How this helps patients and the public
Our first peer-reviewed paper presents a new AI model designed to help detect lung cancer earlier. The model analyses CT scans, highlights small lung nodules that may be cancerous, and provides additional information to support clinical decisions. Early and accurate detection is one of the strongest predictors of survival, and tools like this aim to support clinicians by spotting changes that are easy to miss. The research shows promising results and demonstrates how AI can assist doctors in giving patients faster, more reliable answers.
What this means for clinicians
This publication introduces a multi-task deep learning framework capable of both nodule detection and diagnostic characterisation using lung CT imaging. The model is designed to streamline radiology workflows by reducing false negatives, improving consistency, and offering clinically interpretable outputs such as visual heat-maps of regions of interest. While not a replacement for radiologists, the system has the potential to act as a second-reader support tool, strengthening diagnostic confidence and reducing reporting delays.
Why this matters for partners and investors
Our first published paper represents a major scientific validation milestone for Astronomical AI. The research demonstrates the feasibility of an integrated, multi-task AI model that accurately identifies lung nodules and provides diagnostic cues directly from CT imaging. This supports our wider roadmap of building end-to-end AI tools for lung cancer diagnosis and treatment planning. Peer-reviewed publication strengthens our credibility with clinical partners, accelerates regulatory progress, and positions us competitively within a rapidly expanding market for AI-driven oncology solutions.
2026 – LungDetectNet: a multi-task deep learning framework with enhanced detection and descriptive capabilities, published on ScienceDirect.
Read the paper: https://www.sciencedirect.com/science/article/pii/S1746809425016696