Medical Device Link.

Originally Published IVD Technology January/February 2003

Instrumentation Development

By developing a virtual network for distributing the benefits of its technology among users, an instrument manufacturer expands its market.

Carl W. Apfelbach and Jose de la Torre-Bueno

Manufacturers introducing new diagnostic systems to the marketplace should always expect to encounter obstacles to market acceptance, growth, and penetration—even though their innovative technology may have high value and its advantages over existing methods may be documented. This happened to ChromaVision Medical Systems Inc. (San Juan Capistrano, CA), which manufactures and markets the Automated Cellular Imaging System (ACIS), an instrument developed to help anatomic pathologists analyze specimens on glass slides.

A previous article in IVD Technology detailed the ACIS technology of visual imaging and described how standard-izing and quantifying slide-based test-ing can produce better information for therapeutic decision making.¹ ACIS was readily accepted by healthcare providers and reimbursed by public and private payers. Systems were placed in major hospitals, reference laboratories, and high-volume pathology practices. However, the large potential market of community pathology practices held some challenges.

ACIS is comprised of an automated microscope; a digital camera; and computerized image-processing technology for detecting, counting, and classifying cells based on color, size, and shape. The system is used to analyze immunohistochemically (IHC) stained slides, primarily for cancer-related diagnostics. Studies have shown that the device offers accuracy, precision, and reproducibility of immunostained slide analysis exceeding that possible with manual evaluation, which was the prevailing method.^2,3

A key obstacle in reaching the community pathology practices was that it was impractical for ChromaVision to place a complete system in such low-volume institutions. Additionally, many community practices do not conduct immunohistochemistry testing in their own laboratories because they lack the equipment or personnel to perform quality IHC staining. To overcome these market barriers, ChromaVision developed a remote pathology network based on existing communications technology to make its ACIS technology widely accessible.

This article recounts how the company's solution stimulated adoption of its product in the anatomic pathology market. ChromaVision's experience suggests a model for other instrument manufacturers facing similar challenges.

The Remote Approach

Figure 1. The ACIS-based remote pathology network. (click to enlarge)

The remote-pathology marketing approach allows community pathologists to take advantage of the new automated technology despite being physically removed from it, reviewing low volumes of cases, and having limited in-house IHC staining capability. In the remote network, the pathologist sends a specimen to a ChromaVision–certified reference laboratory for slide preparation. This laboratory prepares the slide with standardized techniques, automated staining equipment, and reagents optimized for image analysis. It exercises strict quality control throughout the process. Using a complete ACIS system equipped with automated microscope and digital camera, the reference laboratory scans and captures digital pictures of the entire slide and then relays the images to the pathologist for remote analysis. The pathologist in turn uses a modified ACIS workstation equipped with a complete suite of software applications to perform the actual image analysis and report results to the treating physicians (see Figure 1).

Creating this remote pathology capability helped ChromaVision to accelerate adoption of its image analysis technology. The company has placed more than 200 ACIS systems and workstations throughout the United States in a relatively short period of time.

The Need for Standardization

Commercial demand for new diagnostic methods frequently awaits the appearance of peer-reviewed studies. Disseminating information about new technology this way takes years. Chroma Vision was able to increase awareness of the need for image analysis technology based on the publication of several studies showing variability in manual IHC testing for a new cancer therapy.^4–6 The studies helped to demonstrate a therapeutic requirement for test quantitation and standardization that ACIS technology could satisfy.

Targeted Therapies. The therapy in question was Herceptin, an anti-HER 2 monoclonal antibody developed by Genentech (South San Francisco) and used to treat a population of breast cancer patients. To qualify for and receive benefit from this therapy, patients must have tumors that overexpress the HER 2 protein, most commonly measured through IHC testing. Since the protein is present in both normal and abnormal cells, it is necessary to measure the degree of overexpression, not just the presence of HER2 in the cell. This is a challenge for manual IHC analysis, which relies on subjective interpretation by the pathologist using a microscope and solely the human eye.

Herceptin is the first of many oncology drugs that will require accurate, precise, and consistent analytical results to ensure that physicians can most appropriately match patients with targeted therapies. More than 400 cancer drugs are now in late-stage clinical trials and another 1400-plus are in other stages of development.⁷ Many of these, like Herceptin, target specific cell be-haviors. They, too, will require more-precise IHC slide analysis than manual methods can provide, and results report-ing among pathologists and laboratories will be required to exhibit greater standardization.

Quantitative IHC Analysis. Biopharmaceutical companies and research organizations conducting drug discovery using image analysis have found that ACIS gives them the ability to quantitate IHC analysis precisely, something not possible with manual evaluation. Because the instruments are likely to be employed in clinical trials testing, image analysis may emerge as the preferred method to qualify patients for certain therapies when the drugs reach the market. Therefore, pathologists using IHC to select patients for therapy, or to monitor that therapy, could use the ACIS image analysis technology to perform such testing accurately.

Concern regarding IHC test variability stems from three problem areas: inconsistency in staining techniques, subjective interpretation of slide contents, and the lack of quantifiable standards for results reporting.^8,9 Results can vary with differences in tissue prepa-ration, reagents, staining techniques (i.e., manual or automated equipment), and the skill of the technologists performing the procedures. Poor staining, the use of "home brew" assays, or inconsistent quality control can lead to variability in pathology interpretation and results.

Depending on the experience level of the pathologists and the volume of slide reading they undertake, different individuals interpreting the same slides manually could arrive at results that diverge considerably, most notably at the treatment/no-treatment decision point. Current manual IHC reporting for HER2 analysis, for example, is only semiquantitative; it relies on subjective scoring, using a scale of 0 to 3+, and produces less than standardized re-sults. As mentioned, studies indicate that the level of accuracy and the degree of reproducibility from pathologist to pathologist increase significantly with the assistance of ACIS.

Achieving Standardization. The ACIS system scores by counting individual pixels of chromogen color and converting the count to one of 256 distinguishable levels of color intensity. Measurement of staining intensity is thus objective. The system offers four modes of function that can be applied individually or in combination:

• Rare-cell (or rare-event) detection, which is used to find positively stained cells.
• Object counting, which is used to analyze and monitor microvessel density in tissue.
• Staining-intensity measurement, which is used to discriminate among 256 precise levels of color.
• Morphometric characterization, which is used in conjunction with visualization to analyze cell size and shape.

Concluding the ACIS-assistance analysis, the pathologist generates a written report that includes graphical and numerical data on each marker analyzed. Results are either reported as the exact percentage of positively stained cells or else scored in decimal one-tenth increments between 0.0 and 4.0 (e.g., 2.4 or 3.1), depending on the test performed and the appropriate reporting range. The report displays representative images of the cells or tissue analyzed and can include the pathologist's interpretive comments. Its comprehensive and graphical nature has proven valua-ble to ChromaVision for marketing the system to pathologists and physicians.

Universal adoption of these standardized scoring and reporting methods would enable detailed comparisons for clinical research and outcomes studies that are not possible with manual assessment techniques.

Adopting New Technology

ChromaVision expected national reference laboratories to adopt its new technology early. These laboratories rely on specimens received from community hospitals that lack the volume, equipment, or pathology expertise to perform certain specialized tests in their own laboratories. But despite studies documenting ACIS-assisted advantages over manual methods, early adoption of the automated imaging system by the reference laboratory segment was slower than expected. The size of the national reference laboratories and the complexity of their decision-making process can delay purchase. Regional reference laboratories tend to embrace the technology more quickly, both because it has advantages over manual evaluation and as a way to differentiate their services in the local market.

Academic centers, also presumed to be early adopters of new technology, posed different challenges for the manufacturer. Adoption in academic contexts can depend on an imposed budget or grant availability, both potentially causing purchase to be delayed.

This segment also frequently requires customized applications to suit specific research needs. Although the system is flexible for both clinical and research needs, customization that requires significant programming modification can be costly and does not fit ChromaVision's current marketing strategy neatly. Nevertheless, ACIS has been placed with many academic and research centers. The company's introduction of an automated tissue microarray application has further stimulated demand in this segment.

The sector that ChromaVision discovered to have the greatest interest in its technology was the community- based pathology groups (see Table I). They are aware of the variability in IHC staining and analysis and value the increased accuracy and standardization the ACIS system can produce among different pathologists and different laboratories. However, though the technology satisfies the diagnostic needs of this receptive market, providing equipment and support services to lower-volume accounts posed challenges of feasibility and profitability. ChromaVision had to find a way to fully meet the needs of this potential user group.

A Virtual Pathology Network

Figure 2. The ACIS cell detection instrument and flat-screen display of image analysis. (click to enlarge)

ChromaVision has devised a progressive program for configuring a virtual pathology network to help make image analysis technology accessible to community pathologists (see Table II). The configuration includes a full ACIS system complete with automated digital microscope and camera, in place at a reference laboratory plus an ACIS workstation (see Figure 2), consisting of a computer, a flat-screen monitor, and image analysis software, installed at the community pathology practice.

Evolution of the Network. Today the specimen is sent to the reference laboratory for staining and image capture. A data tape containing the scanned images is transferred via overnight mail to the remote pathology practice for analysis on a proprietary ChromaVision workstation. Accessioning is performed by the reference laboratory using its own laboratory information systems, with little or no electronic interface with the remote clients. Paper test-request forms and traditional transport media are used.

The next phase of network evolution would be a virtual private network us-ing standard components. The images would no longer be transported via tape, but instead posted to a host server where the remote pathologist could access them through a prearranged private network configuration. Images could be downloaded to the pathologists' Chroma Vision workstation. This type of network makes possible remote accessioning by the pathologist, allows the reference laboratory to receive a manifest to match up with the specimen's arrival, and eliminates the costs, errors, and time delays associated with overnight tape transfer.

The ultimate virtual pathology network will use Web-based applications. The entire process will take place inside a browser with the use of controls or other server pages, and be transparent to the end-user. This final configuration uses active server pages with code embedded in the Web page in the context of a Web-compliant browser. It will eliminate hardware dependency, allowing any PC at any location to become a vehicle for image analysis using proprietary technology and software carried by standard communications devices. Although the components are standard, the method of compression is proprietary and a patent application is pending.

Training and Support. To serve remote ChromaVision clients, the reference laboratory must be certified by the company in the practice of staining using reagents optimized for image analysis. In addition, its personnel must be competent in all image analysis applications. The laboratory's pathologists take a comprehensive training course at ChromaVision's California headquarters and attend periodic update sessions with staff scientists and technical personnel. The reference laboratory also is required to have an efficient logistics network, to provide client support, and to adhere to turnaround schedules.

In the remote pathology laboratory, the ACIS workstation is interfaced with ChromaVision through a modem to allow remote customer support. The pathologists are trained in their own laboratory by ChromaVision technical service personnel who are experienced laboratory professionals highly knowledgeable about ACIS software applications. The remote laboratories are provided with customized supplies such as test request forms, bar code labels, packaging media, and archiving tapes to support shipping to the selected reference laboratory.

Charges for full or partial reference laboratory services provided to remote clients are determined by the reference laboratory. Reimbursement is appropriate for the service performed by each party in the network.

The Network in Practice. A mejor national reference laboratory and Florida pathology practice collaborated with ChromaVision to pilot the remote pathology concept in May 2001. Since the program's introduction, more than 70 participating pathology groups are using the ACIS workstations.

Benefits of using the remote network are spread widely (see Table III). Community pathologists have testified to having more confidence in their ability to meet the needs of physicians and patients by performing a broad range of IHC tests with better accuracy, precision, and reproducibility. They cite local testing capability and the capacity to consult personally with treating physicians as healthcare benefits to their community. Also, the local practice rather than the reference laboratory receives reimbursement for the professional component of image analysis testing, enhancing the practice's revenue stream.

The reference laboratory, for its part, sees using the system for both direct and remote clients as producing the benefits of standardizing IHC analysis and making results reporting consistent for all customers. The highly automated procedure has been integrated into that laboratory's work flow. New accounts are gained through joint efforts between the reference laboratory and ChromaVision sales representatives.

Pathologists sending specimens for slide preparation and staining often request other services, such as flow cytometry, cytogenetic, and other specialized testing. ACIS-assisted remote clients can ask the reference laboratory to perform full analysis during peak periods and employee vacations. Regardless of where the test is performed, the system ensures a standardized interpretation.

The laboratory's image analysis offering has grown from predominately breast cancer testing to include other IHC tests that can benefit from image analysis quantitation. Additional ACIS systems have been placed in the laboratory as the remote pathology program continues to expand.

Conclusion

The prevailing method of manual microscopy provides the principal competition for ACIS slide-analysis technology. In addition, a range of vendors offer static-image capture and transfer via traditional or electronic media for remote pathology applications, but most of their technologies are used for educational or consultation purposes. These systems lack automated image analysis capabilities to enable pathologists to provide quan-titative, standardized results. Competitive image analysis systems are beginning to come to market but have not achieved significant penetration at this time.

Remote pathology using ACIS software gives community practitioners tools to increase analytical accuracy, precision, quantitation, and standardi-zation. As new drugs come to market with associated tests for markers that were used in the clinical trials already defined, community pathologists with automated image analysis capabilities will be able to provide more-accurate test results to better qualify patients for targeted therapies. The electronic distribution of sophisticated image analysis technology to these community practices via the virtual pathology network is a step toward improving medical decision making to benefit patients, healthcare providers, and medical suppliers.

References

1. J Torre-Bueno, KD Bauer, and B Stewart Wagman, "A New Frontier in Pathology: Automating the Reading of Glass Slides," IVD Technology 7, no. 6 (2001): 37–48.

2. S Wang et al., "Automated Cellular Imaging System (ACIS)–Assisted Quantitation of Immunohistochemical Assay Achieves High Accuracy in Comparison with Fluorescence In Situ Hybridization Assay as the Standard," American Journal of Clinical Pathology 116 (2001): 495–503.

3. K Bloom et al., "Comparison of HER-2/neu Analysis Using FISH and IHC When Hercep Test Is Scored Using Conventional Micros-copy and Image Analysis," Breast Cancer Research and Treatment 64 (2000): 99.

4. S Wang et al., "Laboratory Assessment of the Status of HER-2/neu Protein and Oncogene in Breast Cancer Specimens: Comparison of Immunohistochemistry Assay with Fluorescence In Situ Hybridization Assays," Journal of Clinical Pathology 53 (2000): 374–381.

5. MP Hoang et al., "HER-2/neu Gene Amplification Compared with HER-2/neu Protein Overexpression and Interobserver Reproducibility in Invasive Breast Carcinoma," American Journal of Clinical Pathology 113 (2000): 852–859.

6. DC Allred and PE Swanson, "Testing for erbB-2 by Immunohistochemistry in Breast Cancer," American Journal of Clinical Pathology 113 (2000): 171–175.

7. New Medicines in Development for Cancer, 2001 Survey, [on-line] (Washington, DC: Pharmaceutical Research & Manufacturers of America, 2001 [cited 31 October 2002]); available from Internet: http://www.phrma.org/newmedicines/resources/cancer01.pdf.

8. C Taylor, "The Total Test Approach to Standardization of Immunohistochemistry," Archives of Pathology and Laboratory Medicine 124 (2000): 945–951.

9. T Seidal, AJ Balaton, and H Battifora, "Interpretation and Quantification of Immunostains," American Journal of Surgical Pathology 25, no. 9 (2001): 1204–1207.

Carl W. Apfelbach is CEO and president and Jose de la Torre-Bueno, PhD, is vice president for research and development at ChromaVision Medical Systems Inc. (San Juan Capistrano, CA). The authors can be reached via capfelbach@chromavision.com and jdelatorre@chromavision.com, respectively.

Copyright ©2003 IVD Technology