Scholarlink

Background <p>Patients with rectal cancer who have a complete clinical response (cCR) to neoadjuvant chemo/radiotherapy (nCRT) may opt for organ preservation and watch and wait (W&amp;W). This consists of an intense surveillance program including serial endoscopies, pelvic magnetic resonance imaging (MRI), carcinoembryonic antigen (CEA), and computed tomography (CT) scans to detect regrowth at an early stage. However, residual lesions or regrowths can be challenging to identify endoscopically, due to mucosal changes such as friability and neovascularization. We developed a novel deep learning model to assist in the detection of residual or regrown rectal cancer lesions during proctosigmoidoscopy.</p> Methods <p>We trained a convolutional neural network (Wide ResNet-101-2) on a dataset of 1795 annotated frames from proctosigmoidoscopy exams of 97 patients treated at a tertiary referral center. Residual or regrowth was defined by histopathological confirmation. The dataset was split into training and testing cohorts using a 90/10% patient-level split.</p> Results <p>Out of 97 patients, 24 (363 frames) had confirmed residual disease or regrowths, while 73 (1432 frames) presented normal rectal mucosa. The model achieved an overall accuracy of 92.8%, with a sensitivity of 80.0%, specificity of 97.3%, positive predictive value (PPV) of 90.9%, negative predictive value (NPV) of 94.0%, and an area under the receiver operating characteristic curve (AUROC) of 0.886.</p> Conclusions <p>To the best of our knowledge, this is the first deep learning model specifically developed for the detection of residual disease or regrowth following cCR in W&amp;W patients during endoscopic examination. This tool has the potential to aid lesion detection, guide clinical decision-making, and increase opportunities for salvage, curative treatment strategies.</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

Use of artificial intelligence in analysis of endoscopic images to detect residual disease or regrowth in rectal patients with complete clinical response to neoadjuvant chemoradiotherapy

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月14日

查看全文
  • M. A. Javed,
  • M. Mascarenhas,
  • F. Mendes,
  • E. Carvalho,
  • R. Rajan,
  • A. Santos,
  • Z. Khan,
  • I. Blake,
  • S. Ahmed

摘要

Background

Patients with rectal cancer who have a complete clinical response (cCR) to neoadjuvant chemo/radiotherapy (nCRT) may opt for organ preservation and watch and wait (W&W). This consists of an intense surveillance program including serial endoscopies, pelvic magnetic resonance imaging (MRI), carcinoembryonic antigen (CEA), and computed tomography (CT) scans to detect regrowth at an early stage. However, residual lesions or regrowths can be challenging to identify endoscopically, due to mucosal changes such as friability and neovascularization. We developed a novel deep learning model to assist in the detection of residual or regrown rectal cancer lesions during proctosigmoidoscopy.

Methods

We trained a convolutional neural network (Wide ResNet-101-2) on a dataset of 1795 annotated frames from proctosigmoidoscopy exams of 97 patients treated at a tertiary referral center. Residual or regrowth was defined by histopathological confirmation. The dataset was split into training and testing cohorts using a 90/10% patient-level split.

Results

Out of 97 patients, 24 (363 frames) had confirmed residual disease or regrowths, while 73 (1432 frames) presented normal rectal mucosa. The model achieved an overall accuracy of 92.8%, with a sensitivity of 80.0%, specificity of 97.3%, positive predictive value (PPV) of 90.9%, negative predictive value (NPV) of 94.0%, and an area under the receiver operating characteristic curve (AUROC) of 0.886.

Conclusions

To the best of our knowledge, this is the first deep learning model specifically developed for the detection of residual disease or regrowth following cCR in W&W patients during endoscopic examination. This tool has the potential to aid lesion detection, guide clinical decision-making, and increase opportunities for salvage, curative treatment strategies.