Scholarlink

Small Cell Lung Cancer (SCLC) is remarkably aggressive with a significant mortality rate. Detecting the disease early and accurately diagnosing SCLC is very critical and also challenging due to its very swift progression. Thus, there is an urgent need for innovative approaches to improve early detection and treatment outcomes in cases of SCLC. Radiomics offers advancements in diagnosis by extracting detailed features from medical images, giving full insights into tumor heterogeneity by enhancing the ability for personalized treatment planning. Hence, we endeavor to develop a vigorous deep learning model for segmenting SCLC from CT and PET images, facilitating the extraction of radiomics features and enhancing diagnostic and prognostic capabilities. Therefore, in this paper, we compared the performance of four deep learning models such as Mask-RCNN, U-Net, YOLOv8, and nnUNetv2 on a small dataset for SCLC segmentation using PET and CT scans. Experimental analysis showed that the nnUNetv2 outperformed the other models, achieving the highest accuracy of 99.14% and Dice coefficient of 75.81%. This architecture may become pivotal for effective radiomics feature extraction. With the advancement of these methodologies, we aim to substantially improve clinical outcomes and provide a valuable and crucial tool for the early detection and treatment management of SCLC patients.

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

Segmentation of Small Cell Lung Cancer for Radiomics Feature Extraction Using Deep Learning Models

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Zidny Talukder,
  • Munimah Mahreen,
  • Ishrat Jahan,
  • Nasima Begum,
  • Nakiba Nuren Rahman

摘要

Small Cell Lung Cancer (SCLC) is remarkably aggressive with a significant mortality rate. Detecting the disease early and accurately diagnosing SCLC is very critical and also challenging due to its very swift progression. Thus, there is an urgent need for innovative approaches to improve early detection and treatment outcomes in cases of SCLC. Radiomics offers advancements in diagnosis by extracting detailed features from medical images, giving full insights into tumor heterogeneity by enhancing the ability for personalized treatment planning. Hence, we endeavor to develop a vigorous deep learning model for segmenting SCLC from CT and PET images, facilitating the extraction of radiomics features and enhancing diagnostic and prognostic capabilities. Therefore, in this paper, we compared the performance of four deep learning models such as Mask-RCNN, U-Net, YOLOv8, and nnUNetv2 on a small dataset for SCLC segmentation using PET and CT scans. Experimental analysis showed that the nnUNetv2 outperformed the other models, achieving the highest accuracy of 99.14% and Dice coefficient of 75.81%. This architecture may become pivotal for effective radiomics feature extraction. With the advancement of these methodologies, we aim to substantially improve clinical outcomes and provide a valuable and crucial tool for the early detection and treatment management of SCLC patients.