Please do not call it Theranos

To the Editor,

Recently, Snyder et al. published a paper which describes successful measurement of thousands of diverse proteins in 10 µL of blood [1]. This technical development could be used in the future for health-related applications such as diagnosis, prognosis, risk assessment and therapy monitoring. In the interviews to the public media, including “The Stanford Daily” Stanford researchers claim to create ‘Theranos that works’ (stanforddaily.com), Snyder characterized their achievement, as a “Theranos that works” contribution, giving the impression that the recently debunked claims of Theranos and the subsequent highly publicized criminal investigations and convictions [2, 3] were probably premature, but likely achievable. Undoubtedly, the Theranos saga is well known to your readers and our previous PubMed-indexed contributions to the story, and those of very few others, are indelible descriptions of what actually happened [4], [5], [6], [7].

I here draw attention to your readers that the public media-promoted slogan “Theranos that works” is highly misleading. In general, the Theranos Organization had two major objectives. One was the analytical challenge of measuring hundreds or thousands of blood components in a drop of blood, for the purpose of diagnosis and prevention of human disease. The second, and most challenging part, was the demonstration that either all, or some, of these analytes are indeed useful biomarkers for certain diseases. The first challenge has now been met by Snyder et al. [1]. We remind here that previously, other researchers and companies (such as Olink Proteomics; https://olink.com) have commercially available technologies which can quantify with high sensitivity and specificity over 3,000 proteins in a drop of blood (<100 uL). Olink and similar technologies (e.g., https://somagen.com) are now mature, precise and sensitive and can be a used for biomarker discovery [8], [9], [10], [11]. However, this does not mean that these technologies can be used for human diagnostics, and the popular saying “less is more” may be applicable here. In order for one or multiple candidate biomarkers to be useful for disease applications, these biomarkers should exhibit high sensitivity (ability to detect disease, being positive in diseased individuals), high specificity (they should be negative in patients without disease) and more importantly, they should be characterized by high positive predictive value (PPV, which represents the chances of somebody having the disease if the test is positive) thus ruling-in disease, and negative predictive value (NPV, the chances that the person does not have the disease if the test is negative – rule out disease). Please see our previous contributions related to these concepts [12, 13]. Theranos never succeeded in showing that any of the molecules that they could measure have these important biomarker characteristics. Snyder admits that they are in the initial stages of establishing if this is the case with their quantifiable molecules, and they have started companies which will pursue this goal in the future. A similar concept targeting health and disease was proposed a few years ago [12, 13] by pioneering scientist Leroy Hood, though the company he created, Arivale. This company claimed to revolutionize health and disease prevention and diagnosis by using similar concepts to Snyder and Theranos; measurement of hundreds to thousands of proteins in serum, as well as performing other tests such as molecular analyses, whole genome sequencing, profiling the microbiome etc. for wellness testing. We contested the validity of this approach [12, 14, 15]. After a few years in business, Arivale went bankrupt [16]. Simply, the ability of this approach to diagnose and prevent diseases is very limited as we explained [14, 15]. The fact is that many of these candidate tests are producing a lot of false positive results, leading to confusion on what is the best way to follow-up. Consequently, many patients are subjected to additional unnecessary, expensive and potentially dangerous and invasive procedures.

I believe it is premature to support renewed interest on the Theranos principle. This may give the impression that Theranos was in fact visionary, but they didn’t reach the end due to lack of time or funds. The reality, as I emphatically stressed as early as 2015 [5], is that the Theranos technology right from the beginning, was a straight forward approach that has been used in laboratory medicine years earlier. Theranos (and Hood and Snyder) did not manage to show that this kind of diagnostic approach can revolutionize medicine, even if they had analytical/technologically successful tools.

I conclude that the testing of minute amounts of blood from a fingerprick, for thousands of proteins and other components, has been in the forefront of laboratory medicine research for years. While technology now allows for such measurements to be made, their applicability for diagnostics is not a given, until the supporting data are generated. Any relationship of these ongoing efforts with the now debunked Theranos company, should be avoided.