medRxiv · All· Magbanua, M. J. M., Huppert, L. A., Cruz, J., Goudreau, J., Zhou, S., Yin, K., Stejskal, P., Gill, t., Carelli, R., Kim, J., Hsie, L., Wang, M., Lian, Z., Manepalli, S., Phelan, M., Ray, M., Johnson, C., Nieto, J., Schneider, E. C., Mavropoulos, A., Delson, A. L., Simmons, C., Parizadeh, A., Crawford, J., Bruck, M., Her, L., Barnett, V., Cabrera, A. M., Melisko, M., Balassanian, R., Rugo, H. S., van 't Veer, L., Jovic, A., Devine, W. P., Sayaman, R. W., Salek, M., Masaeli, M., Rosenbluth, J., Goodarzi, H. ·· 1 天前AI 评分50
AI 赋能单细胞成像与形态学引导的无标记肿瘤细胞富集，揭示恶性积液异质性、基因组关联性与分化
AI-powered single-cell imaging and morphology-guided label-free enrichment reveal tumor heterogeneity, genomic relatedness, and divergence in malignant effusions
AI 导读
研究团队以预训练深度学习模型分析恶性积液单细胞图像，并用形态学引导无标记微流控富集肿瘤细胞；在10例转移性乳腺癌患者中均识别出候选肿瘤细胞，中位肿瘤细胞比例为5.4%。低深度全基因组测序与全外显子组测序显示，富集细胞与配对积液游离DNA及实体瘤共享部分基因组改变，并揭示患者内及患者间的形态与基因组异质性。
来源：medRxiv · All · medrxiv.org