Journal Title: تابستان 1403
Article title: Evaluation of Response Rate to Chemotherapy in Patients with Breast Cancer using DWI and ADC Sequences of MRI Images by Machine Learning Modeling

Article PDF File:
Page From: 25 To: 37      
Article abstract:
Introduction & Objective: Currently, chemotherapy is considered as one of the standard and important treatments for patients with breast cancer. Locally advanced BCs must receive Neo - adjuvant chemotherapy before surgery. With the advancement of MRI techniques such as DWI sequence, the effects of chemotherapy such as cell apoptosis and necrosis can be investigated. However, ADC images more accurately show the diffusion and movement of water molecules as well as, distinguish between cancerous tissue and normal tissue, thereby investigating the response to treatment. However, the limitations of these techniques have restricted roles and minimal use. With the advent of artificial intelligence and machine learning, besides clinical studies, it is possible to identify more reliable and accurate methods by examining and modeling with the help of imaging data. Therefore, this study aimed to investigate different indices extracted from DWI sequences and draw ADC maps with the help of different algorithms using machine learning to determine the response rate to chemotherapy in patients with breast cancer Materials & Methods: In this study, imaging data related to breast cancer patients were first extracted from the ACRIN website of the American Imaging Society and with the help of different algorithms including decision tree, K-nearest neighbors, logistic regression, support vector machines, linear support vector machines, and random forest. Different indices of DWI images and drawing ADC maps calculated a reliable index to estimate the response to treatment in different stages of chemotherapy (including T0, T1, T2, T3). Results: In this study, the information of 8 patients was collected from the ACRIN site. Then, different indices of DWI images were analyzed with the help of machine learning. The results showed that in the b-value equal to 1000, and with the help of apparent diffusion coefficient (D), ADC maps were drawn and it was shown that an increase in the diffusion coefficient D can indicate a positive response to chemotherapy. The data of 3 patients who responded completely to the treatment were analyzed with different algorithms and indicators and it was shown that the increase of D in ADC images can be seen in all stages of chemotherapy, even before the apparent changes in the images are identified, so it can determine the response rate to treatment. Conclusions: The results of this study showed that the use of D value in ADC images can reliably and with high accuracy determine the response rate to chemotherapy.
Article KeyWords:
Breast Cancer, Machine Learning, Apparent Diffusion Coefficient, Diffusion of Water Molecules, Response to Treatment
Article Authors:

	مهندس زهرا پازوکی	First Author	helenapazoki2@gmail.com
	زهرا طهماسبی	Other Author