Computer-aided demarcation of early gastric cancer: a pilot comparative study with endoscopists
Computer-aided demarcation of early gastric cancer: a pilot comparative study with endoscopists
复制标题
DOI：
10.1007/s00535-023-02001-x
复制
发表时间：
2023-05
影响因子：
6.3
通讯作者：
S. Takemoto;K. Hori;Sakai Yoshimasa;M. Nishimura;K. Nakajo;Atsushi Inaba;Maasa Sasabe;Naoki Aoyama;Takashi Watanabe;Nobuhisa Minakata;H. Ikematsu;H. Yokota;T. Yano
中科院分区：
文献类型：
--
作者：
S. Takemoto;K. Hori;Sakai Yoshimasa;M. Nishimura;K. Nakajo;Atsushi Inaba;Maasa Sasabe;Naoki Aoyama;Takashi Watanabe;Nobuhisa Minakata;H. Ikematsu;H. Yokota;T. Yano
背景早期胃癌（EGC）的准确区域诊断是内镜下可靠切除的关键.计算机辅助诊断（CAD）显示出检测EGC和减少由内镜医师技能差异引起的癌症护理差异的强大潜力。为了在临床实践中使用，CAD应该能够检测和划分病变。本研究提出了一种方案，用于检测和划定EGC下白光内窥镜检查和验证其性能使用1年连续cases.MethodsOnly 300内窥镜图像随机选择68个连续的情况下，用于训练卷积神经网络。所有病例均采用内镜粘膜下剥离术进行治疗，从而能够积累精确确定病变程度的训练数据集。为了验证，使用了来自137个连续病例的462个癌症图像和396个正常图像。从验证结果中，38个随机选择的图像进行了比较，与那些划定由6个endoscients.Results387癌症图像（83.8%）和病变的情况下，307正常图像（77.5%）的EGC成功检测。阳性和阴性预测值分别为81.3%和80.4%。成功检出130例（94.9%）。我们实现了精确的划分EGC与平均交叉超过工会的66.5%，显示病变的范围与平滑的边界;结果是可比的，达到了由specialists.ConclusionsOur计划，验证使用1年连续的情况下，显示潜在的划分EGC。它的性能与专家相匹配;因此，它可能适合未来的临床使用。
BackgroundPrecise area diagnosis of early gastric cancer (EGC) is critical for reliable endoscopic resection. Computer-aided diagnosis (CAD) shows strong potential for detecting EGC and reducing cancer-care disparities caused by differences in endoscopists’ skills. To be used in clinical practice, CAD should enable both the detection and the demarcation of lesions. This study proposes a scheme for the detection and delineation of EGC under white-light endoscopy and validates its performance using 1-year consecutive cases.MethodsOnly 300 endoscopic images randomly selected from 68 consecutive cases were used for training a convolutional neural network. All cases were treated with endoscopic submucosal dissection, enabling the accumulation of a training dataset in which the extent of lesions was precisely determined. For validation, 462 cancer images and 396 normal images from 137 consecutive cases were used. From the validation results, 38 randomly selected images were compared with those delineated by six endoscopists.ResultsSuccessful detections of EGC in 387 cancer images (83.8%) and the absence of lesions in 307 normal images (77.5%) were achieved. Positive and negative predictive values were 81.3% and 80.4%, respectively. Successful detection was achieved in 130 cases (94.9%). We achieved precise demarcation of EGC with a mean intersection over union of 66.5%, showing the extent of lesions with a smooth boundary; the results were comparable to those achieved by specialists.ConclusionsOur scheme, validated using 1-year consecutive cases, shows potential for demarcating EGC. Its performance matched that of specialists; it might therefore be suitable for clinical use in the future.