Published online Dec 28, 2023. doi: 10.4329/wjr.v15.i12.359
Peer-review started: October 3, 2023
First decision: October 9, 2023
Revised: November 13, 2023
Accepted: December 5, 2023
Article in press: December 5, 2023
Published online: December 28, 2023
Processing time: 83 Days and 5.1 Hours
Missing occult cancer lesions accounts for the most diagnostic errors in retro
To achieve appropriate AI model training, a large annotated dataset is necessary to train the AI models. Our goal in this research is to compare two methods for decreasing the annotation time to establish ground truth: Skip-slice annotation and AI-initiated annotation.
We developed a 2D U-Net as an AI second observer for detecting colorectal cancer (CRC) and an ensemble of 5 differently initiated 2D U-Net for ensemble technique. Each model was trained with 51 cases of annotated CRC computed tomography of the abdomen and pelvis, tested with 7 cases, and validated with 20 cases from The Cancer Imaging Archive cases. The sensitivity, false positives per case, and estimated Dice coefficient were obtained for each method of training. We compared the two methods of annotations and the time reduction associated with the technique. The time differences were tested using Friedman’s two-way analysis of variance.
Sparse annotation significantly reduces the time for annotation particularly skipping 2 slices at a time (P < 0.001). Reduction of up to 2/3 of the annotation does not reduce AI model sensitivity or false positives per case. Although initializing human annotation with AI reduces the annotation time, the reduction is minimal, even when using an ensemble AI to decrease false positives.
Our data support the sparse annotation technique as an efficient technique for reducing the time needed to establish the ground truth.
Core Tip: Minimizing diagnostic errors for colorectal cancer may be most effectively performed with artificial intelligence (AI) second observer. Supervised training of AI-observer will require high volume of annotated training cases. Comparing skip-slice annotation and AI-initiated annotation shows that skipping slices does not affect the training outcome while AI-initiated annotation does not significantly improve annotation time.