在囊镜检查期间利用实时智能膀瘤检测的深度学习:一种诊断研究
Zixing Ye1, Yingjie Li1, Yujiao Sun1
1Department of Urology, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China.
Annals of surgical oncology
|March 6, 2025
概括
HRNetV2深度学习模型在检测膀病变时显示出高准确度. 它的性能在高分辨率图像方面显著提高,改善了膀瘤的早期诊断和监测.
科学领域:
- 泌尿器科 泌尿器科 泌尿器科 泌尿器科
- 医疗成像医学成像
- 人工智能的人工智能
背景情况:
- 准确的膀病变检测对于早期癌症诊断和监测至关重要.
- 传统的囊镜视觉检查在检测率方面存在局限性.
- 深度学习有可能提高诊断准确度.
研究的目的:
- 评估HRNetV2深度学习模型用于智能膀病变检测.
- 为了评估模型在不同图像分辨率的性能.
- 确定AI在膀病变识别中的临床实用性.
主要方法:
- 利用HRNetV2语义细分模型对94名患者的102个白光囊镜视频进行了分析.
- 在33657个中手动注释疑似膀病变.
- 使用高分辨率和低分辨率图像的灵敏度,精度和平均Dice (mDice) 评分来评估诊断性能.
主要成果:
- 总体测试组的灵敏度:91.6%,精度:91.3%,mDice:80.3%. 总体测试组的灵敏度:91.6%,精度:91.3%,mDice:80.3%. 总体测试组的灵敏度:91.6%,精度:91.3%,mDice:80.3%.
- 高分辨率图像实现了94.8%的灵敏度,94.4%的精度和84.7%的mDice.
- 低分辨率图像显示性能较差:灵敏度为75.6%,精度为74.8%,mDice为56.6%.
结论:
- HRNetV2 显示出出色的膀病变检测能力,特别是在高分辨率图像中.
- 该模型显示了提高膀瘤临床检测准确性的巨大潜力.
- 建议使用更大,多中心数据集进行进一步验证.
相关概念视频
Imaging Studies VI: Voiding Cystourethrography and Cystography
Voiding Cystourethrography (VCUG) and Cystography are specialized radiographic procedures used to examine the structure and function of the bladder and urethra.Voiding Cystourethrography (VCUG)A Voiding Cystourethrogram (VCUG) is a diagnostic imaging procedure that assesses the anatomy and function of the lower urinary tract. It focuses on the bladder, bladder neck, and urethra, helping detect abnormalities such as vesicoureteral reflux (VUR)—the backward or reverse flow of urine into the...
Urologic Endoscopic Procedure: Cystoscopic Examination
Meaning of Cystoscopic Examination:Cystoscopy is an essential diagnostic tool in urology that is used to assess the structure and function of the genitourinary system. It provides a direct view of the urethra, bladder, and, in some cases, the ureteral openings. This procedure helps detect structural abnormalities, infections, cancers, and blockages in the urinary tract. There are two types of cystoscopy:Flexible cystoscopy is commonly performed in outpatient settings due to its less invasive...


