Interactive web-based platform for automated microscopy analysis of cytotoxic effects in tumour co-culture models

Tools

Fatharani, Annisa and Alsayegh, Ali (2025) Interactive web-based platform for automated microscopy analysis of cytotoxic effects in tumour co-culture models. In: 11th Asian Conference on Tumor Ablation, 2025-10-25 - 2025-10-27, The Grand Hotel Taipei.

[thumbnail of Fatharani-Alsayegh-ACTA-2025-Interactive-Web-Based-Platform-for-Automated-Microscopy-Analysis]
Preview
 Text. Filename: Fatharani-Alsayegh-ACTA-2025-Interactive-Web-Based-Platform-for-Automated-Microscopy-Analysis.pdf
Final Published Version
License: Creative Commons Attribution 4.0 logo
Download (5MB)| Preview

Abstract

Background: Assessing tumour cell morphology in response to cytotoxic treatment is a key part of preclinical cancer research, especially when studying molecular interactions in co-culture models. This process, however, still relies on manual inspection methods that are time-consuming, subjective, and difficult to reproduce across experiments. Although commercial systems like the Operetta exist, they remain inaccessible for many small labs due to cost and technical requirements. We developed an interactive, open-source analysis platform that enables automated, quantitative analysis using microscopy images.   Materials and Methods: The platform was built in Python and deployed via a browser-based Streamlit interface. Users upload microscopy images in common formats (PNG, JPG, TIFF, PNG), adjust Gaussian filter parameters, and select one of seven automated thresholding algorithms (Isodata, Li, Otsu, etc.) plus manual selection. Morphological features (area and eccentricity) are measured for each object detected. Watershed algorithm is included to help separate overlapping cells. Results exported to Excel format with individual methods sheets, colour-coded labelled images, and summary statistics. It is capable of handling batch image data up to approximately thousands of images. Detailed workflow can be seen in Figure 1.   Results: We tested the platform on confocal images of tumour cell lines (A549, HT1080, PANC1), grown alone and in co-culture with neutrophils. In monocultures, the platform accurately detected individual nuclei and identified treatment-related changes, such as reduced nuclear size. In co-culture settings, segmentation became more challenging due to overlap, but the tool still provided informative trends (cell area and object count) that aligned with expected cytotoxic responses.   Conclusions: This lightweight platform offers a practical solution for tumour cell image analysis in labs without access to commercial systems. It supports semi-automated quantification of treatment effects, such as apoptosis-related changes, and is well-suited for ex vivo studies involving co-cultures. Future versions will include machine learning tools to improve accuracy.

ORCID iDs

Fatharani, Annisa and Alsayegh, Ali ORCID logoORCID: https://orcid.org/0000-0001-7083-3639;
CORE (COnnecting REpositories)