CodeCytos: AI-assisted spatial molecular imaging analysis via code-augmented agent action space

2026-05-29

AI Summary

This work addresses the limitations of conventional tissue image analysis tools, which are often manual and lack programmability, thereby hindering customized exploration of complex spatial features. To overcome this, the authors propose CodeCytos—the first agent-based framework that integrates a code-augmented action space into spatial molecular imaging analysis. Centered on a large language model and coupled with image segmentation and spatial feature extraction modules, CodeCytos leverages cross-domain, few-shot code examples for in-context learning, enabling it to interpret natural language instructions and dynamically generate tailored analytical pipelines without requiring expert annotations. Experiments across four tissue-type datasets demonstrate that CodeCytos significantly outperforms existing baselines under minimal prompting, substantially enhancing the efficiency of spatial feature exploration and accelerating biomarker discovery.

Abstract

Conventional tissue image analysis software provides foundational capabilities for cellular analysis, including segmentation, basic morphological feature extraction, and spatial organization analysis. However, these tools often require manual intervention and are not well integrated with code-driven automation, limiting efficiency and scalability for complex spatial tissue studies. In addition, they offer limited flexibility for custom analyses, as they typically support only a fixed set of pre-implemented spatial cellular features. To address these limitations, we propose CodeCytos, a coding-based reasoning agent framework that enables dynamic, programmable interaction with spatial molecular imaging data to improve automation and customization. CodeCytos is designed to streamline the exploration of custom spatial cellular features and adapt to diverse research needs. We demonstrate its utility through case studies on four expert-curated datasets from distinct tissue types: frontal cortex, non-small-cell lung cancer, pancreas, and tonsil. We evaluate CodeCytos under a realistic minimal prompt setting, where bioscientists pose simple questions without task-specific instructions or contextual information about spatial cellular analysis, and benchmark multiple LLM backbones with strong coding capabilities. We further show that incorporating tailored, domain-agnostic few-shot in-context coding-reasoning examples (randomly sampled demonstrations outside the spatial analysis domain) can substantially improve performance without requiring costly, expert-crafted in-domain demonstrations. Overall, CodeCytos outperforms baseline approaches, highlighting the potential of code-action agents to assist with custom feature exploration in spatial molecular imaging and to accelerate biomarker discovery.

Problem

• spatial molecular imaging
• custom spatial features
• automation
• image analysis software
• scalability

Innovation

• code-augmented agent
• spatial molecular imaging
• programmable analysis
• few-shot reasoning
• custom feature exploration