Jonathan Briggs

One source of new products is to transfer an existing technology to a new application. ChromaVision Medical System’s Automated Cellular Imaging System (ACIS) is an example of transferring a military technology to a medical application. ACIS technology was originally developed as part of the Strategic Defense Initiative (SDI), the missile defenses system more popularly known as Star Wars. A key challenge for Star Wars engineers was to develop a system that could quickly and accurately determine which object among a field of many, most of which were decoys, was the missile with the nuclear payload. The answer to this challenge was a computer-based imaging system that used rare event analysis to examine a field of numerous similar objects and accurately discriminate among them. Therein lies the connection between Star Wars and cell analysis.

Dr. Torre Bueno, Chromavision vice president of research and development with the company’s Automated Cellular Imaging System (ACIS)

The idea for a medical application of the Star Wars rare event analysis system originally developed by John Scott at XL Vision Inc. in the late 1990s. Scott designed a system to find microscopic cancer cells in pathology slides. Because only a few of the tens of millions of cells in a slide may be cancerous, micrometastases often go undetected. Not to mention that reviewing slide after slide of mostly non-cancerous cells is both time-consuming and tiring. However, finding those cancer cells, if they are there, can be just as important as finding that missile with the nuclear warhead, at least to the individual whose slide it is. Knowing whether or not there are micrometastases present can increase the accuracy of disease staging and prognosis. Bone marrow metastasis, for example, can be of greater value than node status, tumor stage or tumor grade in assessing the status of breast cancer patients. Determining the presence or absence of occult tumor cells also may be useful for monitoring disease progression and evaluating the effectiveness of different therapies.

Current ACIS technology is able to detect one suspicious cell among 108 normal cells, in a standardly prepared IHC slide. Analyzing cell samples with the ACIS system is very straightforward. First, microscopic slides of tissue samples are prepared using ICH or cytochemistry. Up to 100 slides are then loaded into the machine. At this point, the machine can run unattended. Using an automated microscope, ACIS scans each slide at low magnification. Cells of interest are identified by color and counted. The location of suspect cells is recorded and these cells are further characterized at higher magnification using additional color and morphometric criteria. Digital images of suspect cells are collected and stored. The pathologist reviews the collected cell images visually and with the aid of application-specific quantitative information.

XL Vision formed ChromaVision in 1997 to market the ACIS technology and develop additional applications. Despite the inherent logic of using computer-based rare event analysis to scan pathology slides, it took more than a simple show-and-tell to establish the value of this Star Wars-derived technology to clinical labs.

One of the more formidable aspects of marketing ACIS technology, according Dr. Torre Bueno, ChromaVision vice president of research and development, “has been to demonstrate that ACIS can increase the accuracy and reliability of slide-based diagnostics, which of course means increasing the accuracy and reliability of the person reading the slide.” When you transfer an existing technology to a new application, although you do not have to demonstrate the need for the product, you still have to convince potential customers of the advantage your product has over existing methods. In the case of ACIS, this meant demonstrating a potential improvement in a procedure that is done more than a billion times a year in the United States. Fortunately, for ChromaVision, the basis of the argument in favor of ACIS is a fact of human nature, or rather human vision.

Although IHC staining and interpretation is the standard of care for pathological diagnosis of many diseases, the limited ability of the human eye to discriminate between variations in color has resulted in considerable inter- and intra-laboratory variability in IHC-based test results. For example, in a recent British study that involved 200 laboratories in 26 countries, false-negative rates for the detection of estrogen receptor status in breast cancer tumors ranged from 30 percent to 60 percent in the tumors with low expressing tumors. So the marketing problem was not to convince potential customers of the need for improvement in IHC-based slide analysis, but to convince them a computerized imaging system originally developed to ID nuclear missiles could improve on the current method of slide analysis.

To establish the benefits of ACIS technology, ChromaVision conducted a clinical study comparing the accuracy and reliability of IHC scoring done by pathologists with and without the assistance of ACIS. The study found that 44 percent of the slides initially tagged as negative by traditional manual review were correctly identified as positive by a pathologist using ACIS. In addition, use of the ACIS system resulted in a 300 percent increase in sensitivity to detect tumor cells compared to the traditional manual method, and the rate of accuracy for pathologists in the study using ACIS to identify tumor cells in bone marrow samples was 100 percent.

“It’s simply due to evolution,” said Dr. Torre-Bueno. “Over time, humans developed greater abilities for pattern recognition than for distinguishing levels of color intensity. Consequently, our ability to distinguish color intensity is relative. We interpret color intensities in relation to one another. Whereas, with ACIS, there is a consistent baseline reference and a higher level of acuity for color intensity.” ChromaVision’s ACIS technology uses differences in color intensity to locate and characterize stained cells in IHC slides. Unlike the human eye, which can only detect about seven color levels, ACIS can reliably distinguish up to 255 levels of color. Consequently, ACIS increases the reliability of IHC-based cell analysis. Other ACIS-based marketed by ChromaVision include HER2/neu, p53, KI-67, and estrogen and progesterone receptor assays. So if you have an idea for transferring a technology to a new application, keep in mind the lesson of ChromaVision and ACIS. “When you transfer a technology to a new application, you may save time and money in development, but you may have to spend more effort marketing,” said Dr. Torre Bueno.