DIAGNOSTICS IN INFORMATION BASED Medicine

White Paper: A Point of View

 
 
Version 1.1

 
 
Satish Gambhir

IBM Life Sciences

                                                                                                                                  July 21, 2003 
 

Acknowledgement

Many have contributed to this white paper, in particular, Houtan Aghili, Rich Bakalar, John Cornali, Greg Rutledge, Harley Rodin, Philip Barker, Eric Goodall, George Miller, Alan Kalvin, Patrick Boyle, Bill Kirkland, Kip Peterson, Michael Hehenberger, Rohit Singh, Anwer Khan, Maree McLaren, Neil de Crescenzo, Scott Jenkins, Mark Morrison, Cecyl Hobbs, Cathi Stahlbaum, Doug O’Boyle, Theresa Gaffney, Ron Martin and Brett Davis. I would like to thank the entire team and especially Houtan Aghili for his thorough review and numerous valuable suggestions. 
 

The views presented in the white paper are those of author’s. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Table of Content

Executive Summary

Good diagnostics capabilities are an integral part of Information-based medicine and essential for providing effective patient treatment. Information-based medicine, as defined by the US National Library of Medicine, is the conscientious, explicit and judicious use of current best information in making decisions about the care of individual patients. It integrates individual clinical expertise with the best external clinical evidence from research.

The increasing availability and growth in digital data from Medical imaging and other emerging diagnostics tests is enhancing our ability to provide more effective targeted treatment solutions. Our deeper understanding of biochemicalinteractions at the cellular and molecular levels will allow us to move from curative (e.g., antibiotics) and preventive (e.g., vaccines) medicine to ‘predictive and corrective’ medicine enabling new paradigms in health care. At the same time the volume of complex diverse diagnostics data and post processing requirements are creating new challenges to manage, integrate, reduce and mine the data for improving the accuracy of diagnostic decisions.

These challenges and the shift in our customers’ environment from analog film to digital diagnostics data play to IBM’s strengths. IBM Life Sciences has a unique opportunity to capitalize on this change in customers’ environment by developing and offering new solutions and transforming our customers’ businesses. Transformation to “on demand” diagnostics environment in which the customers’ business processes are integrated across the company and with their partners, suppliers and customers will allow IBM’s customers to respond with speed to any demand, market opportunity or external threat. Moreover, enhanced workflows and more effective governance models together with “pay-as-you-use” utility concept, the hallmarks of transformation to an on demand environment, will improve diagnostics capabilities.

Even though the maturity of the on demand transformation is still in its early stages, many of our key competitors have launched on-demand initiatives: Sun has been actively campaigning implementing services on-demand using the Sun Open Net Environment (Sun ONE); Hewlett-Packard has launched a major advertising and marketing campaign as kickoff to its goal of reshaping itself into a leading on-demand contender, positioning its approach as “adaptive infrastructure” and its ads ask customers to “demand more” — a team-to-beat reference to IBM’s on demand strategy. Emerging PACS (Picture Archiving & Communications Systems) vendors and our potential business partners, such as Stentor and Emageon have incorporated elements of on demand in their offerings. For example, Stentor has developed iSite™, an image and information management system to deliver on-demand diagnostic quality images over existing hospital networks, and "always online" long-term storage. Clearly our competitors are beginning to exploit this paradigm shift and have signaled their commitment with acquisitions and strategic partnerships.

The opportunity to provide I/T infrastructure for Medical imaging diagnostics and systems integration is large. Table 1 depicts Frost & Sullivan’s estimates of worldwide (WW) PACS sites, storage requirements and revenues from H/W, S/W and Systems Integration in 2003-2007 timeframe. Many analysts believe that the technology adoption is likely to be accelerated and actual revenue attainment will be much larger.  
 
 

Table 1. Forecast of PACS I/T Revenue (Source: F&S)

In order to gain market and mind share in this emerging market and to accelerate Life Sciences growth we recommend developing a few showcase solutions with PACS vendors, large hospitals and research centers, learn and develop expertise, and capture a major share of the I/T market.

Introduction

Good diagnostics capabilities are an integral part of Information-based medicine and essential for providing effective patient treatment. Information-based medicine, as defined by the US National Library of Medicine, is the conscientious, explicit and judicious use of current best information in making decisions about the care of individual patients. It integrates individual clinical expertise with the best external clinical evidence from research.

The increasing availability and growth in digital data from Medical imaging and other emerging diagnostics tests is enhancing our ability to provide more effective targeted treatment solutions. Our deeper understanding of biochemicalinteractions at the cellular and molecular levels will allow us to move from preventive (e.g., vaccines) and curative (e.g., antibiotics) medicine to ‘predictive and corrective’ medicine enabling new paradigms in health care. At the same time the volume of this complex diverse data and post processing requirements are creating new challenges to manage, integrate, reduce and mine the data to improve accuracy of diagnostic decisions.

These challenges and the shift in our customers’ environment from analog film to digital diagnostics data play to IBM’s strengths. IBM Life Sciences has a unique opportunity to capitalize on this change in customers’ environment by developing and offering new solutions and transforming our customers’ businesses.

I/T Opportunity

Diagnostics data can be broadly classified into Image and non-Image data. The image data is generally obtained from many different imaging modalities, e.g., digital X-Rays (Radiographs), MRI (Magnetic Resonance Imaging), Ultrasound, CT (Computed Tomography), visible light pictures. Non-Image data such as patient demographic data, clinical reports and gene expression profile data must be integrated with the current image data and pertinent priors to provide enhanced diagnosis and review disease progression.

The images produce large amount of digital data. For example, data acquired from 40 million mammography exams in the US alone can be more than 5 Peta bytes per year. Many PACS vendors provide scanners and an integrated information technology infrastructure for taking, storing, distributing, retrieving and visualizing Medical images. Table 2 shows the estimated number of exams performed per year in the US (A conservative estimate), and the average amount of data produced from various image modalities. In the table, uncompressed size of 128 Mega Bytes for mammography is for 4 views, each 4096x4096 pixels with gray scale range represented by 16 bits. Uncompressed size of 500 M Bytes for multi slice CT scan is for 1000 slices each 512x512 pixels and 16 bit gray scale range.

Table 2. Digital Data from Various Imaging Modalities

The opportunity to provide I/T infrastructure for Medical imaging diagnostics and systems integration is large. Table 3 depicts Frost & Sullivan’s estimates of worldwide (WW) PAC Systems sites, storage requirements and revenues from H/W, S/W and Systems Integration in 2003-2007 timeframe. Many analysts believe that the technology adoption is likely to be accelerated and the actual revenue attainment will be much larger.  
 

Table 3. Forecast of PACS I/T Revenue (Source: F&S)

Diagnostic Value Chain

The diagnostics value chain consists of data acquisition, interpretation, distribution & storage and finally providing effective targeted treatments, as depicted below in Figure 1. Integration with electronic Medical records and effective data mining further enhance diagnostics value.  
 

Figure 1. Diagnostic Value Chain

As indicated, the value chain starts with the acquisition of diagnostics image and non-image data. A physician, in consultation with a radiologist, adds value by incorporating radiologist’s interpretations. Distribution of the data for enhancement, collaboration and storage for future comparisons of results provides additional value. Effective treatment through integration with the entire Medical records and outcomes analysis is at the top of the value chain. Treatment can be improved iteratively with the availability of additional diagnostics information, computer aided diagnostics, query capability to retrieve image examples for a given diagnosis and collaboration with additional sub-specialists. Clearly, IBM Life Sciences can play a role in the entire value chain of diagnostics and treatment.