Streamlining and Accelerating the Molecular Tumor Board Process at the University Medical Center Hamburg-Eppendorf

Abstract

Background

We developed, introduced, and evaluated Molecular ONcology Optimized CLinical Evaluation (MONOCLE), a secure, open-source web application at the University Medical Center Hamburg-Eppendorf (UKE), to optimize the analysis and discussion of complex cancer cases in molecular tumor boards (MTB).

Objective

MONOCLE standardizes and harmonizes documentation, while its integrated Knowledge Connector accelerates literature research for personalized treatment.

Patients and Methods

The system was designed by merging the requirements of the German Network for Personalized Medicine (DNPM), the medical staff involved in the MTB process, and the team developing MONOCLE. The usability was evaluated using the System Usability Scale (SUS) and user tasks. Overall process optimization was measured by the number of automated tasks that can be performed.

Results

MONOCLE, introduced into clinical practice in June 2024, significantly reduces time for documentation, as three manual steps now run automatically, and transfer of the data to the DNPM is possible. Its usability and SUS showed positive results, ranging between 92.5 and 97.5.

Conclusions

As the first open-source and extendable solution for standardized MTB documentation, MONOCLE enables wider adoption by other medical centers.

Supplementary Information

The online version contains supplementary material available at 10.1007/s11523-025-01160-6.

Key Points

The implementation of MONOCLE streamlined and accelerated the workflow of the MTB at the University Medical Center Hamburg-Eppendorf.

The flexible, extendable, open-source, and privacy-preserving MONOCLE web application can be adapted and implemented within other medical centers.

The integrated KC can facilitate complex literature research for MTB cases.

Introduction

In recent years, the field of personalized oncology has seen significant advancements, opening new opportunities for patients who previously had limited options [1]. Advances in sequencing technologies, along with progress in bioinformatics, cancer immunology, and molecular biology, have significantly reduced the time and cost associated with targeted therapies, which is especially relevant given the traditionally labor-intensive process of collecting detailed family histories in clinical practice [2]. Currently, the cost of long-read sequencing technologies (e.g., PacBio, 10X Genomics) is declining, offering improved insights into the structure and order of genomic rearrangements, features often missed by next-generation sequencing (NGS). Broader clinical adoption of these technologies could enhance diagnostic yield for genetic diseases. Continued progress toward pan-ethnic reference genomes, large-scale public variant databases [3], and richer clinical genetic repositories [4, 5] will help overcome remaining technical barriers. Moreover, early sequencing of many cancers, as recommended by guidelines, enables earlier targeted treatment and improved care. As an example, NGS accelerates diagnosis by analyzing multiple genes simultaneously in a single assay. Thus, costs for sequencing technologies have been falling over the last decades and consequently, therapies are no longer exclusive to a small subset of patients but are accessible to a broader patient base [6].

The now possible process of interpreting genetic variants, assessing druggability, and evaluating patient status is inherently complex and time-consuming. It necessitates the expertise of a multidisciplinary team, including, among others, oncologists, pathologists, and bioinformaticians [7]. The multidisciplinary molecular tumor board (MTB) process, therefore, involves multiple steps, diverse expertise, and various data types from different departments at different time points [8]. Identifying personalized treatment options based on solid evidence remains challenging given the rarity of identical genomic alterations and clinical conditions among patients.

The complexity of MTBs extends beyond individual institutions and reflects broader challenges at the European level. The EU’s flagship initiative, 1+ Million Genomes (1+MG), exemplifies this by aiming to enable secure, cross-border access to genomic and clinical data across Europe. This initiative supports cutting-edge research, evidence-based health policy, and the development of personalized treatments, with the overarching goal of improving disease prevention and patient outcomes. As one of the world’s largest genomics projects, 1+MG is instrumental in establishing global standards in precision medicine [9]. At the national level, Germany is actively contributing to this landscape through initiatives such as the Modellvorhaben after §64e SGB V [10], aiming to make molecular diagnostics and treatment billable by health insurance companies, and the German Network for Personalized Medicine (Deutsches Netzwerk für Personalisierte Medizin, DNPM). The DNPM, comprising 28 university medical centers, seeks to enhance the diagnosis and treatment of complex or rare cancers and improve patients’ access to personalized medicine. These national efforts are closely aligned with European goals and help translate large-scale genomic initiatives into improved clinical practice and patient care. The DNPM collaborates with European networks such as Genomic Medicine Sweden (GMS) and ICPerMed to share data and conduct cross-border research and clinical trials. These efforts align with the European Health Data Space (EHDS) initiative, which promotes the exchange of medical data across Europe [11].

To tackle key challenges in personalized medicine—particularly dealing with small, non-standardized datasets—the DNPM aims to expand collective knowledge. This includes collaboration with international cancer centers and the integration of the following steps:

1. Establishment of a standardized clinical and genetic core dataset (cCDS and gCDS, respectively) to be used by all participating university medical centers in Germany that conduct MTBs [12].

2. Combination of the pseudonymized datasets from all participating medical centers into a single node to facilitate collaborative research and registry efforts [13].

Participation in the DNPM network will enable access to pseudonymized patient data across multiple medical centers, significantly expanding the available patient cohort for research. Since MTB data from other medical centers adhere to the same standardized core data structure, with minimal use of free-text fields, the quality and quantity of data available for retrospective research will be greatly enhanced compared with previous MTB datasets.

While the benefits of participating in DNPM are clear, the integration of patient data from diverse clinical domains, such as genomic and clinical data, presents several challenges:

1. Disparate internal clinical IT structures and varying data collection and documentation systems make it more difficult to collect and standardize the data at one central point.

2. A cascading process to obtain genomic data, including several internal and external institutions at a medical center [14].

3. Iterative procedures from patient admission to MTB, involving multiple potential clinical pathways [15].

4. Organizational and data protection challenges that are especially complex within the German law context.

Although software solutions for structured MTB documentation, such as Onkostar [16], are available, they are not suitable for our university medical center owing to infrastructure mismatches. In particular, the pathology and oncology departments operate independently, each using distinct data structures and software systems. In addition, Onkostar is a proprietary solution, while our center advocates for flexible open-source options to support a sustainable research environment.

In addition to the complex data collection and integration process for MTB, which includes many different institutions and previously used data collection systems, finding suitable treatment options can be challenging. Conducting literature research for such multifaceted problems with limited patient data is extensive and time-consuming. To streamline this process, our collaboration partner, the German Cancer Research Center (Deutsches Krebsforschungszentrum, DKFZ), and the National Center for Tumor Diseases Heidelberg (NCT) [17] has developed the Knowledge Connector (KC) [18–20], a tool designed to expedite literature research by integrating genetic knowledge with patient data and relevant literature based on genetic alterations. This tool aims to significantly reduce the time required for literature research, which is a major bottleneck in providing personalized treatment options to more patients [21]. In contrast to cBioPortal [22, 23], which is mainly used to explore patient-unspecific genomic profiles, the KC tailors its genomic information to the given patient-specific molecular data [20].

To solve the above-mentioned pressing issues, this paper presents our own open-source web application, Molecular ONcology Optimized CLinical Evaluation (MONOCLE), developed at the University Medical Center Hamburg-Eppendorf (UKE), Germany, in the Institute for Applied Medical Informatics (IAM). MONOCLE, in conjunction with the KC, enhances the workflow and documentation of MTBs, enabling high-quality treatment for more patients. We detail MONOCLE’s structure and software components and evaluate its performance through dedicated user tasks and an updated version of the system usability scale (SUS) [24].

Methods and Materials

Introduction of MTB and “Traditional” Process Analysis

The University Cancer Center Hamburg (UCC Hamburg) at UKE established the MTB in 2016. Thus, all processes and data security measures are fully compliant with the legal and regulatory framework governing German healthcare institutions, with particular adherence to the specific requirements of the federal state of Hamburg. The detailed process analysis is available in Lauk et al. [25].

Initially, each step of the traditional MTB process was manual, involving multiple individual processes across different departments. Patients were admitted to the UKE and registered in the SAP (SAP Deutschland, Germany) system for accounting and controlling. Clinical documentation was managed using Soarian (Cerner, USA), our clinical workplace system. The hospital information system (HIS) therefore consists of both systems. Molecular diagnostics, conducted separately, produced an analysis file detailing genetic mutations and alterations, which was then manually scanned and uploaded to the patient’s Soarian file. Literature research regarding molecular diagnostics was manually performed by the oncologists, consuming significant time, with checking at least three different databases for information and genetic alterations. The MTB conference documentation and reporting were managed using Soarian forms, which did not allow a dedicated overview of the MTB patients, especially if their cases were complex and included many different internal and external departments.

Given Soarian’s closed system, which lacks data export capabilities, participation in the DNPM—requiring pseudonymized data transmission in a distinct format to the central DNPM node—was infeasible. Thus, participating in DNPM itself would not have been possible. In addition, obtaining an overview of MTB patients and the necessary core dataset (CDS) format, integrating genetic and clinical data, was challenging and time-consuming because the data transfer had to be carried out manually and therefore also contained a high potential for errors. One aim was thus to standardize, streamline, and facilitate this process.

Aims and Requirements

There are several requirements that MONOCLE must fulfil. They are specified by the different involved parties, which were assessed through interviews before the requirements were set. Overall, data protection of the patients is the first and most respected requirement for the development of MONOCLE, therefore, user access has to be restricted. The MTB process needs to be in accordance with national and federal law and must safeguard patient data against unauthorized access. The aim of the UCC Hamburg, conversely, was to facilitate and streamline the process of the MTB and thereby reduce the workload for the oncologists. The requirements profile for the web application thus includes a standardization and centralization of the MTB workflow, which includes automatic collection of data from the HIS, and transfer back to HIS after the MTB is completed. The standardization also included the elimination of free-text fields, which was crucial to improve interoperability. The transfer of genetic patient data from the laboratory to the web application is another requirement. It involves automatic integration of gCDS with the cCDS and the transfer of the combined data to the DNPM network node (DNPM:dip), which facilitates, streamlines, and centralizes MTB documentation and provides a clear and organized overview of patient status, particularly for patients undergoing multiple MTB evaluations or receiving treatment both at UKE and from external oncologists. The improvement of literature research can be fulfilled by integrating the KC into the web application. For the development of MONOCLE, facilitating the transfer of the data to the DNMP is the main challenge.

The findability, accessibility, interoperability, reuse (FAIR) principle of the software is another requirement for MONOCLE. To ensure adjustments in the future due to changing legal or other frameworks, a modular structure for MONOCLE is essential. In order to facilitate the use of the software, the implementation of an easily understandable, structured, and standardized user interface (UI) is needed. As most of the process steps included manual work from the oncologists, we sought to decrease their workload by facilitating and standardizing the workflow to allow medical documentation staff to fill out the forms and perform as many steps as possible. The integration of a user-friendly UI while obeying the requirements of data standardization and formatting of the DNPM was a multilayered process that included iterative work with all participating institutions, but first and foremost, the oncological department of our university medical center. The overall requirements for MONOCLE are depicted in Fig. 1.

Fig. 1.

Project requirements for the improved MTB process from medical informatics perspective (red), the DNPM network (yellow), the data protection perspective (green), and the oncologists at the UCC Hamburg (blue). cCDS clinical core dataset, CDS core dataset, DNPM German Network for Personalized Medicine, FAIR findable, accessible, interoperable, reusable, gCDS genetic core dataset, HIS hospital information system, MTB molecular tumor board, nNGM National Network Genomic Medicine, UCC Hamburg University Cancer Center Hamburg

Development of MONOCLE

The development of MONOCLE was a process that required weekly meetings with the different institutions within our medical center, i.e., oncology, (neuro-)pathology, bioinformatics, the development team, the data protection office, and with external ones, i.e., the developers of the centralized data-sharing-platform and other participants of the DNPM. Moreover, we organized multiple end-user workshops to improve the user interface iteratively and, more importantly, the understanding of the process and the cooperation between the oncologists, our institute (IAM), and the other departments involved.

The software was developed by software developers at UKE, taking agile methods into account, where validation of the software was made by the four-eyes principle. This bottom-up approach facilitates development during operation and leads to a high level of user satisfaction, as they are included in the evaluation at regular intervals. It also avoids errors at an early stage and prevents undesirable developments.

Evaluation

The usability of MONOCLE in combination with the KC was assessed using the English version of the SUS [26] by the main users of the tool: one medical documentarian and two oncologists. Each of them evaluated the new process with five different patients. Moreover, the users were asked to give free-text feedback (in German) after performing the following tasks:

1. I can use MONOCLE in the secure hospital information system.

2. I can add a physician’s name (oncologist or pathologist) to the address book.

3. I can automatically send a previously selected sequencing request to the pathologist by choosing the correct name from the address book.

4. I can create a new patient, including all master patient data (name, birthdate, health insurance, etc.).

5. I can create the complete clinical patient case, including diagnoses, pathology reports, former treatments, patient history, etc.

6. I can create the correct dependencies as provided in the CDS data model.

7. I can obtain the full genetic data set from the bioinformatic pipeline to the requested patient.

8. I can see and edit the complete CDS.

9. I can get an overview of completed and open steps for each patient.

10. I can get an overview of all patients and I can filter/sort for different criteria (name, MTB date, birthdate, status, etc.).

11. I can select names from the address book that get updates about every patient step.

12. I can open the selected patient case in the KC via MONOCLE.

13. I can successfully use the KC for my literature research.

14. I can generate an MTB report in the KC and resubmit it to MONOCLE.

15. I can generate an MTB report in MONOCLE in case I do not need the KC.

16. I can transmit the completed CDS to the DNPM network node.

17. I can leave comments and highlight important ones within one patient case for others.

18. I can create a new treatment episode while keeping the “old” patient data.

19. I can see but not edit the frozen first treatment episode of the patient.

20. I can see if the export to the DNPM network node was successful.

After implementing MONOCLE to the clinical infrastructure, a ticket system was established so any problems that arise can be solved immediately.

Code Availability

The code and the data format were published on GitHub https://github.com/UKEIAM/MONOCLE.

Results

New Process

The transition from the traditional process to the optimized, new one was done during ongoing clinical operations in June 2024. We provided frequently asked questions (FAQ), as well as a user manual to enhance the seamless integration of MONOCLE into the clinical routine. The streamlined and centralized process for the MTB is depicted in Fig. 2. The new MONOCLE-enabled process involves several key steps: creating patient records with admission data; documenting core datasets, including the cCDS and the gCDS; assigning genetic data; transferring data from bioinformatics to the gCDS automatically; creating therapy recommendations and the MTB report; forwarding data to the KC; and transmitting patient data to the DNPM network node.

Fig. 2.

Detailed data flow and process description: the documentation staff or the physician logs in to MONOCLE in our protected clinical environment, creates a patient, and documents its cCDS. The bioinformatics facility sends the already processed genetic data to MONOCLE. Then, the physician completes a few manual adjustments of the gCDS. After the full CDS is obtained, the data is sent to the KC within the MONOCLE platform, where the literature research is performed. The MTB report and therapy recommendation can also be written in the KC and then transmitted to MONOCLE. Having the full CDS and the MTB report, the case is sent to the DNPM:DIP node. The DNPM broker is now able to ask statistical questions about the cohort. cCDS clinical core dataset, DNPM German Network for Personalized Medicine, gCDS genetic core dataset, HIS hospital information system, KC Knowledge Connector, MTB molecular tumor board

The MONOCLE system addresses the extract, transform, load (ETL) process and the transfer to the DNPM network node, while MONOCLE’s linkage to the KC improves time-consuming and complex literature research.

User access is established through an administrative protocol managed by KeyCloak, defining precise usage rights and permissions to safeguard patient data against unauthorized access. KeyCloak is an established and widely used tool for authentication and role management [27]. The frontend, including a user-oriented graphical user interface, is available to medical staff in the HIS, while the backend, database, and pseudonymization service are hosted within our protected, institutional research environment. The frontend was coded in ReactJS and communicates via a REST API with the Java backend. The utilized database was MongoDB. The “Mainzelliste”, a self-developed tool, performs the pseudonymization [28, 29]. Java, ReactJS, and MongoDB allow an easily extendable implementation and are commonly used for similar projects in the coding context. Using the Mainzelliste provides the necessary flexibility to customize the pseudonymization to the project-specific requirements, including the distinction between the role and rights of pathologists, documentation staff, oncologists, and IT administration. Technical implementation details can be found in Fig. 3.

Fig. 3.

Technical overview of the MONOCLE software: KeyCloak manages user authentication, with a ReactJS frontend communicating via REST API to a Java backend. MongoDB stores data. Pseudonymization is handled by the Mainzelliste tool. Pseudonymized core data is sent via the DNPM:DIP node. The DNPM queries key figures and parameters. Lines indicate hosting boundaries: physicians’ view, our institute, and pseudonymized data transfer outside of our medical center. DNPM German Network for Personalized Medicine, IAM Institute for Applied Medical Informatics

Medical and clinical records of MTB patients are generated during the admission process, with pertinent patient information meticulously documented and monitored at each stage using the designated tool (Fig. 2). The capture of clinical and genetic core data adheres to DNPM specifications [30], allowing for the customization of MTB reports and therapy recommendations to meet organizational prerequisites. Additional patient information can be incorporated via comment functions. The CDS, MTB reports, and therapy recommendations can be exported from the software to ensure seamless interoperability. Once the oncologist has all the necessary data to provide a personalized treatment option, the KC can be used for literature research via the KC button within MONOCLE. The KC integrates patient-specific clinical knowledge with molecular diagnostics results and information from various knowledge databases (e.g., CiViC, OncoKB). This integration and combination of “world information” with the patient-specific clinical and genetic data aids in identifying potential treatment options by including genetic alterations, potential drugs, affected molecular pathways, evidence levels, and knowledge sources. In addition, MTB oncologists can create and share therapy recommendations—consisting, for example, of molecular alterations, drugs, and evidence levels—with other medical centers, as these recommendations do not contain patient-specific information. Participating centers, next to our university medical center, are the NCTs Heidelberg and Dresden, and the University Medical Center Augsburg. The KC allows the generation of an MTB report with the gathered molecular and drug information [20]. The KC automatically transfers MTB reports to MONOCLE to complete the patient case. In case the KC is not used, the MTB report can be generated, stored, and exported to the UKE’s internal health archive in MONOCLE. After confirming the completeness and correctness of the patient’s dataset, the patient information, consisting of the pseudonymized CDS, encompassing both genetic and clinical data, is sent to the DNPM. The DNPM can assess it for key figure evaluation.

If the patient has to be discussed again, due to e.g., tumor progress or drug intolerance, a new patient episode can be started while keeping the already provided information. Screenshots of MONOCLE can be found in the Supplementary Figs. S1–S4.

Evaluation

The performance evaluation highlights the acceleration, standardization, and facilitation of the process steps before and after implementing MONOCLE and the KC, as shown in Table 1. In a comparison of the traditional and the new process, we were able to automate three steps of the whole process, which had to be done manually in the traditional process, i.e., merge genetic and clinical data, perform the literature research, and document the recommendations of the tumor board. All process steps were either accelerated or standardized, thereby substantially reducing the required steps involving oncologists and reducing the required time. In addition, the transfer of the data to the DNPM is now possible, which was not feasible within the old process.

Table 1.

Qualitative comparison between the process steps and the proposed setup with MONOCLE (and the KC extension). Standardization refers to both the transfer to a standardized format and the internal standardization of our processes

MTB steps	Old process	Process with MONOCLE and the KC	Process improvement
Patient admission at UKE	Manually Documentation in HIS	Manually Documentation in MONOCLE	Acceleration
Documentation of clinical data	Manually Examination of patient and mainly free-text documentation in HIS	Manually Examination of patient and standardized documentation in MONOCLE	Standardization
Molecular diagnostics and documentation	Manually Communication with oncologists and (neuro-)pathology with fax, mail, and phone calls	Automated Bioinformatics pipeline established, data directly associated with specific patients	Acceleration, standardization
Literature research	Manually Literature search could take up to 1–3 h per patient	Automated The KC is used, and an own knowledge base can be built	Acceleration
Therapy recommendation and MTB conference	Manually No distinct overview of patients and MTB conference is available	Automated Overview of molecular alterations and reasons (knowledge bases) for treatment choice	Acceleration, standardization
Transmission of patient data to DNPM:DIP	Not possible No export from Soarian possible, patients in different systems not matched	Automated Transfer after initialization	Acceleration, standardization

DNPM German Network for Personalized Medicine, HIS hospital information system, KC Knowledge Connector, MTB molecular tumor board, UKE University Medical Center Hamburg-Eppendorf

Members of the oncological department, the three main users of MONOCLE, evaluated the web application (2:1 female:male). The SUS ranged between 92.5 and 97.5, showing an excellent rating. The text feedback after the given tasks, which were described in the methods section, indicated that (1) all requested tasks can be performed with MONOCLE and the KC extension, and (2) the graphical UI is intuitive, easy to use, and meets expectations and requirements. Three evaluators noted that the standardization provided by MONOCLE enhanced the documentation process, and integrating the KC significantly reduced the workload for literature research. However, two evaluators felt that structured documentation according to the DNPM core data set increased the documentation workload compared with the previous free-text documentation method.

Discussion

In this study, we introduced and evaluated MONOCLE, a comprehensive open-source web application designed to optimize, accelerate, and standardize the MTB documentation process in alignment with the DNPM initiative’s CDS requirements. Developed by the IAM at UKE, MONOCLE allows and facilitates seamless documentation within the DNPM network, integrating both genetic and clinical datasets.

While MONOCLE streamlines the documentation workflow, the KC significantly enhances the efficiency of literature searches to identify personalized therapeutic options. Linking MONOCLE to this framework resulted in a substantial reduction of steps from patient admission to MTB report generation, as both the oncologists’ documentation time and the literature review time were markedly reduced.

As reflected in user task performance and free-text feedback, both user groups reported improvements in documentation, standardization, and clarity. In addition, the usability and flexibility of MONOCLE’s user interface were positively highlighted, also supported by high SUS scores. As the main user group of MONOCLE is limited, more diverse evaluation methods could not be applied. While the workload for oncologists in terms of documentation was significantly reduced, the introduction of structured documentation initially increased the burden on the documentation staff. This was expected, as structured documentation typically requires more time compared with free text, particularly during the initial phase. Consequently, the documentation staff expressed concerns in the user interviews about the lack of time savings in their workflow. However, after a period of adaptation, the process became more efficient and widely accepted, as standardized forms proved easier to complete and search through than free-text documentation once users received proper training.

The KC offers clinicians a unique capability to obtain a structured synthesis of patient-specific clinical and molecular data, alongside curated publications that support targeted therapeutic strategies based on identified mutations. Importantly, the KC is the first tool to share therapy recommendations across institutions without compromising patient confidentiality. Unlike cBioPortal, an open-source platform primarily focused on exploratory visualization and large-scale cancer genomics datasets, the KC operates at the individual patient level, streamlining the identification of tailored treatment options. Consequently, the KC enhances the generation of MTB reports, which are then seamlessly transmitted to MONOCLE for integration into clinical workflows. Future cooperation or collaboration between the KC and cBioPortal seems beneficial to use the synergies and full potential of both platforms as cBioPortal provides high-quality access to molecular profiles and clinical attributes from large-scale cancer genomics projects.

Existing tumor documentation systems, such as OnkoStar, were not utilized owing to their proprietary nature, which limits adaptability to our specific clinical environment and hampers flexibility in responding to evolving documentation and user interface requirements. MONOCLE was purposefully designed to integrate seamlessly into our medical center’s IT infrastructure, linking key departments including oncology, pathology, bioinformatics, and IAM. Due to confidentiality constraints, we are unable to disclose further details regarding the medical center structure and the corresponding design decisions.

The correct documentation for DNPM represents only one foundational step for the platform. Future plans include extending MONOCLE’s capabilities to support documentation for the pilot project pursuant to §64e SGB V (Modellvorhaben §64e) [10], the National Network Genomic Medicine (nNGM) [31], and other cancer registries. The former is based on comprehensive genome sequencing (short- and long-read sequencing technologies) as part of a structured clinical treatment process and the subsequent consolidation of clinical and genomic data in a data infrastructure that facilitates analysis of the data obtained to improve medical care. This project is to be carried out uniformly throughout Germany [32]. The latter collects all lung cancer cases treated with a personalized therapy approach. To extend MONOCLE for the documentation of §64e and nNGM, we are currently expanding the forms and providing options for physicians or medical documentarians to select one or multiple reporting networks. We will highlight the required fields for each network and develop reporting-specific exporters and interfaces. Central to this development is our existing metadata repository “M5,” which manages and maps different ontologies [28]. Moreover, we plan to automatically transfer the data to the Gießener Tumordokumentationssystem (GTDS) [33] and, subsequently, to the Hamburgisches Krebsregister (HKR) [34] as requested by federal law.

In addition to standardizing tumor documentation as the aim of participation in DNPM, we aim to harmonize the data sources used for this process. Within the Personalized Medicine for Oncology (PM4Onco) [35] consortium, which involves over 20 German university medical centers, we will adopt the FHIR standard [36] and leverage data integration centers (DICs) [37] to harmonize precision oncology efforts at national level. Future developments will also include the extension of the KC with a natural language processing plugin. This will enhance literature research, automate the updating of therapy recommendations by assessing their similarities, contradictions, and dependencies, and ultimately improve the quality of the therapy recommendation repository. Furthermore, we envision advancing the automatic generation of therapy recommendations through web scraping powered by large language models.

Conclusions

The incorporation of MONOCLE with the KC extension into the clinical protocol at the UKE has significantly streamlined and accelerated the workflow of the MTB for oncologists. This integration has facilitated a standardized and homogeneous approach to documentation, yielding enhanced and expedited interdisciplinary collaboration among healthcare professionals participating in the MTB. Consequently, this streamlined process will lead to more patients benefiting from the MTB’s expertise, as the oncological human power required for documentation and literature research is the bottleneck of personalized medicine. Thus, the benefits are twofold: first and foremost, the patient receives new treatment options after an unsuccessful standard therapy approach. Second, the more unified datasets are, the more research regarding targeted treatment options can be performed to consequently improve the therapy approaches based on clinical and genomic data.

Notably, the utility of MONOCLE extends beyond its application within the UKE’s MTB, as it is open-source and aligns with the documentation standards of the DNPM, in which more than 20 German university medical centers are enrolled. Consequently, this tool can be readily adapted and implemented within other medical centers affiliated with the DNPM network. The ongoing extension of the tool to match other national documentation requirements, such as the ones from nNGM, §64e, and GTDS, will further expand the scope of applications. Moreover, MONOCLE, if translated, can be used in an international precision medicine context.

Supplementary Information

Below is the link to the electronic supplementary material.

Funding

Open Access funding enabled and organized by Projekt DEAL.

Declarations

Funding

Open access funding enabled and organized by Projekt DEAL. This study was funded by the DNPM (Deutsches Netzwerk für personalisierte Medizin) with the grant number: DNPM: 01NVF20006.

Conflict of interest

Layla Tabea Riemann, Maximilian Ataian, Felicia P. S. Hähner, Benjamin Roth, Alexander Knurr, Anne Kamitz, Maximilian Christopeit, Carsten Bokemeyer, and Frank Ückert declare that they have no conflicts of interest that might be relevant to the contents of this manuscript.

Data availability

Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.

Ethics approval

MONOCLE was approved by the ethics committee of the Hamburg Chambers of Physicians (approval number: 2023-101098-BO-ff).

Code availability

The code is available at Github https://github.com/UKEIAM/MONOCLE.

Author contributions

Conceptualization: Layla Tabea Riemann, Felicia P. S. Hähner, Maximilian Christopeit, Frank Ückert, and Carsten Bokemeyer. Methodology: Layla Tabea Riemann, Maximilian Ataian, Anne Kamitz, Alexander Knurr, Benjamin Roth, and Felicia P. S. Hähner. Formal analysis and investigation: Layla Tabea Riemann, Anne Kamitz, Alexander Knurr, and Maximilian Christopeit. Writing—original draft preparation: Layla Tabea Riemann. Writing—review and editing: Layla Tabea Riemann, Maximilian Christopeit, Anne Kamitz, Alexander Knurr, Felicia P. S. Hähner, Carsten Bokemeyer, and Frank Ückert. Funding acquisition: Maximilian Christopeit, Carsten Bokemeyer, and Frank Ückert. Resources: Carsten Bokemeyer and Frank Ückert. Supervision: Carsten Bokemeyer and Frank Ückert.

Consent to participate

Not applicable.

Consent for publication

Not applicable.