Ida Häggström

A frozen CT-CLIP model can outperform traditional clinical baselines in lung cancer survival prediction, even with limited data.

Freeing survival prediction from reliance on expert-derived features, this model learns directly from CT scans and radiology reports to improve accuracy, especially when clinical data is scarce.

Injecting even noisy "exemplar" bounding boxes at inference time dramatically boosts medical object detection performance, without any retraining.