TITLE:
A Brief Discussion on the Theory and Application of Artificial Intelligence in Medical Diagnostics
AUTHORS:
Jiuxin Cui, Kehong Liu, Xuejun Liang
KEYWORDS:
Artificial Intelligence, Medical Laboratory Testing
JOURNAL NAME:
Journal of Signal and Information Processing,
Vol.17 No.1,
December
18,
2025
ABSTRACT: In this new era of rapid technological advancement, Artificial Intelligence (AI) has achieved comprehensive, multi-level development. AI is simultaneously advancing across algorithms, computing power, and data, propelling medical testing from “digitalization” to “digital intelligence”. Technological Level: Deep learning, federated learning, and multimodal fusion have reshaped the entire workflow—from sample preprocessing and feature extraction to quality control and result interpretation—forming a self-iterating knowledge-decision loop. Automated testing models, continuously trained on datasets exceeding one million annotated samples, simultaneously enhance sensitivity, specificity, and reporting timeliness. Application Dimension: AI has been embedded into core scenarios such as cytomorphology, mass spectrometry analysis, gene sequencing, and immunochromatography, enabling real-time “human-machine collaboration” decision-making. Leveraging edge computing, it extends tertiary hospital diagnostic capabilities to county-level facilities. Innovation Dimensions: Small-sample transfer learning, explainable algorithms, and dual-loop feedback mechanisms between laboratory and clinical data are addressing three major bottlenecks: data heterogeneity, scarce annotations, and causal traceability. This paper systematically outlines the technical foundations and practical applications of AI in medical laboratory testing, explores its innovative developments and challenges, and concludes with a vision for new pathways where AI can empower medical laboratory science.