Multi-task Learning Approach for Intracranial Hemorrhage Prognosis

Miriam Cobo; Amaia Pérez del Barrio; Pablo Menéndez Fernández-Miranda; Pablo Sanz Bellón; Lara Lloret Iglesias; Wilson Silva

doi:10.1007/978-3-031-73290-4_2

Multi-task Learning Approach for Intracranial Hemorrhage Prognosis

Miriam Cobo^*, Amaia Pérez del Barrio, Pablo Menéndez Fernández-Miranda, Pablo Sanz Bellón, Lara Lloret Iglesias, Wilson Silva

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

Abstract

Prognosis after intracranial hemorrhage (ICH) is influenced by a complex interplay between imaging and tabular data. Rapid and reliable prognosis are crucial for effective patient stratification and informed treatment decision-making. In this study, we aim to enhance image-based prognosis by learning a robust feature representation shared between prognosis and the clinical and demographic variables most highly correlated with it. Our approach mimics clinical decision-making by reinforcing the model to learn valuable prognostic data embedded in the image. We propose a 3D multi-task image model to predict prognosis, Glasgow Coma Scale and age, improving accuracy and interpretability. Our method outperforms current state-of-the-art baseline image models, and demonstrates superior performance in ICH prognosis compared to four board-certified neuroradiologists using only CT scans as input. We further validate our model with interpretability saliency maps. Code is available at https://github.com/MiriamCobo/MultitaskLearning_ICH_Prognosis.git.

Original language	English
Title of host publication	Machine Learning in Medical Imaging
Subtitle of host publication	15th International Workshop, MLMI 2024, Held in Conjunction with MICCAI 2024, Proceedings
Editors	Xuanang Xu, Zhiming Cui, Kaicong Sun, Islem Rekik, Xi Ouyang
Publisher	Springer
Pages	12-21
Number of pages	10
ISBN (Electronic)	978-3-031-73290-4
ISBN (Print)	978-3-031-73292-8
DOIs	https://doi.org/10.1007/978-3-031-73290-4_2
Publication status	Published - 23 Oct 2024
Event	15th International Workshop on Machine Learning in Medical Imaging, MLMI 2024 was held in conjunction with the 27th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2024 - Marrakesh, Morocco Duration: 6 Oct 2024 → 6 Oct 2024

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	15242 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	15th International Workshop on Machine Learning in Medical Imaging, MLMI 2024 was held in conjunction with the 27th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2024
Country/Territory	Morocco
City	Marrakesh
Period	6/10/24 → 6/10/24

Keywords

Explainable AI
Multi-task learning
Prognosis

Access to Document

10.1007/978-3-031-73290-4_2

Embargoed Document

978-3-031-73290-4_2
Final published version, 823 KB
Licence: Taverne
Embargo ends: 23/04/25
Request copy

Cite this

Cobo, M., Pérez del Barrio, A., Menéndez Fernández-Miranda, P., Sanz Bellón, P., Lloret Iglesias, L., & Silva, W. (2024). Multi-task Learning Approach for Intracranial Hemorrhage Prognosis. In X. Xu, Z. Cui, K. Sun, I. Rekik, & X. Ouyang (Eds.), Machine Learning in Medical Imaging: 15th International Workshop, MLMI 2024, Held in Conjunction with MICCAI 2024, Proceedings (pp. 12-21). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 15242 LNCS). Springer. https://doi.org/10.1007/978-3-031-73290-4_2

Cobo, Miriam ; Pérez del Barrio, Amaia ; Menéndez Fernández-Miranda, Pablo et al. / Multi-task Learning Approach for Intracranial Hemorrhage Prognosis. Machine Learning in Medical Imaging: 15th International Workshop, MLMI 2024, Held in Conjunction with MICCAI 2024, Proceedings. editor / Xuanang Xu ; Zhiming Cui ; Kaicong Sun ; Islem Rekik ; Xi Ouyang. Springer, 2024. pp. 12-21 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{646c1386c3b848c38b65e5f3a958438c,

title = "Multi-task Learning Approach for Intracranial Hemorrhage Prognosis",

abstract = "Prognosis after intracranial hemorrhage (ICH) is influenced by a complex interplay between imaging and tabular data. Rapid and reliable prognosis are crucial for effective patient stratification and informed treatment decision-making. In this study, we aim to enhance image-based prognosis by learning a robust feature representation shared between prognosis and the clinical and demographic variables most highly correlated with it. Our approach mimics clinical decision-making by reinforcing the model to learn valuable prognostic data embedded in the image. We propose a 3D multi-task image model to predict prognosis, Glasgow Coma Scale and age, improving accuracy and interpretability. Our method outperforms current state-of-the-art baseline image models, and demonstrates superior performance in ICH prognosis compared to four board-certified neuroradiologists using only CT scans as input. We further validate our model with interpretability saliency maps. Code is available at https://github.com/MiriamCobo/MultitaskLearning_ICH_Prognosis.git.",

keywords = "Explainable AI, Multi-task learning, Prognosis",

author = "Miriam Cobo and {P{\'e}rez del Barrio}, Amaia and {Men{\'e}ndez Fern{\'a}ndez-Miranda}, Pablo and {Sanz Bell{\'o}n}, Pablo and {Lloret Iglesias}, Lara and Wilson Silva",

note = "Publisher Copyright: {\textcopyright} The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.; 15th International Workshop on Machine Learning in Medical Imaging, MLMI 2024 was held in conjunction with the 27th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2024 ; Conference date: 06-10-2024 Through 06-10-2024",

year = "2024",

month = oct,

day = "23",

doi = "10.1007/978-3-031-73290-4_2",

language = "English",

isbn = "978-3-031-73292-8",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer",

pages = "12--21",

editor = "Xuanang Xu and Zhiming Cui and Kaicong Sun and Islem Rekik and Xi Ouyang",

booktitle = "Machine Learning in Medical Imaging",

address = "Germany",

}

Cobo, M, Pérez del Barrio, A, Menéndez Fernández-Miranda, P, Sanz Bellón, P, Lloret Iglesias, L & Silva, W 2024, Multi-task Learning Approach for Intracranial Hemorrhage Prognosis. in X Xu, Z Cui, K Sun, I Rekik & X Ouyang (eds), Machine Learning in Medical Imaging: 15th International Workshop, MLMI 2024, Held in Conjunction with MICCAI 2024, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 15242 LNCS, Springer, pp. 12-21, 15th International Workshop on Machine Learning in Medical Imaging, MLMI 2024 was held in conjunction with the 27th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2024, Marrakesh, Morocco, 6/10/24. https://doi.org/10.1007/978-3-031-73290-4_2

Multi-task Learning Approach for Intracranial Hemorrhage Prognosis. / Cobo, Miriam; Pérez del Barrio, Amaia; Menéndez Fernández-Miranda, Pablo et al.
Machine Learning in Medical Imaging: 15th International Workshop, MLMI 2024, Held in Conjunction with MICCAI 2024, Proceedings. ed. / Xuanang Xu; Zhiming Cui; Kaicong Sun; Islem Rekik; Xi Ouyang. Springer, 2024. p. 12-21 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 15242 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - Multi-task Learning Approach for Intracranial Hemorrhage Prognosis

AU - Cobo, Miriam

AU - Pérez del Barrio, Amaia

AU - Menéndez Fernández-Miranda, Pablo

AU - Sanz Bellón, Pablo

AU - Lloret Iglesias, Lara

AU - Silva, Wilson

N1 - Publisher Copyright: © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

PY - 2024/10/23

Y1 - 2024/10/23

N2 - Prognosis after intracranial hemorrhage (ICH) is influenced by a complex interplay between imaging and tabular data. Rapid and reliable prognosis are crucial for effective patient stratification and informed treatment decision-making. In this study, we aim to enhance image-based prognosis by learning a robust feature representation shared between prognosis and the clinical and demographic variables most highly correlated with it. Our approach mimics clinical decision-making by reinforcing the model to learn valuable prognostic data embedded in the image. We propose a 3D multi-task image model to predict prognosis, Glasgow Coma Scale and age, improving accuracy and interpretability. Our method outperforms current state-of-the-art baseline image models, and demonstrates superior performance in ICH prognosis compared to four board-certified neuroradiologists using only CT scans as input. We further validate our model with interpretability saliency maps. Code is available at https://github.com/MiriamCobo/MultitaskLearning_ICH_Prognosis.git.

AB - Prognosis after intracranial hemorrhage (ICH) is influenced by a complex interplay between imaging and tabular data. Rapid and reliable prognosis are crucial for effective patient stratification and informed treatment decision-making. In this study, we aim to enhance image-based prognosis by learning a robust feature representation shared between prognosis and the clinical and demographic variables most highly correlated with it. Our approach mimics clinical decision-making by reinforcing the model to learn valuable prognostic data embedded in the image. We propose a 3D multi-task image model to predict prognosis, Glasgow Coma Scale and age, improving accuracy and interpretability. Our method outperforms current state-of-the-art baseline image models, and demonstrates superior performance in ICH prognosis compared to four board-certified neuroradiologists using only CT scans as input. We further validate our model with interpretability saliency maps. Code is available at https://github.com/MiriamCobo/MultitaskLearning_ICH_Prognosis.git.

KW - Explainable AI

KW - Multi-task learning

KW - Prognosis

UR - http://www.scopus.com/inward/record.url?scp=85208421730&partnerID=8YFLogxK

U2 - 10.1007/978-3-031-73290-4_2

DO - 10.1007/978-3-031-73290-4_2

M3 - Conference contribution

AN - SCOPUS:85208421730

SN - 978-3-031-73292-8

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 12

EP - 21

BT - Machine Learning in Medical Imaging

A2 - Xu, Xuanang

A2 - Cui, Zhiming

A2 - Sun, Kaicong

A2 - Rekik, Islem

A2 - Ouyang, Xi

PB - Springer

T2 - 15th International Workshop on Machine Learning in Medical Imaging, MLMI 2024 was held in conjunction with the 27th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2024

Y2 - 6 October 2024 through 6 October 2024

ER -

Cobo M, Pérez del Barrio A, Menéndez Fernández-Miranda P, Sanz Bellón P, Lloret Iglesias L, Silva W. Multi-task Learning Approach for Intracranial Hemorrhage Prognosis. In Xu X, Cui Z, Sun K, Rekik I, Ouyang X, editors, Machine Learning in Medical Imaging: 15th International Workshop, MLMI 2024, Held in Conjunction with MICCAI 2024, Proceedings. Springer. 2024. p. 12-21. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-031-73290-4_2