking them as lab-developed or homebrew tests, and Foundation has said it will fall into line with its parent company. Sooner might be better than later. 

Share This Article