Automated Micronucleus Test Evaluation ～Provided as an automatic analysis service～

Challenge

The micronucleus test is one of the standard assays used to evaluate chromosomal damage–induced toxicity (genotoxicity). It is primarily conducted to assess the risk of cancer and genetic disorders by measuring the frequency of micronuclei—small, extranuclear bodies that form when chromosomal fragments fail to be incorporated into daughter nuclei during cell division—in treated cell populations. Traditionally, this test has relied on manual microscopic observation, where researchers visually evaluate each cell. This approach imposes a heavy workload and can lead to inter-observer variability in results.

LPIXEL’s Solution – AI Technology and Impact

LPIXEL’s Automated Micronucleus Test Assessment Technology uses a deep-learning AI model trained on the evaluation criteria of skilled experimenters to automatically analyze specimen images and generate assessment reports through the two-step process outlined below.

Automatic identification of individual cells

Determine the presence or absence of micronuclei in the detected cells, and quantification of the incidence rate.

Supported tests: May-Giemsa stain, acridine orange stain (other tests available upon consultation)
This service has been employed by pharmaceutical companies, CROs, and manufacturers in the food and cosmetics industries.

＜Achievements / publications＞
Validation of automated observation techniques for in vitro micronucleus assay slides using deep learning
— Joint presentation with Takeda Pharmaceutical Company Limited
（Workshop of the Japanese Environmental Mutagen and Genome Society, 2021）

Examination of automated micronucleus observation using AI-based image analysis — Joint presentation with Shiseido Co., Ltd. (The 49th Annual Meeting of the Japanese Society of Toxicology, 2022）
Investigation of automated analysis of micronucleus tests using a three-dimensional human epidermis model with AI-based analysis
— Joint presentation with Shiseido Co., Ltd. (The 52nd Annual Meeting of the Japanese Society of Toxicology, 2025)