Robustness Investigation on Deep Learning CT Reconstruction for Real-Time Dose Optimization
This study investigates the robustness of a deep learning reconstruction method for real-time dose optimization in CT, which is important for patient safety. The findings suggest that current deep learning methods may not be robust enough for clinical deployment, especially when encountering out-of-distribution data.
This paper investigates the robustness of AUTOMAP for real-time CT reconstruction with limited projections, a critical step for organ-specific automatic exposure control. On the MNIST dataset, AUTOMAP achieved a false rate of 1.6% and 6.8% for 2 and 4 projections respectively when the test set was similar to the training set. However, when tested on an unseen digit, the false rate dramatically increased to 94.4%. For medical images, AUTOMAP achieved an average root-mean-square error of 290 HU, reconstructing coarse body outlines but misshaping some organs.
In computed tomography (CT), automatic exposure control (AEC) is frequently used to reduce radiation dose exposure to patients. For organ-specific AEC, a preliminary CT reconstruction is necessary to estimate organ shapes for dose optimization, where only a few projections are allowed for real-time reconstruction. In this work, we investigate the performance of automated transform by manifold approximation (AUTOMAP) in such applications. For proof of concept, we investigate its performance on the MNIST dataset first, where the dataset containing all the 10 digits are randomly split into a training set and a test set. We train the AUTOMAP model for image reconstruction from 2 projections or 4 projections directly. The test results demonstrate that AUTOMAP is able to reconstruct most digits well with a false rate of 1.6% and 6.8% respectively. In our subsequent experiment, the MNIST dataset is split in a way that the training set contains 9 digits only while the test set contains the excluded digit only, for instance "2". In the test results, the digit "2"s are falsely predicted as "3" or "5" when using 2 projections for reconstruction, reaching a false rate of 94.4%. For the application in medical images, AUTOMAP is also trained on patients' CT images. The test images reach an average root-mean-square error of 290 HU. Although the coarse body outlines are well reconstructed, some organs are misshaped.