MF2C3: Multi-Feature Fuzzy Clustering to Enhance Cell Colony Detection in Automated Clonogenic Assay Evaluation

摘要: A clonogenic assay is a biological technique for calculating the Surviving Fraction (SF) that quantifies anti-proliferative effect of treatments on cell cultures: this evaluation often performed via manual counting colony-forming units. Unfortunately, procedure error-prone and strongly affected by operator dependence. Besides, conventional assessment does not deal with colony size, which generally correlated delivered radiation dose or administered cytotoxic agent. Relying upon direct proportional relationship between Area Covered Colony (ACC) count along growth rate, we propose MF2C3, novel computational method leveraging spatial Fuzzy C-Means clustering multiple local features (i.e., entropy standard deviation extracted from input color images acquired general-purpose flat-bed scanner) ACC-based SF quantification, considering only covering percentage. To evaluate accuracy proposed fully automatic approach, compared SFs obtained MF2C3 against four different lines. The achieved results revealed high correlation ground-truth measurements based counting, outperforming our previously validated using thresholding L*u*v* well images. In conclusion, multi-feature inherently leverages concept symmetry in pixel distributions, might be reliably used studies.