AN INTEGRATED CLINICAL AND SURGICAL DATA COLLECTION PROTOCOL METHOD FOR ARTIFICIAL INTELLIGENCE-BASED PROGNOSIS AND DECISION SUPPORT IN GLIOBLASTOMA MANAGEMENT.

Abstract

The invention relates to an AI-integrated clinical and surgical data collection protocol method for providing data to AI-based prognosis and decision support systems in glioblastoma patients.

Description

AI-BASED PROGNOSIS IN GLIOBLASTOMA MANAGEMENT AND INTEGRATED CLINICAL AND SURGICAL DATA COLLECTION FOR DECISION SUPPORT. PROTOCOL METHOD Technical Area This invention is a standardized solution specifically developed for glioblastoma patients. This relates to clinical and surgical data collection protocol methods. State of the Art Currently, data on glioblastoma patients (clinical, imaging and biological) data), from different sources, on different systems, in different formats and at different times. It is being collected without any standardization and there are data gaps. Therefore, artificial intelligence models often fall short of being "generalizable". Current clinical practices, particularly in the field of neurosurgery, are compatible with AI (Artificial Intelligence). and longitudinal data acquisition processes - from the same patient or the same sample, different Collection of data repeatedly at time points - a protocol that standardizes it However, surgically obtained tissue samples are mostly not available. Used solely for histopathological diagnosis; these samples include genomic and transcriptomic specimens. or is not processed in a standardized way for AI-based analyses. During the clinical follow-up process, neurological examinations are performed irregularly, and neurocognitive However, these assessments are rarely performed. Imaging data are generally used in routine clinical practice. The samples are taken for illustrative purposes and are insufficient for radiographic analysis or data labeling. Blood samples are often not taken for biomarker studies or longitudinal studies. The data is not collected (longitudinally). Therefore, the available data is supported by artificial intelligence. It is not suitable for training models. In the current literature, some multicenter databases and biobanks (e.g., TCGA, Ivy) are discussed. Even though glioblastoma atlas is found; - This data is mostly retrospective, and we aim to collect prospective, surgically focused data. It does not include. - Tissue, blood, imaging, and clinical data were collected synchronously from the same patient. It is not. 1 - Standardization and thorough monitoring required for training artificial intelligence models, for example. Elements are missing. In addition, inventions examined in existing patent databases (for example: AI-assisted brain) tumor classification system, MRI-based progression prediction systems) are mostly single It is based on modality (e.g., display only) and multi-modal data integration, It does not involve clinical follow-up or intraoperative biological sampling. In other words, Existing data sources are fragmented, lack standardization, and are often spatially inconsistent. labeled tissue samples, serial neurocognitive assessments, or liquid biopsy data It lacks critical components such as these. This fragmented structure is not reliable for glioblastoma and This hinders the development of generalizable artificial intelligence models. The current technology addresses the following scientific problems, and solutions to these problems... should be brought:  Fragmented clinical and biological data,  Lack of systematic neurocognitive monitoring,  Failure to integrate biomarkers from imaging, tissue, and peripheral blood,  Data sets of sufficient quality for training artificial intelligence (AI) models absence  Lack of trust in clinical decision support systems. Because of these shortcomings, the artificial intelligence models developed often fail to meet the standards of real clinical practice. under these conditions, it lacks generalizability and is not at a sufficient level in terms of clinical safety. It is not accessible. Existing publications and patents generally focus on a single modality (e.g., MRI only). (imaging or genomic analysis) focuses on and is integrated with the surgical process. A multimodal, longitudinal clinical data collection protocol has not been defined. Here are some similar examples:  TCGA Glioblastoma Dataset: This is a comprehensive genetic dataset, but clinical follow-up, It does not include cognitive assessment and tissue sampling standards.  Ivy Glioblastoma Atlas Project: Includes radiological and molecular data, but It is not integrated with longitudinal clinical follow-up and neurocognitive data. 2  Multimodal data integration in brain tumors, especially glioblastoma, AI-powered diagnostic and prognosis systems and bio-digital data harmonization There are several important institutions and companies working in these fields. However, these The work of institutions generally begins with surgery, as offered by the invention. AI that integrates longitudinal clinical, cognitive, biomarker, and imaging data. It does not offer a structure at the clinical protocol level that is compatible with it. Academic / Research Institutions:  DKFZ (German Cancer Research Center) – Particularly glioma biomarkers and Large-scale data projects working on radiogenomics (e.g., EPI-GBM) It is carrying out.  Allen Institute (USA) – Molecular-anatomical study with the Ivy Glioblastoma Atlas Project It conducts mapping studies.  Massachusetts General Hospital – Martinos Center for Biomedical Imaging – MR It focuses on the integration of images and genomic data with AI. Companies and Startups:  Owkin (France / USA): Specializes in AI modeling with multimodal data, It works with glioma datasets.  Aiforia (Finland): AI-based diagnostic support systems in pathology images. It is being developed and is also used in some brain tumor studies.  Synaptive Medical (Canada): AI-based imaging specifically for neurosurgery. It offers systems and intraoperative guidance solutions. However: None of these institutions and companies have yet expanded beyond surgery to include tissue and fluid biomarkers, synchronize radiological imaging, neurocognitive monitoring, and clinical course data. prospectively collecting data and directly using this data for training AI systems and clinical decision-making. a protocol or system that standardizes in a way that can be transferred to support tools It has not been defined. In this respect, the invention goes beyond existing studies and is innovative. It offers. 3 In summary, current clinical follow-up approaches include clinical examination findings and imaging. data and molecular biomarkers at different time points, in different systems, and in different data is obtained in various formats, and this data is presented in a holistic and comparative manner. Evaluating it presents technical difficulties. This is especially true for artificial intelligence. In data-based analyses, problems such as temporal discrepancies, data loss, and non-standard inputs can occur. This leads to... Brief Description of the Invention The primary aim of this invention is to improve the diagnosis and monitoring of high-grade brain tumors such as glioblastoma. and to support treatment planning; intraoperative tissue sampling, longitudinal clinical follow-up, neurocognitive assessments, imaging, and peripheral standardized system that enables integrated collection of biomarker data in blood. The goal is to develop a clinical and surgical protocol method. In other words, this invention, A standardized clinical and surgical approach specifically developed for glioblastoma patients. This relates to the data collection protocol. This protocol is for AI-powered diagnosis and decision support. This provides a foundation for training these systems with high accuracy and their safe application in the clinic. prepares. Technical Solutions Provided by the Invention  Radiological examinations at specific times (0, 3, 6, 12 months) during the pre-operative and post-operative periods. images, peripheral blood biomarker analyses, and clinical follow-up data are synchronized. timely and organized collection,  All data is labeled in a way that is compatible with AI (Artificial Intelligence) and research data It is transferred to the base. Innovations it brought to the sector  One of the first clinical protocols to provide multidisciplinary and multimodal data integration. one of them.  An innovative approach that makes clinical monitoring and biological sample collection AI-compatible. It is an approach.  Surgical, oncological, neurological, radiological, and psychometric assessments all in one place. It combines under the chart. 4 The goal is to develop high-quality, AI-based diagnostic and prognostic models to support these technologies. The goal is to obtain integrated datasets. The protocol involves intraoperative tumor tissue analysis. sampling, biomarkers obtained from peripheral blood, radiological imaging, including patients' life expectancy, neurological examinations, and neurocognitive assessments. This involves synchronizing longitudinal tracking. The fundamental technical problem that this invention solves is the monitoring of glioblastoma patients from diagnosis to follow-up. Holistic, high-quality, and multimodal clinical approach encompassing the entire patient journey. The problem is the lack of data sets. This invention has been developed:  The tumor tissue surgically removed is verified using radiological and molecular data. It enables timely matching.  Radiological imaging, peripheral blood sampling, neurocognitive and neurological Standardize assessment timelines (e.g., pre-operative, 3rd, 6th, and 12th months) does,  Imaging (MRI, CT), pathology, genomics, and biomarkers in peripheral blood It offers a structured template that facilitates the integration of data,  It enables the integration of data into AI models for diagnostic and prognostic purposes. This protocol ensures data integrity, clinical traceability, and technical compatibility with AI. In this way, a significant gap in neuro-oncology is being closed and brought into clinical practice. This lays the groundwork for the development of transferable and explainable AI solutions. In this context, the invention is a systematic and longitudinal progression of data that begins during surgery. It is the first multidisciplinary protocol methodology targeting integration and clinical impact. Currently available. Going beyond solutions, an AI-compatible, prospective and multimodal patient dataset. It makes its production possible. Therefore, the discovery is particularly relevant for high-risk brain tumors, especially glioblastoma. Developed for use in the diagnosis, treatment, and follow-up processes of grade 1 glial tumors. It is a multimodal clinical data collection and integration protocol method. The invention is technically directly related to the following areas:  Neurosurgery (Intraoperative tissue sampling and postoperative patient follow-up)  Neuro-oncology (Holistic evaluation of the biological and clinical characteristics of the tumor)  Artificial Intelligence-Assisted Medicine (Standardized data infrastructure for training AI models) (creation)  Biobank Management (Prospective archiving of tissue, blood, and data samples)  Radiomics and Image Processing Systems (Data from pre- and post-surgical images) (included in the set)  Diagnostic and Decision Support Systems (AI-based algorithms to be used in the clinic) nutrition) Advantages:  Provides standardization: Applicable in different centers, data that does not change per patient. It creates collection processes.  Performs multimodal data integration: Tissue, blood, imaging, and neurocognitive data. The data is collected simultaneously and completely from the same patient.  Optimized for Artificial Intelligence (AI): Labeling, scheduling, formatting, and integrity. From this perspective, it produces high-quality data compatible with artificial intelligence algorithms.  Patient-centered: Prioritizing clinical benefit throughout the entire process from diagnosis to follow-up, genuine It has an approach that incorporates life data.  Adaptable and scalable: Suitable for other brain tumors, pediatric tumors, and different types of tumors. They can be easily integrated into data collection systems in the centers. Unique Elements:  The first clinical protocol to directly link intraoperative sampling with AI modeling. It is its structure.  Neurocognitive follow-up data (MoCA, KPS, etc.) in time series from tissue samples and samples. It is the first systematic model to integrate with biomarker data.  Tumor biology + clinical course + cognitive status + imaging + liquid biopsy, etc. There are very few examples of synchronously collecting five different modalities for the same patient. one of them.  Enables the integration of artificial intelligence outputs into clinical decision support systems. It provides a data infrastructure that ensures ethical and legal compliance.  Key clinical data of international initiatives with high impact potential It has the potential to form the infrastructure. 6 The most striking feature of the invention is its ability to improve intraoperative surgical procedures in glioblastoma patients. encompassing the entire patient journey, from initial assessment to neurocognitive and biological monitoring. It is the first to offer an integrated clinical data protocol method. Thanks to this protocol method... In summary:  Tissue, blood, imaging, and cognitive data from the same patient, on the same timeline. They gather together in a harmonious way.  This data is in a format that can be directly transferred to systems that will be trained with artificial intelligence, and It is prepared in the standardization process.  In this way, the artificial intelligence models that will be developed will be accurate not only in terms of technical correctness. not only that, but it also becomes reliable and explainable for clinical decision-makers. In short, this breakthrough directly links clinical reality with artificial intelligence modeling. It is the first structured protocol method that connects prospective and patient-centered individuals. Detailed Description of the Invention Data for AI-based prognosis and decision support systems in glioblastoma patients. This protocol method, developed to provide this, is applicable in clinical, surgical, molecular, imaging and Prospective, time-mapped analysis of liquid biopsy data with standard labeling structure. It envisages the collection of data used in artificial intelligence algorithms. Data preprocessing is improved by increasing the completeness and multimodal integration level of the sets. The aim is to reduce the need for analysis and increase the efficiency of analysis. The invention's method generally involves the following steps: The first step is to inform the patient before the operation and obtain written consent. followed by clinical performance (KPS) and neurocognitive (MoCA) assessments This involves obtaining pre-operative MRI / CT images and analyzing this data, at least processed by at least one control unit that includes a microprocessor, memory and data interface It will be digitally recorded on at least one electronic device. In the second step, tissue samples are taken during surgery, with at least one neuronavigation guided by a probe; Sampling time, anatomical location, imaging coordinates, pathological diagnosis, and Time-locked, spatial data that includes molecular analysis results. contains coordinate information and is related to clinical, imaging and molecular data. 7 It is labeled using a layered, structured, and standardized data model. The technical novelty of the invention lies in the fact that the tissue samples mentioned in the second step can only be placed in a biobank. Not only is it limited to purposeful storage, but it is also artificially enhanced by associating it with neuronavigation data. standard, machine-readable and designed to provide direct input to intelligence algorithms. It is the labeling in a format compatible with the algorithm. In the third step, miRNA, cell-free DNA (cfDNA), and metabolomics were administered preoperatively. Plasma and serum samples for analysis, integrated with the clinical process and prospective liquid biopsy. This will be done within the scope of the approach. Blood samples taken will be analyzed for miRNA, cfDNA, and metabolomics. predefined molecular processing protocols suitable for the analyses will be processed accordingly. In this context, the examples in question are; centrifugation, Following fractionation and stabilization steps, isolation and stabilization of miRNA and cfDNA were performed. purification, and for metabolomic analyses, small molecule extraction and normalization. They will be subjected to processing. The obtained miRNA, cfDNA and metabolomic data will be analyzed numerically. and standardized digital data formats while preserving their structural characteristics will be controlled in a way that can be transformed and processed by artificial intelligence algorithms. It will be recorded in an electronic data medium via the unit. An integration software system; the subject of the invention is a clinical data collection method. To implement this, you need a clinical data collection module, a biobank sample management module, and It includes an artificial intelligence data labeling module. The fourth step is in the early post-operative period, preferably 0–1 month following the operation. During this procedure, the patient undergoes MRI imaging and a neurological examination. Postoperative MRI images obtained at this stage for pathology and molecular analysis. The results are not limited to evaluation for clinical follow-up purposes only, but also include pre-operative assessments. compared with clinical, imaging and neurocognitive data obtained during the period AI-powered information by at least one control unit for analysis. It will be technically integrated into the system. The integration mentioned in the fourth step involves post-operative MRI images. Digital data of pathology and molecular analysis outputs (e.g., in DICOM format). The data is received by the control unit and processed using predefined data processing methods. Standardization according to instructions, patient identification information and time parameters 8 by pairing it with, and making it processable by artificial intelligence algorithms, centrally It involves the steps of recording data in an electronic medium. Thanks to this technical integration, multiple results are obtained in the early post-operative period. Parameterized data are used not only for individual assessments but also for different time points. AI-based integration between acquired clinical, imaging, and molecular data. to create an infrastructure that will enable comparative analysis will be structured. The fifth step involves post-operative longitudinal clinical and cognitive follow-up. During this period, specifically at months 3, 6, and 12, the same patient again tested positive for KPS, MoCA, and neurological abnormalities. Examinations should be conducted according to a standard protocol, and MRI and blood tests should be performed at the same time. Coordinated re-taking of samples, In the sixth step, all clinical, imaging, tissue, and blood data belonging to the patient are collected; different Although they were obtained from different sources and at different times, at least one check was performed. It will be time-locked and connected to a common time reference via the unit, and thus Cognitive tests, MRI images, and biological sample data represent the same point in time. The data will be synchronized in such a way that artificial intelligence will be used to process it. will be labeled in standardized data formats compatible with algorithms and clinical, biological and imaging data in a single integrated data structure in a multimodal way. It will be integrated within. This integration only involves storing the data together. No, it's a comparative analysis based on temporal consistency between different data modalities. It will be structured in a way that will enable analysis to be performed. This step is carried out. Time-locked multimodal integration; cognitive test data, MRI imaging data, and simultaneously for AI-based analysis processes of blood / tissue-derived biological data In terms of creating an infrastructure that technically enables it to connect to the window, It differs from the single or sequential data evaluation approaches found in the literature. Thus, these datasets are processed temporally by artificial intelligence algorithms. The inputs will be processed as compatible and mutually related inputs. The method For reliable and repeatable operation, in terms of each data modality AI-based analysis when a predefined minimum number of data points is reached. The processes will be initiated, and the system will be scalable to accommodate the increasing number of patients and data. It will be structured in this context; the method will involve initial analysis with limited data sets. 9 This allows for the performance of the model, but as the amount of data increases, the model's performance improves. It will operate on a gradual learning and updating basis, increasing its effectiveness. The control unit mentioned in the first step is suitable for running artificial intelligence algorithms. a hardware architecture containing a microprocessor and / or graphics processing unit (GPU), used in clinical and It is configured for processing and recording imaging data. In the third step, miRNA, cell-free DNA (cfDNA), and metabolomics were administered preoperatively. Plasma and serum samples for analysis, integrated with the clinical process and prospective liquid biopsy. This will be done within the scope of the approach. Blood samples taken will be analyzed for miRNA, cfDNA, and metabolomics. predefined molecular processing protocols suitable for the analyses will be processed accordingly. In this context, the examples in question are; centrifugation, Following fractionation and stabilization steps, isolation and stabilization of miRNA and cfDNA were performed. purification, and for metabolomic analyses, small molecule extraction and normalization. They will be subjected to processing. The obtained miRNA, cfDNA and metabolomic data will be analyzed numerically. and standardized digital data formats while preserving their structural characteristics will be controlled in a way that can be transformed and processed by artificial intelligence algorithms. It will be recorded in an electronic data medium via the unit. Therefore, in glioblastoma patients, different solid tumor types and hematological Data for AI-based prognosis and decision support systems in patients with malignancies The invention aims to provide clinical and surgical data collection integrated with artificial intelligence. protocol method; - The first step is informing the patient before the operation and obtaining written consent. Clinical performance (KPS) and neurocognitive (MoCA) after administration evaluations are performed and pre-operative MRI / CT images are obtained and recording by at least one control unit to at least one electronic device ensuring that patient examinations are video recorded, - In the second step, during surgery, the procedure is performed with the assistance of at least one neuronavigation device. tissue samples; sampling time, anatomical location, imaging including coordinates, pathological diagnosis and molecular analysis results, time-mapped, containing spatial coordinate information, and clinical, imaging a multilayered, structured, and standardized data model associated with molecular data labeling using, - In the third step, miRNA, cell-free DNA (cfDNA) and Plasma and serum samples for metabolomic analyses are integrated into the clinical process. Within the scope of a prospective liquid biopsy approach, the blood samples taken, Pre-prepared to be suitable for miRNA, cfDNA and metabolomic analyses. processing in accordance with defined molecular processing protocols, The samples, following the centrifugation, fractionation and stabilization steps, Isolation and purification of miRNA and cfDNA, small for metabolomic analyses. The resulting molecules are subjected to extraction and normalization processes. The numerical and structural characteristics of miRNA, cfDNA, and metabolomic data will be preserved. converting them into standardized digital data formats and artificial intelligence via the control unit in a way that can be processed by algorithms recording in an electronic data medium, - In the fourth step, in the early post-operative period, preferably 0–1 month following the operation including the patient's MRI scan and neurological examination performed and with the post-operative MRI images obtained at this stage Pathology and molecular analysis results are used solely for clinical follow-up purposes. evaluation and clinical findings obtained in the preoperative period, to be analyzed comparatively with imaging and neurocognitive data technically an AI-powered information system controlled by a small number of control units integration, - In the fifth step, longitudinal clinical and cognitive follow-up in the post-operative period. That is, repeat KPS, MoCA and neurological examinations in the same patient at 3, 6 and 12 months. implementation according to standard protocol and at least one MRI and blood test within the same time frame. the sample being taken again in a coordinated manner, - In the sixth step, all clinical, imaging, tissue and blood data of the patient are collected; in different Although they were obtained from different sources and at different times, the most important ones in question by connecting to a common time reference through a small control unit and thus cognitive tests, MRI images, and biological sample data are all considered simultaneously. synchronization to represent the point in question, data standardized in a way that is compatible with artificial intelligence algorithms. labeling in various formats and the distribution of clinical, biological and imaging data. integrated in a modular way within a single integrated data structure and this integration based on temporal consistency between different data modalities structured in a way that will enable comparative analysis 11 the steps include at least one clinical data collection module and / or at least one biobank sample. using a management module and / or at least one AI data labeling module It includes the steps for implementation. An integration software system; the subject of the invention is a clinical data collection method. To implement this, you need a clinical data collection module, a biobank sample management module, and It includes an artificial intelligence data labeling module. These modules enable multimodal integration. It is sufficient for implementation. However, the system must be reliable on a clinical scale and For sustainable operation, preferably workflow / scheduling, data integration (PACS, etc.), additional techniques such as identity concealment and security, data quality control, and analytics / AI implementation. Modules can also be used. In the presented method, the timing of data collection in clinical practice depends on the individual physician. by taking into consideration the patient's clinical condition, treatment plan and follow-up requirements This is determined. However, the longitudinal data collection proposed in the method... In order to standardize the approach, the early postoperative period (0–1 Reference time points such as the 3rd, 6th, and 12th months are included in the protocol. These time points are defined. These are not a mandatory and automatic calendar mechanism, These are recommended follow-up windows that facilitate compliance in clinical practice. The integration software system described in the method is for clinical appointment scheduling or not for automatic timing; but for clinical, imaging, tissue and recording, labeling, and managing biological data in a standard format at different times Data from various points are linked to a common time reference in a multimodal manner. It is designed for integration. Therefore, automation, data management and a clinic that is at the integration level and determines the patient call or sample collection time. This is not a decision support or appointment system. This approach is different from the method. increasing its applicability in centers and different clinical settings; at the same time while preserving physicians' clinical discretion, the data obtained is processed by artificial intelligence algorithms. It ensures that it is standardized in a way that is compatible with [the standards]. The presented method is not limited to a specific tumor type; it incorporates clinical, imaging, and molecular approaches. and time-locked and multimodal integration of biological data based on artificial intelligence. because it offers a technical infrastructure based on different solid tumor types and hematological It is adaptable for malignancies. 12 In this context, the imaging modalities used depend on the type of tumor to be examined. Although molecular markers and biological sample types may vary, the method... data collection, standardization, time matching and artificial intelligence-based The analysis steps will be carried out within the framework of the same technical principles. The proposed method prevents data loss and incomplete data generation through a multi-stage technique. This will be ensured through a control and verification mechanism. Within this scope, the clinic, Image and molecular data are checked by at least one control unit at the time they are obtained. Data will be collected digitally and recorded in a central electronic data environment. Data collection During the process, each dataset includes patient identification information, data type, and time parameters. The control unit will automatically match any missing, incorrect, or incompatible data entries. It will be identified and marked by [the relevant authority]. In case of missing data, for the relevant data type... Remeasurement, resampling, or the use of alternative data sources will be implemented. The system will be configured in this way. Furthermore, by artificial intelligence algorithms... It can work in situations where there is missing or partial data during the analyses performed. Flexible analysis mechanisms based on data completion, weighting, or modality. This will be used to ensure the method's functionality and analysis despite data loss. continuity will be preserved. In summary, this invention is a step forward in the development of treatments for glioblastoma and similar high-grade brain tumors. It is a multidisciplinary clinical protocol that begins with the surgical procedure and is long-term. the first to include tracking and be directly integrated with AI-powered analytics This is one example. Explanations of the technical terms mentioned in the invention method:  KPS (Karnofsky Performance Scale): An indicator of the level of functional independence. Neuro-oncological performance score.  MoCA (Montreal Cognitive Assessment): Measures memory, attention, and executive functions. cognitive screening test.  Time-Locked Data: Data from different sources are displayed at the same point in time. alignment. Industrial Applicability of the Invention 13 The invention is a clinically standardized solution specifically developed for glioblastoma patients. and is related to surgical data collection protocol methods, and is applicable to the pharmaceutical industry. It is of a certain quality. The developed protocol method can be implemented with at least one electronic device; hospitals, private healthcare institutions, clinical research centers, AI-powered medical software It can be directly adapted and used by companies and biotechnology companies. Application Format: Application as a Clinical Protocol Method in Hospitals:  By the neurosurgery, neurology, oncology and radiology departments together integrated as a point-of-care clinical practice protocol method that can be used It is done.  Surgical teams collect samples from the biobank using standard intraoperative sampling methods. It collects suitable tissue.  Neurocognitive assessments are systematically conducted by the neuropsychology unit. It is applied. Integrated with Data Collection and Integration Software:  The invention integrates into clinical decision support systems and electronic patient record software. It can be done.  Healthcare technology companies can use this protocol for AI-based systems. It can create high-quality datasets. Licensing to Biotechnology and Artificial Intelligence Companies:  This protocol method, which will provide data for training the algorithms to be developed, AI It can be licensed to companies that produce diagnostic / treatment software based on diagnostic principles.  It also provides data standardization for companies conducting liquid biomarker research. Use in Multicenter Clinical Trials:  Multicenter data from prospective clinical trials at national and international levels It can be applied as a collection standard. 14  It can be integrated into framework programs such as EUCAIM, Horizon Europe, and similar programs. The method described in the invention is suitable not only for academic research but also for medical devices. manufacturers, artificial intelligence software companies, biotechnology companies, and hospital information management companies. It is also adaptable by systems. The protocol method that is the subject of the invention is research-based. It can be used in clinical projects, as well as by healthcare technology companies developing artificial intelligence. It can also be used by university hospitals and biotechnology companies. The invention is not limited to the above descriptions; a person skilled in the field can easily create the invention. It can present different applications. These are the claims of the invention and the protection sought. should be evaluated within this context.

Claims

1. Glioblastoma patients may have different types of solid tumors and hematological malignancies. To provide data for AI-based prognosis and decision support systems in patients. in order to; - The first step is informing the patient before the operation and obtaining written consent. Clinical performance (KPS) and neurocognitive (MoCA) after administration evaluations are performed and pre-operative MRI / CT images are obtained and recording by at least one control unit to at least one electronic device ensuring that patient examinations are video recorded, - In the second step, during surgery, the procedure is performed with the assistance of at least one neuronavigation device. tissue samples; sampling time, anatomical location, imaging including coordinates, pathological diagnosis and molecular analysis results, time-mapped, containing spatial coordinate information, and clinical, imaging a multilayered, structured, and standardized data model associated with molecular data labeling using, - In the third step, miRNA, cell-free DNA (cfDNA) and Plasma and serum samples for metabolomic analyses are integrated into the clinical process. Within the scope of a prospective liquid biopsy approach, the blood samples taken, Pre-prepared to be suitable for miRNA, cfDNA and metabolomic analyses. processing in accordance with defined molecular processing protocols, The samples, following the centrifugation, fractionation and stabilization steps, Isolation and purification of miRNA and cfDNA, small for metabolomic analyses. The resulting molecules are subjected to extraction and normalization processes. The numerical and structural characteristics of miRNA, cfDNA, and metabolomic data will be preserved. converting them into standardized digital data formats and artificial intelligence via the control unit in a way that can be processed by algorithms recording in an electronic data medium, - In the fourth step, in the early post-operative period, preferably 0–1 month following the operation including the patient's MRI scan and neurological examination performed and with the post-operative MRI images obtained at this stage Pathology and molecular analysis results are used solely for clinical follow-up purposes. evaluation and clinical findings obtained in the preoperative period, to be analyzed comparatively with imaging and neurocognitive data 16 technically an AI-powered information system controlled by a small number of control units integration, - In the fifth step, longitudinal clinical and cognitive follow-up in the post-operative period. That is, repeat KPS, MoCA and neurological examinations in the same patient at 3, 6 and 12 months. implementation according to standard protocol and at least one MRI and blood test within the same time frame. the sample being taken again in a coordinated manner, - In the sixth step, all clinical, imaging, tissue and blood data of the patient are collected; in different Although they were obtained from different sources and at different times, the most important ones in question by connecting to a common time reference through a small control unit and thus cognitive tests, MRI images, and biological sample data are all considered simultaneously. synchronization to represent the point in question, data standardized in a way that is compatible with artificial intelligence algorithms. labeling in various formats and the distribution of clinical, biological and imaging data. integrated in a modular way within a single integrated data structure and this integration based on temporal consistency between different data modalities structured in a way that will enable comparative analysis the steps include at least one clinical data collection module and / or at least one biobank sample. using a management module and / or at least one AI data labeling module Clinical and surgical data integrated with artificial intelligence, characterized by its application. collection protocol method.

2. The integration mentioned in the fourth step involves post-operative MRI images and pathology and the molecular analysis outputs are received as digital data by the control unit, this data in accordance with predefined data processing instructions standardization, matching with patient identification information and time parameters, and by being made processable by artificial intelligence algorithms and centralized electronically Request 1 is characterized by including the steps for recording data to the data environment. a clinical and surgical data collection protocol method integrated with artificial intelligence, for example. 17