MICROSERVICE ARCHITECTURE OF THE IMMUNOHISTOCHEMICAL IMAGE PROCESSING SYSTEM

Abstract

The process of diagnosing cancer is complex and time-consuming. In addition, this process requires the organization of interaction between several teams of doctors - diagnosticians. There are also several approaches to making a diagnosis, so the development of a decision support system that allows several participants to interact is an urgent task. Currently, diagnosis based on immunohistochemical and histological images using predetermined diagnostic rules is widely used. A key role in this process is played by a diagnostician who has the necessary knowledge, skills and competence. The development of an automated system for determining the quantitative characteristics of immunohistochemical images, such as the relative area of cell nuclei and the intensity of staining of these cells, will allow speeding up the image stack processing process. The development of a user-friendly graphical interface for doctors will make the decision support system more convenient and intuitive. A large number of constraints must be taken into account when developing computer systems for image-based disease diagnosis, including workstation capacities, shared secure access mechanisms, etc. The use of microservice architecture has been actively implemented recently, especially in systems that require significant computing and memory resources. The principle of microservice architecture consists in dividing programs or systems into a set of small services that are connected to each other using network technologies.

In this work, the authors developed and presented a microservice architecture of a web system for viewing studies based on immunohistochemical images, consisting of separate modules for continuous and reliable work within the laboratory. The key feature of this development is the presence of a modern graphical interface for doctors without the need for special skills to manage studies and the presence of a module for automatic processing of immunohistochemical images.