..
/ cscw / home /
 

PAIRADS

– Prostate Artificial Intelligence Reporting and Data System

PAIRADS Logo
Laufzeit: 2021-2024

Förderung: Bundesministerium für Forschung und Bildung im Rahmen von STARTinteraktiv
BMBF Logo
Ansprechpartner: Sheree Saßmannshausen

Projektwebseite: www.pairads.ai

Presseartikel:
https://www.interaktive-technologien.de/projekte/pairads
https://www.uni-siegen.de/start/news/forschungsnews/952525.html
https://www.wp.de/staedte/siegerland/uni-siegen-mit-hightech-gegen-prostatakrebsvorgehen-
id233210415.html

Beschreibung

Im Forschungsprojekt PAIRADS arbeiten wir gemeinsam mit dem Unternehmen Gemedico GmbH sowie mit Assoziativpartner*innen aus der Radiologie zusammen. Ziel ist es, einen Demonstrator für die Prostatakrebsdiagnostik in der Radiologie zu
entwickeln, der mit Hilfe einer Künstlichen Intelligenz (KI) die Radiolog*innen entlastet und bereits eigenständig durch eine Bilderkennung in den MRTs Prostatakarzinome identifiziert und lokalisiert.

Die Wissenschaftler*innen der CSCW-Gruppe erforschen dabei den Bereich der Human-Centered-AI. Zentral ist dabei welche Bedürfnisse und Erwartungen die Benutzer an die KI sowie an die Interaktion mit der KI haben, sodass sie unterstützt werden können und eine gewünschte User Experience aufgebaut wird. Es sollen Anforderungen erhoben werden, die festlegen wie die Ergebnisse der KI aufbereitet werden sollten und wie die menschlichen Benutzer*innen und das technische System
wechselseitig interagieren können – auf funktionaler, aber auch emotionaler Basis. Wichtig ist dabei, dass die KI lediglich als Assistenz, nicht als eigenständige Instanz dient. Dafür wird die aktuelle Vorgehensweise der Radiolog*innen in ihrer Befundung eines Prostatakarzinoms erhoben, indem u.a. vor Ort kontexuelle Beobachtungen und
Interviews durchgeführt werden, um konkrete Anforderungen an die Künstliche Intelligenz und ihren Trainingsprozess erheben zu können. Beispielsweise gilt die Befundung eines Prostatakarzinoms als recht komplex, sodass häufig zwei
Radiolog*innen eingesetzt werden. Dabei werden in der Medizin oft bestimmte Schemata verwendet. In diesem Falle handelt es sich um den PI-RADS-Standard für die Befundung eines Prostatakarzinoms, an dem sich die Radiolog*innen orientieren.
Wie sie genau vorgehen, soll im Rahmen der Feldforschung betrachtet, reale Daten erhoben und schließlich analysiert werden. Der Forschungsansatz des Human-in-the-loop gilt ebenfalls als Forschungsgegenstand, sodass der Input aus dem aktuellen Stand sowie das Feedback der Radiolog*innen bezüglich des KI-Ergebnisses stets mit berücksichtigt wird und kontinuierlich in den Algorithmus und den weiteren Trainingsprozess der Künstlichen Intelligenz mit einfließt.
So soll ganzheitlich eine menschzentrierte Entwicklung sichergestellt werden. Durch den Einbezug von Domänenexpert*innen und weiteren Rechts- und Ethikexpert*innen werden Workshops durchgeführt, um Datenschutzaspekte sowie ethisch-rechtliche und soziale Aspekte zu beleuchten, die beispielsweise eine gerechte Wahrnehmung (fairness) und auch Vertrauen (trust) gegenüber der KI sicherstellen soll.

Publikationen

2023

  • Ontika, N., Saßmannshausen, S., Pinatti, A. F. D. C. & Pipek, V. (2023)Embedding Artificial Intelligence into Healthcare Infrastructure for Prostate Cancer Diagnosis

    [BibTeX] [Abstract] [Download PDF]

    Early detection and diagnosis of prostate cancer are of utmost significance for effective treatment, and artificial intelligence (AI) has the potential to assist radiologists in this area by analyzing medical images and improving diagnostic accuracy, especially given the scarcity of radiologists. This article outlines our ongoing research, focusing on designing a human-centered AI system to aid radiologists in detecting and diagnosing prostate cancer and integrating it into the existing infrastructure. Through qualitative field research involving observations, contextual inquiries, and …

    @article{ontika_embedding_2023,
title = {Embedding {Artificial} {Intelligence} into {Healthcare} {Infrastructure} for {Prostate} {Cancer} {Diagnosis}},
issn = {2510-2591},
url = {https://dl.eusset.eu/handle/20.500.12015/5020},
abstract = {Early detection and diagnosis of prostate cancer are of utmost significance for effective treatment, and artificial intelligence (AI) has the potential to assist radiologists in this area by analyzing medical images and improving diagnostic accuracy, especially given the scarcity of radiologists. This article outlines our ongoing research, focusing on designing a human-centered AI system to aid radiologists in detecting and diagnosing prostate cancer and integrating it into the existing infrastructure. Through qualitative field research involving observations, contextual inquiries, and ...},
language = {en},
urldate = {2023-10-20},
author = {Ontika, Nazmun and Saßmannshausen, Sheree and Pinatti, Aparecido Fabiano De Carvalho and Pipek, Volkmar},
year = {2023},
note = {Publisher: European Society for Socially Embedded Technologies (EUSSET)},
keywords = {pairads},
}

  • Ontika, N. N., Sassmannshausen, S. M., Pinatti De Carvalho, A. F. & Pipek, V. (2023)PAIRADS: Hybrid Interaction Between Humans and AI in Radiology

    IN HHAI 2023: Augmenting Human Intellect doi:10.3233/FAIA230108
    [BibTeX] [Download PDF] 

    @incollection{ontika_pairads_2023,
title = {{PAIRADS}: {Hybrid} {Interaction} {Between} {Humans} and {AI} in {Radiology}},
shorttitle = {{PAIRADS}},
url = {https://ebooks.iospress.nl/doi/10.3233/FAIA230108},
urldate = {2023-07-31},
booktitle = {{HHAI} 2023: {Augmenting} {Human} {Intellect}},
publisher = {IOS Press},
author = {Ontika, Nazmun Nisat and Sassmannshausen, Sheree May and Pinatti De Carvalho, Aparecido Fabiano and Pipek, Volkmar},
year = {2023},
doi = {10.3233/FAIA230108},
keywords = {pairads},
pages = {395--397},
}

2022

  • Ontika, N. N., Saßmannshausen, S. M., Syed, H. A. & Pinatti De Carvalho, A. F. (2022)Exploring Human-Centered AI in Healthcare: A Workshop Report

    IN IRSI Report, Vol. 19, Pages: 1–54
    [BibTeX] [Abstract] [Download PDF]

    As a technique of improving the quality of life, AI has the potential to take a significant part in healthcare worldwide. However, in order to facilitate the widespread use of AI systems, we must first better comprehend the influence of AI on the healthcare sector. To create an acceptable intelligent system for healthcare, a comprehensive evaluation of ethically driven design, technology that effectively addresses human intellect, and human aspects of design is required. Our two-day workshop at the European Conference on CSCW in 2022 focused on Human-centered AI in the healthcare domain. In the workshop, we brought together researchers and practitioners in health informatics to accelerate conversations about developing usable and efficient intelligent systems that are more understandable and reliable for users.

    @article{ontika_exploring_2022-1,
series = {International reports on socio-informatics},
title = {Exploring {Human}-{Centered} {AI} in {Healthcare}: {A} {Workshop} {Report}},
volume = {19},
issn = {1861-4280},
url = {https://www.iisi.de/wp-content/uploads/2022/10/IRSI_V19I2.pdf},
abstract = {As a technique of improving the quality of life, AI has the potential to take a significant part in healthcare worldwide. However, in order to facilitate the widespread use of AI systems, we must first better comprehend the influence of AI on the healthcare sector. To create an acceptable intelligent system for healthcare, a comprehensive evaluation of ethically driven design, technology that effectively addresses human intellect, and human aspects of design is required. Our two-day workshop at the European Conference on CSCW in 2022 focused on Human-centered AI in the healthcare domain. In the workshop, we brought together researchers and practitioners in health informatics to accelerate conversations about developing usable and efficient intelligent systems that are more understandable and reliable for users.},
language = {English},
number = {2},
journal = {IRSI Report},
author = {Ontika, Nazmun Nisat and Saßmannshausen, Sheree May and Syed, Hussain Abid and Pinatti De Carvalho, Aparecido Fabiano},
editor = {Pipek, Volkmar and Rohde, Markus},
month = oct,
year = {2022},
keywords = {pairads},
pages = {1--54},
}

  • Ontika, N. N., Saßmannshausen, S. M., Syed, H. A., de Carvalho, A. F. P. & Pipek, V. (2022)‪Towards Human-Centered AI: Learning from Current Practices in Radiology‬

    [BibTeX] [Download PDF] 

    @inproceedings{ontika_towards_2022,
title = {‪{Towards} {Human}-{Centered} {AI}: {Learning} from {Current} {Practices} in {Radiology}‬},
shorttitle = {‪{Towards} {Human}-{Centered} {AI}},
url = {https://scholar.google.com/citations?view_op=view_citation&hl=de&user=3f5u4_kAAAAJ&citation_for_view=3f5u4_kAAAAJ:2osOgNQ5qMEC},
urldate = {2022-11-15},
author = {Ontika, Nazmun Nisat and Saßmannshausen, Sheree May and Syed, Hussain Abid and Carvalho, Aparecido Fabiano Pinatti de and Pipek, Volkmar},
year = {2022},
keywords = {pairads},
}

  • Ontika, N. N., Syed, H. A., Saßmannshausen, S. M., Harper, R. H., Chen, Y., Park, S. Y., Grisot, M., Chow, A., Blaumer, N., Pinatti de Carvalho, A. F. & Pipek, V. (2022)Exploring Human-Centered AI in Healthcare: Diagnosis, Explainability, and Trust

    doi:10.48340/ecscw2022_ws06
    [BibTeX] [Abstract] [Download PDF]

    AI has become an increasingly active area of research over the past few years in healthcare. Nevertheless, not all research advancements are applicable in the field as there are only a few AI solutions that are actually deployed in medical infrastructures or actively used by medical practitioners. This can be due to various reasons as the lack of a human-centered approach for the or non-incorporation of humans in the loop. In this workshop, we aim to address the questions relevant to human-centered AI solutions associated with healthcare by exploring different human-centered approaches for designing AI systems and using image-based datasets for medical diagnosis. We aim to bring together researchers and practitioners in AI, human-computer interaction, healthcare, etc., and expedite the discussions about making usable systems that will be more comprehensible and dependable. Findings from our workshop may serve as ‘terminus a quo’ to significantly improve AI solutions for medical diagnosis.

    @article{ontika_exploring_2022,
title = {Exploring {Human}-{Centered} {AI} in {Healthcare}: {Diagnosis}, {Explainability}, and {Trust}},
issn = {2510-2591},
shorttitle = {Exploring {Human}-{Centered} {AI} in {Healthcare}},
url = {https://dl.eusset.eu/handle/20.500.12015/4409},
doi = {10.48340/ecscw2022_ws06},
abstract = {AI has become an increasingly active area of research over the past few years in healthcare. Nevertheless, not all research advancements are applicable in the field as there are only a few AI solutions that are actually deployed in medical infrastructures or actively used by medical practitioners. This can be due to various reasons as the lack of a human-centered approach for the or non-incorporation of humans in the loop. In this workshop, we aim to address the questions relevant to human-centered AI solutions associated with healthcare by exploring different human-centered approaches for designing AI systems and using image-based datasets for medical diagnosis. We aim to bring together researchers and practitioners in AI, human-computer interaction, healthcare, etc., and expedite the discussions about making usable systems that will be more comprehensible and dependable. Findings from our workshop may serve as ‘terminus a quo’ to significantly improve AI solutions for medical diagnosis.},
language = {en},
urldate = {2022-06-27},
author = {Ontika, Nazmun Nisat and Syed, Hussain Abid and Saßmannshausen, Sheree May and Harper, Richard HR and Chen, Yunan and Park, Sun Young and Grisot, Miria and Chow, Astrid and Blaumer, Nils and Pinatti de Carvalho, Aparecido Fabiano and Pipek, Volkmar},
year = {2022},
note = {Accepted: 2022-06-22T04:34:51Z
Publisher: European Society for Socially Embedded Technologies (EUSSET)},
keywords = {pairads},
}