Back

AI-based histologic scoring enables automated and reproducible assessment of enrollment criteria and endpoints in NASH clinical trials

Iyer, J. S.; Pokkalla, H.; Biddle-Snead, C.; Carrasco-Zevallos, O. M.; Lin, M.; Shanis, Z.; Le, Q.; Juyal, D.; Pouryahya, M.; Pedawi, A.; Hoffman, S.; Elliott, H.; Leidal, K.; Myers, R. P.; Chung, C.; Billin, A. N.; Watkins, T. R.; Resnick, M.; Wack, K.; Glickman, J.; Burt, A. D.; Loomba, R.; Sanyal, A. J.; Montalto, M. C.; Beck, A. H.; Taylor-Weiner, A.; Wapinski, I.

2023-04-25 pathology
10.1101/2023.04.20.23288534 medRxiv
Show abstract

Clinical trials in nonalcoholic steatohepatitis (NASH) require histologic scoring for assessment of inclusion criteria and endpoints. However, guidelines for scoring key features have led to variability in interpretation, impacting clinical trial outcomes. We developed an artificial intelligence (AI)-based measurement (AIM) tool for scoring NASH histology (AIM-NASH). AIM-NASH predictions for NASH Clinical Research Network (CRN) grades of necroinflammation and stages of fibrosis aligned with expert consensus scores and were reproducible. Continuous scores produced by AIM-NASH for key histological features of NASH correlated with mean pathologist scores and with noninvasive biomarkers and strongly predicted patient outcomes. In a retrospective analysis of the ATLAS trial, previously unmet pathological endpoints were met when scored by the AIM-NASH algorithm alone. Overall, these results suggest that AIM-NASH may assist pathologists in histologic review of NASH clinical trials, reducing inter-rater variability on trial outcomes and offering a more sensitive and reproducible measure of patient therapeutic response.

Matching journals

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

1
Nature Communications
○ 5641 papers in training set
Top 13%
13.2%
Similar papers in this journal
2
Gastroenterology
○ 42 papers in training set
Top 0.1%
11.1%
Similar papers in this journal
3
eBioMedicine
○ 183 papers in training set
Top 0.1%
8.0%
Similar papers in this journal
4
Nature Medicine
○ 125 papers in training set
Top 0.2%
7.3%
Similar papers in this journal
5
The American Journal of Pathology
○ 32 papers in training set
Top 0.1%
5.5%
Similar papers in this journal
6
npj Digital Medicine
○ 118 papers in training set
Top 1%
5.2%
50% of probability mass above
7
Communications Medicine
○ 113 papers in training set
Top 0.4%
4.1%
8
Hepatology
○ 22 papers in training set
Top 0.1%
4.1%
9
eLife
● 5828 papers in training set
Top 33%
3.5%
10
Nature
○ 645 papers in training set
Top 5%
2.4%
11
Cell Reports Medicine
○ 153 papers in training set
Top 1%
2.4%
12
JCI Insight
● 277 papers in training set
Top 3%
2.4%
13
Cancer Discovery
● 66 papers in training set
Top 1%
2.1%
14
Journal of Hepatology
○ 21 papers in training set
Top 0.2%
1.7%
15
Science Translational Medicine
● 127 papers in training set
Top 2%
1.7%
16
Cell Reports
○ 1498 papers in training set
Top 20%
1.5%
17
iScience
○ 1154 papers in training set
Top 26%
1.1%
18
Scientific Reports
○ 3612 papers in training set
Top 62%
1.1%
19
Genome Medicine
○ 183 papers in training set
Top 4%
1.1%
20
Gastro Hep Advances
○ 11 papers in training set
Top 0.2%
1.0%
21
Gut
● 40 papers in training set
Top 0.7%
1.0%
22
The Journal of Pathology
○ 26 papers in training set
Top 0.6%
1.0%
23
Modern Pathology
○ 22 papers in training set
Top 0.5%
0.6%
24
Nature Genetics
○ 286 papers in training set
Top 5%
0.6%
25
PLOS ONE
● 5266 papers in training set
Top 61%
0.6%
26
JHEP Reports
○ 11 papers in training set
Top 0.3%
0.6%
27
Cell Reports Methods
○ 165 papers in training set
Top 4%
0.6%
28
Journal of Clinical Investigation
● 179 papers in training set
Top 6%
0.6%

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.