¹Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany
²Friedrich-Alexander-Unviersität Erlangen-Nürnberg, Erlangen, Germany
³SRH University of Applied Sciences, Fürth, Germany

Text

Introduction: Precision oncology has increasingly shifted toward individualized treatment strategies, driven by vast molecular datasets from next-generation sequencing (NGS) and other high-throughput technologies. Molecular Tumor Boards (MTBs), as interdisciplinary panels, are key in translating this data into therapeutic recommendations, especially for rare or treatment-resistant cancers [1]. Yet, manually handling and integrating diverse and growing data sources often strain workflows and limit patient throughput. To address this, the German PM4Onco project aims to standardize and digitally support decision-making in MTBs [2]. A central component is the implementation of cBioPortal, an open-source platform designed to visualize and interpret molecular and clinical data. Originally developed for research, cBioPortal has been extended in German clinical contexts to provide MTB-specific functionalities, including trial matching and therapy documentation [3]. Despite its potential, a clinical decision tool aid to support patient care is ultimately useless unless it is put into practice [4]. This study analyzes existing implementations in German university hospitals to derive best-practice recommendations and support a more structured, scalable integration.

Methods: A qualitative process evaluation was conducted through semi-structured interviews with 21 participants from 11 university hospitals involved in the PM4Onco project. Participants included oncologists, bioinformaticians, IT specialists, and MTB coordinators. Prior to each interview, two questions were used to assess site-specific implementation success and choose the appropriate interview guide. Data were transcribed verbatim, anonymized, and analyzed using Mayring’s qualitative content analysis [5]. This process yielded four thematic domains – (1) current use, (2) benefits and prospects, (3) challenges and limitations, and (4) implementation process – derived deductively from the predefined interview guide. Subcategories were then developed inductively to capture emerging themes within each domain.

Results: The evaluation showed varying (1) cBioPortal implementation: only one hospital had fully integrated the tool into routine MTB processes. Six institutions used it mainly for case preparation, while others reported limited or no usage. Key advantages (2) comprised easier data access and visual clarity, supporting preparation and collaboration. Main barriers (3) included the lack of automated data integration, limited system interoperability with existing hospital information systems, and insufficient IT support. Unclear responsibilities, weak coordination, and poor documentation were compounded by legal demands like parallel reporting and strict data protection. Most sites lacked structured implementation strategies (4), rollout often exceeded a year and remained user-driven rather than strategically managed.

Conclusion: This study revealed heterogeneous cBioPortal implementation across German university hospitals, shaped by differences in infrastructure, stakeholder engagement, and resource allocation. Successful sites emphasized early stakeholder involvement, structured testing, and iterative planning with IT support. A stepwise rollout and tailored training were essential for sustainable use. Communication gaps among developers and users highlighted the need for structured feedback and shared learning formats. Based on these findings, a flexible best-practice framework was developed to support site-specific adaptation while promoting standardized implementation, underscoring the importance of accompanying implementation science. Future options to address manual workload issues could include AI-based decision support. This work guides implementation and underpins evaluation strategies in precision oncology. It forms the basis for a context-sensitive decision guide for MTB software.

The authors declare that they have no competing interests.

The authors declare that a positive ethics committee vote has been obtained.

Literatur

[1] Tamborero D, Dienstmann R, Rachid MH, Boekel J, Lopez-Fernandez A, Jonsson M, et al. The Molecular Tumor Board Portal supports clinical decisions and automated reporting for precision oncology. Nat Cancer. 2022 Feb;3(2):251–61. DOI: 10.1038/s43018-022-00332-x
[2] Pflugradt L, Metzger P, Gräßel L, Strantz C, Blaumeiser A, Busch H, et al. PM4Onco: personalisierte Medizin für die Onkologie: Effektive Datennutzung für molekulare Tumorboards. Onkologie. 2024 Oct;30(10):898–904. DOI: 10.1007/s00761-024-01593-0
[3] Strantz C, Böhm D, Ganslandt T, Börries M, Metzger P, Pauli T, et al. Empowering personalized oncology: evolution of digital support and visualization tools for molecular tumor boards. BMC Med Inform Decis Mak. 2025 Jan 16;25(1):29. DOI: 10.1186/s12911-024-02821-8
[4] Markowetz F. All models are wrong and yours are useless: making clinical prediction models impactful for patients. npj Precis Onc. 2024 Feb 28;8(1):54. DOI: 10.1038/s41698-024-00553-6
[5] Rädiker S, Kuckartz U. Analyse qualitativer Daten mit MAXQDA: Text, Audio und Video. Wiesbaden: Springer Fachmedien Wiesbaden; 2019. DOI: 10.1007/978-3-658-22095-2