Scholarlink

<p>Nasopharyngeal cancer (NPC) exhibits distinct histopathological characteristics when compared to other malignancies of the head and neck. Collecting information regarding the prognosis of tumors in patients is incredibly costly and challenging, particularly for NPC patients. Our study aims to develop a prognostic strategy that integrates machine learning and deep learning techniques specifically to predict the NPC prognosis. The study uses fuzzy c-means clustering (FCM) on MR images of the NPC region to automatically label and group NPCs, categorizing them into low- and high-risk groups based on specific structures within the NPC region. A deep learning model was introduced that integrates a convolutional neural network (CNN) to extract important features from MRI data and long-short-term memory (LSTM) to recognize patterns and connections across MR slices. The proposed CNN–LSTM framework eliminates a significant deficit of existing NPC prognostic models that fail to recognize 3D tumor dynamics by simultaneously learning tumor-specific spatial features and within-slice temporal progression patterns. To improve the model’s adaptability and reduce the calculated loss, stochastic gradient descent (SGD) and adaptive moment estimation (Adam) were employed. The experimental results show that the proposed system, using SGD, has outperformed the previous state-of-the-art, achieving an F1 score of 0.97 and a Matthews Correlation Coefficient (MCC) of 0.93. The proposed methodology is effective in predicting the prognosis of nasopharyngeal carcinoma, which can help radiation oncologists create more personalized therapeutic strategies for patients with NPC.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Automated labeling and prognostic prediction of nasopharyngeal carcinoma based on fuzzy c-mean and integrated CNN-LSTM model

  • 期刊论文

  • 发表日期：

    2026年6月25日

查看全文
  • Hassan Ali Khan,
  • Gong XueQing,
  • Muhammad Ahtsam Naeem,
  • Zeeshan Bin Siddique

摘要

Nasopharyngeal cancer (NPC) exhibits distinct histopathological characteristics when compared to other malignancies of the head and neck. Collecting information regarding the prognosis of tumors in patients is incredibly costly and challenging, particularly for NPC patients. Our study aims to develop a prognostic strategy that integrates machine learning and deep learning techniques specifically to predict the NPC prognosis. The study uses fuzzy c-means clustering (FCM) on MR images of the NPC region to automatically label and group NPCs, categorizing them into low- and high-risk groups based on specific structures within the NPC region. A deep learning model was introduced that integrates a convolutional neural network (CNN) to extract important features from MRI data and long-short-term memory (LSTM) to recognize patterns and connections across MR slices. The proposed CNN–LSTM framework eliminates a significant deficit of existing NPC prognostic models that fail to recognize 3D tumor dynamics by simultaneously learning tumor-specific spatial features and within-slice temporal progression patterns. To improve the model’s adaptability and reduce the calculated loss, stochastic gradient descent (SGD) and adaptive moment estimation (Adam) were employed. The experimental results show that the proposed system, using SGD, has outperformed the previous state-of-the-art, achieving an F1 score of 0.97 and a Matthews Correlation Coefficient (MCC) of 0.93. The proposed methodology is effective in predicting the prognosis of nasopharyngeal carcinoma, which can help radiation oncologists create more personalized therapeutic strategies for patients with NPC.