Metraipolis › archive › dc_5e5a2d1de445

How Deep is the Signal? Time-Context up to 30 Years for Colorectal Cancer Prediction in a Primary-Care EHR Model — arXiv:2611.07265

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.07265
License (as archived)
CC0-1.0
Archived
2026-11-06T14:42:17.077Z
SHA-256 of the archived text
bbf4f704f5c60abd43b1fb96c5142c94abf16cafda68cf60dc2e3b9318182913
Stable URL of this page
https://metraipolis.com/documento/dc_5e5a2d1de445

How to cite

How Deep is the Signal? Time-Context up to 30 Years for Colorectal Cancer Prediction in a Primary-Care EHR Model — arXiv:2611.07265 — arXiv (en), https://arxiv.org/abs/2611.07265 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_5e5a2d1de445 (Metraipolis archive dc_5e5a2d1de445, sha256 bbf4f704f5c60abd43b1fb96c5142c94abf16cafda68cf60dc2e3b9318182913)
BibTeX
@misc{dc5e5a2d1de445,
  title = {How Deep is the Signal? Time-Context up to 30 Years for Colorectal Cancer Prediction in a Primary-Care EHR Model — arXiv:2611.07265},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.07265},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_5e5a2d1de445; sha256 bbf4f704f5c60abd43b1fb96c5142c94abf16cafda68cf60dc2e3b9318182913},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.07265. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

EHR foundation models read medical events within a fixed context window, and long context has so far been measured only in tokens. How far back in time a model must look to capture the earliest predictive signal is unknown. We introduce time-context, the calendar lookback window before the prediction point, and provide, to our knowledge, the first systematic evaluation of decade-scale time-context for an EHR foundation model. We pretrained a decoder-only transformer (34.6M parameters) on 1.4B coded events from 3.6M primary-care patients in North West London and fine-tuned it to predict incident colorectal cancer under a 12-month pre-diagnostic buffer. A frozen cohort of 1,008 cases and 9,582 matched controls, all registered for at least 30 years, was evaluated at 15 nested lookback windows from 2 to 30 years, so that every truncation is compared within the same patients. Discrimination rose from an AUROC of 0.551 at 2 years to 0.590 at 18 years, then plateaued, with about 80% of the gain reached within 10 years; an XGBoost baseline converged to the same level (peak AUROC 0.589). A simulation further suggests that rotary position embeddings (RoPE) over elapsed time alias beyond about 20 years at the default base frequency, motivating a retuned $\theta$. Time-context is a measurable design axis for clinical prediction models, distinct from token-context, and should be reported.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI); Quantitative Methods (q-bio.QM).
Pagina del lavoro: https://arxiv.org/abs/2611.07265 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-11-06T14:18:42.671Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).