CLARITY: Medical World Model for Guiding Treatment Decisions by Modeling Context-Aware Disease Trajectories in Latent Space

arXiv 25.12 2025 43.2 method, application

TLDR

CLARITY is a medical world model that forecasts disease trajectories in latent space, integrating temporal and clinical context for treatment decisions.

Reasoning

The paper introduces a novel approach to medical world modeling by explicitly incorporating temporal and clinical contexts, and linking predictions to treatment decisions, which addresses key limitations of prior work. Its strength lies in the structured latent space and prediction-to-decision framework, but the evaluation is limited to a single dataset (MU-Glioma-Post) and the abstract does not provide details on broader generalization or ablation studies.

Read-first score

Read-first score 43.2, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 41.

Recency 6%

86.7

Uses a gentle age decay so recent papers surface without erasing older foundations. 2025

Methodology quality 18%

Screens visible abstract and analysis fields for experiment, dataset, baseline, metric, and limitation evidence. markers=baseline,dataset,evaluation

Topical relevance 29%

58.6

Uses existing LLM keyword relevance scores normalized to 0-100. world model,world simulator,generative world model,interactive world model,video world model,world dynamics prediction,model-based reinforcement learning world model

Reproducibility 18%

Screens links and visible text for paper, code, dataset, artifact, and repository signals. pdf=True; code=False; dataset=False; markers=dataset

Citation impact 18%

Uses OpenAlex-shaped citation metadata as a bibliometric attention signal, separate from paper quality. cited_by_count=0

Citation velocity 12%

Citation velocity estimates citations per publication-year to reduce old-paper bias. velocity=0.00

Field roles

FrontierMethodology anchor

Rank sensitivity

Stability: volatile; rank range: 383.

Keyword Scores

world model

world dynamics prediction

generative world model

world simulator

interactive world model

model-based reinforcement learning world model

video world model

Deep Analysis

Innovations

Forecasting disease evolution directly within a structured latent space rather than relying on stochastic diffusion models for visual reconstruction
Explicit integration of time intervals (temporal context) and patient-specific data (clinical context) to model treatment-conditioned progression as smooth, interpretable trajectories
Novel prediction-to-decision framework that translates latent rollouts into transparent, actionable treatment recommendations

Methodology

CLARITY is a medical world model that operates in a structured latent space to forecast disease evolution. It incorporates temporal context (time intervals) and clinical context (patient-specific data) to model treatment-conditioned trajectories. The model uses a prediction-to-decision framework to convert latent rollouts into actionable treatment plans, and is evaluated on the MU-Glioma-Post dataset against baselines including MeWM and medical-specific large language models.

Key Results

On the MU-Glioma-Post dataset, CLARITY outperforms the recent MeWM by 12% and significantly surpasses all other medical-specific large language models in treatment planning performance.

Limitations

Evaluation is limited to a single dataset (MU-Glioma-Post), so generalizability to other cancer types or clinical settings is not established
The latent space modeling may not fully capture all causal physiological transitions, potentially limiting physiological faithfulness
The approach requires detailed patient-specific temporal and clinical data, which may not be available in all clinical scenarios

Resources

Paper