Combining chest X-rays and CT can improve diagnostic accuracy, but CT requires selective use because of its cost and radiation exposure.
We propose a modality-aware alert framework that determines the need for additional CT acquisition using only an X-ray.
The framework disentangles CT representations into general features shared with X-rays and CT-specific features.
Differences between these representations help identify cases likely to benefit from the additional information provided by CT.
By recommending CT when diagnostic gains are expected, the framework reduces unnecessary imaging and improves diagnostic efficienc

Jaehyeong Bae is a PhD student at Chung-Ang University.
His research focuses on medical imaging and multimodal learning.
He studies complementary information in chest X-rays and CT scans.
His interests include predicting when additional imaging is useful.
He aims to support clinical decisions and efficient imaging use.