PD-1/PD-L1 pathological picture recognition method and device based on deep learning

Abstract

The invention relates to the technical field of image recognition and medical treatment, and provides a PD-1 / PD-L1 pathological picture recognition method based on deep learning, which comprises the following steps: S1, constructing a deep residual network model; s2, acquiring a PD-1 / PD-L1 immunohistochemical staining picture which is subjected to artificial marking; s3, constructing a PD-1 / PD-L1 pathological staining picture recognition model on the deep residual network model by using an immunohistochemical staining picture; and S4, identifying the PD-1 / PD-L1 pathological picture of the patient to be detected by using the identification model. The invention further provides a corresponding PD-1 / PD-L1 pathological picture recognition device based on deep learning, computer equipment and a computer readable storage medium. The recognition model can replace manual judgment, and whether positive expression exists in the PD-1 / PD-L1 in the body of a patient or not is correctly, rapidly and stably judged.

Description

technical field [0001] The present invention relates to the fields of image recognition technology and medical technology, in particular to a method and device for recognizing PD-1 / PD-L1 pathological pictures based on deep learning. Background technique [0002] Tumor is the number one disease in the world. Tumor immunotherapy is a treatment method that restores the body's normal anti-tumor immune response by restarting and maintaining the body's tumor immune cycle, thereby controlling and eliminating tumors. Immune cells (such as T cells, B cells, and myeloid cells) express programmed death 1 (PD-1) against tumor cells, and tumor cells produce this protein in order to resist the attack of immune cells The corresponding ligand is called programmed death ligand-1 (programmed death ligand 1, PD-L1). The combination of PD-L1 and PD-1 will generate a molecular signal, which will reduce the activity of immune cells, thereby blocking the attack of immune cells on tumor cells, and...

AI Technical Summary

Problems solved by technology

[0004] In view of the progress of deep learning in image recognition in recent years, the purpose of the present invention is to provide a PD-1 / PD-L1 pathological picture recognition method and device based on deep learning, which solves the subjectivity and personal spirit of human recognition and interpretation. Problems that affect conditions such that results may be erroneous or unstable

Images