Enter your email address
Submit
Write your message
Iranian Journal of Breast Diseases
Back to the articles list |
Back to browse issues page
Ethics code: IR.LUMS.REC.1402.174
Pooria Afsharifard1 
, Mohammad Bagher Dowlatshahi2 
, Mahshid Abbasi3 , Abbas Rezaeian4 , Morteza Amraei * 5
, Mohammad Bagher Dowlatshahi2 , Mahshid Abbasi3 , Abbas Rezaeian4 , Morteza Amraei * 5
1- Department of Health Information Technology, School of Allied Medical Sciences, Lorestan University of Medical Sciences, Khorramabad, Iran
2- Department of Computer Engineering, Faculty of Engineering, Lorestan University, Khorramabad, Iran
3- Department of Radiation Oncology, School of Medicine, Lorestan University of Medical Sciences, Khoramabad, Iran
4- Department of Radiology Technology, School of Allied Medical Sciences, Lorestan University of Medical Sciences, Khorramabad, Iran
5- Department of Health Information Technology, School of Allied Medical Sciences, Lorestan University of Medical Sciences, Khorramabad, Iran ,morteza.amraei@yahoo.com
2- Department of Computer Engineering, Faculty of Engineering, Lorestan University, Khorramabad, Iran
3- Department of Radiation Oncology, School of Medicine, Lorestan University of Medical Sciences, Khoramabad, Iran
4- Department of Radiology Technology, School of Allied Medical Sciences, Lorestan University of Medical Sciences, Khorramabad, Iran
5- Department of Health Information Technology, School of Allied Medical Sciences, Lorestan University of Medical Sciences, Khorramabad, Iran ,
Abstract: (214 Views)
Introduction: The heart and lungs are among the organs at risk of receiving additional radiation during radiation therapy of breast cancer patients. In recent years, artificial intelligence and machine learning have brought about significant advancements in the field of medicine. This study aimed to predict the radiation dose received by the heart and lungs in breast cancer patients undergoing radiotherapy, taking into account the anatomical characteristics of these organs through the application of machine learning techniques.
Methods: This applied study was conducted by reviewing medical records in 2023 and extracting anatomical features present in chest computed tomography scans of 210 female patients with left breast cancer who had undergone lumpectomy surgery. Patient data were extracted from the Picture Archiving and Communication Syste, and multi-label classification algorithms were employed to predict the radiation dose received by the heart and lungs. The performance of the algorithms was further evaluated using metrics such as accuracy, precision, recall, F1-Score, and Hamming loss.
Results: Based on the performance evaluation results of 7 multi-label classification algorithms and considering 16 anatomical variables influencing the amount of radiation received by the heart and lungs, the Random Forest (RF) algorithm achieved the best performance among other algorithms with an accuracy of 41.9%, precision of 73.3%, recall of 70.6%, F1 score of 73.1%, and Hamming loss of 27.4%.
Conclusion: The use of machine learning algorithms and considering anatomical features make it possible to identify suitable patients for 3D wedge pair radiotherapy. More advanced techniques, such as Intensity-Modulated Radiation Therapy or Deep Inspiration Breath Hold, can be recommended for other patients at risk of receiving high doses of radiation to the heart and lungs.
Methods: This applied study was conducted by reviewing medical records in 2023 and extracting anatomical features present in chest computed tomography scans of 210 female patients with left breast cancer who had undergone lumpectomy surgery. Patient data were extracted from the Picture Archiving and Communication Syste, and multi-label classification algorithms were employed to predict the radiation dose received by the heart and lungs. The performance of the algorithms was further evaluated using metrics such as accuracy, precision, recall, F1-Score, and Hamming loss.
Results: Based on the performance evaluation results of 7 multi-label classification algorithms and considering 16 anatomical variables influencing the amount of radiation received by the heart and lungs, the Random Forest (RF) algorithm achieved the best performance among other algorithms with an accuracy of 41.9%, precision of 73.3%, recall of 70.6%, F1 score of 73.1%, and Hamming loss of 27.4%.
Conclusion: The use of machine learning algorithms and considering anatomical features make it possible to identify suitable patients for 3D wedge pair radiotherapy. More advanced techniques, such as Intensity-Modulated Radiation Therapy or Deep Inspiration Breath Hold, can be recommended for other patients at risk of receiving high doses of radiation to the heart and lungs.
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
