Routine Usage of AI -based Chest X -ray Reading Support in a M ulti-site Medical Supply Center Authors Karsten Ridder1 MD Alexander Preuhs2 PhD Axel Mertins1 Clemens Joerger2 PhD Affiliations 1 MVZ UhlenbrockPartner Dortmund Germany

2025-05-03 0 0 409.69KB 6 页 10玖币

侵权投诉

Routine Usage of AI-based Chest X-ray Reading Support in a Multi-site Medical Supply Center

Authors: Karsten Ridder1, MD / Alexander Preuhs2, PhD / Axel Mertins1 / Clemens Joerger2, PhD

Affiliations: 1: MVZ Uhlenbrock&Partner, Dortmund, Germany

2: Siemens Healthineers, Erlangen, Germany

Keywords: Artificial Intelligence, Chest X-ray, Radiology

Key information:

1. Research question: How can we establish an AI support for reading of chest X-rays in clinical

routine and which benefits emerge for the clinicians and radiologists. Can it perform 24/7

support for practicing clinicians?

2. Findings: We installed an AI solution for Chest X-ray in a given structure (MVZ Uhlenbrock &

Partner, Germany). We could demonstrate the practicability, performance, and benefits in 10

connected clinical sites.

3. Meaning: A commercially available AI solution for the evaluation of Chest X-ray images is able to

help radiologists and clinical colleagues 24/7 in a complex environment. The system performs in

a robust manner – supporting radiologists and clinical colleagues in their important decisions– in

practises and hospitals regardless of the user and X-ray system type producing the image-data.

Introduction

As one of the worldwide most taken X-ray procedures, chest X-ray (CXR) is one of the most important

and also most demanding modalities in daily medical imaging. Due to its accessibility and fast

acquisition, chest radiographs are frequently used as a first diagnostic step, deciding on further patient

treatment [1,2]. During normal working times there is a high workload for every radiologist with a

certain risk of volatile errors, missed diagnosis, misinterpretation, and late diagnosis. While most of the

research assessing AI-based clinical decision support systems is focused on quantitative evaluations in a

retrospective and synthetic environment [3,4], this abstract is focused on the implementation of an AI-

based chest evaluation system in a real-world clinical setup. Therefore, we deployed an AI system,

capable of analysing CXR, in a large clinical setup including 10 clinical sites and qualitatively assess the

benefits of such a system to the clinicians and radiologists.

Material and methods

We installed an AI-based CXR reading support in clinical routine in a German multi-site medical supply

centre (MVZ Uhlenbrock and Partner). The MVZ for radiology connects 14 institutions including hospitals

and private practices. 10 of these institutions produce CXR images in a clinical routine set-up 24/7. All

acquired posterior anterior (PA) images from stationary X-ray systems of different vendors (Siemens

Healthineers, Carestream Health, Oehm und Rehbein GmbH, Philips) are sent to a central server where

they are uploaded to the cloud for evaluation. We use the AI Rad Companion Chest X-ray from Siemens

Healthineers (Figure 1) [5]. The AI system identifies five pre-specified findings on CXRs (p.a. view):

Pneumothorax, Pleural Effusion, Pulmonary Lesions, Consolidation, and Atelectasis. To assess the impact

of the deployed AI system we retrospectively analysed the AI-detected findings, and we conducted

several interviews with radiologists and connected physicians (e.g. surgeons and internal doctors) about

the benefits.

Results

Within one month about 1400 frontal images, received from the 10 institutions, were analysed. Figure 2

shows the percentage of CXR images with one or more findings. Depending on the institution the images

containing radiographic findings vary between 22% and 53% (Figure 3). The radiologists at the respective

sites reassured that those numbers are realistic and are due to the different patient population that are

treated (out-patient vs hospital). Network performance did not slow down the clinical flow of reading

and the AI result matches the physician’s opinion in 8 out of 10 cases. Due to missing ground-truth, a

quantitative evaluation is currently not possible. Quality improvement by a kind of second read and

diagnostic support during extra hours (e.g. nightshifts) were mentioned as the most important benefits

during the interviews. The AI was considered especially helpful by the radiologists, if it detects difficult

findings, e.g., as depicted in Figure 4, where the AI highlights a lesion behind the heart, which was

confirmed later by a CT scan as shown in Figure 5. We received mixed feedback about preferred positive

predictive value (PPV) and negative predictive value (NPV), with a slight tendency towards high NPV for

radiologists and high PPV for clinicians. Finally, we learnt that training and communication of how to use

the AI in daily routine by the users is an important success factor.

Discussion and Conclusion

文档加载中……请稍候！
如果长时间未打开，您也可以点击刷新试试。

下载文档到电脑，查找使用更方便

10 玖币 0人已下载

立即下载

摘要：

RoutineUsageofAI-basedChestX-rayReadingSupportinaMulti-siteMedicalSupplyCenterAuthors:KarstenRidder1,MD/AlexanderPreuhs2,PhD/AxelMertins1/ClemensJoerger2,PhDAffiliations:1:MVZUhlenbrock&Partner,Dortmund,Germany2:SiemensHealthineers,Erlangen,GermanyKeywords:ArtificialIntelligence,ChestX-ray,Radiology...

展开>> 收起<<

Routine Usage of AI -based Chest X -ray Reading Support in a M ulti-site Medical Supply Center Authors Karsten Ridder1 MD Alexander Preuhs2 PhD Axel Mertins1 Clemens Joerger2 PhD Affiliations 1 MVZ UhlenbrockPartner Dortmund Germany.pdf

共6页,预览2页

还剩页未读，继续阅读

声明：本站为文档C2C交易模式，即用户上传的文档直接被用户下载，本站只是中间服务平台，本站所有文档下载所得的收益归上传人(含作者)所有。玖贝云文库仅提供信息存储空间，仅对用户上传内容的表现方式做保护处理，对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私，请立即通知玖贝云文库，我们立即给予删除！

Routine Usage of AI -based Chest X -ray Reading Support in a M ulti-site Medical Supply Center Authors Karsten Ridder1 MD Alexander Preuhs2 PhD Axel Mertins1 Clemens Joerger2 PhD Affiliations 1 MVZ UhlenbrockPartner Dortmund Germany

相关推荐

开通VIP享超值会员特权

作者详情

相关内容

热门标签

举报选择: