The Trouble with Precision in 'Precision Medicine'

(First published on LinkedIn, January 2017)

I am not sure who coined the term 'Precision Medicine,' but it is hard to escape its trendiness - particularly in oncology. Precision is both evocative and catchy. 'Precision Medicine' has, by now, made its way not only into the speeches of health policy makers (a ‘mantra’ in many high-profile presentations) but also to the high echelons of some pharmaceutical companies. Managers who work on biomarkers and patient selection have started to describe their jobs as ‘doing precision medicine.'

The psychology behind this is evident and the aspiration laudable, of course. Adding the word 'precision' carries the promise of radically improving clinical outcomes using genomics in a field that is famously complex and in urgent in need of better results.

The strong impetus is real and based on a multitude of genomic technologies that have enabled the collection of massively parallel information on cancer patients' tumor-DNA. Most of this information we do not understand very well, yet. Thus, there is plenty of potential for innovation after the successful approvals of the first wave of relatively 'simple' companion diagnostic tests. The more complex next-generation genomic technologies have become ubiquitous in research, clinical trials, and, increasingly, also in clinical practice.

However, are we putting the cart before the horse when using the term 'precision medicine' to describe the current clinical practice? After all, many medical disciplines have had considerable success without early evocations of 'precision' term. Who has heard about 'precision infectious disease medicine,' or 'precision vaccines,' or 'precision cardiology' or 'precision transplantation?'

Currently, 'Precision Medicine' targets mostly relatively rare 'actionable' mutations. Consequently, most cancer patients, unfortunately, are unlikely to profit from such a treatment right now. Spelling this out does not diminish the importance of testing every lung cancer or melanoma patient (examples), and should not reduce the intensive continuation of molecular and clinical research using genomics. The promise is great indeed. I have high hopes that more people will profit from this research, particularly as it may enable the identification of neoantigens and the design of highly personalized cancer vaccines. I also do not want to minimize the value testing has provided to patients who live longer because they carried these rare but actionable mutations. However, it is always good to keep the balance between vision and reality. Moreover, there are still open technical issues.

In a recent letter to JAMA Oncology, the authors compared two commercially available next-generation sequencing (NGS) platforms in their ability to identify mutations from samples derived from the same patients. They looked at a subset of nine patients that had overall 45 genetic alterations measurable by either platform. Below an excerpt from the journal:

"One patient had no identified genetic alterations using either platform. The remaining eight patients had 45 alterations, only 10 (22%) of which were concordant between the platforms. For 2 of these remaining eight patients, there was no concordance among the described alteration. For these eight patients, a total of 36 drugs were mentioned. However, only nine drugs (25 %) were recommended for the same patients by both platforms, and in 5 patients, there was no overlap between the drugs recommended by the 'first' test and those recommended by the 'second' test. Concordance among drugs improved to 8 of 13 (62%) when reported mutations were also concordant".

It is not the first report pointing to discordant results between next-generation sequencing platforms performing complex testing (reference).

For those not familiar with the practical aspects of the diagnostic laboratory work, it is essential to understand that differences in the sequencing platforms’ ability to detect mutations are not solely a function of the sequencing technologies' differing designs and sensitivity. There is also a significant contribution of variation introduced by the methods that precede the actual DNA readout (sequencing) and the bioinformatics. These processes are technically not as 'glitzy' or high-tech as the back-end technology and the computational analysis. The steps consist of tumor tissue sampling (timing and quantity of the biopsy), fixation, the DNA/RNA extraction and the tumor sample’s molecular composition (intra- and inter-tumor heterogeneity). Sample fixation is in many cases not controlled by the testing laboratory as collection and fixation can occur at other clinical sites, at different times, and often using different processing protocols. All these processes impact the quality of the diagnostic end-result. Their standardization and quality control, while relevant, has remained difficult. Integrating and managing these processes also determines how a test perform - not only analytically but also clinically.

Increasing routine patient testing for driver mutations needs perhaps a more detailed diagnostic workup to ensure that the information it provides and the potential therapy the results recommend become more reliable.

Let me be clear. I am not minimizing the technological and clinical achievements brought about by genomic technologies. I have seen very directly the insights they provide. Moreover, yes, I feel as excited about the future as everyone else. Clinically, however, we are still in the early stages, and our understanding of functional genomics is still in its infancy.

Cancer is a complex disease, and we should be a bit more careful when we characterize the current clinical practice as 'precision medicine.'

In summary, it would be more accurate to say that medicine is making progress in extending the lives of some patients whose tumors' carry relatively rare but actionable mutations. The results point in the direction of a potential 'Precision Medicine' in the hopefully near future. Without this qualification, 'Precision Medicine' tastes a bit of marketing driven by very impressive technological advances that are still pending their clinical validation to extend patients' lives from months to years.