360° OncoImaging Trajectory (OIT): Reframing Oncology Monitoring as a Multimodal, Time‑Aligned Trajectory (Preprint)

BACKGROUND Oncology monitoring traditionally treats imaging, molecular assays, and patient signals as discrete snapshots. This paradigm can delay detection of deterioration, misclassify transient effects, and obscure actionable transitions. OBJECTIVE To introduce the 360° OncoImaging Trajectory (OIT), a framework that fuses multimodal signals onto a common temporal axis to infer interpretable trajectory states and transitions. METHODS OIT comprises four modules: Multimodal Inputs, Temporal Harmonization, Trajectory Inference Engine, and Clinical Decision Outputs. It defines a four-state taxonomy (Stable, Early Deterioration, PseudoResponse, Recovery) and uses probabilistic models (changepoint + HMM) with cross-modal corroboration and attribution. A staged validation roadmap includes simulation, synthetic cohorts, and public dataset testing. RESULTS OIT is designed to detect deterioration earlier than snapshot rules by leveraging multimodal corroboration. Simulated experiments will quantify lead time gains, alert performance, and attribution fidelity. Prototype algorithms emphasize interpretability and uncertainty quantification. CONCLUSIONS OIT reframes oncology monitoring from isolated measures to trajectories, enabling earlier, explainable, and actionable decision support. The staged validation plan supports reproducible progress without premature clinical commitments.