Nanochip designs for high-precision guidance of subcellular deformations
Irregularities in nuclear shape are common characteristic features of cancer cells. However, assessment of such irregularities still mainly relies on the visual inspection of individual pathologists in clinical practice and is suffered from low accuracy and reproducibility. Thus, an automatic and high-throughput system for quantitative and objective evaluation of nuclear shape irregularities is highly needed.
This method automatically isolates individual nuclei, recognizes the subnuclear morphological features on nanostructures, quantitatively characterizes the subnuclear shape irregularities, and performs Machine Learning based classification for cell grading. It affords malignancy classification in different cancer types, assessing anti-cancer drug effects as well as probing heterogeneity in cancer cell populations.
