1. Field of the Invention
The field of the invention is network- and software-based diagnostic pathology.
2. Description of the Background
Methods for the diagnosis and treatment of serious medical conditions are increasingly important. By way of example, cancer is currently a major cause of death in the world, and is overtaking cardiovascular disease as the leading cause of death in Western countries. The optimization of tools for the diagnosis and treatment of cancer is paramount. One important tool in the diagnosis of cancer is immunohistochemistry. Immunohistochemistry is the process of localizing proteins in cells or a tissue section by exploiting the principle of antibodies—acting as reagents—binding specifically to antigens in biological tissues and is frequently used in the diagnosis and treatment of cancer. More particularly, immunohistochemical staining (“immunostaining”) is a form of testing that allows pathologists to conduct immunohistochemistry for diagnostic purposes. Through immunostaining, specific molecular markers detected by antibodies indicate to pathologists the presence of particular cancer types. Accordingly, pathologists dealing with potentially cancerous tumors are required to have a broad understanding of antibodies and immunohistochemistry to select groups of antibodies for use in immunostaining in order to make differential diagnoses of similarly appearing tumors.
In addition to immunohistochemistry for the diagnosis of cancer, many other types of reagents or reagent-like substances or representations are used for the purpose of differential diagnosis or prediction of medical conditions. These include but are not limited to immunofluorescense, histochemistry, and molecular pathology, to name a few. Pathologists diagnosing and treating these conditions are equally tasked with having a broad understanding of the various reagents used in the differential diagnosis of these conditions.
Significantly, in recent years there has been a commercial explosion of available antibody reagents for the purposes of diagnosing cancer, and this trend may be presumed to extend to the reagents available for diagnosis of other medical conditions. Concurrent with this trend, determining the best panels of antibodies for immunostaining has become increasingly challenging and time consuming for physicians. Despite the trend, and the problems associated with it, there have been very few efforts to provide an easy way to determine the best groups of antibodies for immunostaining that will aid in the differential diagnosis of tumors. One internet-based system which has been developed in the recent past was ImmunoQuery®. ImmunoQuery® was a successful immunochemistry database query system that would list antibodies for differentiation purposes, rank the antibodies in terms of their ability to differentiate between cancerous conditions, and provide instant references to medical references relating to the antibodies. However, ImmunoQuery® was generally a diagnostic toolbox, and did not utilize a step-by-step or workflow optimized diagnostic system. Moreover, ImmunoQuery® did not use digital images for diagnostic purposes.
Therefore, there is a significant and ongoing need in the diagnosing and treating medical community for a universal step-by-step tool which incorporates a comprehensive database of reagents, optimizes the process of selecting proper groups of reagents for further testing in connection with differential diagnosis, and which allows pathologists to correctly and easily interpret the results of the testing using these reagents by providing a reference to digital images in addition to important numerical measures which correspond to the testing results.
In accordance with the above, an innovative diagnostic system and method is provided. A network-based system is used to collect preliminary diagnosis hypotheses from treating physicians and staff, to recommend customized groups of reagents for further testing based on pathologist input and a wide range of medical reference data, and to provide digital images of positive and negative testing results in addition to numerical measures of reagent behavior to assist pathologists in interpreting the results of laboratory testing based on the recommended reagents. The invention results in an accurate, step-by-step, work flow optimized system and method for making differential diagnoses of medical conditions through the use of reagents or reagent-like materials or representations.
A portion of the disclosure of this patent document contains material which is subject to copyright protection. The copyright owner has no objection to the facsimile reproduction by anyone of the patent document or the patent disclosure, as it appears in the Patent and Trademark Office patent file or records, but otherwise reserves all copyrights whatsoever.
The present invention in its various embodiments is a network-based system and method for recommending reagents or reagent-like materials or representations to a pathologist or other individual (hereinafter referred to as “user”) for the differential diagnosis of, among other diagnoses, medical conditions. More specifically, as defined herein and as used for the purposes of the present invention, the term “reagent” may encompass any substance, material, or representation which is used to indicate the presence or possible presence of another substance, material, or representation.
Once a user suspects potential diagnoses, these diagnoses are entered into the network-based system. Based on this data, the system generates a list of potential reagents for use in further testing required for conclusive differential diagnosis. The recommended reagents, or a subset thereof, are made subject to reaction or further testing and the results of the testing are compared with stored digital images and corresponding numerical measures of reagent behavior in the network-based system. The user then makes a selection from among those diagnoses in the system whose images and corresponding numerical measures of each reagent's behavior are most comparable with the results of the testing. Using this selection, the system then generates a diagnosis, diagnoses, or inconclusive results, the latter of which requires additional meta-analysis in the system or the repetition of certain steps in the method to obtain a conclusive diagnosis.
The method begins once a user determines that reagents will be required for diagnosis of a medical condition [100]. At this point, the user will likely have some idea as to potential diagnoses. Using this preliminary information, the user will access the network-based system. Generally, and according to one embodiment of the invention, the network-based system utilizes a software program maintained or hosted on a central server or servers which are remotely accessible by a user over the internet. However, the system could also consist of a smaller and more localized network, or even the use of software installed on a single computer terminal.
The preferred embodiment is a web-based product named STATdx™PathIQ™, and consists of subscription-based software accessible via login at the STATdx™PathIQ™ internet webpage [101], as shown in
STATdx™PathIQ™ utilizes a Java servlet-based back end with an HTML and CSS front end running on Resin's professional web server. STATdx™PathIQ™ further utilizes a Postgres database to store reference data and a Windows file system for storing images. The Resin server runs on Microsoft Windows 2003. Content is assembled in STATdx™PathIQ™ by using an XML-based authoring system which creates links between images, antibodies, diagnoses, and cases. An example of a suitable XML-based authoring system is that disclosed in patent application Ser. No. 10/723,018, which is hereby incorporated by reference. After content is assembled in an XML database, it is exported into a form which is compatible with the STATdx™PathIQ™ system. Content is also assembled in STATdx™PathIQ™ by use of a relational or table-based authoring system for entering diagnosis and reagent relations which have been manually parsed from peer reviewed medical references. This table is stored in a Postgres database that is compatible with STATdx™PathIQ™. Images and other content in STATdx™PathIQ™, which are discussed in more detail below, are served up from a web server and accessed from a web browser.
Upon accessing the network-based component, a user is presented with a query page or module [106] which allows the user to enter the suspected diagnosis or diagnoses [200]. In the preferred embodiment, this page is shown as
Once the suspected diagnoses are entered into the network-based system, the meta-analysis is initiated by the user, and the system, in response, generates a list of potential reagents and their properties, as reflected by various numerical measures, which would, through further testing or reaction, potentially differentiate among the diagnoses previously entered [300]. In the preferred embodiment, the results of this meta-analysis are displayed on a new page in the sequence [114], which is shown in
In the preferred embodiment, each of the Panels displays, for each antibody listed, various numerical measures of antibody behavior including a percentage [117] representing the correlation between an entered diagnosis and a positive immunostaining result, the total number of reported cases [118] from which the percentage is derived, as well as images [119] of positive and/or negative immunostain results. The images may be enhanced by the selection thereof initiated by a user, resulting in a pop-up window including an enlargement of the image along with its thumbnail image [120], as shown in
In the preferred embodiment, once the network-based system has generated a list of possible and/or recommended antibodies for immunostaining, and any adjustments to the list have been made by the user, the user may click “Save Panel for Analysis” [124] as shown in
Once a user has obtained a list of reagents from the network-based system, the user then orders testing or reaction of the reagents using some or all of the recommended reagents [500]. The testing may be conducted by any number of means, although, in the preferred embodiment, immunostaining using antibodies is conducted in the traditional laboratory environment [600]. After testing is completed, the user then obtains the results from the testing for each of the reagents [700], and visually examines them, which ordinarily, but not necessarily, entails microscopy examination of slides of tissue samples.
After the user has examined the results of the testing, the user again accesses the network-based system. In the preferred embodiment, the user, as discussed above, logs into the system, and selects the relevant saved list or Panel under “Open Cases” [125] as shown in
It is also possible that the input by the user into the system initially fails to result in a diagnosis. At this stage, the user has two basic options. First, the user may initiate further meta-analysis in the system without regard to the initial diagnoses hypothesis [200], but rather, with respect to the reagents previously obtained from the system and tested [1000]. In other words, the user may enter the reagents previously tested into the system, whereby the system, using meta-analysis, generates a list of possible discrete diagnoses from which a user may select for differential diagnosis purposes. The user accomplishes this by comparing the actual testing results of the reagents previously selected and tested with the various permutations of hypothetical test results and their corresponding diagnoses in the generated list [1100]. This allows a user to significantly narrow possible diagnoses, and possibly attain a definitive diagnosis, without having to repeat steps [200]-[1000]. However, because of the possibility that the tested reagents are not sufficient to conclusively differentiate between diagnoses, it may result that a user only narrows possible diagnoses through this step and must obtain a conclusive diagnosis by repeating steps [200]-[1000] using the newly generated list of diagnoses [1200].
In the preferred embodiment, further meta-analysis of the reagents tested [1000] [1100] is accomplished by using the “Build Ab Panel” feature [130] shown in
Second, if despite following all of the above steps the user does not obtain a diagnosis, the user may then return and repeat previous steps [200]-[1000], but making various changes in the input to, or variables of the network-based system or re-examining and repeating steps outside the system in an attempt to obtain a definite diagnosis or diagnoses [1100]. Such changes may include inputting different suggested diagnoses to the system, further customizing the generated list of reagents in the system, conducting further reagent testing outside the system using different reagents, or re-examining the results of the reagent testing outside the system and re-comparing these results with the digital images and/or numerical measures inside the system.
Finally, several other features and tools of the system may assist the user in making diagnoses during any of the above-mentioned steps, and thereby constitute additional ancillary steps in the method. For example, the preferred embodiment incorporates several other features and tools to assist the user in making diagnoses during any of the above-mentioned steps. These features include an “Analyze Results” tab [138] which allows a user to compare up to five antibodies with a suspected diagnosis group as shown in