Prediction bias in biological ageing markers

@misc{lin2026prediction,
  title = {{Prediction bias in biological ageing markers}},
  author = {Yiqun Lin and Ariel Yuhan Ong and Matthew Yu Heng Wong and Yilan Wu and Wenyi Hu and Shubhank Shobhit Verma and Broder Poschkamp and Lie Ju and Akshay Narayan and Sophie A. Martin and Maitrei Kohli and Fares Antaki and Alexander C. Heatley and Josef Huemer and Meng Wang and Ke Zou and Zheyuan Wang and Eden Ruffell and Dominic Williamson and Justin Engelmann and Hyunmin Kim and Takahiro Ninomiya and Rahul A. Jonas and Yansong Liu and Zijie Cheng and Hanyuan Zhang and Hejie Cui and Jiayang Xu and Hualiang Wang and Zuozhu Liu and Bin Pu and Chubo Liu and Kenli Li and Xinpeng Ding and Guokai Zhang and Weiru Wang and Jiaxiang Wang and Huazhu Fu and Zongyuan Ge and Xujia Liu and Mohammad Eslami and Milen Raytchev and Tobias Elze and Michael G. Morley and Neil P. Oxtoby and Daniel C. Alexander and Anthony P. Khawaja and Ching-Yu Cheng and Lisa Zhuoting Zhu and Yih Chung Tham and James H. Cole and Carol Y. Cheung and Pearse A. Keane and Siegfried K. Wagner and Yukun Zhou},
  howpublished = {arXiv},
  year = {2026},
  month = sep,
  doi = {10.48550/arxiv.2609.30322},
  eprint = {2609.30322},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2609.30322},
  note = {CC BY-NC-ND 4.0},
}

TY  - UNPB
TI  - Prediction bias in biological ageing markers
AU  - Yiqun Lin
AU  - Ariel Yuhan Ong
AU  - Matthew Yu Heng Wong
AU  - Yilan Wu
AU  - Wenyi Hu
AU  - Shubhank Shobhit Verma
AU  - Broder Poschkamp
AU  - Lie Ju
AU  - Akshay Narayan
AU  - Sophie A. Martin
AU  - Maitrei Kohli
AU  - Fares Antaki
AU  - Alexander C. Heatley
AU  - Josef Huemer
AU  - Meng Wang
AU  - Ke Zou
AU  - Zheyuan Wang
AU  - Eden Ruffell
AU  - Dominic Williamson
AU  - Justin Engelmann
AU  - Hyunmin Kim
AU  - Takahiro Ninomiya
AU  - Rahul A. Jonas
AU  - Yansong Liu
AU  - Zijie Cheng
AU  - Hanyuan Zhang
AU  - Hejie Cui
AU  - Jiayang Xu
AU  - Hualiang Wang
AU  - Zuozhu Liu
AU  - Bin Pu
AU  - Chubo Liu
AU  - Kenli Li
AU  - Xinpeng Ding
AU  - Guokai Zhang
AU  - Weiru Wang
AU  - Jiaxiang Wang
AU  - Huazhu Fu
AU  - Zongyuan Ge
AU  - Xujia Liu
AU  - Mohammad Eslami
AU  - Milen Raytchev
AU  - Tobias Elze
AU  - Michael G. Morley
AU  - Neil P. Oxtoby
AU  - Daniel C. Alexander
AU  - Anthony P. Khawaja
AU  - Ching-Yu Cheng
AU  - Lisa Zhuoting Zhu
AU  - Yih Chung Tham
AU  - James H. Cole
AU  - Carol Y. Cheung
AU  - Pearse A. Keane
AU  - Siegfried K. Wagner
AU  - Yukun Zhou
PY  - 2026
DA  - 2026/09/29
PB  - arXiv
DO  - 10.48550/arxiv.2609.30322
UR  - https://arxiv.org/abs/2609.30322
AB  - Biological ageing markers have attracted growing interest, with models estimating age from organ imaging or blood biomarkers. An estimated age above chronological age or the age-specific population expectation is assumed to reflect accelerated ageing and poorer health. Previous research has supported this assumption through positive associations between disease and age gaps or acceleration. However, in this study, we identified a widespread health-dependent prediction bias in ageing markers that affects their key interpretation and application. Specifically, we investigated five ageing markers derived from retinal images, brain MRI, chest radiographs, abdominal CT, and blood tests, and evaluated them using association analyses. We observed the well-recognised phenomenon of regression to the mean (RTM) in the four organ-image based markers, whereby estimated ages were shifted towards the mean age of the training cohort. More importantly, we revealed that the strength of RTM varied with health status, with stronger RTM in unhealthy than in healthy individuals. This differential RTM introduced a health-dependent prediction bias that persisted after calibration and systematically altered associations across age subgroups, suggesting that whole cohort associations may not reflect those observed within individual age subgroups. Additionally, we showed that the tested ageing markers, including PhenoAge derived from blood biomarkers, had limited ability to distinguish health status at the individual level. These findings call for careful interpretation of biological ageing markers and their use in clinical studies, and highlight the need for further development and validation before these ageing markers can reliably inform individual health assessments.
ER  -