Driving · 2026-10-07
Federated Bayesian Surveillance of Mechanical Thrombectomy Adverse Events: A Population Risk Layer for Surgical Digital Twins
Damini Rijhwani
AI-assisted summary based on the paper. Consult the linked source for methods and limitations. Benchmark numbers appear after editor verification.
TL;DR
The authors propose a population risk layer for surgical digital twins that models adverse-event rates with per-class Gamma–Poisson posteriors learned via a federated Bayesian protocol under Rényi differential privacy. They benchmark on the complete FDA MAUDE cohort for thrombus-retrieval catheters (NRY) across K=8 manufacturer sites, reporting held-out Poisson log-likelihood improvements for their conjugate federated method at matched privacy budgets.
Why it matters
World-model simulators and learned surgical predictors lack grounded estimates of how often devices harm patients. This work evaluates a privacy-preserving way to learn manufacturer-specific adverse-event rate beliefs from fragmented post-market data, enabling digital twins to condition simulations on population-level risk structure without transferring raw records.
Method
- Model complication-class event counts at each site k with conditionally Poisson counts y_{z,k} ~ Poi(n_{z,k} θ_z) and a Gamma prior over rates θ_z; update is conjugate via natural-parameter sufficient statistics exchanged across sites.
- Use a federated Bayesian surveillance protocol where sites exchange only Rényi-differentially-private natural-parameter updates (per-round 2Z scalars per site) with norm clipping, Gaussian noise, and secure aggregation.
- Benchmark on a manufacturer-federated partition of the complete FDA MAUDE MAUDE cohort for product code NRY; evaluate held-out Poisson log-likelihood per test event for multiple federated and centralized baselines.
Limitation
MAUDE is voluntary and under-reports true incidence; reporting propensities vary by manufacturer and over time, and MAUDE query reproducibility is a recognized challenge. Keyword classification is transparent but coarse (25% match no class, and narrative quality varies). Privacy operates at the event level assuming an honest-but-curious server without collusion, and the evaluation scores per-site posteriors.
Abstract (from arXiv)
Learned surgical simulators and world models can roll out plausible procedural futures, but they carry no grounded estimate of how often interventional devices actually harm patients. We propose treating population-scale adverse-event surveillance as a distinct belief layer of the surgical digital twin, and we evaluate a federated Bayesian protocol for learning it under formal privacy guarantees. Each site holds per-class Gamma-Poisson posteriors over adverse-event rates and exchanges only R\'enyi-differentially-private natural-parameter updates. We benchmark on the complete FDA MAUDE cohort for thrombus-retrieval catheters (product code NRY): 8,617 reports, of which 6,491 are classified by transparent keyword rules into five thrombectomy complication classes and partitioned across $K=8$ manufacturer sites. At a matched privacy budget of $(\\varepsilon \\approx 2.09, \\delta = 10^{-5})$, the conjugate protocol attains a held-out Poisson score of -5.78 per test event versus -26.58 for FedAvg with differential privacy. The non-private federated model also outperforms centralized pooling (+3.19 vs +2.93), evidence that manufacturer-specific complication profiles are real and that federation preserves them. Because MAUDE lacks procedure denominators, outputs are relative rate orderings rather than absolute risks, and we report all privacy-utility operating points.
Related papers
- Understanding Trajectory Heterogeneity in Federated World Model Learning
- Same Predictions, Different Harms: Causal Auditing of Patient World Models
- Repeated-Measure Leakage, Distribution Shift, and Reliability under Partial Observation in Patient World Models
- PatientWM: Towards Generalizable Clinical World Models via Unified Patient States
- A Belief-State World Model for Catheter Navigation under Sparse Fluoroscopy: A Planar Proof of Concept
Publication record
Bibliographic evidence describes where and when a record appeared. It does not validate scientific claims.
First public release: 2026-10-06 · Indexed by Pub-AI: 2026-10-07 UTC
Federated Bayesian Surveillance of Mechanical Thrombectomy Adverse Events: A Population Risk Layer for Surgical Digital Twins
Preprint posted: 2026-10-06 · Original record ↗
- arxiv
- 2610.08464
Explicit source versions (1)
- v1 · 2026-10-06
- Preprint available
Source date: 2026-10-06 · First detected: 2026-10-25 UTC · Historical baseline
Source evidence (1)
- Official evidence ↗ · atom:published (official metadata)
Abstract · arxiv
Learned surgical simulators and world models can roll out plausible procedural futures, but they carry no grounded estimate of how often interventional devices actually harm patients. We propose treating population-scale adverse-event surveillance as a distinct belief layer of the surgical digital twin, and we evaluate a federated Bayesian protocol for learning it under formal privacy guarantees. Each site holds per-class Gamma-Poisson posteriors over adverse-event rates and exchanges only Rényi-differentially-private natural-parameter updates. We benchmark on the complete FDA MAUDE cohort for thrombus-retrieval catheters (product code NRY): 8,617 reports, of which 6,491 are classified by transparent keyword rules into five thrombectomy complication classes and partitioned across $K=8$ manufacturer sites. At a matched privacy budget of $(\\varepsilon \\approx 2.09, \δ= 10^{-5})$, the conjugate protocol attains a held-out Poisson score of -5.78 per test event versus -26.58 for FedAvg with differential privacy. The non-private federated model also outperforms centralized pooling (+3.19 vs +2.93), evidence that manufacturer-specific complication profiles are real and that federation preserves them. Because MAUDE lacks procedure denominators, outputs are relative rate orderings rather than absolute risks, and we report all privacy-utility operating points.
