Scholarlink

Background <p>The application of artificial intelligence in gynecological imaging has gained momentum in recent years. Deep learning which uses convolutional neural networks takes a shorter time to train with average-sized available datasets in contrast to classic or traditional learning methods. Radiomics on the other hand helps in the extraction of data or features from images. This article intends to review the advances made in the diagnosis of pelvic adnexal masses using AI-based models, their drawbacks, and their future implications.</p> Main text <p>Artificial intelligence-based models have shown capabilities comparable to that of radiologists in distinguishing benign from malignant lesions, especially in the case of epithelial ovarian tumors in ultrasound, CT as well as MRI. It has also shown promise in predicting surgical outcomes and prognosis in patients with pelvic masses. It has also proved to be a useful assisting tool for radiologists and clinicians.</p> Conclusion <p>The ultimate aim of AI, however, is the complete integration of patient history, clinical data, and laboratory parameters with radiological and pathological parameters to achieve at a proper diagnosis and plan management guidelines for the clinician.</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

Artificial Intelligence in Imaging of Pelvic Adnexal Masses: A Narrative Review

  • 期刊论文

  • 发表日期：

    2025年5月21日

查看全文
  • Nirupam Konwar Baishya,
  • Kangkana Baishya

摘要

Background

The application of artificial intelligence in gynecological imaging has gained momentum in recent years. Deep learning which uses convolutional neural networks takes a shorter time to train with average-sized available datasets in contrast to classic or traditional learning methods. Radiomics on the other hand helps in the extraction of data or features from images. This article intends to review the advances made in the diagnosis of pelvic adnexal masses using AI-based models, their drawbacks, and their future implications.

Main text

Artificial intelligence-based models have shown capabilities comparable to that of radiologists in distinguishing benign from malignant lesions, especially in the case of epithelial ovarian tumors in ultrasound, CT as well as MRI. It has also shown promise in predicting surgical outcomes and prognosis in patients with pelvic masses. It has also proved to be a useful assisting tool for radiologists and clinicians.

Conclusion

The ultimate aim of AI, however, is the complete integration of patient history, clinical data, and laboratory parameters with radiological and pathological parameters to achieve at a proper diagnosis and plan management guidelines for the clinician.