Lifecycle tissue engineering: An AI-enabled multimodal sensing and digital twin roadmap
- 1School of Nursing, Zhejiang Chinese Medical University, Hangzhou, Zhejiang, China
- 2Department of Mechanical Engineering, University of Manitoba, Winnipeg, R3T 2N2, Canada
- 3College of Basic Medical Sciences, Zhejiang Chinese Medical University, Hangzhou, Zhejiang, China
Abstract
The clinical translation of tissue engineering (TE) is critically hampered by the absence of real-time, non-invasive monitoring throughout the entire lifecycle of tissue fabrication and repair. Current reliance on end-point assessments obscures dynamic biological events, making early detection of deviations and failures challenging. While multimodal sensing offers rich data streams, inherent issues such as long-term drift, noise accumulation, and difficulties in cross-modality comparability impede their transformation into reliable process metrics and actionable insights. This review introduces a novel, comprehensive four-layer framework consisting of Problem, Data, Model, and Application Evidence, to systematically integrate artificial intelligence (AI)-enabled multimodal sensing across the full TE lifecycle. The Problem Layer delineates critical biological uncertainties and monitoring requirements. The Data Layer surveys diverse in vitro and in vivo observables, sensing modalities, and readout strategies. The Model Layer bridges data acquisition with decision-making, emphasizing AI's role in signal processing, feature extraction, data fusion, and predictive analytics for early failure identification and outcome prediction. The Application Evidence Layer synthesizes current evidence, showcasing AI and sensing applications in donor or cell screening, in vitro process optimization, post-implantation quantitative readout, longitudinal assessment, and diagnostic prediction. This framework moves beyond fragmented approaches by proposing a unified architecture for lifecycle TE, addressing a significant gap in existing literature. We further delineate a strategic roadmap for clinical translation, underscoring the imperative for standardization, long-term reliability, unified evaluation protocols, and responsible AI development. By distinguishing demonstrated technologies from integration steps that remain conceptual or require TE-specific validation, this review outlines a structured roadmap for moving TE monitoring from isolated readouts toward more predictive, traceable, and verifiable translational workflows.
Related
- From digital reminders to artificial intelligence–enabled understanding: can personalized treatment predictions improve therapeutic adherence?
- AI-enabled and comparator digital interventions for adolescent mental health (2020–2025): early detection, human–AI collaboration, and ethical governance—a PRISMA-ScR scoping review
- Metabolic engineering of Salmonella enterica for coupled kynurenine sensing and depletion enhances antitumor efficacy
- Isolating the exposomic impact of acute war and displacement within chronic crisis using multimodal, AI-enabled phenotyping
- Toward comprehensive multimodal health monitoring in heat: A digital biomarker exploration study
BibTeX
@article{liu2026lifecycle,
title = {{Lifecycle tissue engineering: An AI-enabled multimodal sensing and digital twin roadmap}},
author = {Xing Liu and Yitao Zhou and Cuiyi Wang and Lewei Tu and Xiaoxuan Zhao and Jingjing Gu and Qiuhua Sun and Malcolm Xing and Yeqin Yang},
journal = {Bioactive Materials},
year = {2026},
month = oct,
doi = {10.1016/j.bioactmat.2026.06.044},
url = {https://doi.org/10.1016/j.bioactmat.2026.06.044},
note = {CC BY-NC-ND 4.0},
}
RIS
TY - JOUR
TI - Lifecycle tissue engineering: An AI-enabled multimodal sensing and digital twin roadmap
AU - Xing Liu
AU - Yitao Zhou
AU - Cuiyi Wang
AU - Lewei Tu
AU - Xiaoxuan Zhao
AU - Jingjing Gu
AU - Qiuhua Sun
AU - Malcolm Xing
AU - Yeqin Yang
PY - 2026
DA - 2026/10/27
JO - Bioactive Materials
DO - 10.1016/j.bioactmat.2026.06.044
UR - https://doi.org/10.1016/j.bioactmat.2026.06.044
AB - The clinical translation of tissue engineering (TE) is critically hampered by the absence of real-time, non-invasive monitoring throughout the entire lifecycle of tissue fabrication and repair. Current reliance on end-point assessments obscures dynamic biological events, making early detection of deviations and failures challenging. While multimodal sensing offers rich data streams, inherent issues such as long-term drift, noise accumulation, and difficulties in cross-modality comparability impede their transformation into reliable process metrics and actionable insights. This review introduces a novel, comprehensive four-layer framework consisting of Problem, Data, Model, and Application Evidence, to systematically integrate artificial intelligence (AI)-enabled multimodal sensing across the full TE lifecycle. The Problem Layer delineates critical biological uncertainties and monitoring requirements. The Data Layer surveys diverse in vitro and in vivo observables, sensing modalities, and readout strategies. The Model Layer bridges data acquisition with decision-making, emphasizing AI's role in signal processing, feature extraction, data fusion, and predictive analytics for early failure identification and outcome prediction. The Application Evidence Layer synthesizes current evidence, showcasing AI and sensing applications in donor or cell screening, in vitro process optimization, post-implantation quantitative readout, longitudinal assessment, and diagnostic prediction. This framework moves beyond fragmented approaches by proposing a unified architecture for lifecycle TE, addressing a significant gap in existing literature. We further delineate a strategic roadmap for clinical translation, underscoring the imperative for standardization, long-term reliability, unified evaluation protocols, and responsible AI development. By distinguishing demonstrated technologies from integration steps that remain conceptual or require TE-specific validation, this review outlines a structured roadmap for moving TE monitoring from isolated readouts toward more predictive, traceable, and verifiable translational workflows.
SN - 2452-199X
ER -