Chronic lymphocytic leukemia tumor cell recognition method based on machine learning

Abstract

The invention discloses a chronic lymphocytic leukemia tumor cell recognition method based on machine learning. The method comprises the following steps: (1) detecting an antigen related to chronic lymphocytic leukemia, and calculating a confidence interval boundary of fluorescence intensity of the antigen; (2) constructing a neural network model, wherein model input is a confidence interval boundary, and the model output is expected output for judging whether chronic lymphocytic leukemia exists or not; (3) training the neural network model by utilizing the collected cell fluorescence intensity data of the chronic lymphocytic leukemia patient and the non-chronic lymphocytic leukemia patient; and (4) collecting cell fluorescence intensity data of a new patient needing to be diagnosed, and giving a reference recognition result by utilizing the neural network model. The clinician can be assisted in accurately identifying whether the tumor cells belong to chronic lymphocytic leukemia tumorcells or not by means of machine learning and a large amount of historical data, and thus the efficiency and the quality of clinical diagnosis are improved.

Description

technical field [0001] The invention relates to a tumor cell-assisted identification method, in particular to a machine learning-based identification method for chronic lymphocytic leukemia tumor cells, which belongs to the technical field of medical testing. Background technique [0002] Chronic lymphocytic leukemia (CLL) is a hematopoietic malignancy that occurs in middle-aged and elderly people, and its main feature is that malignant tumors with specific immunophenotypic characteristics can be detected in peripheral blood or bone marrow Mature B lymphocytes. In clinical medicine, CD19 can be detected by flow cytometry + Multiple antigens on tumor cells, by artificially analyzing the expression characteristics of these antigens to determine whether they are CLL tumor cells. [0003] The typical immunophenotype of CLL tumor cells is: CD19 + , CD5 + , CD23 + 、CD200 + , CD43 + , CD10 - , FMC7 - ; surface immunoglobulin light chain (kappa / lambda), weak expression of C...

AI Technical Summary

Problems solved by technology

[0004] Since CLL is a chronic lymphoproliferative disorder, in addition to CLL, this type of disease also includes mantle cell lymphoma, hairy cell leukemia, marginal zone lymphoma, follicular cell lymphoma, immature cell lymphoma and lymphoplasmacytic lymphoma / giant cell lymphoma. Gammaglobulinemia, etc., the immunophenotype of tumor cells in such diseases has many intersecting expression features, which makes it difficult to identify abnormalities, and some CLL itself also exhibits atypical immunophenotypic features

Therefore, to determine whether such cells are CLL tumor cells, even a professional physician who has been trained for many years has more than 10% identification errors or identification uncertainties

Images