Advancements and future directions of artificial intelligence in tumor imaging: A comprehensive review of techniques and applications
Jinglei Xue, Jing Yu, Qianqian Mao, Xiaochun Gu
Abstract
Medical images captured using various imaging technologies can reveal the internal structures and functions of the human body. Tumor images depict the location, size, shape, and biological characteristics of tumors, which are crucial for diagnosis, staging, treatment planning, and outcome assessment. These images are obtained through modalities such as radiography, computed tomography, magnetic resonance imaging, ultrasound, positron emission tomography, single-photon emission computed tomography, digital mammography, digital breast tomosynthesis, histological imaging, and molecular imaging. The application of artificial intelligence (AI) in healthcare, particularly in tumor image recognition, is vital for early diagnosis, treatment planning, and prognostic assessment. Techniques such as AI, machine learning (ML), neural networks (NNs), and deep learning (DL) enhance the accuracy and efficiency of tumor recognition. This review introduces the fundamental principles and interrelationships among AI, ML, NNs, and DL, explores their applications and characteristics in tumor imaging, discusses their limitations in clinical settings and future research directions, and provides a comprehensive overview of technological advancements in this field and their potential for future medical applications.