文献速递皮肤肿瘤

基于小样本临床图像开发的深度学习计算机辅助分类器在皮肤肿瘤诊断中超越认证皮肤科医生

Deep-learning-based, computer-aided classifier developed with a small dataset of clinical images surpasses board-certified dermatologists in skin tumour diagnosis.
PubMed 文献 · Br J Dermatol · 2018-09-19

研究利用2003至2016年筑波大学医院1842例患者的4867张临床图像训练深度卷积神经网络(DCNN),涵盖14种恶性与良性诊断,并与13名认证皮肤科医生和9名皮肤科进修医生对比。DCNN整体分类准确率76.5%,恶性/良性分类准确率92.4%±2.1%,高于认证皮肤科医生的85.3%±3.7%和进修医生的74.4%±6.8%(P<0.001)。

要点

  • DCNN整体分类准确率76.5%,恶性/良性分类准确率92.4%±2.1%。
  • 认证皮肤科医生恶性/良性分类准确率85.3%±3.7%,进修医生74.4%±6.8%。
  • DCNN恶性/良性分类准确率高于认证皮肤科医生(P<0.001)。

核心结论

  • 使用相对小样本数据集训练的DCNN分类皮肤肿瘤图像比认证皮肤科医生更准确。
  • 当前系统可能具有在全科医疗实践中用于筛查的能力,尤其因为分类仅需单张临床图像。

研究设计与样本

  • 使用2003至2016年筑波大学医院1842例皮肤肿瘤患者的4867张临床图像训练深度卷积神经网络(DCNN),图像涵盖14种诊断,包括恶性和良性病变。
  • 以13名认证皮肤科医生和9名皮肤科进修医生为对照测试DCNN性能。

主要发现

  • 训练后DCNN整体分类准确率为76.5%。
  • DCNN恶性/良性分类敏感度为96.3%(将恶性正确分类为恶性),特异度为89.5%(将良性正确分类为良性)。
  • 认证皮肤科医生恶性/良性分类准确率为85.3%±3.7%,皮肤科进修医生为74.4%±6.8%,前者统计学上高于后者(P<0.01)。
  • DCNN恶性/良性分类准确率达92.4%±2.1%,高于认证皮肤科医生(P<0.001)。

临床怎么用

  • 作者认为当前系统可能具有在全科医疗实践中用于筛查的能力,尤其因为分类仅需单张临床图像。

局限性

  • 原文未报告局限性

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

原文

BACKGROUND: Application of deep-learning technology to skin cancer classification can potentially improve the sensitivity and specificity of skin cancer screening, but the number of training images required for such a system is thought to be extremely large. OBJECTIVES: To determine whether deep-learning technology could be used to develop an efficient skin cancer classification system with a relatively small dataset of clinical images. METHODS: A deep convolutional neural network (DCNN) was trained using a dataset of 4867 clinical images obtained from 1842 patients diagnosed with skin tumours at the University of Tsukuba Hospital from 2003 to 2016. The images consisted of 14 diagnoses, including both malignant and benign conditions. Its performance was tested against 13 board-certified dermatologists and nine dermatology trainees. RESULTS: The overall classification accuracy of the trained DCNN was 76·5%. The DCNN achieved 96·3% sensitivity (correctly classified malignant as malignant) and 89·5% specificity (correctly classified benign as benign). Although the accuracy of malignant or benign classification by the board-certified dermatologists was statistically higher than that of the dermatology trainees (85·3% ± 3·7% and 74·4% ± 6·8%, P < 0·01), the DCNN achieved even greater accuracy, as high as 92·4% ± 2·1% (P < 0·001). CONCLUSIONS: We have developed an efficient skin tumour classifier using a DCNN trained on a relatively small dataset. The DCNN classified images of skin tumours more accurately than board-certified dermatologists. Collectively, the current system may have capabilities for screening purposes in general medical practice, particularly because it requires only a single clinical image for classification.

来源:Br J Dermatol,查看原文。版权归原作者所有。

引用

  1. Fujisawa Y, Otomo Y, Ogata Y, et al. Deep-learning-based, computer-aided classifier developed with a small dataset of clinical images surpasses board-certified dermatologists in skin tumour diagnosis. Br J Dermatol. 2018. PMID: 29953582. doi:10.1111/bjd.16924

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。