TY - JOUR
T1 - HINT
T2 - Hierarchical interaction network for clinical-trial-outcome predictions
AU - Fu, Tianfan
AU - Huang, Kexin
AU - Xiao, Cao
AU - Glass, Lucas M.
AU - Sun, Jimeng
N1 - This work was supported by IQVIA, NSF award SCH-2014438 , PPoSS 2028839 , IIS-1838042 , NIH award R01 1R01NS107291-01 , and OSF Healthcare.
PY - 2022/4/8
Y1 - 2022/4/8
N2 - Clinical trials are crucial for drug development but often face uncertain outcomes due to safety, efficacy, or patient-recruitment problems. We propose the Hierarchical Interaction Network (HINT) to predict clinical trial outcomes. First, HINT encodes multi-modal data (drug molecule, target disease, trial eligibility criteria) into embeddings. Then, HINT trains knowledge-embedding modules using drug pharmacokinetic and historical trial data. Finally, a hierarchical interaction graph connects all of the embeddings to capture their interactions and predict trial outcomes. HINT was trained and validated on 1,160 phase I trials, 4,449 phase II trials, and 3,436 phase III trials. It obtained 0.665, 0.620, and 0.847 F1 scores on separate test sets of 627 phase I, 1,653 phase II, and 1,140 phase III trials, respectively. HINT significantly outperforms the best baseline method on most metrics. The benchmark dataset and codes are released at https://github.com/futianfan/clinical-trial-outcome-prediction.
AB - Clinical trials are crucial for drug development but often face uncertain outcomes due to safety, efficacy, or patient-recruitment problems. We propose the Hierarchical Interaction Network (HINT) to predict clinical trial outcomes. First, HINT encodes multi-modal data (drug molecule, target disease, trial eligibility criteria) into embeddings. Then, HINT trains knowledge-embedding modules using drug pharmacokinetic and historical trial data. Finally, a hierarchical interaction graph connects all of the embeddings to capture their interactions and predict trial outcomes. HINT was trained and validated on 1,160 phase I trials, 4,449 phase II trials, and 3,436 phase III trials. It obtained 0.665, 0.620, and 0.847 F1 scores on separate test sets of 627 phase I, 1,653 phase II, and 1,140 phase III trials, respectively. HINT significantly outperforms the best baseline method on most metrics. The benchmark dataset and codes are released at https://github.com/futianfan/clinical-trial-outcome-prediction.
KW - clinical-trial-outcome prediction
KW - data mining
KW - deep learning
KW - graph neural network
KW - neural network
KW - trial outcome prediction benchmark
UR - http://www.scopus.com/inward/record.url?scp=85124873295&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85124873295&partnerID=8YFLogxK
U2 - 10.1016/j.patter.2022.100445
DO - 10.1016/j.patter.2022.100445
M3 - Article
C2 - 35465223
AN - SCOPUS:85124873295
SN - 2666-3899
VL - 3
JO - Patterns
JF - Patterns
IS - 4
M1 - 100445
ER -