February 17 2023
Reading time 7 minutes
Laboratory diagnostics ;generates huge amounts of data and systematizing all of this information requires gigantic amount of work. To simplify routine processes, laboratories actively use computer and information technologies, namely the capabilities provided by artificial intelligence (AI)1, enabling broad prospects in data processing, analysis and research1.
Laboratory information ;systems (LIS) are programs used for optimizing laboratory diagnostics and automating information acquisition2. They are developed to simplify scientific, information, business operations in medical laboratories2. This is necessary for efficiently operating medical institutions that are capable of accurate and timely management decisions and analysis.
Information processing used to take significant amount of time before LIS were introduced2. Manual entry of patient data and tracking of analyzes are labor-intensive processes. The situation was further complicated and delayed by the fact that after a medical analysis, the staff had to systematize patient data, issue invoices and provide various other reports2. Furthermore, various errors could occur during these steps. ;
The introduction of information technology ;has noticeably sped up laboratory ;processes2. Data entry and automatic quality control have become better and more simplified. All results are checked and archived automatically, payment information is transmitted electronically and reporting is streamlined using templates2. Also, artificial intelligence has become part of laboratory ;testing. All this has a positive effect on the quality of the results. ;
Modern medical laboratories generally use two types of software1:
Programs designed to manage the entire informationflow circulating in the laboratory. Necessary for automation of operational and managerial actions.
Programs based on artificial intelligence which participate in the laboratory diagnostic processes.
Let’s take a closer look at how they work:
The main purpose of this type of LIS is to automate the work of employees, increase their efficiency, reduce the number of errors and manual operations, and help manage many other aspects of medical laboratory operation. ;For example, when a patient is referred for blood sampling and analysis, information technology controls all stages of their visit. This data set contains3:
date of admission;
information about the attending physician;
patient information, including demographic data;
type of sample taken;
doctor’s orders for the analysis;
the date the test result was sent to the doctor.
The software unites the entire laboratory under its control, creating and maintaining a centralized database for all departments and routine processes2. This includes1,2:
Automatic processing and transfer of referral forms.
Automatic transfer of tests, tracking results for each individual patient.
Optimization of service classification and billing.
Collection and storage of secure data (age, gender, ethnicity, blood type, etc.).
Interaction with third parties, including doctors and insurance companies.
Accounting, including the possibility of audit.
Tracking orders received through websites, applications, telecom.
Inventory management of reagents and consumables.
In general, these laboratory computer systems can be classified as service management systems.
Another branch of laboratory software is based on artificial intelligence that optimizes research activities. It controls the processing of biomaterial, analyzes images obtained from a digital microscope, interprets the data, and issues preliminary diagnosis1.
AI learns to do this through machine learning algorithms and analysis of scientific data: scientific publications, diagnostic ;methods, image libraries, clinical recommendations1. All information received by the AI when analyzing a biomaterial is compared with its data set and the accumulated “experience” of the whole healthcare system. The AI’s final diagnosis is checked and approved by a doctor.
Algorithms are time-tested. For example, IRIS analyzers have been used for more than 20 years for automated urine microscopy1. They identify objects by analyzing photomicrographs with neural networks trained on a library of photographs of urine elements4. Another example is the use of AI-based image analysis programs to detect pathological sites in histological slides5.
Today, AI technologies are implemented to identify hidden, previously unknown patterns within a comprehensive patient examination, such as new biomarkers and suspected pathologies. In one study, AI categorized subjects into three prostate cancer risk categories based on an analysis of biomarker levels, ultrasound or tomography, age, and other relevant characteristics6. The AI allowed identifying risk groups with clinically significant or clinically insignificant cancer, separate people with benign adenoma from healthy ones.
Digital FAP from SberMedAI is one of the modern AI applications in Russian laboratory diagnostics7. It is a medical computer with software that allows automating and optimizing the interaction between the doctor and the patient with the support of AI algorithms.
The basic set includes equipment for assessing the condition of the main organs and systems of the patient: an electrocardiograph, a tonometer, a pulse oximeter, a spirometer, blood and urine analyzers.
Digital FAP provides tools for primary diagnostics: recording ECG, checking sugar and cholesterol levels, and deciphering the results of a general urine test. The workstation allows the user to automatically interpret data received from medical instruments. If a value deviates from the norm, it is indicated on the screen by an arrow pointing up or down.
Digital FAP uses AI in two algorithms: a cardiograph and TOP-3. Here’s how it works:
During electrocardiography, the results are decoded and processed by an AI, after which the cardiographer makes an unconfirmed automatic diagnosis, which the doctor comments on.
A computer program receives data from laboratoryand instrumental diagnostics and combines them with the manually entered results of the initial appointment and the patient’s comments.
While making the final diagnosis, the TOP-3 algorithm offers the doctor three options for the most probable diagnoses in accordance with the ICD-10 codes. The final decision remains with the doctor.
The whole process takes just a few minutes. The program creates a patient’s medical record, which stores his data, diagnosis and appointments with detailed information on the procedures and their results.
Digital FAP can work both autonomously and in tandem with the information system of a medical institution in order to verify the diagnosis. The patient’s data is transmitted using secure communication channels.
While today’s information technology has become very effective, it is still no more than a tool. Artificial intelligence can’t do without human knowledge, skills and control. As the experience of stationary laboratories has shown, one of the main points of quality control remains the taking of biomaterial samples8. For example, some preanalytical factors, such as the used anticoagulant or the volume ratio of anticoagulant to biomaterial, can have a pronounced effect on laboratory results8.
For an examination using the Digital FAP, data collection is performed by a paramedic or a doctor who receives and examines the patient, based on the recommendations of the program and using the attached tools. Although the AI issues a preliminary diagnosis, the final decision always remains with the specialist.
On the other hand, management information technologies are perfect for simplifying routines and complementing the laboratory computer ecosystem. They help streamlining data flow and speeding up the diagnosis process, which is vital for the success of the laboratory that can take in hundreds of patients a day.
Evgina SA, Gusev AV, Shamanskiy MB, Godkov MA. Artificial intelligence on the doorstep of the laboratory. Laboratory Service. 2022;11(2):18‑26. (In Russ.). https://doi.org/10.17116/labs20221102118
Lukić V. Laboratory Information System – Where are we Today? J Med Biochem. 2017 Jul 14;36(3):220-224. doi: 10.1515/jomb-2017-0021. PMID: 30564059; PMCID: PMC6287214. doi: 1515/jomb-2017-0021
Volkova Irina Aleksandrovna, Talan A.E., Buchneva E.A., and Shcherbo S.N. Counting the formed urine elements with the automatic analyzer Iris IQ 200 tm. Clinical laboratory diagnostics, vol. 59, no. 11, 2014, pp. 37-39.
Mukhopadhyay S, Feldman M, Abels E et al. Whole Slide Imaging Versus Microscopy for Primary Diagnosis in Surgical Pathology. American Journal of Surgical Pathology. 2018;42(1):39-52. https://doi.org/10.1097/pas.0000000000000948
Gruson D, Bernardini S, Dabla P, Gouget B, Stankovic S. Collaborative AI and Laboratory Medicine integration in precision cardiovascular medicine. Clinica Chimica Acta. 2020;509:67-71. https://doi.org/10.1016/j.cca.2020.06.001
Human DNA is an extremely important dataset which determines structure and functions of cells. This huge amount...
February 17, 2023
A surgical operation is a complex task that requires utmost concentration and precise movement. Technology supports the...
February 17, 2023