Automated Blood Cell Counting and Classification using Deep Learning
Medical image processing is becoming one of the most efficient and reliable medical engineering applications due to the major advances in segmentation and computational techniques. Image processing is used to analyze x-ray and blood smear images for early diseases diagnosis such as anemia, kidney tumors, and lung diseases. Clinically, to measure the number of blood cells and evaluate their characteristics, a complete blood count (CBC) is required. In biomedicine, CBC is used to analyze patient’s overall health and detect blood diseases or abnormal immunological reaction by measuring the number and/or morphology of red blood cells (RBCs) and white blood cells (WBCs). Traditional processes for CBC rely on human assistance, where counting and classifying cells are done manually through a microscope. Those techniques are usually inaccurate, time consuming, and may lead to erroneous results as the whole process is subject to human factors. Currently, some hospitals and laboratories start applying sophisticated analyzers which use laser flow cell sensors to detect and count different blood elements and cells automatically. However, those instruments are very expensive and most of laboratories and hospitals in rural places are unable to afford such sophisticated and efficient equipment. Hence, a computer-based solution is required to automate the CBC and reduce its cost.
Recently, microscopic image processing has received more attention and has shown impressive improvements in biomedicine to enable diagnostic tests for diverse healthcare applications. Different approaches have been suggested in literature in order to automate RBCs and WBCs counting. Most of those studies focused on images filtering and pre-processing to enhance cells segmentation and counting. Meanwhile, artificial intelligence (AI) is also witnessing an increasing proliferation in biomedicine and researchers are investigating its use in real world and healthcare applications through machine learning techniques and neural network computing. In fact, AI proved its efficiency in bioclinical medicine, molecular medicine, and medical imaging especially in pattern recognition and image segmentation. In this research, a novel framework that automatically counts and classifies different blood cells, i.e., RBCs and WBCs, in a given microscopic blood smear image using a combination of convolutional neural network (CNN), transfer learning, and mask R-CNN techniques. The objective is to apply image segmentation techniques in order to locate, predict boundaries, classify, and count RBCs and WBCs in blood smear images. Unlike previous studies which relied on traditional image processing techniques and image segmentation, in this work, we advocate the use of mask R-CNN for not only its efficiency in object detection and classification but also for its performance in instance segmentation. As a result, it can be utilized to detect different cells blood smears and differentiate between overlapped cells and objects belonging to the same class. In this context, we propose the use of Resnet-101 as a backbone for feature pyramid network model and Microsoft common objects in context (MS-COCO) pre-trained model to initiate the neural network model weights. In addition, data augmentation and regulation techniques are applied to enhance the model detection and reduce the counting error. The obtained results reveal a highly detection rate of different blood cells. In addition, unlike other state-of-the-art techniques, our proposed method has the ability to identify overlapped and faded cells.
An Optimized Drug Regimen and Chemotherapy Scheduling for Cancer Treatment
Neoplastic diseases and cancers are known as the principle causes of human deaths around the world. In fact, the latest statistics in cancer morality and incidence indicate that cancer burden has reached an alarming level worldwide. It is expected to reach 22 million new cases of people suffering from cancer per year in the next decade. To treat cancer and neoplastic diseases, different procedures such as chemotherapy, radiotherapy, hormone therapy, and immunotherapy can be used and applied according to different factors like tumor type, size, and stage. Chemotherapy is considered as the main therapy applied to fight against those diseases, where oral or intravenous chemotherapy agents are used to control not only the tumor size and growth but also to attack and destroy carcinogen cells. Physicians resort to Randomized Clinical Trials (RCTs) as standardized treatment methods to evaluate the chemotherapy efficiency and toxicity during the treatment period. Despite their impressive results in treating cancer, those standards are costly and time-consuming and require the testing of different chemotherapy plans. In fact, since they are based on empirical evidence and clinical record carried out during the drug development, various cancer types are still incurable. To overcome the limitation of human intervention in protocol selection and treatment scheduling, mathematical models and optimized scheduling are employed to understand the different interactions between the chemotherapy agent, tumor cells, and different cells present in human body.
In this research, we develop a joint drug regimen and chemotherapy scheduling for cancer treatment. We propose an optimized chemotherapy scheduling and drug regimen protocol aiming at minimizing the tumor cell size of cancer patients, drug consumption, and the total therapy period. We simulate the dynamics and interactions between tumor cells, effector-immune cells, circulating lymphocytes, and chemotherapy agents using ordinary differential equations. We then adopt a successive alternation between treatment and relaxation sessions during the therapy. In the treatment session, we inject an optimal amount of chemotherapy drug in order to shrink tumor cells without excessively harming the effector-immune cells while, in the relaxation session, we set a period of time to allow the body to recover. During this phase, tumor and other natural cells have both the possibility to grow again. To determine the duration of each session as well as the amount of drug to be injected in order to minimize the therapy duration, we proceed with a joint optimization approach where the optimal drug control at each treatment session is determined using a linear optimal control solver, while the duration of each session is optimized using swarm intelligence. The meta-heuristic optimization algorithm has exposed considerable efficiency in treating tumors and scheduling chemotherapy sessions compared to other existing approaches. The obtained results show that the optimized system is more effective in killing tumor cells, scheduling the chemotherapy sessions, and optimizing the drug consumption during the same number of cycles.