Non-Invasive Skin Cancer Diagnosis Using NIR Multi-Spectral Image Analysis

9724991康这项建议的研究是建立一种安全、非侵入性和快速的计算机辅助诊断技术，利用近红外(NIR)多波长信息(从450-800 nm之间的8个波长，间隔50 nm)检测恶性皮肤肿瘤。研究的目标是：(1)早期发现恶性皮肤肿瘤，如果没有活检，医生无法确诊；(2)从不典型葡萄胎综合征或发育不良痣综合征患者的多个可疑葡萄胎(数百个)中识别恶性肿瘤。已经完成了一项初步研究，使用了四种不同血红蛋白浓度的皮肤肿瘤模型的实验性组织模型。实验结果表明，近红外图像的多光谱分析可以区分四种实验模式。规划拨款旨在将这项研究进行到可以提交定期提案的程度。***

9724991 Kang This proposal is on research to establish a safe, non- invasive, and rapid computer aided diagnostic technique for the detection of malignant skin tumors utilizing near infrared (NIR) multiple wavelength information (8 wavelengths from 450-800 nm at a 50 nm interval). The research objectives are: (1) early detection of malignant skin tumors, which can not be confirmed by the physician without biopsy, and (2) malignancy identification from multiple suspicious moles (hundreds) from patients with atypical mole syndrome or with dysplatic nevus syndrome. A preliminary study has been completed using experimental tissue models with skin tumor models at four different hemoglobin concentrations. The experimental results suggest that multi-spectral analysis of NIR images can differentiate the four experimental models. The planning grant seeks to carry this research to a point where a regular proposal can be submitted. ***