Landscape of rapid-detection systems for the food industry
Part of the demand for new solutions has developed from the evolution of the regulatory landscape over the last 20 years, while a majority of the demand has come from companies striving for continuous improvement of their food-safety systems.
Justin Ransom, Ph.D., OSI Group LLC
In general, the greatest challenge for all testing methods is their applied sensitivity and specificity. Although all of the test kits have validated claims for superior performance in these areas — and even have AOAC approval — the systems do not always perform as expected in the production environment or food matrix. In general, a precautionary principle in accepting a higher likelihood of false-positives and little to no likelihood of false-negatives in the testing systems is commonly used; however, one has to decide what level of uncertainty can be tolerated.
Interestingly, in almost all of the rapid technologies currently available, the undeniable challenge of enrichment still exists. In a perfect world, testing for pathogens would be as easy as measuring pH with a piece of litmus paper.
The reality is that in today’s food-processing world, we are looking for a needle in a haystack. It is no longer common to find high populations of target organisms in the environment, much less in the product; therefore, the tests must include an enrichment step that allows adequate growth such that the target organism can be detected by the test. Traditionally, this step can take days.
Rapid-detection systems commonly use patented selective enrichment broths for the targeted organism, whereas other methods use a less selective enrichment broth. Instead, they use incubation time and temperature controls to select for growth of the targeted organism. Both techniques have proven to be effective through extensive validation.
The use of molecular technology for the amplification of target DNA has become relatively common and easy to use. Unlike methods used in popular crime shows on TV, there are challenges that exist within these detection systems. The problems principally exist because the population of target organisms carrying the specific DNA is so low.
Another challenge is that other species of bacteria carrying the same or similar genetic targets cause complications within the testing systems. As a result, it is becoming more common for rapid-detection systems to employ multiple steps to prevent this from occurring. These steps add little time to achieve the test result, but add great value in improving the sensitivity and specificity of that result.
Another issue with testing, and with rapid-detection systems in general, is that companies can spend millions of dollars on these tests every year, and oftentimes more than 99 percent of the results are negative. Yet, anyone who has spent time with statistical process control (SPC) knows, when 99 percent of the observations are “zero” and 1 percent of the observations are “one,” then by using SPC, your process is “out of control” when you have a positive test result.
In reality, this is not necessarily the case. Ideally, with improving technology, rapid-detection systems will get to a point where they can not only produce a positive or negative result from a target organism, but they can also give an indication of other target organisms (pathogenic or non-pathogenic) in the processing environment. By this means, through more holistic sampling and analysis, food processors would be able to more accurately characterize whether or not their process is “in control” for a specific pathogen, and also note other generic markers that might provide more insight into the process.
Everyone may have a different idea regarding what the ultimate rapid-detection system could detect. One idea might be that the ultimate technology would look for multiple pathogens in the same sample, at the same time, using the same enrichments. In addition to that, it would also be able to detect pathogens at the highest applied sensitivity and specificity on the market. Ideally, test results would be received within 24 hours of sample collection, and while positive test results are uncommon, process-control data could be generated to give an indication of the processing environment and/or resulting product.
We are looking for a magical test, the ever-elusive silver bullet, which actually may emerge not too far in the distant future. What was once complicated science is now becoming increasingly simple as technology progresses, so a test of this caliber may not be considered science fiction after all, but become science fact.
There may not currently be one single rapid-detection system that can meet every demand across all food-processing environments or food matrices; however, there are systems that are superior in specific environments or specific matrices.
It is up to the food processor to understand the testing needs for today, then strive for — and embrace — continuous improvement in the future.