Scholarlink

<p>In epithelial ovarian cancer (EOC), accurate assessment of the lesions is crucial for determining the appropriate treatment strategy. We evaluated artificial intelligence (AI) for detecting peritoneal and small bowel dissemination in EOC using preoperative contrast-enhanced computed tomography (CT). We retrospectively analyzed 227 patients (254 CT scans) with histologically confirmed ovarian, fallopian tube, or peritoneal carcinoma who underwent surgery between October 2019 and December 2024 at two affiliated institutions. The mean patient age was 56.6&#xa0;years. The FIGO stage was early in 118 cases and advanced in 136 cases. Histology included serous (n = 123) and non-serous (n = 131). The mean CA125 level and peritoneal cancer index (PCI) score were 2,243.9 U/mL and 5.8, respectively. Two AI models were developed: one for peritoneal dissemination (P-Model) and one for small bowel dissemination (SB-Model). Diagnostic performance was evaluated by sensitivity, specificity, and accuracy. For the P-Model, clinical factors were compared between high- and low-accuracy groups. The P-Model achieved 68.75% sensitivity, 85.84% specificity, and 77.30% accuracy; the SB-Model, 86.40%, 77.47%, and 81.93%, respectively. Low AI accuracy was linked to ascites volume, CA125 level, and PCI score. AI-based analysis of preoperative CT demonstrated favorable diagnostic performance for detecting dissemination in patients with EOC.</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

Diagnostic performance of artificial intelligence for detecting peritoneal and small bowel dissemination in epithelial ovarian cancer using preoperative contrast-enhanced CT imaging

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年3月23日

查看全文
  • Ria Kim,
  • Toshiyuki Seki,
  • Katsuhiko Noda,
  • Kaname Yoshida,
  • Kota Yokosu,
  • Rintaro Hamada,
  • Erika Habuchi,
  • Masahiro Takahashi,
  • Hirokuni Takano,
  • Aikou Okamoto

摘要

In epithelial ovarian cancer (EOC), accurate assessment of the lesions is crucial for determining the appropriate treatment strategy. We evaluated artificial intelligence (AI) for detecting peritoneal and small bowel dissemination in EOC using preoperative contrast-enhanced computed tomography (CT). We retrospectively analyzed 227 patients (254 CT scans) with histologically confirmed ovarian, fallopian tube, or peritoneal carcinoma who underwent surgery between October 2019 and December 2024 at two affiliated institutions. The mean patient age was 56.6 years. The FIGO stage was early in 118 cases and advanced in 136 cases. Histology included serous (n = 123) and non-serous (n = 131). The mean CA125 level and peritoneal cancer index (PCI) score were 2,243.9 U/mL and 5.8, respectively. Two AI models were developed: one for peritoneal dissemination (P-Model) and one for small bowel dissemination (SB-Model). Diagnostic performance was evaluated by sensitivity, specificity, and accuracy. For the P-Model, clinical factors were compared between high- and low-accuracy groups. The P-Model achieved 68.75% sensitivity, 85.84% specificity, and 77.30% accuracy; the SB-Model, 86.40%, 77.47%, and 81.93%, respectively. Low AI accuracy was linked to ascites volume, CA125 level, and PCI score. AI-based analysis of preoperative CT demonstrated favorable diagnostic performance for detecting dissemination in patients with EOC.