Study On Cancer Cell Recognition Method Based On Atomic Force Nanomanipulation

Malignant tumor is a disease with high incidence rate and mortality worldwide.Every day,tens of thousands of families suffer from cancer.If we can find tumors in time at the early stage of cancer and treat them,we can greatly improve the survival rate of cancer patients.In the early stages of cancer patients,it is difficult to screen out all tumors,leading to missed diagnosis and worsening of the condition.In order to avoid this and improve the survival rate of cancer patients,it is of great significance to study the recognition method of cancer cells and the correct judgment of cancer grade.Currently,in the micro world,nanomanipulation based on an atomic force microscope(atomic force nanomanipulation)is one of the important methods for researchers to understand and study the micro world.In this dissertation,we first used a home-made atomic force microscope to image and test the tumor cells(Human liver cancer cells: SMMC-7721,Hep G2 and HCC-LM3,human gastric cancer cells: SGC-7901,human skin squamous cell carcinoma cells: A431,human lung cancer cells: A549 and H1299,human esophageal cancer cells: KYSE150,KYSE30,TE-1and esophageal 70,human breast cancer cells: MCF-7,human lympholeukemic cells: Hal-01,colorectal cancer cells: SW480,human cervical cancer cells: Hela,human neuroblastoma cells: SH-SY5 Y,human multiple myeloma cells: U266).The morphology and mechanical property data of the cells were extracted and characterized.The extracted data were analyzed to prepare for the modeling of cancer cells and image recognition of cancer cells.Secondly,the obtained cancer cell images were preprocessed and segmented using the home-made atomic force microscope,and different cancer cell images were studied and trained using the improved convolutional neural network model.Normal gastric cells(GES-1)and gastric cancer cells(SGC-7901)were selected for the experiment,and then processed with the phellinus igniarius culture solution.The home-made atomic force microscope was used for imaging,The morphology and mechanical properties of the two cell surfaces were characterized.The experimental results were used to validate the improved convolutional neural network model,and a convolutional neural network model with a higher recognition rate and more suitable for cell images was obtained.The obtained convolutional neural network model identified any type of cancer cells in the experiment.Finally,the finite element analysis software ABAQUS was used to define cell properties.Data processing was performed on the morphology and mechanical parameters of all cancer cells obtained from the experiment.The viscoelastic material properties were assigned to the cells,and the cancer cells were modeled.The resulting cancer cell model can more intuitively present the morphology and mechanical parameter characteristics of various cancer cells in the microscopic world.The various parameter ranges of cancer cells and the differences in parameters between different cancer cells can be more intuitively presented.