Kasra Arbabi - Student Research Committee, School of Nursing and Midwifery, Iranshahr University of Medical Sciences, Iranshahr, Iran
Mohsen Marzban - Student Research Committee, School of Nursing and Midwifery, Iranshahr University of Medical Sciences, Iranshahr, Iran
Amin Yarmohammadi - Student Research Committee, School of Nursing and Midwifery, Iranshahr University of Medical Sciences, Iranshahr, Iran
Arefeh Nikpendar - Student Research Committee, School of Nursing and Midwifery, Iranshahr University of Medical Sciences, Iranshahr, Iran
Zahra Alizadeh - Student Research Committee, School of Nursing and Midwifery, Iranshahr University of Medical Sciences, Iranshahr, Iran
Alireza Gholamnezhad - Student Research Committee, School of Nursing and Midwifery, Iranshahr University of Medical Sciences, Iranshahr, Iran

Background & aims: Advances in technology have been able to affect all aspects of human life,especially in the field of medical sciences. Advances in medical imaging techniques, artificial intelligence,machine learning, and computer vision offer new opportunities for building intelligentdecision support tools. Due to the expensive and invasive nature of cancer diagnosis, the need forcost-effective and non-invasive methods has increased. This study is an overview of the diagnosisand prediction of one of the most common and deadly diseases, brain tumor, using machinelearning and deep learning methods.Method: We obtained this title by searching the scientific database of PubMed using the keywordsbrain tumor, MRI, GAN, central nervous system cancer, CT scan, deep learning, machinelearning and artificial intelligence. References and related articles were cited.Results: The use of artificial intelligence reduces the percentage of errors compared to humandiagnosis. Also, compared to machine learning, deep learning provides better performance fordiagnosis and subdivision.Conclusion: Cancer grading is an important aspect of targeted therapy. In this article, we examinedthe application of artificial intelligence in early tumor diagnosis, classification of prognosis,metastasis, prediction, challenges and potential of these techniques. Given that magnetic resonanceimaging (MRI) is the most common method for diagnosing brain tumors. In this article, wehave made efforts to apply different types of deep learning methods on MRI data and identifiedthe challenges in the field in search of potential future paths. One of the branches of deep learningthat has been successful in image processing is CNN. In this review, we have also worked ondifferent architectures of CNN.

MRI, brain tumor, Diagnosis, artificial intelligence, Deep learning, Machine learning