Search

Arian Ranjbar,1 Eilin Wermundsen Mork,1 Jesper Ravn,1 Helga Brøgger,2 Per Myrseth,2 Hans Peter Østrem,3 Harry Hallock2 1Medical Technology and E-Health, Akershus University Hospital, Lørenskog, Norway; 2Group Research and Development, DNV AS, Høvik, Norway; 3Business Assurance, DNV AS, Høvik, NorwayCorrespondence: Arian Ranjbar, Akershus University Hospital, Sykehusveien 25, 1478, Lørenskog, Norway, Tel +46700436768, Email arian.ranjbar@medisin.uio.noAbstract: Artificial intelligence (AI) provides a unique opportunity to help meet the demands of the future healthcare system. However, hospitals may not be well equipped to handle safe and effective development and/or procurement of AI systems. Furthermore, upcoming regulations such as the EU AI Act may enforce the need to establish new management systems, quality assurance and control mechanisms, novel to healthcare organizations. This paper discusses challenges in AI implementation, particularly potential gaps in current management systems (MS), by reviewing the harmonized standard for AI MS, ISO 42001, as part of a gap analysis of a tertiary acute hospital with ongoing AI activities. Examination of the industry agnostic ISO 42001 reveals a technical debt within healthcare, aligning with previous research on digitalization and AI implementation. To successfully implement AI with quality assurance in mind, emphasis should be put on the foundation and structure of the healthcare organizations, including both workforce and data infrastructure.Keywords: artificial intelligence, management systems, quality assurance, risk management, implementation

Introduction: This study aimed to evaluate the consistency of lung cancer case assessments across multidisciplinary team (MDT) sites in Denmark. The goal was to appraise the comparability of outcomes between hospitals in a real-world context., Methods: We prepared sixty comprehensive, fictitious lung cancer case stories, complete with images, and distributed them to the four primary lung cancer MDT conferences in Denmark. These cases were subsequently evaluated as had they been ordinary patients during regular MDT meetings. We compared the conclusions on assigned TNM stage and proposed treatment intent using Kappa statistics., Results: The consensus on assigned stage (Stages IA-B, IIA-B, IIIA-B, IV, and undetermined) corresponded to a Fleiss' Kappa-value of 0.62 (95% CI: 0.52-0.71). The overall assessment of curability, categorized as Curable, Incurable, and Undetermined, corresponded to a Kappa-value of 0.72 (CI: 0.61-0.84). However, for cases unanimously judged by all MDT sites to be Stage III, the concordance on treatment intent was poor, with an agreement coefficient of only 0.32 (95% CI: -0.27-0.97)., Conclusion: In detail, the level of agreement on assigned stages was less than desired. In consequence, comparative analyses of treatment results from different hospitals or centres may be prone to bias caused by systematic differences in stage assessment or intent of treatment. The least consensus was observed for cases in Stage III, indicating a need for quality improvement efforts to ensure a higher degree of consistency in MDT decisions., Competing Interests: Declarations Ethics approval and consent to participate Not applicable. The study is based on fictitious case stories and associated imaging used in the study are fully and irrevocably anonymized. The Danish National Medical Research Ethics Committee (https://researchethics.dk/) waived the need for ethical approval and referred to the local Institutional Review Board at Aarhus University Hospital (IRB-AUH) for permission to access selected imaging for full and irrevocable anonymization and subsequent use in the study, for which permission was granted. The need for consent to participate was waived by IRB-AUH as it was deemed unnecessary for the intended use according to Danish national regulations (Danish Health Act, § 42 d (https://www.retsinformation.dk/eli/lta/2018/1286#P42d)). The use of fictitious case stories and fully and irrevocable anonymized data for research are permissible under Danish and European General Data Protection Regulation (the GPDR rules – REGULATION (EU) 2016/679 OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL of 27 April 2016, article 26 (https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32016R0679)). Consent for publication Not applicable. The study is based on fictitious case stories and fully and irrevocably anonymized images, the use of which is permissible under Danish and European GDPR rules as stated and explained above for ‘Ethics approval and consent to participate’. Competing interests The authors declare no competing interests. Authors information Each author is a senior consultant with extensive experience in lung cancer diagnostics at university hospitals or is conducting research into the function of MDT in lung cancer., (© 2024. The Author(s).)

Monitoring of artificial intelligence (AI)-based algorithms is necessary for safe implementation and will be required in upcoming regulations. This study investigates the potential for monitoring of AI in hospitals. First, by reviewing regulatory requirements and state of the art of monitoring. Then, by conducting a gap analysis of ISO42001, containing industry agnostic requirements harmonized with the EU AI Act. The analysis illustrates the need for comprehensive monitoring capable of capturing deviations in input, performance drifts and unintended interactions. However, hospitals often suffer from a technical debt, and the gap analysis provides qualitative indications on implementation challenges, including data quality, infrastructure and limitations in continuous improvement.

Xylans' unique properties make it attractive for a variety of industries, including paper, food, and biochemical production. While for some applications the preservation of its natural structure is crucial, for others the degradation into monosaccharides is essential. For the complete breakdown, the use of several enzymes is required, due to its structural complexity. In fact, the specificity of enzymatically-catalyzed reactions is guided by the surface, limiting or regulating accessibility and serving structurally encoded input guiding the actions of the enzymes. Here, we investigate enzymes at surfaces rich in xylan using surface plasmon resonance spectroscopy. The influence of diffusion and changes in substrate morphology is studied via enzyme surface kinetics simulations, yielding reaction rates and constants. We propose kinetic models, which can be applied to the degradation of multilayer biopolymer films. The most advanced model was verified by its successful application to the degradation of a thin film of polyhydroxybutyrate treated with a polyhydroxybutyrate-depolymerase. The herein derived models can be employed to quantify the degradation kinetics of various enzymes on biopolymers in heterogeneous environments, often prevalent in industrial processes. The identification of key factors influencing reaction rates such as inhibition will contribute to the quantification of intricate dynamics in complex systems., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2024 The Author(s). Published by Elsevier Ltd.. All rights reserved.)

Objective: This study aims to assess the ability of state-of-the-art machine learning algorithms to detect valvular heart disease (VHD) from digital heart sound recordings in a general population that includes asymptomatic cases and intermediate stages of disease progression., Methods: We trained a recurrent neural network to predict murmurs from heart sound audio using annotated recordings collected with digital stethoscopes from four auscultation positions in 2,124 participants from the Tromsø7 study. The predicted murmurs were used to predict VHD as determined by echocardiography., Results: The presence of aortic stenosis (AS) was detected with a sensitivity of 90.9%, a specificity of 94.5%, and an area under the curve (AUC) of 0.979 (CI: 0.963-0.995). At least moderate AS was detected with an AUC of 0.993 (CI: 0.989-0.997). Moderate or greater aortic and mitral regurgitation (AR and MR) were predicted with AUC values of 0.634 (CI: 0.565-703) and 0.549 (CI: 0.506-0.593), respectively, which increased to 0.766 and 0.677 when clinical variables were added as predictors. The AUC for predicting symptomatic cases was higher for AR and MR, 0.756 and 0.711, respectively. Screening jointly for symptomatic regurgitation or presence of stenosis resulted in an AUC of 0.86, with 97.7% of AS cases ( n  = 44) and all 12 MS cases detected., Conclusions: The algorithm demonstrated excellent performance in detecting AS in a general cohort, surpassing observations from similar studies on selected cohorts. The detection of AR and MR based on HS audio was poor, but accuracy was considerably higher for symptomatic cases, and the inclusion of clinical variables improved the performance of the model significantly., Competing Interests: Medsensio AS is a company that provides products for automatic detection of abnormal lung sounds. JR and MJ are employed by Medsensio. HM was working for Medsensio during this research project, and PW was employed for some time by Medsensio after the research project was mostly completed. JR, MJ, and LB owns shares in Medsensio. PW, HM, HS, MJ, TD, JR, SA, AD, JS, and LB have filed two patents for the algorithms outlined in this paper., (© 2024 Waaler, Malbye, Schirmer, Johnsen, Donnem, Ravn, Andersen, Davidsen, Aviles Solis, Stylidis and Bongo.)