Where next with liquid biopsy as a screening test for cancer – a global insurer’s (Swiss Re) inquiry

(First published on LinkedIn, 11/2016)

Insurance companies are natural 'Futurists.'

They are good at anticipating the future; their financial bets depend on it.

Not surprisingly, they pay strong attention to social, technical and environmental developments that could impact their business. Global insurers are often at the leading edge of assessing future scenarios across very diverse fields of human activity (see global warming etc.).

Insurers who provide benefits payable upon diagnosis of cancer have recently identified a new significant (financial) risk: The widespread adoption of screening procedures that could indicate the presence of disease in the absence of clinical manifestations. A key question they now ask is whether liquid biopsy technologies will change how and when we diagnose cancer.

Detecting cancer without any clinical signs

Swiss Re is one of the largest re-insurance companies underwriting, among other things, lifetime cancer risk in several regions of the globe. Countries such as China and South Korea have become leading adaptors of screening technologies. Recently, Swiss Re organized an Expert Forum on Cancer Diagnostics in Zürich, Switzerland. The insurer wanted to understand better the new technologies potential of achieving a much earlier detection of cancer. Current definitive diagnosis relies on tissue pathology and imaging. Too often, unfortunately, they can only confirm the advanced clinical stage of the disease. From an insurer's perspective, Swiss Re’s concern is that a putative early detection result could compel clients to ask for payout in the absence of a clinical presentation. Underlying this, Swiss Re experts reported that in some Asian countries wealthy asymptomatic individuals are already requesting liquid biopsy screens for ‘early detection.’

So, what to do with negative or positive results in this situation? If such tests are analytically validated, a negative result does not mean that there is no future cancer risk. A positive test may trigger additional, follow-up investigations such as imaging. The detection of oncogenic targets in otherwise healthy individuals is not entirely new. Significant numbers of healthy individuals carry bcr-abl translocations without having concurrent chronic myelogenous leukemia (CML) presentation (reference). The considerable risk scenario for Swiss Re (and some attending clinicians) was the possibility that in some, perhaps less regulated regions of the globe, the definition diagnosis of cancer may be changed based on clinically unvalidated (intended use) liquid biopsy screens.

Clinicians at the meeting appeared somewhat skeptical about the potential of liquid biopsy screens to achieve early detection of cancer - but the technology is here to stay for therapeutic monitoring.

What was the take-home message from this meeting? They expressed strong objections to moving away too quickly from pathology and imaging-based diagnosis without more clinical evidence. The argument centered around clinical sensitivity (particularly for CTCs and EVs). They worried that shedding spurious CTCs, ctDNA or EVs into the bloodstream could be part of a normal process. Some clinicians argued that not all small solid tumors would develop into metastatic disease. However, there was a consensus that liquid biopsy-based methods will become standard tools for therapeutic monitoring, and the understanding of the genetic evolution of the patient’s tumor after a traditional cancer diagnosis. The recent approval of the COBAS EGFR mutation test by the FDA to identify mutations from blood in lung cancer is a case in point.

The group also discussed the use of whole-genome sequencing methods for clinical decision-making. There was strong support of whole-genome sequencing of CTCs and FFPE samples for research. However, most participants did not believe that there is sufficient clinical evidence for improved clinical outcomes beyond analyzing the known actionable genes (i.e., EGFR, BRAF, NRAS, ALK, etc.). The consensus was that the understanding of the vast information generated by genome-wide sequencing runs is still too limited to be clinically useful. The participants preferred targeted panels, particularly in situations where FFPE tissue is limiting. From their perspective, it would be difficult to justify billing the patient/healthcare system for genome-wide exploratory analysis services without more clinical validation. As sequencing and analysis costs come down, this may change, of course. An exception to this could be immune or vaccine-based therapies, which will likely require massively parallel neoantigen identification for personalized vaccine production.

We need to define better what 'early' detection means

Ultimately, it will be important to define better what ‘early’ detection means. It is likely that liquid biopsy will lead to earlier detection of cancer for some tumor types (i.e., ovarian). However, it may be very challenging to use it for ‘population screening’ (perhaps based on age group) if clinical sensitivity remains limited. In the short run, the risk to insurance companies underwriting lifetime cancer risk appears somewhat limited.

Link to Swiss Re Forum on Cancer Diagnostics: