Back

Analysis and Mitigation of Equipment-induced Shortcuts in AI Models for Laparoscopic Cholecystectomy

Protserov, S.; Repalo, A.; Mashouri, P.; Hunter, J.; Masino, C.; Madani, A.; Brudno, M.

2026-04-24 surgery
10.64898/2026.04.22.26351545 medRxiv
Show abstract

Machine learning models have seen a lot of success in medical image segmentation domain. However, one of the challenges that they face are confounders or shortcuts: spurious correlations or biases in the training data that affect the resulting models. One example of such confounders for surgical machine learning is the setup of surgical equipment, including tools and lighting. Using the task of identification of safe and dangerous zones of dissection in laparoscopic cholecystectomy images and videos as a use-case, we inspect two equipment-induced biases: the presence of surgical tools in the field of view and the position of lighting. We propose methods for evaluating the severity of these biases and augmentation-based methods for mitigating them. We show that our tool bias mitigations improve the models' consistency under tool movements by 9 percentage points in the most inconsistent cases, and by 4 percentage points on average. Our lighting bias mitigations help reduce fraction of true dangerous zone pixels that may be predicted as safe under light changes from 5% to 1.5%, without compromising segmentation quality.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

1
PLOS Computational Biology
1633 papers in training set
Top 0.9%
19.5%
2
npj Digital Medicine
97 papers in training set
Top 0.4%
10.9%
3
PLOS ONE
4510 papers in training set
Top 23%
7.5%
4
Biology Methods and Protocols
53 papers in training set
Top 0.1%
6.7%
5
BMC Medical Informatics and Decision Making
39 papers in training set
Top 0.5%
5.1%
6
Scientific Reports
3102 papers in training set
Top 21%
5.1%
50% of probability mass above
7
Medical Image Analysis
33 papers in training set
Top 0.3%
3.7%
8
IEEE Access
31 papers in training set
Top 0.2%
2.9%
9
Human Brain Mapping
295 papers in training set
Top 2%
2.7%
10
Nature Machine Intelligence
61 papers in training set
Top 2%
1.9%
11
Heliyon
146 papers in training set
Top 2%
1.7%
12
Nature Communications
4913 papers in training set
Top 53%
1.6%
13
iScience
1063 papers in training set
Top 19%
1.4%
14
Cancers
200 papers in training set
Top 3%
1.3%
15
Journal of Medical Imaging
11 papers in training set
Top 0.2%
1.3%
16
PLOS Digital Health
91 papers in training set
Top 2%
1.3%
17
Proceedings of the National Academy of Sciences
2130 papers in training set
Top 39%
1.0%
18
Neurocomputing
13 papers in training set
Top 0.4%
0.9%
19
Journal of Biomedical Informatics
45 papers in training set
Top 1%
0.8%
20
Nature Methods
336 papers in training set
Top 6%
0.8%
21
IEEE Transactions on Medical Imaging
18 papers in training set
Top 0.4%
0.8%
22
Diagnostics
48 papers in training set
Top 2%
0.8%
23
Nature Medicine
117 papers in training set
Top 4%
0.8%
24
Frontiers in Plant Science
240 papers in training set
Top 5%
0.8%
25
Frontiers in Medicine
113 papers in training set
Top 6%
0.8%
26
eBioMedicine
130 papers in training set
Top 4%
0.8%
27
Science Advances
1098 papers in training set
Top 29%
0.8%
28
Frontiers in Computational Neuroscience
53 papers in training set
Top 2%
0.7%
29
Brain Communications
147 papers in training set
Top 4%
0.5%
30
Bioengineering
24 papers in training set
Top 2%
0.5%