Method and system for providing observable measurements indicative of medical diagnostics

Abstract

Computer-implemented methods and systems are provided for providing observable measurements indicative of medical diagnostics. The method / system implements multiple steps. One step is directed to obtaining an image data sequence of a patient, the image data sequence having a plurality of medical image data sets which are recorded in each case at different times in a time period. Another step is directed to extracting a temporal sequence from a sequence of medical image data. Another step is directed to determining observable measurements based on the time series. Another step is directed to providing a considerable measurement.

Description

Technical Field

[0001] Embodiments of the present invention relate to methods and systems configured for deriving observables that can indicate a medical diagnosis based on time series analysis. Furthermore, embodiments of the present invention relate to methods and systems for deriving such time series from raw data, such as, for example, medical image data or other patient data, selecting and comparing or associating such time series with each other in order to provide observables. Furthermore, embodiments of the present invention relate to the use of observables in further clinical processes, such as, for example, in creating a medical assessment report, (partially) automatically making a medical diagnosis or finding treatment options, respectively, based on the observables. Background Art

[0002] In medical diagnosis, the time course is of great importance. Therefore, often not only the current value of a parameter is decisive, but also how the value develops. For example, a lesion that shrinks will have a completely different diagnostic indication than a lesion that grows or is detected for the first time.

[0003] However, since each value is often visible in the context of other values, correlation analysis becomes complicated. For example, a patient's weight loss may be classified as non-critical based on other values, or indicate a worsening of the patient's condition. More generally, for example, an episode of an increased or decreased first value that coincides with an episode of an increased or decreased second value may indicate a completely different disease pattern than in the case of a constant second value.

[0004] The problem in the practical implementation of this knowledge is that this information is usually not accessible precisely when evaluating medical image data. Therefore, the doctor / doctor who performs the evaluation is usually limited to simultaneously (nebeneinander) viewing image data sets recorded at different times. Quantitative measurements are usually not acquired. Incorporation of other time series outside of radiology usually fails due to the accessibility of this data and the only limited time available for the evaluation. Therefore, in the clinical routine, there is often not enough time to load the individual documents from the patient's file in order to check whether they contain information important for the evaluation.

[0005] Another problem arises because data from different medical disciplines, such as radiology, laboratory or pathology, are largely acquired and viewed separately. Although knowledge from different disciplines is collected in the case of complex diseases, this usually takes the form of individual presentations for different aspects. That is, the radiologist reports the radiological assessment, the pathologist reports the pathological assessment, and the internist reports the patient's symptoms. This makes a systematic, i.e. quantitative and possibly comparative analysis of the time course of the individual values ​​difficult. Summary of the invention

[0006] Against this background, the object is to provide a method and a device which assists the user in the assessment of a patient and which in particular simplifies the consideration of the time course in clinical routines.

[0007] The above-mentioned and other objects are achieved by means of a computer-implemented method for providing an observable quantity indicative of a medical diagnosis, a system for providing an observable quantity indicative of a medical diagnosis, a computer program product or a computer-readable storage medium according to the present application. Advantageous refinements are described in the following description.

[0008] The solution according to the invention for the purpose is described below not only with respect to the claimed system, but also with respect to the claimed method. The features, advantages or alternative embodiments / aspects mentioned here can also be transferred to other claimed subjects, and vice versa. In other words, the physical embodiments (which are for example directed to the system) can also be improved by combining the features described or claimed in the method. The corresponding functional features of the method can be formed by corresponding physical modules.

[0009] According to one aspect, a computer-implemented method for providing an observable quantity indicative of a medical diagnosis is provided. The method has a plurality of steps. One step is directed to obtaining an image data sequence of a patient, wherein the image data sequence has a plurality of medical image data sets, the medical image data sets being recorded at different moments in a time period. Another step is directed to extracting a time series from the medical image data sequence. Another step is directed to determining an observable quantity based on the time series.

[0010] Another step is directed to providing observables.

[0011] Medical image data sets may generally be data sets recorded with the aid of a medical imaging modality. A medical image data set comprises image data for imaging a patient's body, a body part or a body portion. The medical image data sets of an image data sequence may each essentially show the same body, the same body part or the same body portion, but recorded at different times.

[0012] The image data may relate to medical image data having two or three spatial dimensions. A medical image data set comprises image data in the form of a two-dimensional or three-dimensional arrangement of pixels or voxels, for example. Such an array of pixels or voxels may represent color, intensity, absorption or other parameters as a function of two-dimensional or three-dimensional position and may be obtained, for example, by suitable processing of measurement signals obtained by a medical imaging modality.

[0013] Imaging modalities may include, for example, computed tomography scanners, magnetic resonance scanners, x-ray scanners, ultrasound scanners, etc. Image data acquired with the aid of such or similar modalities are also referred to as radiological image data.

[0014] The medical image data set may be formatted in a standard image format, such as the Digital Imaging and Communications in Medicine (DICOM) format, for example, according to the standard DICOM PS3.12020c (or a later or earlier version of the standard). The medical image data set may be stored in an electronic archive or storage device, such as, for example, a picture archiving and communication system (English: Picture Archiving and Communication System, PACS).

[0015] The time period may be preset and may be, for example, weeks, months or days. According to some examples, the time period may be preset based on the background information described herein. The image data sequence has at least two, preferably more than two, image data sets recorded within the time period.

[0016] According to some examples, obtaining the image data sequence may include calling up an image data set from an electronic storage device. In particular, this may include calling up a further image data set based on the image data set to be evaluated. The image data set to be evaluated may be the temporally latest image data set of the image data sequence. The further image data set may, for example, be automatically selected from a plurality of image data sets of the patient in the storage device. According to some examples, this may be done with the following provision: the further image data set "matches" the image data set to be evaluated in terms of content and / or in terms of time, for example in such a way that the further image data set essentially shows the same body, the same body part or the same body portion as the image data set to be evaluated or is recorded within the time period.

[0017] The time series may include a series of measured values ​​detected at different times. In particular, the measured values ​​of the time series may be of the same type. "Same type" may mean here in particular that the measured values ​​are measured values ​​of the same measured variable. In particular, the time series may be a time series of measured variables associated with a patient (or a patient's image data sequence). In particular, the measured variables may be medical or physiological measured variables, such as, for example, abnormal variables or body composition of the patient's body. The measured values ​​are based on an image data set or are derived from an image data set. Here, each measured value of the time series may correspond to one, in particular exactly one, image data set or may be derived therefrom.

[0018] The measured values ​​of the first time series can be derived in particular from the image data of the image data set. In particular, the measured values ​​can be derived or extracted from the image data set by means of image processing or image data analysis.

[0019] According to some examples, the step of extracting comprises applying an image data analysis function to a medical image data set of the medical image data sequence, wherein the image data analysis function is configured to extract image data based measurements from the medical image data set, and the time series has image data based measurements.

[0020] According to some examples, the observable comprises or is based on the temporal behavior of the first time series. In particular, the observable may be a measure for the temporal behavior of the first time series. In particular, the observable may comprise or be a measure for a rise or fall of the first time series. Alternatively, the observable may comprise or be based on or be a measure for the temporal behavior of the first time series relative to one or more further time series.

[0021] According to some examples, providing the observable comprises outputting the observable to a user. The user may be, for example, a doctor performing the assessment, in particular a radiologist. The output may be performed, for example, via, in particular, a graphical user interface.

[0022] In addition, providing the observables may include providing the observables for further processing. The further processing may include generating results that are available to the user in a diagnostic process based on the observables. For example, this may include a medical assessment report or a portion of such a medical assessment report or a treatment recommendation.

[0023] By automatically extracting the time series from a series of medical image data sets, the user is firstly relieved of the work of comparing the individual image data sets and manually extracting the measured values ​​for comparison. By determining the observables on this basis, the time series is also automatically analyzed as to whether a medical diagnosis is indicated in the time series. As a result, the user is not only supported in the assessment of the patient, but also the consideration of the time course in the clinical routine is facilitated.

[0024] According to one aspect, the step of determining the observable comprises obtaining a second time series of measured variables associated with the patient, which is different from the time series and covers a second time period that at least partially overlaps with the time series, wherein the observable is additionally ascertained based on the second time series.

[0025] The second time series can be automatically extracted from the patient's electronic patient file or patient data, for example. The patient data can include image data, in particular image data sequences, and non-image data. Alternatively, the second time series can also be selected by the user from a plurality of available second selection time series and obtained therefrom. According to some examples, the available second time series can be automatically extracted from the patient's patient data, in particular. Like the time series, the second time series can include a series of measured values ​​detected at different times. In particular, the measured values ​​of the second time series can be of the same type. "Same type" can in particular mean here that the measured values ​​are measured values ​​of the same measurement variable.

[0026] The measured variables of the second time series may be different from the measured variables of the first time series.

[0027] By taking into account the second time series, different temporal developments can be considered in relation to one another. This makes it possible to identify correlations that would not be apparent if the individual measured variables were considered individually.

[0028] According to some examples, the step of obtaining the second time series includes extracting the second time series from the medical image data sequence.

[0029] The second time series then corresponds to the time series in its basic design, but can in particular be based on different measured variables. In particular, the explanations and definitions given here for the time series then also apply to the second time series. In particular, extracting the second time series can include applying a further image data analysis function to the medical image data set of the medical image data sequence, wherein the further (different from the image analysis function) image data analysis function is designed to extract further (different from the image data-based measured values ​​of the time series) image data-based measured values ​​from the medical image data set, and the time series has further image data-based measured values.

[0030] Thereby, complementary information may be extracted and compared from the image data, which may improve the diagnostic effectiveness of the observables.

[0031] According to some examples, the step of obtaining the second time series includes extracting the second time series from a second medical image data sequence different from the medical image data sequence.

[0032] The second image data sequence can have a plurality of medical image data sets which were respectively acquired at different times in the second time period.

[0033] According to some examples, the image data set of the image data sequence may be recorded by a first imaging modality, and the image data set of the second image data sequence may be recorded by a second imaging modality, which second imaging modality differs from the first imaging modality in its imaging type. For example, the first imaging modality may be a magnetic resonance modality, and the second imaging modality may be a computed tomography modality.

[0034] By utilizing different sequences of image data, complementary information can be extracted and compared, which can improve the diagnostic power of the observables.

[0035] According to one aspect, the step of determining the observable includes: determining correlation information based on the first time series and the second time series, wherein the correlation information describes a measure for the temporal correlation between the first time series and the second time series; and obtaining the observable based on the correlation information.

[0036] The correlation information may indicate what time relationship the time series and the second time series are in. To evaluate the correlation information, for example, a known correlation function may be applied to the time series and the second time series.

[0037] According to some examples, the observable comprises a correlation of the time series with a second time series, and in particular comprises a correlation of the time series with a temporal behavior of the second time series, or is based on such a correlation. In particular, the observable may comprise a correlation or anti-correlation of the time series with the second time series or a measure thereof.

[0038] By evaluating the correlation information, the measured variables can be analyzed systematically and in relation to one another. This allows correlations that are not available when observing the individual measured variables individually to be quantified. This automatically provides information to the user in order to quickly make an appropriate assessment of the patient.

[0039] According to one aspect, the measured variables of the second time series are not based on medical image data.

[0040] In other words, the second time series is not extracted from the image data set. In particular, the second time series may be extracted from a data set of the patient that is not image data.

[0041] Non-image data are patient data that are not image data. The non-image data may include, for example, one or more medical assessment reports, for example from radiology, pathology, laboratory or other disciplines. The non-image data may include, in particular, longitudinal data, which contain one or more medical values ​​of the patient and / or elements of the patient's medical history. This may involve laboratory data, vital signs and / or other patient-related measurements or examinations.

[0042] The non-image data may be contained, for example, in an electronic patient file (English: Electronic Medical Record (electronic medical record), EMR, or Electronic Health Record (electronic health record), EHR). In this case, the electronic patient file can be called up from one or more storage devices, which can be connected to a medical information network. For example, the storage device can be queried for the patient's non-image data or electronic patient file. For this purpose, for example, an electronic identification such as, for example, a patient ID or an access number can be used. Correspondingly, the non-image data can be received from one or more storage devices of the available storage devices, in which at least a part of the electronic patient file is respectively stored. In this case, the storage device can be, for example, part of a medical information system, such as a hospital information system and / or a PACS system and / or a laboratory information system, etc.

[0043] All available information about the patient, in particular non-image data, but also image data, may be stored in the electronic patient file.

[0044] By taking into account non-imaging-based measured variables, other information outside of imaging diagnosis can be taken into account. Thus, the corresponding observables can be associated with different time developments. This can provide the user with a better decision basis for diagnosis without requiring significant additional effort on the part of the user.

[0045] According to one aspect, obtaining the second time series includes accessing an electronic patient file of the patient and extracting the second time series from the electronic patient file.

[0046] According to one aspect, the method includes extracting multiple time series and / or obtaining multiple second time series from an image data sequence, wherein in a determining step, for multiple different combinations (pairs) of time series selected from the time series and / or the second time series, correlation information and intermediate observables based on the correlation information are determined, each combination (pair) is selected from the time series based on the intermediate observables, and the selected pairs are provided.

[0047] The selected combination (pair) can be displayed to the user during the provision process (e.g. in a corresponding user interface). In particular, the selected combination (pair) can be displayed together with the associated intermediate observables in order to provide an explanation of the selection. For example, the selection can include a check as to whether the intermediate observable exceeds a threshold value. The selection can also include a comparison with an observable pattern indicating a medically relevant fact.

[0048] By calculating the correlations of different combinations of time series (which can be done in the background), anomalies that are important for diagnosis can be systematically sought. The user can then be automatically made aware of such anomalies. Thus, the user does not have to manually form and compare different time series, but receives relevant correlations that are automatically identified and displayed.

[0049] According to one aspect, the measured variables of the second time series include laboratory values ​​of the patient, in particular blood count values.

[0050] Data supplemented by imaging can be systematically included by taking into account laboratory values. It has been demonstrated that (small) abnormalities in imaging are usually accompanied by changes in laboratory values ​​or blood values. Individual observations of individual values ​​usually do not indicate disease, whereas comparisons of these values ​​can fully indicate significant disease.

[0051] According to some examples, the measured values ​​include oxygen saturation, ECOG value, inflammation value, alkaline phosphatase level (AP value), and / or red blood cell level in the patient's blood.

[0052] The oxygen saturation allows conclusions to be drawn about the efficiency of breathing or oxygen transfer and thus about the patient's physical condition. A poor oxygen saturation can indicate other problems with the patient.

[0053] The ECOG value or performance status value describes the physical condition of a cancer patient and is used to quantify the overall health status and limitations in activities of daily life. Based on the value, the progression of the disease can be estimated, appropriate treatment can be sought, and a prognosis can be created.

[0054] Inflammation values, such as the C-reactive protein (CRP) value, provide information about inflammation in the body. Inflammation values ​​rise in the event of infection, inflammation and tissue damage.

[0055] The determination of alkaline phosphatase (AP value) is used as an indicator for diseases, for example diseases of the liver and bile duct, tumors or for changes in bone metabolism.

[0056] Too many or too few red blood cells in the blood can also indicate a specific disease. For example, a small value of red blood cells can indicate a disease of the bone marrow, such as leukemia, or indicate a tumor disease.

[0057] The mentioned values ​​can therefore complement the imaging well and contribute to a better estimation of the measured values ​​from the imaging.

[0058] According to some examples, the measured variables of the second time series include medications taken by the patient.

[0059] Dosing can include the manner and amount of a drug administered to a patient.

[0060] By taking medication into consideration, it is possible to determine, for example, whether the physical condition of the patient being imaged was caused by the medication.

[0061] According to some instances, the step of determining an observable includes: obtaining a third time series that is different from the time series and the second time series of another measured variable associated with the patient, wherein the third time series covers a third time period that at least partially overlaps with the first time period and the second time period, determining correlation information based on the first time series, the second time series, and the third time series, wherein the correlation information describes a measure for the temporal correlation between the first time series, the second time series, and the third time series, and obtaining the observable based on the correlation information.

[0062] By taking into account the third time series, further information can be taken into account. In particular, the measured variables of the third time series can be the medication of the patient. Thus, for example, the measured variables in the imaging (e.g., the patient's physical characteristics, such as water retention) can be related to the measured variables in the laboratory (such as, for example, AP values) and the medication.

[0063] According to one aspect, the step of determining an observable includes: a step of obtaining a single event associated with a patient within a time period; a step of determining correlation information based on a time series and a single event, wherein the correlation information describes a measure for a temporal correlation between the time series and the single event; and a step of obtaining an observable based on the correlation information.

[0064] According to one aspect, the step of determining an observable includes: a step of obtaining a single event associated with a patient within a second time period; a step of determining correlation information based on a second time series and the single event, wherein the correlation information describes a measure for a temporal correlation between the second time series and the single event; and a step of obtaining an observable based on the correlation information.

[0065] A single event can, for example, relate to a therapy or treatment of a patient. In particular, a single event can relate to radiotherapy, surgery, chemotherapy and / or immunotherapy.

[0066] In addition, a single event can be related to the patient's health status. In particular, a single event can include an accident, an attack (such as a stroke), the beginning of an inflammatory disease, etc.

[0067] According to some examples, the step of obtaining a single event may include extracting a single event in an electronic patient file or from the patient data of the patient based on a time series. In particular, this may include searching for a single event in an electronic patient file or from the patient data of the patient based on a time period.

[0068] The relevance information can indicate the temporal relationship of the time series and in particular the temporal behavior of the time series and the individual events. To evaluate the relevance information, for example, the temporal behavior of the time series can be determined and linked to the extracted individual events.

[0069] The time series can be linked to possible individual events by means of the correlation information. It can thus be automatically determined whether the temporal development in the time series coincides with an individual event or is temporally related. This not only supports the user in the assessment of the patient, but also makes it easier to take into account the temporal course in the clinical routine.

[0070] According to one aspect, the method further comprises the step of obtaining background information related to the patient's condition, wherein the step of extracting comprises selecting an image data-based measurement variable from a plurality of different image data-based measurement variables based on the background information, and the first time series comprises image data-based measurement values ​​of the selected measurement variable.

[0071] In particular, the background information can describe framework conditions that are important for the user's diagnostic activity. Examples of framework conditions are: "Evaluation of chest CT recordings after trauma", "Follow-up examination within the scope of cancer treatment of organ X", "Analysis of MR recordings of the lungs", "Confirmation of suspected diagnosis Y", etc. The background information can be obtained, for example, based on corresponding user input and / or an electronic patient file and / or an electronic worklist.

[0072] Based on the background information, the measurement variables that are particularly important for the current case can be automatically selected. For example, if a cancer patient should be assessed based on the background information, the proportion of muscle tissue can be used as a measurement variable to provide information about the progression and development of the disease. Similarly, the size of the lesion can be used as a measurement variable to show whether the treatment is effective.

[0073] According to some examples, observables for unselected measured variables may first be ascertained in the background.

[0074] According to some instances, the extracting step further includes selecting an image data analysis function from a plurality of provided image data analysis functions, wherein the selected image data analysis function is configured to extract image data-based measurement values ​​of selected measurement variables from the medical image dataset, and applying the selected image data analysis function to the medical image dataset of a medical image data sequence.

[0075] According to some examples, the step of obtaining a second time series includes selecting a non-image data-based measurement variable from a plurality of different non-image data-based measurement variables based on the background information, and extracting a non-image data-based measurement value of the selected measurement variable from the non-image data of the patient.

[0076] By selecting the measured variables based on the background information, the measured variables that are particularly important for the current case can be automatically selected. For example, if a cancer patient should be assessed, the AP value or the red blood cell value can provide an indication about the progression of the disease.

[0077] According to some examples, the step of obtaining the image data sequence of the patient includes selecting a plurality of image data sets for the image data sequence from a plurality of available image data sets of the patient based on the context information.

[0078] According to some examples, obtaining the second time series includes accessing an electronic patient file of the patient, selecting a data set from the patient file based on the background information, and extracting the second time series from the selected data set.

[0079] Thus, for example, for screening the lungs of a patient, data sets or image data sets can be selected in a targeted manner from the patient's file which image a region of the lungs or contain important information.

[0080] According to some examples, the selection may be made based on rules, for example as follows: a certain type of image dataset or dataset is always selected for the context. The datasets may be, for example, individual documents, such as reports or forms. According to some examples, the selection may be made based on metadata stored in the image dataset, such as, for example, so-called DICOM tags, or based on document titles or keywords in the documents.

[0081] According to one aspect, the medical image data set comprises a record of the entire body part or a full body record, respectively, of a patient.

[0082] A body part can be a region of a patient's body that includes a plurality of body parts or organs. The region can be, for example, the patient's thorax, chest, or upper body region.

[0083] By using image data sets showing entire body parts or even the entire body as a full-body recording, the time series can give an indication of the overall condition of the patient as it emerges from the imaging.

[0084] The measured variables of the time series can in particular describe the condition of a body part or the entire body of the patient. Thus, values ​​related to the physical condition can be quantified in the measured variables, which values ​​can only be displayed to the user with difficulty through a simple inspection of the image data.

[0085] According to some embodiments, the measured variables of the time series are selected from:

[0086] - body composition analysis values, BCA values, including in particular muscle fraction and / or body fat fraction and / or water fraction,

[0087] - size of the lesion,

[0088] - the tumor burden in a body part of the patient's body or in the patient's entire body, and / or

[0089] - Distribution of FDG in a body part of the patient's body or in the entire body of the patient.

[0090] BCA stands for "Body Composition Analysis" and denotes a method of extracting values ​​that characterize the physical condition in general from medical image data. It is therefore not targeted at individual lesions, but rather at overall properties of the body. Examples of this are the composition of muscle and fat tissue or the extent of water retention. Changes in the values ​​can, for example, indicate the progression of a cancer disease, but are difficult for the user himself to read from the image data. This is due to the fact that although human perception is trained to detect various changes, such as lesions, it is difficult to make proportional estimates. This applies more and more because the human body is a three-dimensional object. Therefore, the image data set is also a three-dimensional image data set that the user can only have a difficult overview of in a two-dimensional screen view. The inventors have also recognized that, in particular, the correlation with blood picture values, such as AP values ​​or red blood cell values, enables well-founded conclusions about the progression of the disease process and is a valuable observable.

[0091] In order to provide such a BCA value, a correspondingly designed image analysis function can be applied to the image data set. In particular, such an image analysis function can be designed to provide a BCA value based on a three-dimensional image data set for the entire body or a body part.

[0092] Detecting the size of a lesion across multiple image datasets is also challenging or very time consuming for the user. The lesion must be identified and measured by the user from image dataset to image dataset. The user can be relieved of this work by an image analysis function that can track and measure the lesion from image dataset to image dataset. For this purpose, a number of known detection algorithms are available.

[0093] Tumor burden (English: Tumor Burden) indicates how strongly the patient's body or body part is affected by tumors. This is again a value that is only difficult to estimate or even quantify in its entirety for the user.

[0094] For example, tumor burden can be detected by administering a suitable contrast agent in a PET-CT examination with the aid of image data analysis subsequently performed by a correspondingly constructed image data analysis function. A PET-CT examination is a combination of positron emission tomography (PET) and computed tomography (CT). PET can make metabolic processes in the body visible in an image manner. To this end, a very small amount of radiolabeled substance is administered to the patient. The substance is distributed in the body and accumulates in specific tissues, such as tumors. Here, FDG (F-18 deoxyglucose) can be used as a marker. FDG is a glucose molecule labeled with radioactive fluorine. Because cancer cells have an increased glucose consumption than healthy cells, FDG accumulates in diseased cells in an enhanced manner.

[0095] With the help of a PET camera, it is possible to visualize the different distributions of cancer cells in the body. Thus, even primary cancers measuring a few millimeters can be traced.

[0096] Combined PET-CT devices enable computed tomography to be performed almost simultaneously. By combining the two methods, cell areas with high glucose metabolic activity can be detected, thus providing information about the tumor burden.

[0097] The inventors have recognized that in particular the correlation of tumor burden or FDG values ​​with blood values, such as AP values ​​or erythrocyte values, enables well-founded conclusions about the progression of the disease process and is a valuable observable.

[0098] According to some examples, the providing step includes comparing the observables with a preset observable pattern, and displaying the time series and / or the observables based on the comparison.

[0099] If the association with the second time series is performed according to some aspects and examples described herein, the second time series may also be displayed together with the first time series accordingly.

[0100] An observable pattern may, for example, include one or more conditions for one or more observables, such as an anti-correlation for two measured variables or a threshold for a drop or rise in a measured variable. An observable pattern may include a combination of multiple conditions for different observables. For example, an observable pattern may be considered satisfied when all conditions are true.

[0101] If there are multiple time series or multiple second time series, this selection can also be performed. Then, only the time series or second time series in which the associated (intermediate) observable falls into the observable mode (for example, by the observable being above a threshold value) are displayed.

[0102] By selectively displaying, the relevant observables can be pointed out to the user in a targeted manner. Therefore, the display can serve as a warning to the user about the deterioration of the disease image. At the same time, the user is notified only when an abnormality actually exists, which reduces the burden on the user.

[0103] According to some examples, the method further comprises the steps of determining a medical diagnosis based on the observables and providing the medical diagnosis.

[0104] The medical diagnosis can include a prediction of a medical assessment, such as confirmation or rejection of a suspected diagnosis. In addition, the medical diagnosis can indicate the probability that the medical assessment exists.

[0105] According to some examples, determining the medical diagnosis may include selecting the medical diagnosis from a plurality of suspected diagnoses.

[0106] For example, multiple suspected diagnoses can be provided based on background information. The multiple suspected diagnoses are preselected, and the diagnosis that best matches the case based on the observables can be selected from the preselected. Here, the suspected diagnoses can be associated with one or more indicators showing the suspected diagnoses, respectively. Thereby, the diagnosis can be determined based on the comparison between the observables and the indicators.

[0107] If there are multiple observables (eg in multiple (second) time series), all of them may be used to determine a medical diagnosis or to select from multiple suspected diagnoses or to compare with an indicator.

[0108] By providing a diagnosis, a possible diagnosis is suggested to the user based on the observables. In other words, a diagnosis is predicted. Thus, the user is effectively supported in the assessment of the patient without having to manually create and evaluate the time series.

[0109] According to some examples, providing a medical diagnosis includes generating a medical assessment report based on (and / or utilizing) the observable and / or the diagnosis.

[0110] For example, the diagnosis can be optionally converted into a template for a medical assessment report together with the observables and / or (multiple) time series or (multiple) second time series. In particular, it can also be proposed that a template is selected from a plurality of available templates based on the observables (and / or the diagnosis).

[0111] The medical assessment report can be the result of a diagnostic process, which is intended to determine the patient's condition with respect to one or more clinical aspects based on medical data important thereto. The medical assessment report can have a document. In particular, the medical assessment report can have a structured document or a structured document portion. In addition, the medical assessment report can include an unstructured document or an unstructured document portion. In addition, the medical assessment report can be constructed by one or more templates, which are input with patient-specific information within the scope of the assessment.

[0112] By providing a medical assessment report, the user can be provided with a treatment result which can be directly used further.

[0113] According to one aspect, the step of determining a medical diagnosis includes: selecting a disease pattern from a plurality of selected disease patterns based on an observable, querying an electronic patient file of the patient from a database, matching the selected disease pattern with entries in the electronic patient file, and determining a medical diagnosis based on the matching.

[0114] The disease pattern may have one or more signs or indicators that indicate the disease. The signs or indicators may be, for example, a value range for a measured variable and / or an observable quantity based on one or more measured variables. The measured variable may be a measured variable based on image data or other measured variables based on non-image data.

[0115] The selected disease mode can be preset. In addition, the selected disease mode can be pre-selected from a plurality of available disease modes based on background information.

[0116] In addition, the selected disease pattern can be extracted from an electronic compendium or textbook. An example of such an electronic textbook is eRef by Thieme. Disease patterns can be extracted, for example, with the aid of a so-called Large Language Model (LLM) as described herein. In addition, the selected disease pattern can be obtained based on comparing the disease patterns of patients.

[0117] The matching can include searching for information in the electronic patient file and comparing it with the disease model. In particular, in the electronic patient file, signs or indicators can be searched from the disease model. In addition, measured variables and / or observables can be searched in the electronic patient file, and the associated values ​​can be compared with the value range in the disease model.

[0118] By using disease patterns, a more accurate prediction of the diagnosis can be achieved, since it does not have to be based solely on observables. Thus, the user obtains a reliable and well-utilizable output.

[0119] According to some examples, matching includes: determining that observables and / or measured variables are missing in the electronic patient file compared to the disease pattern, and optionally, applying an analysis function to the electronic patient file to provide the missing observables and / or measured variables, and / or creating an examination protocol for examining the patient to provide the missing observables and / or measured variables.

[0120] The examination protocol may include general advice or instructions for a user or other person performing the examination. In addition, the examination protocol may include control instructions for manipulating a modality for performing the examination. The examination may be, for example, an imaging examination using an imaging modality.

[0121] By determining the missing measured variables, on the one hand, conclusions can be drawn about how reliable the provided diagnosis is. On the other hand, the user receives information about how he can further ensure the diagnosis. The optional automatic application of analysis functions or the optional automatic creation of examination protocols can further relieve the user's burden based on this, because the next steps for ensuring the diagnosis are automatically started.

[0122] According to some examples, matching the selected disease pattern with the electronic patient record can be performed using a transformer network and in particular a large language model (LLM).

[0123] The transformer network is a neural network architecture that generally has an encoder, a decoder, or an encoder and a decoder. In some cases, the encoder and / or decoder consists of multiple corresponding encoding layers or decoding layers. In each encoding layer and decoding layer there is a so-called attention mechanism.

[0124] Attention mechanisms, sometimes also called self-awareness, relate data elements (e.g., words) within a series of data elements to other data elements within the series. For example, a self-awareness mechanism can enable a model to, for example, examine a group of words within a sentence and determine the relative meaning that another group of words within the sentence has with respect to the examined group of words.

[0125] In particular, the encoder can be configured such that it converts the input (elements of a disease pattern or a document in an electronic patient file) into a digital representation. The digital representation can include a vector for each input token (e.g., each word). The encoder can be configured such that it implements an attention mechanism so that each vector of a token is influenced by other tokens in the input. Thus, tokens can be linked to each other and, for example, when there is no word-for-word correspondence, the tokens can also be identified as being important for the disease pattern.

[0126] Furthermore, the converter network may comprise a classification module or unit configured such that it associates the output of the encoder or decoder with a set of learned outputs, eg elements of a disease pattern.

[0127] According to some examples, the training of the converter model can be performed in two stages: pre-training and fine training. In the pre-training stage, the converter network can be trained with a large data corpus in order to learn the underlying semantics of the problem. Such pre-trained converter networks can be used for different languages. For the specific applications described herein, fine training can be performed based on medical texts with verified (annotated) meanings and / or medical ontologies (medizinischen Ontologien), such as RADLEX and / or SNOMED. Thus, the converter network can learn typical relations and synonyms of medical expressions.

[0128] For additional information about converter networks, refer to the document "Attention Is All You Need", arXiv:1706.03762, June 12, 2017, the contents of which are incorporated by reference into this application in their entirety.

[0129] LLM can be based on transformer networks. LLM is a language model characterized by its size. Artificial neural networks built with it can contain tens of millions to billions of weights and are pre-trained with the help of self-supervised learning and semi-supervised learning. Transformer networks contribute greatly to faster learning. Exemplary LLMs that can be used to match disease patterns with electronic patient records are the GPT series (GPT-3.5 and GPT-4 used in ChatGPT), the LLaMa series, PaLM (used in Google Bard), BLOOM, Ernie 3.0Titan or Claude.

[0130] The advantage of the transformer network is that, due to the attention mechanism, a wide range of correlations in the input data can be effectively considered. In addition, the encoder is able to process data in parallel, which saves computational resources. Moreover, due to autoregression, the decoder is able to iteratively generate a sequence of output values ​​with high reliability.

[0131] According to one aspect, the method comprises the steps of comparing an observable with corresponding observables of a plurality of reference patients different from the patient; selecting a comparison patient from the plurality of reference patients based on the comparing step; and providing information about the comparison patient as comparison information.

[0132] According to some examples, the information about the comparison patient may include an electronic patient file of the comparison patient or a selection from the electronic patient file, such as an assessment report.

[0133] The electronic patient files of the reference patients can be stored in a storage device of the medical information system, for example. The electronic patient files can respectively contain the corresponding observables of the corresponding reference patients. According to some examples, the corresponding observables can be obtained using a converter network, in particular an LLM.

[0134] According to one aspect, the method further comprises the step of selecting a comparison patient from a plurality of reference patients based on the observable.

[0135] According to some examples, the selection of the comparison patient may be performed based on the selected disease pattern. According to some examples, the selection may include matching the selected disease pattern with the electronic patient files of the reference patients, respectively.

[0136] According to one aspect, the step of comparing comprises, for each corresponding observable, determining a similarity measure to the observable, wherein in the step of selecting, the selection is performed according to the similarity measure.

[0137] According to one aspect, providing the comparison information includes outputting a warning if a critical condition is indicated for the comparison patient.

[0138] By automatically selecting a comparison patient, the user can be informed of similar cases / patients. The user can then follow the similar cases when assessing and selecting treatment options. Thus, the user can be effectively supported when assessing and treating a patient.

[0139] According to some examples, the method further comprises the steps of determining a treatment option based on the observable and providing the treatment option.

[0140] According to some examples, the step of determining a treatment option includes selecting a treatment option from a plurality of selected treatment options based on the observable.

[0141] According to some examples, the step of selecting a treatment option is additionally performed based on the background information and / or the selected disease model and / or the specific medical diagnosis.

[0142] According to some examples, the method further comprises pre-selecting a selected treatment option from the plurality of treatment options based on the background information and / or the selected disease mode and / or the specific medical diagnosis.

[0143] According to some examples, the step of determining the treatment option further includes: obtaining an indication pattern for the selected treatment option; querying the patient's electronic patient file from a database; matching the indication pattern with entries in the electronic patient file; and verifying the treatment option based on the match.

[0144] The indication pattern may include indications and / or contraindications for a treatment option.The indication pattern may be associated with a selected treatment option, for example, based on one or more medical criteria.

[0145] According to some instances, matching of the indication pattern includes: determining that observables and / or measured variables are missing in the electronic patient record compared to the indication pattern; and optionally, applying an analysis function to the electronic patient record to provide the missing observables and / or measured variables; and / or creating an examination protocol for examining the patient to provide the missing observables and / or measured variables.

[0146] According to one aspect, a system for providing observables indicative of a medical diagnosis is provided. The system has a computing device and an interface. The interface is configured to obtain an image data sequence of a patient, wherein the image data sequence has a plurality of medical image data sets, and the medical image data sets are recorded at different times in a time period. The computing device is configured to extract a time series from the medical image data sequence. The computing device is also configured to determine an observable based on the time series. The computing device is also configured to provide the observable via the interface.

[0147] The computing device can be configured as a back-end computing device. In particular, the computing device can be configured as a server system. The computing device can have a cluster or a group of computing devices and a data storage device. The back-end computing device can be data-connected to a front-end computing device via a medical information network, and the front-end computing device hosts a user interface for a user. The computing device can be data-connected to multiple different (but especially the same type) front-end computing devices via a medical network. (Multiple) front-end computing devices can belong to a medical institution, such as, for example, a clinic, a hospital, or a hospital consortium. The back-end computing device can also belong to a medical institution or be configured outside a medical institution. The back-end computing device can be data-connected to multiple different front-end computing devices that belong to different medical institutions, respectively, via a medical information network.

[0148] Alternatively, the computing device may also be distributed on the back-end computing device and the front-end computing device, or include only the front-end computing device.

[0149] According to an embodiment, the medical information network can be based on the HL7 standard. Health Level 7 (HL7) is a set of international standards for exchanging data between institutions in the health sector and their computer systems. In particular, the communication and / or data connection can be based on the FHIR standard. Fast Healthcare Interoperability Resources (FHIR) is a standard developed by HL7. It assists in the exchange of data between software systems in the health sector. By using the HL7 or FHIR standards, data can be transmitted in a structured manner and without the need for reformatting.

[0150] The interface can generally be configured for data exchange between the computing device and other components. The interface can be implemented in the form of one or more separate data interfaces, which can have hardware ports and / or software ports, such as a PCI bus, a USB interface, a Fire-Wire interface, a ZigBee interface or a Bluetooth interface. The interface can also have an interface to a communication network, wherein the communication network can have a local area network (LAN), such as an intranet or a wide area network (WAN) or the Internet. Correspondingly, one or more data interfaces can have a LAN interface or a wireless LAN interface (WLAN or WiFi).

[0151] Furthermore, the system can have a data memory connected to the interface, in which image data sets and / or electronic patient files and / or analysis functions and / or image data analysis functions and / or converter networks / LLMs are stored and can be called up via the interface and thus made available. The data memory can be designed as a central or discrete memory unit or cloud storage.

[0152] Additionally, the system may include one or more imaging modalities, such as a computed tomography system, a magnetic resonance system, an angiography system, an X-ray system, a positron emission tomography system, a mammography system, and / or the like.

[0153] The advantages of the proposed system correspond substantially to the advantages of the proposed method. Features, advantages or alternative embodiments / aspects of the method can also be transferred to other claimed subjects and vice versa.

[0154] In another aspect, the present invention relates to a computer program product, which comprises a program and can be directly loaded into the memory of a programmable control device and has a program structure, such as libraries and auxiliary functions, so that when the computer program product is executed, the method for providing observable quantities indicating a medical diagnosis, in particular according to the above-mentioned embodiment / aspect, is performed.

[0155] In addition, in another aspect, the present invention relates to a computer-readable storage medium, on which a readable and executable program segment is stored, so that when the program segment is executed by a control device, all steps of the method for providing observable quantities indicating a medical diagnosis according to the above-mentioned embodiment / aspect are performed.

[0156] The computer program product can include software with source code, which still has to be compiled and linked or only has to be interpreted, or an executable software code, which only has to be loaded into a processing unit for execution. The method can be carried out quickly, also reproducibly and robustly by means of a computer program product. The computer program product is configured so that it can carry out the method steps according to the invention by means of a computing unit. The computing unit must have the prerequisites, such as, for example, a corresponding working memory, a corresponding processor, a corresponding graphics card or a corresponding logic unit, so that the corresponding method steps can be carried out efficiently.

[0157] The computer program product is stored, for example, on a computer-readable storage medium or saved on a network or server, from which the computer program product can be loaded into a processor of a corresponding computing unit, which processor can be directly connected to the computing unit or constitute a part of the computing unit. In addition, the control information of the computer program product can be stored on a computer-readable storage medium. The control information of the computer-readable storage medium can be constituted so that when a data carrier is used in the computing unit, the control information executes the method according to the present invention. An example of a computer-readable storage medium is a DVD, a tape or a USB stick on which electronically readable control information, especially software, is stored. If the control information is read from the data carrier and stored in the computing unit, all embodiments / aspects of the above-mentioned method according to the present invention can be executed. Therefore, the present invention can also be based on the computer-readable medium and / or the computer-readable storage medium. The advantages of the proposed computer program product or the associated computer-readable medium basically correspond to the advantages of the proposed method. BRIEF DESCRIPTION OF THE DRAWINGS

[0158] Other features and advantages of the present invention can be seen from the following description of the exemplary embodiments according to the schematic diagram. The modifications mentioned in the above context can be combined with each other to form new embodiments. In different drawings, the same reference numerals are used for the same features.

[0159] The accompanying drawings show:

[0160] Figure 1 A schematic diagram illustrating one embodiment of a system for providing an observable indicative of a medical diagnosis,

[0161] Figure 2 A flow chart showing a method for providing an observable indicative of a medical diagnosis according to one embodiment,

[0162] Figure 3 a diagram showing optional method steps in a method for providing an observable indicative of a medical diagnosis according to one embodiment,

[0163] Figure 4 A data flow diagram illustrating a method for providing an observable indicative of a medical diagnosis according to one embodiment,

[0164] Figure 5 a schematic diagram showing a graphical user interface in a method for providing an observable indicative of a medical diagnosis,

[0165] Figure 6 a diagram showing optional method steps in a method for providing an observable indicative of a medical diagnosis according to one embodiment,

[0166] Figure 7 a diagram showing optional method steps in a method for providing an observable indicative of a medical diagnosis according to one embodiment,

[0167] Figure 8 a diagram showing optional method steps in a method for providing an observable indicative of a medical diagnosis according to one embodiment,

[0168] Fig. 9 Diagram showing optional method steps in a method for providing an observable indicative of a medical diagnosis according to one embodiment. DETAILED DESCRIPTION

[0169] exist Figure 1 FIG. 1 shows a system 1 for providing observables OBS indicative of a medical diagnosis according to one embodiment. The system 1 has at least one front-end computing device 10 with a user interface.

[0170] (or also just “user interface 10 ”), a medical information system 40 and a backend computing device 20 (or also just “computing device 20 ”), which are communicatively connected to each other via a medical network or data interface 26 .

[0171] Medical image data set SET and other information, such as, for example, electronic patient records EPA, can be provided to the front-end computing device 10 by the medical information system 40 via a suitable interface 26. Figure 1The system shown in the figure has a plurality of front-end computing devices 10, all of which access the same medical information system 40 or exchange data with the back-end computing device 20. In the embodiment shown, the back-end computing device 20, the medical information system 40 and (multiple) front-end computing devices 10 are part of the same medical institution. For example, the medical institution can be a clinic, a clinic consortium, a hospital or a hospital consortium. Correspondingly, the network connecting the components via the interface 26 can be configured as an internal network of the institution and, for example, include an intranet (e.g., a local area network and / or a wireless local area network).

[0172] The front-end computing device 10 can be configured as an assessment station or an assessment workplace, where a user N can view and analyze patient data or a medical image data set SET of an electronic patient archive EPA, and create, check, change, verify and determine a medical diagnosis DIG and a treatment option BHO. Therefore, the front-end computing device 10 can also be referred to as a user interface 10. The front-end computing device 10 has a physical user interface, such as a physical user interface including a display and / or an input device. The front-end computing device 10 can have a processor. According to an embodiment, the processor can have a central processing unit (CPU), a graphics processing unit (GPU), a digital signal processor (DSP), an image processing processor, an integrated (digital or analog) circuit or a combination of the above components, and other devices for providing observables OBS, diagnosis DIG and / or treatment options BHO. The front-end computing device 10 can include, for example, a desktop PC, a laptop or a tablet computer.

[0173] The medical information system 40 can generally be configured to detect and / or store and / or forward medical image data sets SET and electronic patient archives EPA. For example, the medical information system 40 can have one or more (not shown) databases. In particular, the database can be implemented in the form of one or more cloud storage modules. Alternatively, the database can be implemented as a local memory or a distributed memory, for example, as a PACS (Picture Archiving and Communication System), a hospital information system (KIS), a laboratory information system (LIS), an "electronic medical record (EMR)" information system and / or other medical information systems. According to some examples, the medical information system 40 can also include one or more (not shown) medical imaging modalities, such as, for example, a computed tomography system, a magnetic resonance system, an angiography system, a C-arm X-ray system, a positron emission tomography system, an X-ray system, etc.

[0174] The medical image data set SET may include medical image data. In this context, the image data may be medical image data with two or three spatial dimensions. The image data may be generated, for example, by means of an imaging medical modality, such as an X-ray device, a computer tomography device, a magnetic resonance device, a positron emission tomography device or an angiography device or other devices. Such image data may also be referred to as radiological image data.

[0175] The image data contained in the medical image data set SET can be formatted, for example, in accordance with the DICOM format. DICOM (=Digital Imaging and Communications in Medicine) is an open standard for the communication and management of medical image data and associated data.

[0176] The patient's image data set SET may be contained in the patient's electronic patient archive EPA.

[0177] In addition to the medical image data set SET, the electronic patient archive EPA may also include non-image data. For example, non-image data may be examination results that are not based on medical imaging. This may include laboratory data, vital data, spirometry data or neurological examination logs. In addition, non-image data may include text data sets, such as structured and unstructured medical assessment reports (English: medical report). In addition, non-image data may also be patient-related data. This may, for example, include demographic information about the patient, such as his age, gender or weight. Non-image data may, for example, be incorporated into the image data as metadata. Alternatively or in addition, non-image data may also be stored in a separate document.

[0178] The back-end computing device 20 may have a processor. According to an embodiment, the processor may have a central processing unit (CPU), a graphics processing unit (GPU), a digital signal processor (DSP), an image processing processor, an integrated (digital or analog) circuit or a combination of the above components and other devices for providing observables OBS. The back-end computing device 20 may be implemented as a separate component or have a computer group, such as a cluster. Such a system may be referred to as a server system. According to an embodiment, the back-end computing device 20 may be configured as a local server. In addition, the back-end computing device 20 may have a working memory, such as a RAM, in order to, for example, temporarily store an image data set SET, an electronic patient archive EPA or a time series ZR1, ZR2. For example, by computer-readable instructions, by design and / or hardware, the back-end computing device 20 is configured so that it can perform one or more method steps according to an embodiment of the present invention.

[0179] The back-end computing device 20 can be connected to the front-end computing device 10 and / or the storage device RD and / or the medical information system 40 via the interface 26. Via the interface 26, the back-end computing device 20 can receive a medical image dataset SET, an electronic patient archive EPA and / or a user input INP, and based on the image dataset SET, the electronic patient archive EPA and / or the user input INP, can automatically select the time series ZR1, ZR2 and create and provide the observable OBS in a computer-supported manner.

[0180] In order to provide the observables OBS, the backend computing device 20 may have different modules.

[0181] The module 21 is designed as a data query module. The data query module may be designed to access the medical information system 40 and search for a medical image data set SET or an electronic patient archive EPA. The module 21 may be designed in particular to formulate a search request and adapt the search request to the medical information system 40 .

[0182] The module 22 may be configured as a user interaction module or unit. The module 22 may be configured to provide the time series ZR1, ZR2 and observables OBS, as well as the diagnosis DIG and / or treatment options BHO based thereon, to the user N. In addition, the module 22 may be configured to detect one or more user inputs INP and provide them for processing in the backend computing device 20. Such user input INP may include, for example, voice, gestures, eye movements, operation of an input device such as a computer mouse, etc. The user input INP may be directed to interaction with the medical image dataset SET.

[0183] Module 23 can be understood as an analysis module. Module 23 is configured to extract time series ZR1, ZR2 and analyze them. For example, module 23 can be configured to extract time series ZR1 from an image data set SET. Such time series ZR1 is a time series of measurement variables BD-MG based on image data. In addition, module 23 can be configured to extract time series ZR2 from an electronic patient file EPA involving measurement variables NBD-MG based on non-image data. Module 23 can be configured to execute a corresponding analysis function AF. For example, one or more such analysis functions AF can be configured to extract measurement values ​​(image data analysis functions) for measurement variables BD-MG based on image data from an image data set SET through image data processing. In addition, one or more analysis functions AF can be configured to extract measurement values ​​for measurement variables NBD-MG based on non-image data from an electronic patient file EPA.

[0184] In order to analyze the extracted time series ZR1, ZR2 and to provide the observable quantity OBS, the module 23 can also be configured to execute the time series analysis function ZR-AF. The time series analysis function ZR-AF is then configured to determine the time development of one or more time series ZR1, ZR2. In particular, the time series analysis function ZR-AF can include a correlation algorithm, which is configured to determine the correlation information KOR of the two time series ZR1, ZR2, and the observable quantity OBS is provided based on the correlation information KOR.

[0185] Module 24 can be understood as a further module or LLM module. The module can be configured to convert the observable OBS provided by module 23 into a result that can be used by the user N. For example, module 24 can be configured to provide the observable OBS as a text in a natural language. In addition, module 24 can be configured to provide a diagnosis DIG and / or a treatment option BHO based on the observable OBS, optionally in the case of further matching with the electronic patient archive EPA. In addition, module 24 can be configured to search for a comparison patient with a case situation similar to that of the patient in the medical information system 40 based on the observable OBS.

[0186] The module 24 may be configured to host and execute a network of translators or a large language model (LLM). The LLM may be used to convert observables OBS into natural language and / or to match with electronic patient records EPA.

[0187] The division of the backend computing device 20 into modules 21-24 is only used to more simply explain the working method of the backend computing device 20 and should not be understood as limiting. The modules 21-24 or their functions can also be combined in one element. The modules 21-24 can also be understood here in particular as computer program products or computer program segments, which implement one or more of the method steps described herein when executed in the backend computing device 20.

[0188] exist Figure 2 A schematic flow chart of a method for providing an observable variable OBS is shown in FIG. The sequence of the method steps is not limited by the order shown nor by the selected numbering. Therefore, if necessary, the sequence of the steps can be exchanged and individual steps can be omitted. In addition, one or more steps, in particular a series of steps, and optionally the entire method can be performed repeatedly. Figure 4 The corresponding diagram is shown in Figure 2 Unless otherwise stated, Figure 2 and Figure 4 Optional steps and data flows in are shown with dashed lines.

[0189] In an optional step S0, background information KI may first be obtained. The background information KI includes the medical background of the patient to be assessed. This may be, for example, an assessment task that the user N must complete based on the available patient data. Additionally or alternatively, the background information KI may include a description of the patient's clinical stage. Additionally or alternatively, the background information KI may include a description of the patient's possible disease or diagnosis.

[0190] The background information KI can be derived, for example, from the electronic patient archive EPA of the patient. Additionally or alternatively, inputs of the user via the user interface 10 can be taken into account in the background information KI.

[0191] In step S10, an image data sequence SE of a patient is obtained. The image data sequence SE has a plurality of individual image data sets SET of a patient, which are recorded at different times. The image data set SET can be loaded in step S10, for example, from a medical information system 40. In this case, background information KI can optionally be taken into account, so that only image data sets SET corresponding to the background information KI are loaded. The loading of the image data sequence SE can be triggered, for example, by a user N manually selecting the corresponding case / patient via the user interface 10.

[0192] In step S20, one or more time series ZR1 are extracted from the image data set SET. The time series ZR1 is also referred to as a first time series ZR1 or as a time series ZR1 based on image data. The time series ZR1 has a longitudinal sequence of measured values ​​for the measured variables BD-MG. The measured variables BD-BG can be extracted from the image data set in step S20, for example, by applying a corresponding image analysis function AF to the image data set SET.

[0193] The measured variable(s) BD-MG can be predefined. Alternatively, the measured variable BD-MG can be selected specifically for the respective patient based on a plurality of measured variables BD-MG that are possible in principle. This can be done, for example, based on the background information KI.

[0194] The first time series ZR1 basically has measured values ​​which can be extracted from the medical image data set SET, ie the image data. For example, BCA values ​​such as body fat percentage, muscle percentage etc., which are derived from recordings of the patient's body, should be mentioned here.

[0195] In step S30, observables OBS are determined from one or more first time series ZR1, which are important for the diagnosis or treatment of the patient or indicate the diagnosis or treatment. In particular, the observables OBS may include the time development of one or more time series ZR1, ZR2. For example, the observables OBS may include the time derivatives of the time series ZR1, ZR2.

[0196] In step S40, the observable OBS is provided. This may include displaying the observable OBS in the user interface 10. Optionally, the underlying time series ZR1, ZR2 may be displayed here.

[0197] Furthermore, the providing in step S40 may include forwarding the observables for optional further processing. Some possible further processing steps are listed below with reference to steps S50-S100. The steps are optional and may be connected to step S40 individually or in any combination.

[0198] In an optional step S50, the medical diagnosis DIG can be predicted based on the observable OBS. Thus, for example, the temporal development of the time series ZR1, ZR2 (either individually or in comparison with other time series ZR1, ZR2) can indicate the medical diagnosis DIG. Optionally, other information, such as background information KI or information in the patient's electronic patient file EPA, can be taken into account when predicting the medical diagnosis DIG in step S50.

[0199] The prediction of the medical diagnosis DIG may be made based on rules. Alternatively, the prediction in step S50 may be determined by applying a trained algorithm to the observables OBS, optionally to the background information KI and / or the electronic patient archive EPA

[0200] In step S60 , the diagnostic DIG is provided. This may include displaying the diagnostic DIG in the user interface 10 . Optionally, the underlying time series ZR1 , ZR2 and other information important for the decision may be displayed here.

[0201] In step S70, based on the observable quantity OBS, one or more comparison patients having medical similarity with the patient can be determined. To this end, a plurality of reference patients can be provided in step S70, from which comparison patients having increasing similarity with the patient (based on the observable quantity OBS) can be selected based on the observable quantity OBS.

[0202] In addition, other information, such as background information KI or information from the electronic patient file EPA of the patient / reference patient, may also be taken into account when selecting the comparison patient.

[0203] Then, in step S80, information VI about the comparison patient (also referred to as comparison information VI) may be provided. For example, the comparison information VI may be extracted from the electronic patient profile EPA of the comparison patient. The comparison information VI may include, for example, diagnosis, treatment performed, disease progression, etc. Alternatively, the comparison information VI may include the corresponding electronic patient profile EPA of the comparison patient or a portion thereof.

[0204] The comparison information VI can be provided to the user N by being displayed in the user interface 10. For example, a link to the electronic patient file EPA of the comparison patient can be provided to the user N.

[0205] In step S90, a treatment option BHO for the patient can be determined based on the observable OBS. Thus, for example, the temporal development of the time series ZR1, ZR2 can indicate the treatment option BHO. Optionally, other information, such as the medical diagnosis DIG, background information KI or information in the patient's electronic patient file EPA, can be taken into account when determining the treatment option BHO in step S90. In a similar manner to when predicting the medical diagnosis DIG, the treatment option BHO can be determined based on rules or using a trained algorithm.

[0206] In step S100, a treatment option BHO is provided. This may include displaying the treatment option BHO in the user interface 10. Optionally, the underlying time series ZR1, ZR2 and other information relevant for the decision may be displayed here.

[0207] exist Figure 3 A schematic flow chart of optional sub-steps for determining the observable variable OBS is shown in FIG. The sequence of the method steps is not limited by the order shown nor by the selected numbering. Therefore, the sequence of the steps can be exchanged and individual steps can be omitted if necessary. In addition, one or more steps, in particular a series of steps, can be performed repeatedly. Figure 3 The sub-steps shown in can be performed in step S30. Figure 4 The corresponding diagram is shown in Figure 3 The data flow associated with the method is shown in FIG.

[0208] For some diseases, it is important that the time series ZR1, ZR2 are considered not only individually but also in relation to one another. Figure 3 The embodiments shown in FIG. 1 focus on the described aspects.

[0209] First, in step S31, one or more second time series ZR2 are obtained, which are different from the (first) time series ZR1, but overlap with the first time series ZR1 at least partially in time. The second time series ZR2 can also be based on image data. Alternatively, the second time series ZR2 may not be based on image data. For example, the measured variable NBD-MG reproduced in the second time series ZR2 can be based on the laboratory and in particular on the patient's blood value. Such a second time series ZR2 can be extracted, for example, from the patient's electronic patient file EPA.

[0210] Which second time series ZR2 is to be provided in step S31 can be ascertained, for example, based on the background information KI.

[0211] According to some examples, the second time series ZR2 may also include only one single event. The single event may, for example, be related to the treatment and / or health condition of the patient.

[0212] In step S32 , the first time series ZR1 and the second time series ZR2 are mathematically linked to determine the association information KOR. The selected first time series ZR1 and second time series ZR2 can be linked to each other, wherein the selection can be made from a plurality of first time series ZR1 or second time series ZR2 based on the background information KI, for example.

[0213] Alternatively, all possible combinations of the first time series ZR1 and the second time series ZR2 may first be correlated to obtain corresponding correlation information KOR or observables OBS, respectively. Then, the relevant combinations of the time series ZR1, ZR2 may be identified and provided based on the correlation information KOR.

[0214] One or more observables OBS can then be determined from one or more pieces of association information KOR. This can include, for example, a selection based on association information KOR of medically significant combinations of time series ZR1, ZR2. This can be a combination of time series that shows a significant temporal association.

[0215] Furthermore, step S33 can include transmitting the correlation information KOR in a format that can be easily interpreted by the user N and that is subsequently provided to the user N as an observable OBS. For example, the correlation information KOR itself can be transmitted in a time series that can be displayed together with the time series ZR1, ZR2 on which it is based. Furthermore, the observable OBS obtained based on the correlation information KOR can include a text that describes the temporal correlation encoded in the correlation information KOR to the user N (for example, "the muscle share has decreased while the AP value has increased over the past two months"). For example, such a text can be provided with the aid of LLM.

[0216] exist Figure 5 A graphical user interface GUI according to one embodiment is shown in FIG. The graphical user interface GUI may be implemented, for example, in the user interface 10. The graphical user interface GUI may provide one or more results of the processing described herein to the user N.

[0217] The time series ZR1 , ZR2 may be shown / displayed in a section of the graphical user interface GUI. The time series ZR1 , ZR2 may be automatically selected for display, for example based on the observable OBS. Alternatively, the time series ZR1 , ZR2 may be selected by a user N.

[0218] exist Figure 5 In the example shown in , a time series ZR1 is shown, which is based on measured variables BD-MG based on image data. Here, BCA values ​​are shown by way of example, which can include individual measured variables BD-MG related to the body composition of the patient, such as body fat content or muscle content, or a combination of such measured variables BD-MG. In this case, each measured value can correspond to an imaging examination (e.g. a whole-body CT recording).

[0219] In addition, Figure 5 , three (second) time series ZR2 are shown by way of example, which are based on measured variables NBD-MG based on non-image data. In the present case, these are the AP values, the red blood cell values ​​and the MCV values ​​of the patient, as they can be obtained, for example, from a blood test of the patient. The individual measured values ​​can here relate to a blood test of the patient. The MCV value stands for "mean corpuscular volume". The blood value indicates the average volume of a single blood cell (red blood cell). For example, the value is important for finding the cause of anemia. The values ​​are determined within the scope of a blood test, but are not very convincing individually.

[0220] Furthermore, in the section of the time series ZR1, ZR2, the correlation of the time series ZR1 and ZR2 is shown as the observable quantity OBS. First, the time series is constantly stretched in a similar manner. From a certain moment, a decrease in the BCA value and an increase in the AP value or the MCV value can be observed. This becomes apparent in the decrease in the observable quantity OBS, which indicates the anti-correlation.

[0221] Additionally or alternatively, observables OBS can be provided as text output. This is shown in the lower left section of the graphical user interface GUI. Compared with the associated drawing, the text output can be more easily accessible to the user N. In addition, the text output can be directly integrated into the assessment report. Additionally, possible diagnosis DIG and / or treatment options BHO can be output in other sections of the graphical user interface.

[0222] exist Figure 6 A schematic flow chart of optional sub-steps for providing / displaying observables OBS is shown in . The sequence of the method steps is not limited by the order shown nor by the selected numbering. Therefore, the sequence of the steps can be exchanged and individual steps can be omitted if necessary. In addition, one or more steps, in particular a series of steps, can be performed repeatedly. Figure 6 The sub-steps shown in can be performed within step S40.

[0223] Depending on the available data, a plurality of observables OBS are possible. In particular, a plurality of time series ZR1, ZR2 opens up a plurality of possible combinations for determining the relevance information KOR. In order to support the user N in finding relevant observables OBS or time series ZR1, ZR2, it can be proposed in step S41 that a selection is made depending on the observable OBS.

[0224] For this purpose, the observable OBS can be compared, for example, with a threshold value or a specific pattern of the observable. This makes it possible to check whether the observable OBS, for example, indicates a medically relevant fact. Based on the comparison, the observable OBS and thus the associated time series ZR1, ZR2 can be selected.

[0225] Based on the selection in step S41, the selected observable OBS and / or time series ZR1, ZR2 can be displayed in step S42 (see Figure 5 ).

[0226] exist Figure 7 A schematic flow chart of optional sub-steps for determining a medical diagnosis DIG is shown in FIG. The sequence of the method steps is not limited either by the order shown or by the selected numbering. Thus, the sequence of the steps can be exchanged if necessary and individual steps can be omitted. In addition, one or more steps, in particular a series of steps, can be performed repeatedly. Figure 7 The sub-steps shown in can be performed within step S50.

[0227] In order to determine the medical diagnosis DIG, a disease pattern may be first identified in step S51. To this end, a disease pattern corresponding to one or more observables OBS may be selected from a plurality of disease patterns. In particular, the most likely disease pattern based on the observable OBS may be selected.

[0228] A plurality of disease patterns may be pre-selected from a plurality of predetermined disease patterns based on the background information KI. A disease pattern may generally include indications and contraindications for a particular disease to be diagnosed. At least one of the indications may be an observable OBS as described herein.

[0229] In step S52, the electronic patient file EPA of the patient is retrieved from the medical information system 40. This can be done, for example, with the aid of an electronic patient identifier, such as a patient ID, with which the patient is registered in the medical information system 40.

[0230] In step S53, the disease model can then be matched to the electronic patient file. Here, a check can be performed as follows: to what extent the indications and contraindications of the disease model are reflected in the electronic patient file. According to some examples, this can be done using a converter network or LLM. Indications and contraindications in the electronic patient file EPA that do not have a direct (verbatim) correspondence in the electronic patient file EPA can also be identified.

[0231] In step S54, a medical diagnosis DIG can be determined based on the match. For example, if it can be verified in step S53 whether the electronic patient profile EPA reproduces the indications and contraindications of the disease pattern, it can be inferred that there is a disease on which the disease pattern is based. This can be provided as a diagnosis DIG. Conversely, when there are no indications or contraindications, the disease can be ruled out. This can also be provided as a medical diagnosis DIG.

[0232] Optionally, step S54 may be followed by step S51 and step S53 again, so as to match other disease patterns with the electronic patient archive EPA. Here, in step S51, the next most likely disease pattern based on the observable OBS may be selected respectively.

[0233] exist Figure 8 A schematic flow chart of optional sub-steps for finding a comparison patient is shown in FIG. The sequence of the method steps is not limited by the order shown nor by the selected numbering. Therefore, if necessary, the sequence of the steps can be exchanged and individual steps can be omitted. In addition, one or more steps, in particular a series of steps, can be performed repeatedly. Figure 8 The sub-steps shown in can be performed within step S70.

[0234] Step S70 processes obtaining a comparison patient from a plurality of reference patients. The reference patients may be registered in the medical information system 40, respectively. For each reference patient, an electronic patient file EP may be placed in the medical information system 40. The electronic patient file EPA may contain observables corresponding to the observables OBS determined for the patient, respectively. Alternatively, such corresponding observables for the reference patient may be generated by a process similar to that in the patient.

[0235] In step S71, the observables OBS can be compared with corresponding observables of a reference patient. Here, one or more observables OBS can be used for the comparison. Optionally, the reference patient can be preselected from a larger group of patients registered in the medical information system 40 based on the background information KI.

[0236] Then in step S72, based on step S71, a reference patient having a certain similarity with the patient based on the observable OBS can be identified as a comparison patient. This can be in particular a reference patient showing a corresponding observable similar to the observable OBS.

[0237] The comparison patient can be informed to the user N. In addition, in step S73, it can be proposed to obtain information about the selected comparison patient (comparison information VI). The comparison information VI can be extracted from the corresponding electronic patient file EPA of the comparison patient. The comparison information VI can include, for example, the diagnosis DIG or the treatment option BHO. In addition, the comparison information VI can include one or more assessment reports of the comparison patient.

[0238] exist Fig. 9 A schematic flow chart of optional sub-steps for providing / displaying treatment options BHO is shown in FIG. The sequence of the method steps is not limited either by the order shown or by the selected numbering. Thus, if necessary, the sequence of the steps can be exchanged and individual steps can be omitted. In addition, one or more steps, in particular a series of steps, can be performed repeatedly. Fig. 9 The sub-steps shown in can be performed within step S90.

[0239] First, in step S91, a disease mode can be selected based on one or more observables OBS. In principle, the process can be carried out as in step S51. In the embodiment, the disease mode is respectively associated with at least one treatment option BHO. The treatment option BHO can be preset, for example, based on the medical criteria for the disease on which the disease mode is based. The treatment option BHO can be respectively associated with other indicators that indicate whether the treatment option BHO is suitable for the corresponding patient. Such indicators can include, for example, allergies to specific drugs or age restrictions, etc.

[0240] In step S92, the electronic patient file EPA of the patient can be retrieved from the medical information system 40. In principle, this can be done as in step S52.

[0241] In step S93, the disease pattern is matched with the electronic patient archive EPA. In principle, this can be done as in step S53.

[0242] Based on the matching in step S93, the treatment options BHO can be determined in step S94. If, for example, the presence of a disease pattern is verified in step S93, the associated treatment options BHO can be provided. In order to ensure the treatment options BHO, for indications or contraindications of the disease pattern described or indicators of the treatment options BHO not found in the electronic patient archive EPA, further information about the patient can be automatically queried (for example at the user N) and a further examination of the patient can be suggested or initiated in the medical information system 40.

[0243] In particular, the embodiments have been described in detail with reference to the accompanying drawings, but it should be noted that a number of modifications are possible. It should also be noted that the exemplary embodiments are merely examples, which should not limit the scope of protection, application and configuration in any way. More precisely, the above description provides a person skilled in the art with a guide for implementing at least one embodiment, wherein various modifications can be made according to the wishes of a person skilled in the art, in particular alternative or additional features and / or modifications of the functions and / or arrangements of the described components, without deviating from the subject matter and its legal equivalents respectively specified in the embodiments and / or without departing from its scope of protection. Regardless of the grammatical gender of a particular term, people with male or female gender identity are included.

Claims

1. A computer-implemented method for providing an observable (OBS) indicative of a medical diagnosis, the method having the following steps: - obtaining (S10) a medical image data sequence (SE) of a patient, wherein the medical image data sequence (SE) comprises a plurality of medical image data sets (SET), which are respectively recorded at different times in a time period, - extracting (S20) a time series (ZR1) from said sequence of medical image data (SE), - determining (S30) said observable (OBS) based on said time series (ZR1), and - Providing (S40) said observable (OBS).

2. The method according to claim 1, wherein the step (S30) of determining the observable quantity (OBS) comprises: - obtaining (S31) a second time series (ZR2) of measured variables (NBD-MG, BD-MG) associated with the patient, said second time series (ZR2) covering a second time period at least partially overlapping said time period, and - additionally determining ( S33 ) the observable (OBS) based on the second time series ( ZR2 ).

3. The method according to claim 2, wherein the step (S30) of determining the observable quantity (OBS) comprises: - determining (S32) relevance information (KOR) based on said time series (ZR1) and said second time series (ZR2), wherein said relevance information (KOR) indicates a measure for a temporal relevance between said time series (ZR1) and said second time series (ZR2), and - obtaining (S33) the observable quantity (OBS) based on the correlation information (KOR).

4. The method according to claim 2, wherein the measured variables (NBD-MG) of the second time series (ZR2) are not based on medical imaging. 5 . The method according to claim 4 , wherein the measured variables (NBD-MG) of the second time series ( ZR2 ) include one or more laboratory values ​​of the patient, in particular blood count values.

6. The method according to any one of the preceding claims, wherein the step (S30) of determining the observable quantity (OBS) comprises: - obtaining (S31) a single event associated with the patient within said time period, - determining (S32) relevance information (KOR) based on the time series (ZR1) and the single event, wherein the relevance information (KOR) specifies a measure for the temporal relevance between the time series (ZR1) and the single event, and - obtaining (S33) the observable quantity (OBS) based on the correlation information (KOR).

7. The method according to any one of the preceding claims, further comprising the steps of: - obtaining (S0) background information (KI) related to the patient's condition, in, - the step of extracting (S20) comprises selecting an image data based measurement variable (BD-MG) from a plurality of different image data based measurement variables (BD-MG) based on said background information (KI), and The time series (ZR1) is a time series of selected image data-based measurement variables (BD-MG).

8. The method according to claim 1, wherein the medical image data set (SET) comprises a recording of an entire body part or a whole-body recording of the patient.

9. The method according to any one of the preceding claims, wherein the measured variables (BD-MG) of the time series (ZR1) are selected from: - body composition analysis values, BCA values, including in particular muscle content and / or body fat content, - size of the lesion, and / or - The tumor burden in a body part or body region of the patient or in the entire body of the patient.

10. The method according to any one of the preceding claims, wherein the step of providing (S40) comprises: - comparing the observable quantity (OBS) with a preset observable quantity pattern (S41), and - displaying (S42) said time series (ZR1) and / or said observables (OBS) based on said comparison.

11. The method according to any one of the preceding claims, further comprising the steps of: - determining (S50) a medical diagnosis (DIG) based on said observable (OBS), and - providing (S60) said medical diagnosis (DIG).

12. The method according to claim 11, wherein the step of determining (S50) the medical diagnosis (DIG) comprises: - selecting (S51) a disease pattern from a plurality of selected disease patterns based on said observable (OBS), - querying (S52) the patient's electronic patient record (EPA) from the database, - matching the selected disease pattern with entries in the electronic patient record (EPA) (S53), and - determining (S54) said medical diagnosis (DIG) based on said matching (S53).

13. The method according to any one of the preceding claims, further comprising the steps of: - comparing said observable (OBS) with corresponding observables of a plurality of reference patients different from said patient (S71), - selecting (S72) a comparison patient from said plurality of reference patients based on the step of comparing, - providing (S73) information about the comparison patient (VI).

14. The method according to any one of the preceding claims, further comprising the steps of: - determining (S80) a treatment option (BHO) based on said observable (OBS), and - providing (S90) said treatment option (BHO).

15. A system for providing an observable (OBS) indicative of a medical diagnosis, the system having a computing device (20) and an interface (26), wherein the interface (20) is configured to: - obtaining an image data sequence (SE) of a patient, wherein the image data sequence (SE) comprises a plurality of medical image data sets (SET), which are respectively recorded at different times in a time period, and The computing device (20) is configured to: - extracting a time series (ZR1) from said sequence of medical image data (SE), - determining said observable (OBS) based on said time series (ZR1), and - providing said observable (OBS) via said interface (26). 16 . A computer program product comprising a program and being directly loadable into a memory of a programmable computing unit, the computer program product having program means for carrying out the method according to claim 1 when the program is executed.

17. A computer-readable storage medium having a readable and executable program segment stored thereon, so that when the program segment is executed, all the steps of the method according to any one of claims 1 to 14 are performed.

Images