Automated Cell Survival Colony Counting for Clonogenic Assays
Keywords: Computer vision, segmentation, cell counting, clonogenic assay
TL;DR: Computer vision-based automated cell survival colony counting in clonogenic assays for assessing radiosensitivity of cancer cells.
Abstract: Assessment of cancer cell radiosensitivity is essential for understanding the effectiveness
of radiotherapy. The clonogenic assay is the gold standard for quantifying radiosensitiv-
ity by enumerating survived cell colonies in vitro post-radiation exposure. However, the
technique is time-consuming and subject to observer variability. In this study, we present
a computer vision pipeline to automate colony counting, tested on 787 HCT116 assay im-
ages. Our method addresses challenges: (i) lack of annotated data, (ii) variability in data
collection, and (iii) unavailability of microscopic images of individual colonies. Using our
augmented marker-based watershed algorithm, the pipeline achieved an average of 78.5%
accuracy compared to the ground-truth colony count. Ongoing work focuses on improving
the robustness of the pipeline and validating the approach across more cell lines.
Submission Number: 96
Loading