Duan Lian, Marvdashti Tahereh, Lee Alex, Tang Jean Y, Ellerbee Audrey K
E.L. Ginzton Laboratory and Department of Electrical Engineering Stanford University, Stanford, CA 94305, USA.
Department of Dermatology Stanford University, Stanford, CA 94305, USA.
Biomed Opt Express. 2014 Sep 22;5(10):3717-29. doi: 10.1364/BOE.5.003717. eCollection 2014 Oct 1.
We report an automated classifier to detect the presence of basal cell carcinoma in images of mouse skin tissue samples acquired by polarization-sensitive optical coherence tomography (PS-OCT). The sensitivity and specificity of the classifier based on combined information of the scattering intensity and birefringence properties of the samples are significantly higher than when intensity or birefringence information are used alone. The combined information offers a sensitivity of 94.4% and specificity of 92.5%, compared to 78.2% and 82.2% for intensity-only information and 85.5% and 87.9% for birefringence-only information. These results demonstrate that analysis of the combination of complementary optical information obtained by PS-OCT has great potential for accurate skin cancer diagnosis.
我们报告了一种自动分类器，用于在通过偏振敏感光学相干断层扫描（PS-OCT）获取的小鼠皮肤组织样本图像中检测基底细胞癌的存在。基于样本散射强度和双折射特性的组合信息的分类器的灵敏度和特异性显著高于单独使用强度或双折射信息时。组合信息的灵敏度为94.4%，特异性为92.5%，而仅强度信息的灵敏度和特异性分别为78.2%和82.2%，仅双折射信息的灵敏度和特异性分别为85.5%和87.9%。这些结果表明，对PS-OCT获得的互补光学信息进行组合分析在准确诊断皮肤癌方面具有巨大潜力。
