Development of a Machine Learning Model to Detect Abnormal Plasma Cells in Peripheral Blood of Patients Suffering from Plasma Cell Proliferative Disorders

Rohit K, Shekhar A, Hiranmay M, Amit S and Garima J

Department of Health Research, Ministry of Health and Family Welfare, India

^*Correspondance to: Garima Jain 

 PDF  Full Text Research Article | Open Access

Abstract:

Background: Plasma cell leukemia is an uncommon but fierce plasma cell neoplasm. Plasma cell leukemia is classified into two categories: Primary PCL and secondary PCL. While sPCL is the leukemic transformation of an already identified multiple myeloma, pPCL is a malignant plasma cell proliferation discovered during the leukemic phase. The goal of this work was to create an AI model using Python that could detect PCL in cell slide images. Methodology: The study was carried out using a small PCL dataset, which we also utilized to create mask images with QuPath. We then used deep learning techniques and the Python programming language to build a model. The PyTorch package, ResNet50, and U-Net artificial neural networks were used to build and train the model. Matplotlib was then used to evaluate and visualize the model's output. Result: An AI model has been developed that can detect plasma cells in hematopathology slides. The model performed well in identifying cancerous cells, with an accuracy of 0.98. The model correctly divided the cell images into two separate classes, 'No cancerous cell discovered' and 'cancerous cell detected', using a binary classification algorithm. Conclusion: This study demonstrates that our approach can effectively detect PCL in cell slide images, while also paving the road for future research and practical application. The exciting results emphasize AI's potential to change medical diagnostics and the need for further research in this field.

Keywords:

Plasma Cell Leukemia; Deep learning; U-Net; PyTorch; Accuracy; Python

Cite the Article:

Rohit K, Shekhar A, Hiranmay M, Amit S, Garima J. Development of a Machine Learning Model to Detect Abnormal Plasma Cells in Peripheral Blood of Patients Suffering from Plasma Cell Proliferative Disorders. Clin Oncol. 2024;9:2092..