本研究分析四種癌症（乳癌、子宮內膜癌、肺癌、胰腺癌）的人體血漿，利用表面增強拉曼光譜(Surface-Enhanced Raman Spectroscopy, SERS)結合機器學習，目的是要開發高速、精準的血液感測晶片。我們採用的SERS結構含有多層InGaN量子井，根據69例血漿檢體(乳癌10例、子宮內膜癌10例、肺癌29例、胰臟癌10例、健康對照組10例)的分析結果，本技術能有效檢測出這四種癌，準確率可達96%。;This research focuses on analyzing human plasma associated with four cancers (breast, endometria, lung, pancreas), using the surface-enhanced Raman spectroscopy (SERS) integrated with machine learning for prediction and classification. The SERS structure comprises multiple InGaN quantum wells. According to the test results of 69 clinical cases (10 cases of breast cancer, 10 cases of endometrial cancer, 29 cases of lung cancer, 10 cases of pancreatic cancer, and 10 cases of health control), our technique can quickly identify the types of cancer with the average prediction accuracy of 96%.
