What is the fallacy of modern molecular diagnostics?

What is the fallacy of modern molecular diagnostics?

Molecular diagnostics, which includes biomarkers and genetic tests, is a very rapidly growing and continuously developing field. Click here to Buy Kapake 30mg/500mg Tablets from NHS Heroes in the UK Numerous tests are now being applied for mass population screening to identify the early onset of cancer (i.e. the PSA test), or monitor the progression of chronic diseases. Genetic and gene expression testing are the newest additions to the portfolio of clinical diagnostics. However, most of the tests have limited clinical uptake. Worse, after investing substantial time and resources to obtain FDA approval, the majority of molecular diagnostics become rapidly obsolete, failing to recoup initial investments. There are several potential explanations for this phenomenon. Market fragmentation is an important factor because molecular diagnostic tests are being developed by small companies, which are trying to market the tests using their own sales teams which results in limited market penetration. In the best of cases, these minor companies license their tests to larger diagnostic corporations such as Abbot and Qiegen. Another reason for the poor uptake of these tests is the multitude of proprietary test platforms as developers try to rip off higher profits by bundling instruments with consumables. Such a disarray also limits uptake by clinical diagnostic labs due to the capital spending requirements associated with the introduction of new test platforms, the need to train personnel, as well as physical bench space limitations. However, the main reason hindering clinical utilization of the molecular tests is their inherently poor sensitivity and specificity that limits their predictive value.

Many clinicians argue against the use of PSA testing as even this remarkably popular test does not lead to improved survival rates from prostate cancer, which is the main goal of predictive molecular testing. Furthermore, in addition to the direct expenses of PSA screening, the healthcare system must cover the costs associated with prostate biopsies, the use of which is a direct consequence of the PSA results. All of these deficiencies stem from PSA’s dismal performance as a predictive test (AUC ranges 0.68-0.86)[1],[2]. Similarly, poor performance characterizes other cancer biomarkers such as CA-125 for ovarian cancer.

In general, tests based on single analytes suffer from the same deficiencies. Therefore, in recent years companies have pushed for the discovery of multiplexed sets of biomarkers including soluble proteins, RNA-based expression arrays, and DNA variants that are based in a relatively small number of markers. Once the genome-wide phase of screening is complete, a smaller subset of 3-50 markers is used for test validation. However, again the use of a limited number of markers reduces tests’ specificity and the sensitivity. Developers believe that it is easier to replicate smaller numbers of markers, thus they are willing to use tests with smaller AUCs.

What could remedy the dismal performance of molecular diagnostics? One possibility is the joined use of multiple independent marker sets. Recently, deCode was able to demonstrate that genetic stratification with four genetic markers could substantially improve the predictive performance of the PSA test[3]. The combination of serum biomarkers with genetic variation analysis indeed has tremendous potential, however this approach requires the use of two independent diagnostic platforms, posing a challenge for developers of the diagnostic devises.

Genetic stratification is the most promising area of molecular diagnostics, but it suffering from negative “publicity” that stems from years of overpromising and under delivering. Companies like 23andMe that launched “predictive” testing using inappropriate tests that have minimal predictive values, contributed to public distrust and bashing by the scientific community. Nevertheless, complete genome re-sequencing has enormous potential for solving the conundrum of underpowered testing. However, how is the FDA is going to regulate the sequencing based tests?

The entire approach of molecular diagnostics will be turned on its head as the FDA will not be able to “fix” or “freeze” a set of predictive markers because of our knowledge of these markers expanding from day to day at an exponential rate. One cannot simply mask and ignore the 30 million variants present in every individual, and none can intelligently “cherry pick” 3-50 markers that they can interpret. The new era of genetic testing will be forced into the development of continuously re-iterating and re-analyzing algorithms that will provide a small incremental improvement over the previous generation. I personally believe, the “digital genomic testing” approach will be the ultimate edge of molecular diagnostics and it is not too far stretched a concept – I think the reality is not more than five years away…