A Study of Age and Sex Bias in Multiple Instance Learning Based Classification of Acute Myeloid Leukemia Subtypes

Ario Sadafi; Matthias Hehr; Nassir Navab; Carsten Marr

doi:10.1007/978-3-031-45249-9_25

A Study of Age and Sex Bias in Multiple Instance Learning Based Classification of Acute Myeloid Leukemia Subtypes

Ario Sadafi
, Matthias Hehr
, Nassir Navab
, Carsten Marr

Informatics 16 - Chair of Computer Aided Medical Procedures

Helmholtz Zentrum München German Research Center for Environmental Health
Technical University of Munich
Ludwig-Maximilians-Universität München
Johns Hopkins University

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

1 Scopus citations

Abstract

Accurate classification of Acute Myeloid Leukemia (AML) subtypes is crucial for clinical decision-making and patient care. In this study, we investigate the potential presence of age and sex bias in AML subtype classification using Multiple Instance Learning (MIL) architectures. To that end, we train multiple MIL models using different levels of sex imbalance in the training set and excluding certain age groups. To assess the sex bias, we evaluate the performance of the models on male and female test sets. For age bias, models are tested against underrepresented age groups in the training data. We find a significant effect of sex and age bias on the performance of the model for AML subtype classification. Specifically, we observe that females are more likely to be affected by sex imbalance dataset and certain age groups, such as patients with 72 to 86 years of age with the RUNX1::RUNX1T1 genetic subtype, are significantly affected by an age bias present in the training data. Ensuring inclusivity in the training data is thus essential for generating reliable and equitable outcomes in AML genetic subtype classification, ultimately benefiting diverse patient populations.

Original language	English
Title of host publication	Clinical Image-Based Procedures, Fairness of AI in Medical Imaging, and Ethical and Philosophical Issues in Medical Imaging - 12th International Workshop, CLIP 2023 1st International Workshop, FAIMI 2023 and 2nd International Workshop, EPIMI 2023, Proceedings
Editors	Stefan Wesarg, Cristina Oyarzun Laura, Esther Puyol Antón, Andrew P. King, John S.H. Baxter, Marius Erdt, Klaus Drechsler, Moti Freiman, Yufei Chen, Islem Rekik, Roy Eagleson, Aasa Feragen, Veronika Cheplygina, Melani Ganz-Benjaminsen, Enzo Ferrante, Ben Glocker, Daniel Moyer, Eikel Petersen
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	256-265
Number of pages	10
ISBN (Print)	9783031452482
DOIs	https://doi.org/10.1007/978-3-031-45249-9_25
State	Published - 2023
Event	12th International Workshop on Clinical Image-Based Procedures, CLIP 2023, 1st MICCAI Workshop on Fairness of AI in Medical Imaging, FAIMI 2023, held in conjunction with MICCAI 2023 and 2nd MICCAI Workshop on the Ethical and Philosophical Issues in Medical Imaging, EPIMI 2023 - Vancouver, Canada Duration: 12 Oct 2023 → 12 Oct 2023

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	14242 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	12th International Workshop on Clinical Image-Based Procedures, CLIP 2023, 1st MICCAI Workshop on Fairness of AI in Medical Imaging, FAIMI 2023, held in conjunction with MICCAI 2023 and 2nd MICCAI Workshop on the Ethical and Philosophical Issues in Medical Imaging, EPIMI 2023
Country/Territory	Canada
City	Vancouver
Period	12/10/23 → 12/10/23

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 3 Good Health and Well-being

Keywords

Acute Myeloid Leukemia
Age Bias
Fairness
Multiple Instance Learning
Sex Bias

Access to Document

10.1007/978-3-031-45249-9_25

Cite this

Sadafi, A., Hehr, M., Navab, N., & Marr, C. (2023). A Study of Age and Sex Bias in Multiple Instance Learning Based Classification of Acute Myeloid Leukemia Subtypes. In S. Wesarg, C. Oyarzun Laura, E. Puyol Antón, A. P. King, J. S. H. Baxter, M. Erdt, K. Drechsler, M. Freiman, Y. Chen, I. Rekik, R. Eagleson, A. Feragen, V. Cheplygina, M. Ganz-Benjaminsen, E. Ferrante, B. Glocker, D. Moyer, & E. Petersen (Eds.), Clinical Image-Based Procedures, Fairness of AI in Medical Imaging, and Ethical and Philosophical Issues in Medical Imaging - 12th International Workshop, CLIP 2023 1st International Workshop, FAIMI 2023 and 2nd International Workshop, EPIMI 2023, Proceedings (pp. 256-265). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 14242 LNCS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-45249-9_25

Sadafi, Ario ; Hehr, Matthias ; Navab, Nassir et al. / A Study of Age and Sex Bias in Multiple Instance Learning Based Classification of Acute Myeloid Leukemia Subtypes. Clinical Image-Based Procedures, Fairness of AI in Medical Imaging, and Ethical and Philosophical Issues in Medical Imaging - 12th International Workshop, CLIP 2023 1st International Workshop, FAIMI 2023 and 2nd International Workshop, EPIMI 2023, Proceedings. editor / Stefan Wesarg ; Cristina Oyarzun Laura ; Esther Puyol Antón ; Andrew P. King ; John S.H. Baxter ; Marius Erdt ; Klaus Drechsler ; Moti Freiman ; Yufei Chen ; Islem Rekik ; Roy Eagleson ; Aasa Feragen ; Veronika Cheplygina ; Melani Ganz-Benjaminsen ; Enzo Ferrante ; Ben Glocker ; Daniel Moyer ; Eikel Petersen. Springer Science and Business Media Deutschland GmbH, 2023. pp. 256-265 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{4f0d68fd67d44947875b3b1a31c2e892,

title = "A Study of Age and Sex Bias in Multiple Instance Learning Based Classification of Acute Myeloid Leukemia Subtypes",

abstract = "Accurate classification of Acute Myeloid Leukemia (AML) subtypes is crucial for clinical decision-making and patient care. In this study, we investigate the potential presence of age and sex bias in AML subtype classification using Multiple Instance Learning (MIL) architectures. To that end, we train multiple MIL models using different levels of sex imbalance in the training set and excluding certain age groups. To assess the sex bias, we evaluate the performance of the models on male and female test sets. For age bias, models are tested against underrepresented age groups in the training data. We find a significant effect of sex and age bias on the performance of the model for AML subtype classification. Specifically, we observe that females are more likely to be affected by sex imbalance dataset and certain age groups, such as patients with 72 to 86 years of age with the RUNX1::RUNX1T1 genetic subtype, are significantly affected by an age bias present in the training data. Ensuring inclusivity in the training data is thus essential for generating reliable and equitable outcomes in AML genetic subtype classification, ultimately benefiting diverse patient populations.",

keywords = "Acute Myeloid Leukemia, Age Bias, Fairness, Multiple Instance Learning, Sex Bias",

author = "Ario Sadafi and Matthias Hehr and Nassir Navab and Carsten Marr",

note = "Publisher Copyright: {\textcopyright} The Author(s), under exclusive license to Springer Nature Switzerland AG 2023.; 12th International Workshop on Clinical Image-Based Procedures, CLIP 2023, 1st MICCAI Workshop on Fairness of AI in Medical Imaging, FAIMI 2023, held in conjunction with MICCAI 2023 and 2nd MICCAI Workshop on the Ethical and Philosophical Issues in Medical Imaging, EPIMI 2023 ; Conference date: 12-10-2023 Through 12-10-2023",

year = "2023",

doi = "10.1007/978-3-031-45249-9\_25",

language = "English",

isbn = "9783031452482",

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

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "256--265",

editor = "Stefan Wesarg and \{Oyarzun Laura\}, Cristina and \{Puyol Ant{\'o}n\}, Esther and King, \{Andrew P.\} and Baxter, \{John S.H.\} and Marius Erdt and Klaus Drechsler and Moti Freiman and Yufei Chen and Islem Rekik and Roy Eagleson and Aasa Feragen and Veronika Cheplygina and Melani Ganz-Benjaminsen and Enzo Ferrante and Ben Glocker and Daniel Moyer and Eikel Petersen",

booktitle = "Clinical Image-Based Procedures, Fairness of AI in Medical Imaging, and Ethical and Philosophical Issues in Medical Imaging - 12th International Workshop, CLIP 2023 1st International Workshop, FAIMI 2023 and 2nd International Workshop, EPIMI 2023, Proceedings",

}

Sadafi, A, Hehr, M, Navab, N & Marr, C 2023, A Study of Age and Sex Bias in Multiple Instance Learning Based Classification of Acute Myeloid Leukemia Subtypes. in S Wesarg, C Oyarzun Laura, E Puyol Antón, AP King, JSH Baxter, M Erdt, K Drechsler, M Freiman, Y Chen, I Rekik, R Eagleson, A Feragen, V Cheplygina, M Ganz-Benjaminsen, E Ferrante, B Glocker, D Moyer & E Petersen (eds), Clinical Image-Based Procedures, Fairness of AI in Medical Imaging, and Ethical and Philosophical Issues in Medical Imaging - 12th International Workshop, CLIP 2023 1st International Workshop, FAIMI 2023 and 2nd International Workshop, EPIMI 2023, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 14242 LNCS, Springer Science and Business Media Deutschland GmbH, pp. 256-265, 12th International Workshop on Clinical Image-Based Procedures, CLIP 2023, 1st MICCAI Workshop on Fairness of AI in Medical Imaging, FAIMI 2023, held in conjunction with MICCAI 2023 and 2nd MICCAI Workshop on the Ethical and Philosophical Issues in Medical Imaging, EPIMI 2023, Vancouver, Canada, 12/10/23. https://doi.org/10.1007/978-3-031-45249-9_25

A Study of Age and Sex Bias in Multiple Instance Learning Based Classification of Acute Myeloid Leukemia Subtypes. / Sadafi, Ario; Hehr, Matthias; Navab, Nassir et al.
Clinical Image-Based Procedures, Fairness of AI in Medical Imaging, and Ethical and Philosophical Issues in Medical Imaging - 12th International Workshop, CLIP 2023 1st International Workshop, FAIMI 2023 and 2nd International Workshop, EPIMI 2023, Proceedings. ed. / Stefan Wesarg; Cristina Oyarzun Laura; Esther Puyol Antón; Andrew P. King; John S.H. Baxter; Marius Erdt; Klaus Drechsler; Moti Freiman; Yufei Chen; Islem Rekik; Roy Eagleson; Aasa Feragen; Veronika Cheplygina; Melani Ganz-Benjaminsen; Enzo Ferrante; Ben Glocker; Daniel Moyer; Eikel Petersen. Springer Science and Business Media Deutschland GmbH, 2023. p. 256-265 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 14242 LNCS).

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

TY - GEN

T1 - A Study of Age and Sex Bias in Multiple Instance Learning Based Classification of Acute Myeloid Leukemia Subtypes

AU - Sadafi, Ario

AU - Hehr, Matthias

AU - Navab, Nassir

AU - Marr, Carsten

N1 - Publisher Copyright: © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023.

PY - 2023

Y1 - 2023

N2 - Accurate classification of Acute Myeloid Leukemia (AML) subtypes is crucial for clinical decision-making and patient care. In this study, we investigate the potential presence of age and sex bias in AML subtype classification using Multiple Instance Learning (MIL) architectures. To that end, we train multiple MIL models using different levels of sex imbalance in the training set and excluding certain age groups. To assess the sex bias, we evaluate the performance of the models on male and female test sets. For age bias, models are tested against underrepresented age groups in the training data. We find a significant effect of sex and age bias on the performance of the model for AML subtype classification. Specifically, we observe that females are more likely to be affected by sex imbalance dataset and certain age groups, such as patients with 72 to 86 years of age with the RUNX1::RUNX1T1 genetic subtype, are significantly affected by an age bias present in the training data. Ensuring inclusivity in the training data is thus essential for generating reliable and equitable outcomes in AML genetic subtype classification, ultimately benefiting diverse patient populations.

AB - Accurate classification of Acute Myeloid Leukemia (AML) subtypes is crucial for clinical decision-making and patient care. In this study, we investigate the potential presence of age and sex bias in AML subtype classification using Multiple Instance Learning (MIL) architectures. To that end, we train multiple MIL models using different levels of sex imbalance in the training set and excluding certain age groups. To assess the sex bias, we evaluate the performance of the models on male and female test sets. For age bias, models are tested against underrepresented age groups in the training data. We find a significant effect of sex and age bias on the performance of the model for AML subtype classification. Specifically, we observe that females are more likely to be affected by sex imbalance dataset and certain age groups, such as patients with 72 to 86 years of age with the RUNX1::RUNX1T1 genetic subtype, are significantly affected by an age bias present in the training data. Ensuring inclusivity in the training data is thus essential for generating reliable and equitable outcomes in AML genetic subtype classification, ultimately benefiting diverse patient populations.

KW - Acute Myeloid Leukemia

KW - Age Bias

KW - Fairness

KW - Multiple Instance Learning

KW - Sex Bias

UR - https://www.scopus.com/pages/publications/85175824232

U2 - 10.1007/978-3-031-45249-9_25

DO - 10.1007/978-3-031-45249-9_25

M3 - Conference contribution

AN - SCOPUS:85175824232

SN - 9783031452482

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

SP - 256

EP - 265

BT - Clinical Image-Based Procedures, Fairness of AI in Medical Imaging, and Ethical and Philosophical Issues in Medical Imaging - 12th International Workshop, CLIP 2023 1st International Workshop, FAIMI 2023 and 2nd International Workshop, EPIMI 2023, Proceedings

A2 - Wesarg, Stefan

A2 - Oyarzun Laura, Cristina

A2 - Puyol Antón, Esther

A2 - King, Andrew P.

A2 - Baxter, John S.H.

A2 - Erdt, Marius

A2 - Drechsler, Klaus

A2 - Freiman, Moti

A2 - Chen, Yufei

A2 - Rekik, Islem

A2 - Eagleson, Roy

A2 - Feragen, Aasa

A2 - Cheplygina, Veronika

A2 - Ganz-Benjaminsen, Melani

A2 - Ferrante, Enzo

A2 - Glocker, Ben

A2 - Moyer, Daniel

A2 - Petersen, Eikel

PB - Springer Science and Business Media Deutschland GmbH

T2 - 12th International Workshop on Clinical Image-Based Procedures, CLIP 2023, 1st MICCAI Workshop on Fairness of AI in Medical Imaging, FAIMI 2023, held in conjunction with MICCAI 2023 and 2nd MICCAI Workshop on the Ethical and Philosophical Issues in Medical Imaging, EPIMI 2023

Y2 - 12 October 2023 through 12 October 2023

ER -

Sadafi A, Hehr M, Navab N, Marr C. A Study of Age and Sex Bias in Multiple Instance Learning Based Classification of Acute Myeloid Leukemia Subtypes. In Wesarg S, Oyarzun Laura C, Puyol Antón E, King AP, Baxter JSH, Erdt M, Drechsler K, Freiman M, Chen Y, Rekik I, Eagleson R, Feragen A, Cheplygina V, Ganz-Benjaminsen M, Ferrante E, Glocker B, Moyer D, Petersen E, editors, Clinical Image-Based Procedures, Fairness of AI in Medical Imaging, and Ethical and Philosophical Issues in Medical Imaging - 12th International Workshop, CLIP 2023 1st International Workshop, FAIMI 2023 and 2nd International Workshop, EPIMI 2023, Proceedings. Springer Science and Business Media Deutschland GmbH. 2023. p. 256-265. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-031-45249-9_25