Method and system for automated generation of conclusion based on mr tomography data of mammary glands with classification according to bi-RADS scale

Abstract

FIELD: medical informatics; radiation diagnostics.SUBSTANCE: methods and systems for automated processing of magnetic resonance imaging (MRI) data and the formation of a standardized conclusion. The proposed method consists of providing the physician with a structured interface for entering descriptors of MR tomograms of the mammary glands, automatic analysis of the entered data, assignment of points according to established rules, calculation of the Bi-RADS category by summing the points and automatic generation of a text report. The proposed system also contains a user interface module, a data processing module and a report generation module.EFFECT: increased standardization, objectivity, and speed of generating conclusions on MRI of the mammary glands by automating the classification process according to the Bi-RADS scale and eliminating the subjective factor.3 cl, 1 dwg

Description

[0001] Technical field

[0002] The invention relates to the field of medical informatics and radiation diagnostics, namely to methods and systems for automated processing of magnetic resonance imaging (MRI) data and the formation of a standardized report for a radiologist.

[0003] Technology Level

[0004] There are known methods for interpreting breast MRI scans, in which a radiologist manually analyzes the images, describes the detected changes in free form, and assigns a Bi-RADS classification based on their experience. The disadvantages of these methods include subjectivity, lack of standardization, time-consuming, and the risk of errors.

[0005] The closest analogue (prototype) is the manual method of description and assessment according to the Bi-RADS scale, performed by a physician without the use of specialized software implementing a formalized scoring system.

[0006] Disclosure of invention

[0007] The objective of the invention is to eliminate the aforementioned shortcomings, namely: increasing the objectivity, standardization, speed and accuracy of forming conclusions on MRI of the mammary glands.

[0008] The technical result is the creation of an automated tool that minimizes the subjective factor, ensures the uniformity of conclusions and reduces the time for their preparation due to the automatic classification of findings according to formalized criteria.

[0009] The specified technical result is achieved using the claimed method for automated generation of a conclusion based on magnetic resonance imaging (MRI) of the mammary glands, which contains the following stages:

[0010] - provide the user with an interface for data entry, containing fields for selecting descriptors related to: gland symmetry, type of glandular tissue structure, condition of milk ducts, condition of nipples, presence and characteristics of implants, condition of muscles, presence and characteristics of lymph nodes, presence, localization and morphological features of formations;

[0011] - accepts data entered from the user that characterize the specified descriptors;

[0012] - based on the entered data, the presence or absence of formations is automatically determined;

[0013] - if no lesions are present, assign a final Bi-RADS category based on the pre-defined first set of rules;

[0014] - if lesions are present, automatically assign quantitative scores to the entered morphological descriptors for each lesion based on a pre-defined second rule base corresponding to the Bi-RADS scale;

[0015] - automatically calculates the final Bi-RADS category for each formation by summing the assigned quantitative points and comparing the resulting sum with a pre-defined scoring table;

[0016] - automatically generates a text report in the format of a structured document, containing a description of the identified changes based on the entered descriptors and an indication of the calculated final Bi-RADS category.

[0017] Brief description of drawings

[0018] Fig. 1 shows an example of the user interface of the system, containing a structured form with fields for entering descriptors of MR tomograms of the mammary gland.

[0019] Implementation of the invention

[0020] An example of the method and system implementation is explained as follows. A radiologist launches the program. A user interface appears on the screen, containing a structured form with fields for entering data from a breast MRI examination.

[0021] When a lesion is detected, the user selects its characteristics from drop-down lists. The program, based on a built-in rule base, assigns a score to each selected descriptor. For example: "Shape: Irregular" - 2 points, "Edge: Unclear" - 1 point, "Contrast Pattern: Heterogeneous" - 1 point.

[0022] After entering all the data, the program sums up the points. The sum is compared with the evaluation table: 0 points - BI-RADS 1; 1-2 points - BI-RADS 2; 3 points - BI-RADS 3; 4-5 points - BI-RADS 4; 6-8 points - BI-RADS 5.

[0023] Based on the entered descriptors and the calculated BI-RADS category, the program automatically generates a text report in Word document format.

[0024] The system for implementing the method is implemented in the form of a software module written in the C# language and includes: a user interface module, a data processing module, and a report generation module.

[0025] Explanatory note to the drawings

[0026] Fig. 1. Example of the user interface of the system for automated generation of a conclusion based on MR tomography data of the mammary glands.

[0027] Drawing designations:

[0028] 1. General data input area

[0029] • Symmetry of the mammary glands

[0030] • Type of glandular tissue structure

[0031] • Condition of nipples and milk ducts.

[0032] 2. Implant Characteristics Area

[0033] • Shape and location of implants

[0034] • Capsule condition

[0035] • Presence of effusion and folds.

[0036] 3. Lymph node assessment area

[0037] • Size and condition of the lymph nodes.

[0038] 4. Field of morphological assessment of formations

[0039] • Form of education

[0040] • Edge character

[0041] • Contrast parameters

[0042] • Vascularization

[0043] • Kinetics of the curve (early and delayed phases).

[0044] 5. Control element

[0045] • "Generate document" button to generate a conclusion.

Claims

1. A method for automated generation of a report on magnetic resonance imaging (MRI) of the mammary glands, comprising the following stages: • provide the user with an interface for data entry, containing fields for selecting descriptors related to: gland symmetry, type of glandular tissue structure, condition of milk ducts, condition of nipples, presence and characteristics of implants, condition of muscles, presence and characteristics of lymph nodes, presence, localization and morphological features of formations; • accept data entered from the user that characterize the specified descriptors; • based on the entered data, the presence or absence of formations is automatically determined; • if no lesions are present, a final Bi-RADS category is assigned based on a pre-defined first set of rules; • if lesions are present, for each lesion, automatically assign quantitative scores to the entered morphological descriptors based on a pre-defined second rule base corresponding to the Bi-RADS scale, where said descriptors include at least the shape of the lesion, the edge of the lesion, the nature of the contrast, vascularization and the kinetics of the contrast curve, and the rule base for assigning scores contains rules according to which: – the form descriptor “incorrect” is assigned 2 points; – the edge descriptor “radiant” is assigned 2 points, and the descriptor “fuzzy” is assigned 1 point; – the contrast character descriptor “ring-shaped” is assigned 2 points, and “heterogeneous” is assigned 1 point; • automatically calculate the final Bi-RADS category for each lesion by summing the assigned quantitative points and comparing the resulting sum with a pre-defined scoring table, whereby the scoring table for calculating the Bi-RADS category establishes a correspondence according to which: – a score of 0 corresponds to Bi-RADS category 1; – a score of 1-2 corresponds to the Bi-RADS 2 category; – a score of 3 corresponds to Bi-RADS category 3; – a score of 4-5 corresponds to the Bi-RADS category 4; – a score of 6-8 corresponds to the Bi-RADS category 5; • automatically generate a text conclusion in the format of a structured document, containing a description of the identified changes based on the entered descriptors and an indication of the calculated final Bi-RADS category.

2. The method according to paragraph 1, characterized in that the stage of generating a text conclusion includes the formation of a document in Word format.

3. A system for automated generation of a conclusion on magnetic resonance imaging (MRI) of the mammary glands for implementing the method according to paragraph 1, comprising: • a user interface module capable of providing fields for entering descriptors of detected changes; • a data processing module configured to: receive input descriptors, assign them quantitative scores based on a rule base corresponding to the Bi-RADS scale, and calculate a final Bi-RADS category, wherein the rule base for assigning scores contains rules according to which: – the form descriptor “incorrect” is assigned 2 points; – the edge descriptor “radiant” is assigned 2 points, and the descriptor “fuzzy” is assigned 1 point; – the contrast character descriptor “ring-shaped” is assigned 2 points, and “heterogeneous” is assigned 1 point; • a report generation module configured to generate a text report in a structured document format based on the entered descriptors and the calculated Bi-RADS category, whereby the assessment table for calculating the Bi-RADS category establishes a correspondence according to which: – a score of 0 corresponds to Bi-RADS category 1; – a score of 1-2 corresponds to the Bi-RADS 2 category; – a score of 3 corresponds to Bi-RADS category 3; – a score of 4-5 corresponds to the Bi-RADS category 4; – a score of 6-8 corresponds to Bi-RADS category 5.

Images