Automatic analysis of 2D medical image data with additional objects

Abstract

This invention relates to the automatic analysis of 2D medical image data with an additional object. A method for automatically analyzing 2D medical image data (MID) including an additional object (EO) is described. According to this method, medical image data (MID) including the additional object (EO) is acquired from the examination area (ROI) of the patient (P) by a first modality (2), and additional image data (AID) is acquired from the examination area (ROI) using a different modality (C). Automatic image analysis applicable to the additional object (EO) is performed based on the acquired medical image data (MID) and the acquired additional image data (EID). Furthermore, analysis devices (40, 50) are described. A medical imaging system (1) is also described.

Description

Technical Field

[0001] This invention relates to a method for automatically analyzing 2D medical image data including additional objects. The invention also relates to an analysis apparatus. Furthermore, the invention relates to medical imaging systems. Background Technology

[0002] The placement and insertion of external devices are frequently used in clinical practice for life support purposes and patient monitoring. For example, endotracheal tubes are commonly used for ventilation, while intravenous catheters are typically inserted for medication or pressure monitoring. A major challenge in analyzing two-dimensional medical images using external devices is projecting both external and inserted objects onto a single image plane, which superimposes the density of the external device onto the internal tissue structures. This effect complicates image interpretation by obscuring important findings by overlaying the medical device.

[0003] Under the assumption that a sufficiently large training dataset can be used for image analysis applications that essentially aim to learn to distinguish between in vitro objects and inserted devices, it may be feasible to use deep learning to detect and identify devices. A prominent example is the assessment of central venous catheter position, explored by researchers in the following literature: Subramanian et al., “Automated Detection and Type Classification of Central Venus Catheters in Chest X-rays”, https: / / arxiv.org / abs / 1907.01656; and Hansen et al., “Radiographic Assessment of CVC Malpositioning: How can AI best support clinicians?”, https: / / openreview.net / pdf?id=ImcP8kkqtfZ.

[0004] However, this strategy automatically leads to the need to utilize large training datasets to capture any type of variable generated by in vitro devices, which appears particularly challenging in intensive care unit situations with high intrinsic image diversity, for example.

[0005] A second possibility for improving image analysis applications is segmenting any type of external object and inserted object. For example, the authors in the following paper also identified the tip location of a peripherally inserted central catheter by simultaneously segmenting and classifying other inserted devices and external devices: Lee et al., “A Deep-Learning System for Fully-Automated Peripherally Inserted Central Catheter (PICC) Tip Detection”, https: / / dx.doi.org / 10.1007 / s10278-017-0025-z. Of course, this comes at the cost of very high annotation complexity.

[0006] Another possibility for considering various in vitro devices without using large training datasets and annotation work is to manually create synthetic cases. Such an approach is described in Yi et al., “Automatic Catheter and Tube Detection in Pediatric X-ray Images Using a Scale-Recurrent Network and Synthetic Data”, https: / / arxiv.org / pdf / 1806.00921.pdf. The risk of this approach is the potential mismatch between training and testing conditions due to concept drift.

[0007] Therefore, there is a challenge in achieving high quality when analyzing medical image data from examination subjects, including both external and internal devices. Summary of the Invention

[0008] The aforementioned problems are solved by the method for automatically analyzing 2D medical image data including additional objects according to the technical solution of the present invention, by the analysis device according to the technical solution of the present invention, and by the medical imaging system according to the technical solution of the present invention.

[0009] According to the method for automatically analyzing 2D medical image data including additional objects, 2D medical image data is obtained from the patient's examination portion (also known as the region of interest).

[0010] Typically, an additional object is defined as an object intended to be distinguishable, but which, based solely on information from acquired 2D medical image data, cannot be distinguished from, or is difficult to distinguish from, a part of the patient's body or other parts of the patient's body located at the same or near the location of the additional object in the 2D medical image data. For example, an additional object may include a foreign body that is not typically part of a healthy person's body, or a foreign body comprising materials different from the tissues of a healthy person.

[0011] In this context, it must be clearly stated that the term "additional object" also includes cases where there is more than one additional object. Specifically, the term "additional object" includes various external objects, i.e., objects located outside the patient's body. For example, typically, multiple external objects may be present on the patient's body. Such external objects may include catheter ports, ECG devices (ECG = electrocardiogram), cables, endotracheal tubes, or tubing, etc. The patient can be human or animal. Medical images may include, for example, X-ray images, or 2D projections of CT or MR images.

[0012] However, different modalities are also used to acquire supplementary image data from the examination section. These different modalities involve different types of technology, and therefore the image information acquired through different modalities differs from the image information in the medical image data. For example, different modalities used to acquire supplementary image data use different physical principles compared to those used to acquire 2D medical image data. Due to these different modalities, the information about the supplementary object extracted from the supplementary image data may differ from the information about the supplementary object derived from the 2D medical image data. Supplementary image data may also include 2D image data, but it is not limited to this. For example, supplementary image data may also include 3D image data acquired by a 3D image acquisition unit.

[0013] Preferably, the 2D medical image data includes information about the interior of the human body, and additional image information in different modalities includes information about the exterior of the human body. However, the invention is not limited to the preferred variations. As discussed later, different modalities may also include acquisition techniques that demonstrate the intensity of metabolism within the human body for identifying and locating tumors. In this case, the "additional object" includes tumors that must be distinguished from the bones or other parts of the patient's body.

[0014] Furthermore, automated image analysis applicable to additional objects is performed based on the acquired 2D medical image data and the acquired supplementary image data. As discussed later, the results of the analysis may include the identification or location of supplementary objects (preferably external objects) in the 2D medical image data and / or supplementary image data. Also as discussed later, information from both modalities may be used individually in subsequent analysis sub-steps or in combination in a single analysis step. The results of the analysis may also include segmentation, annotation, or information of examination portions that can be used for evaluation or diagnosis.

[0015] Advantageously, additional data sources, including information about the examination portion and, in particular, additional information about additional objects, are used to improve medical image analysis applications. Specifically, the identification and differentiation of objects located at the same or similar locations in 2D medical images can be achieved based on the additional image information. As detailed later, such medical image analysis applications can include improved applications for the identification, localization, and extended use of examination procedures and automated post-processing algorithms, suitable for the identification of additional objects.

[0016] The analysis apparatus according to the invention includes a first input interface for acquiring 2D medical image data from the patient's examination portion.

[0017] The examination includes an additional object intended to be distinguished, but based solely on information from the acquired 2D medical image data, it is not possible or very difficult to distinguish the additional object from a part of the patient's body or other parts of the patient's body located at the same or near the location of the additional object in the 2D medical image data.

[0018] As described above, acquired 2D medical image data can be obtained from various types of medical imaging systems, such as X-ray imaging systems, CT systems, or MR systems. The analysis device also includes a second input interface for acquiring additional image data from the examination area using different modalities. Also as mentioned above, the different modalities preferably include, but are not limited to, techniques for acquiring image data of the external lateral aspect of the human body. Furthermore, the analysis device includes an analysis unit for performing automated image analysis suitable for the additional object based on the acquired 2D medical image data and the acquired additional image data. This analysis device shares the advantages of methods for automatically analyzing 2D medical image data.

[0019] The medical imaging system (preferably an X-ray imaging system) according to the invention includes a scanning unit for acquiring measurement data from an examination portion of a patient, a reconstruction unit for reconstructing image data based on the acquired measurement data and generating 2D medical image data, and an analysis device. Furthermore, the medical imaging system includes an additional mode for acquiring image data from the examination portion, wherein this additional mode is technically different from the main mode of the medical imaging system. The medical imaging system shares the advantages of the analysis device according to the invention.

[0020] Most of the basic components of the analysis apparatus according to the invention can be designed as software components. This is particularly applicable to the analysis unit, but also to a portion of the input interface. However, in principle, especially when particularly fast computations are involved, some of these components can also be implemented in the form of software-supported hardware (e.g., FPGAs). Similarly, the required interfaces—for example, if it is simply a matter of transferring data from other software components—can be designed as software interfaces. However, the required interfaces can also be designed as hardware-based interfaces controlled by suitable software. Furthermore, some of the aforementioned components can be distributed and stored in a local network, regional network, global network, or a combination of network and software (especially cloud systems).

[0021] The largely software-based implementation offers the advantage that existing medical imaging systems can be easily retrofitted via software updates to operate in accordance with the invention. In this regard, this objective is also achieved through a corresponding computer program product having a computer program that can be directly loaded into, for example, the memory device of a medical imaging system. This computer program has program segments that, when executed in the medical imaging system, perform all the steps of the method according to the invention. In addition to the computer program, such a computer program product may also include additional components (such as documentation and / or add-ons) including hardware components for using the software, such as hardware keys (dongles, etc.).

[0022] For transmission to and / or storage on or within a medical imaging system, program segments of a computer program that can be read and executed by a computer unit of the medical imaging system are stored on a computer-readable medium, such as a memory stick, hard disk, or some other transferable or permanently mounted data carrier. The computer unit may include, for example, one or more cooperating microprocessors for this purpose.

[0023] The dependent claims and the following description each contain particularly advantageous embodiments and developments of the invention. In particular, a claim in one claim class may be further developed in a manner similar to a dependent claim in another claim class. Furthermore, within the scope of the invention, various features of different exemplary embodiments and claims may be combined to form new exemplary embodiments.

[0024] In a variation of the method according to the invention for automatically analyzing 2D medical image data including external objects, the analysis is performed based on artificial intelligence. Preferably, the artificial intelligence includes a deep learning-based analysis method for evaluating the 2D medical image data. By using additional information from supplementary image data, the training effort and amount of training data required to train the AI-based analysis can be significantly reduced.

[0025] Preferably, based on the results of the first analysis, the automated image analysis includes a separate first analysis of the supplementary image data and a separate second analysis of the 2D medical image data. In this preferred variant, the analysis is structured as a sequential process, wherein first, the supplementary image data is analyzed and the results of the first analysis are used for the second analysis of the actual 2D medical image data. Compared to a comprehensive analysis, the sequential approach allows for monitoring of intermediate results during the analysis process, which improves the transparency and validation of the analysis. Furthermore, this variant leverages the possibility of better visibility of external objects in the supplementary image data to improve subsequent medical image analysis applications based on the 2D medical image data. Additionally, the scope of the training data base for dedicated image analysis applications can be reduced based on the pre-analysis of the supplementary image data. Furthermore, the 2D image data can be divided into different sub-regions based on the analysis of the supplementary image data, and then these sub-regions can be analyzed using different analysis applications in the second analysis.

[0026] In a variation of the method for automatically analyzing 2D medical image data including external objects using a sequential analysis process, the first analysis includes the step of identifying and / or locating the external objects in supplementary image data. Advantageously, the identification and location of the external objects in the supplementary image data can be used to selectively adjust a second analysis applied to the medical image data itself.

[0027] In a variation of the method according to the invention for automatically analyzing 2D medical image data including external objects, the additional image data includes external image data. This means that the external image data provides information about the exterior of the patient to be examined. Preferably, an optical image acquisition device—particularly preferably a camera—is used to acquire the external image data. Advantageously, the external image data is complementary and enhances the information about the interior of the patient to be examined. Furthermore, using a camera utilizes the visibility of the external object on the camera recording to improve medical image analysis applications. Moreover, advantageously, the invention can be implemented in future product versions (e.g., mobile X-ray imaging systems) that include a camera as a basic component without additional hardware effort.

[0028] In a variation of the method according to the invention for automatically analyzing 2D medical image data including external objects, the first analysis includes the step of extracting information about the type and / or location of the external objects from the external image data.

[0029] Then, identify the type of external object and / or locate the external object in the external image data.

[0030] Based on the preceding identification and / or localization, select a specialized medical image analysis application to identify and / or localize external objects in the medical image data.

[0031] Furthermore, the second analysis includes a step of identifying and / or locating external objects in the medical image data based on a medical image analysis application. Advantageously, the second analysis step can be tailored based on the first analysis, which reduces the workload of training the medical image analysis application used for the second analysis.

[0032] In another variation of the method according to the invention for automatically analyzing 2D medical image data including external objects, the extraction step includes detecting the route of the external object in the external image data. Furthermore, the step of selecting a dedicated medical image analysis application includes adjusting the analysis application to focus on the detected route. Advantageously, regions where artifacts caused by external objects must be suppressed can be defined.

[0033] This adjustment may include taking into account the route of the external object through at least one of the following steps:

[0034] -Consider the route of external objects through the following operations:

[0035] - Suppress overlapping external object regions in medical images, and / or

[0036] - Weight the segmentation results or heatmap regression results to avoid confusion caused by imitators, and / or

[0037] - Process medical image sub-regions with overlapping external objects separately. For example, locate broken bones in a region captured by a plaster model.

[0038] We can use segmentation or heatmap results to selectively suppress the detection of external objects that might be confused with internally inserted objects. For example, a line detector algorithm might confuse an external ECG line with an internal CVC line (CVC = central venous catheter). With such an additional modality (i.e., a camera), we can detect the ECG line and suppress it to avoid false detection. Mimics are objects or radioactive markers that appear similar to the object of interest.

[0039] Two different analytics applications can be used to process medical image sub-regions with overlapping external objects: one analytics application trained on image sub-regions with overlapping objects and another analytics application trained on image sub-regions without overlapping objects.

[0040] Advantageously, the modified image analysis application can be limited to a finite sub-section of the inspection section, which helps to reduce the workload of using the modified image analysis application.

[0041] In another variation of the method according to the invention for automatically analyzing 2D medical image data including external objects, the first analysis includes the detection of the insertion point and / or type of a portion of the external device. For example, the basic image analysis application focuses only on segmenting the internal wiring / tube routes and tips of the human body, without considering wiring / tube routes outside the patient's body. This additional information can be used to adjust the image analysis application so that the external portion of the external device is also used for image analysis.

[0042] In another variation of the method according to the invention for automatically analyzing 2D medical image data including external objects, a portion of the external device includes a port catheter having a patterned cover that varies slightly from the tip to the port device, and the first analysis includes at least one of the following:

[0043] -Based on the type of catheter identified by the overlay pattern

[0044] -The insertion position of the port conduit is located based on the overlay pattern.

[0045] - Estimate insertion depth based on coverage pattern.

[0046] In a variation of the method according to the invention for automatically analyzing 2D medical image data including additional objects, the analysis is performed as a multi-model, artificial intelligence-based analysis based on a combination of two different data sources (i.e., additional image data and 2D medical image data). Advantageously, an end-to-end learning approach is implemented, meaning that only a single model needs to be trained for analyzing both the 2D medical image data and the additional image data. Deep neural networks are typically used as the analysis unit, which is used to process various types of information in the multi-model approach.

[0047] Additional image data may also include PET images from the patient, which are used to analyze metabolism in the examined area. For example, in 2D medical image data, an object resembling a lesion, possibly caused by a clavicle fracture or tumor, is detected in a portion of the clavicle. Because tumors emit metabolic products, the additional PET images can be used to correctly analyze the 2D medical image data. This means that if the object resembling a lesion emits metabolic products, it proves that the 2D image data includes a tumor. If the object resembling a lesion does not emit any metabolic products, the object resembling a lesion is likely caused by a clavicle fracture. Attached Figure Description

[0048] The invention will now be described again with reference to the accompanying drawings. In the different drawings, the same parts are provided with the same reference numerals. The drawings are generally not drawn to scale.

[0049] Figure 1 An X-ray imaging system according to an embodiment of the present invention is shown;

[0050] Figure 2 A flowchart illustrating a method for automatically analyzing 2D medical image data including external objects according to an embodiment of the present invention is shown;

[0051] Figure 3 A flowchart illustrating a method for automatically analyzing 2D medical image data including external objects according to a second embodiment of the present invention is shown;

[0052] Figure 4 A schematic diagram of an analysis apparatus according to an embodiment of the present invention is shown;

[0053] Figure 5 A schematic diagram of an analysis apparatus according to a second embodiment of the present invention is shown;

[0054] Figure 6 A flowchart illustrating a method for automatically analyzing 2D medical image data including external objects according to a third embodiment of the present invention is shown;

[0055] Figure 7A flowchart illustrating a method for automatically analyzing 2D medical image data including external objects according to a fourth embodiment of the present invention is shown. Detailed Implementation

[0056] Figure 1 An X-ray imaging system 1 is shown, which includes... Figure 4 The analysis device 40 is shown in detail. The X-ray imaging system 1 mainly consists of a conventional scanning unit 2, which includes an X-ray detector 2a and an X-ray source 2b opposite to the X-ray detector 2a. Additionally, a patient stage 3 is present, on which the upper part of the patient stage 3, on which the patient P is located, can be moved to the scanning unit 2 to position the patient P below the X-ray detector 2a. An external object EO (e.g., an ECG system) is placed on the patient P. The scanning unit 2 and the patient stage 3 are controlled by a control device 4, which acquires control signals ( Figure 1 (Not shown) The data comes from the control device 4 for controlling the imaging process, and the control device 4 receives measurement data MD from the X-ray detector 2a. The control device 4 also includes a reconstruction unit 4a for reconstructing 2D medical image data MID based on the received measurement data MD.

[0057] Furthermore, the X-ray imaging system 1 also includes a camera C, which acquires external image data (EID) from the patient P and transmits this additional external image data (EID) to the control device 4. The control device 4 also includes an analysis device 40 according to the invention described above, which is used for evaluation based on 2D medical image data (MID) and external image data (EID).

[0058] The components of the analysis device 40 can be implemented primarily or entirely as software components on a suitable processor. In particular, the interfaces between these components can also be designed purely in software. All that is required is access to a suitable storage area where data can be temporarily stored and retrieved and updated at any time.

[0059] Figure 2 A flowchart illustrating a method for automatically analyzing 2D medical image data MID including external objects according to an embodiment of the present invention is shown.

[0060] In step 2.I, the 2D medical image data MID is generated as follows: Figure 1 The X-ray imaging system 1 described herein is obtained from the examination area ROI of patient P.

[0061] In step 2.II, external image data EID is acquired from the ROI of the patient P using camera C.

[0062] In step 2.III, the external object (EO) is identified and located based on the analysis of the additional external image data (EID) and the 2D medical image data (MID). Segmentation of the 2D medical image data (MID) is performed based on AI (Artificial Intelligence) and using additional information from the external image data (EID).

[0063] exist Figure 3 The diagram shows a flowchart 300, which illustrates a method for automatically analyzing 2D medical image data MID including external objects (EO) according to a second embodiment of the present invention.

[0064] In step 3.I, external image data EID is acquired from the ROI of the patient P using camera C.

[0065] In step 3.II, the 2D medical image data MID is generated as follows: Figure 1 The X-ray imaging system 1 described herein is obtained from the examination area ROI of patient P.

[0066] In step 3.III, information about the type T and location POS of the external object EO is obtained from the acquired external image data EID, and the type T of the external object EO is identified and the external object EO in the external image data EID is located.

[0067] In step 3.IV, a dedicated medical image analysis application (MIAA) is selected based on the previous identification and location. For example, if a plaster cast has been identified as an external object, a different dedicated medical image analysis application is selected than the one used for cases without a plaster cast.

[0068] In step 3.V, the selected medical analysis application MIAA is applied to the 2D medical image data MID to segment the 2D medical image data MID and perform AI-based annotation and analysis on the region of interest (ROI).

[0069] exist Figure 4 The diagram shows an analysis apparatus 40 according to a first embodiment of the present invention. The analysis apparatus 40 includes a first input interface unit 41 for receiving 2D medical image data (MID) of an object to be examined (EO), such as X-ray image data of a patient's chest region. The analysis apparatus 40 also includes a second input interface unit 42 for receiving additional external image data (EID) acquired by a camera from the patient's chest region. The analysis apparatus 40 further includes an analysis unit 43 for identifying and / or locating the external object (EO) based on analysis of the additional external image data (EID) and the medical image data (MID).

[0070] exist Figure 5 The diagram shows an analysis apparatus 50 according to the second embodiment.

[0071] Similar to the first embodiment, the analysis device 50 includes an input interface unit 41 for receiving 2D medical image data (MID) of the object being examined (EO), such as CT image data of the patient's chest region. The analysis device 50 also includes a second input interface unit 42 for receiving additional external image data (EID) from a camera and acquired from the patient's chest region.

[0072] Compared with the first embodiment, the analysis device 50 according to the second embodiment includes an object detection and localization unit 51, which is used to obtain information about the type T and location POS of an external object EO from the acquired external image data EID.

[0073] In addition, the analysis device 50 includes an application selection unit 52 for selecting a dedicated medical image analysis application MIAA based on the type T of the object EO and the previous identification of the location POS.

[0074] In addition, the analysis device 50 includes an analysis unit 53, which is used to apply MIAA to the acquired medical image data MID and output some results (such as fragments, annotations, etc.).

[0075] exist Figure 6 The diagram shows a flowchart 600, which illustrates a method for automatically analyzing 2D medical image data including an external object (EO) according to a third embodiment of the present invention.

[0076] In step 6.I, external image data EID is acquired from the ROI of the patient P using camera C.

[0077] In step 6.II, the 2D medical image data MID is generated as follows: Figure 1 The X-ray imaging system 1 described herein is obtained from the examination area ROI of patient P.

[0078] In step 6.III, information about the type T and route CS of the external object EO (ECG cable in this embodiment) is extracted from the acquired external image data EID.

[0079] In step 6.IV, a dedicated medical image analysis application (MIAA) is selected based on the previous identification and location. The dedicated medical image analysis application (MIAA) is modified such that overlapping medical image regions of external devices are suppressed during analysis.

[0080] In step 6.V, a modified medical analysis application, MIAA, is applied to the 2D medical image data MID to segment the MID and perform AI-based annotation and analysis on the examined portions (i.e., regions of interest, ROIs). An alternative approach is to weight the results of the medical image analysis application, specifically by weighting the segmentation results or heatmap regression results to avoid confusion caused by imitators.

[0081] exist Figure 7 The diagram shows a flowchart 700, which illustrates a method for automatically analyzing 2D medical image data including an external object (EO) according to a fourth embodiment of the present invention.

[0082] In step 7.I, external image data EID is acquired from the ROI of the patient P using camera C.

[0083] In step 7.II, the medical image data MID is generated as follows: Figure 1 The X-ray imaging system 1 described herein is obtained from the examination area ROI of patient P.

[0084] In step 7.III, information about the overlay pattern PT of the port conduit is extracted from the external image data EID, which varies slightly from the tip to the port device.

[0085] In step 7.IV, the type T of the port catheter is automatically identified based on the unique overlay pattern PT. Furthermore, the insertion position IP of the port catheter is located based on external image data EID and the determination of the visible end of the port catheter, which also has a unique pattern. Additionally, since the overlay pattern PT varies slightly with the length of the catheter from port to tip, the insertion depth ID is estimated based on the overlay pattern PT at the insertion position IP.

[0086] In step 7.V, a medical analysis application focused on segmenting the internal suture / tube pathways and tips is selected (MIAA). In step 7.IV, information about the suture / tube pathways outside the patient's body is received from the analysis of external image data (EID). In this way, the entire port catheter from port to tip is automatically segmented and identified in detail in step 7.VI based on the 2D image data (MID) and the information received in step 7.IV. Steps 7.III through 7.VI can be implemented automatically, preferably using an AI-based analysis application, allowing the method to be executed automatically without any human intervention.

[0087] The above description is merely a preferred embodiment of the present disclosure and is not intended to limit the present disclosure. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present disclosure should be included within the scope of protection of the present disclosure.

[0088] Furthermore, the use of the indefinite article "a" or "one" does not preclude the possibility that the mentioned feature may appear multiple times. Similarly, the terms "unit" or "device" do not preclude the possibility that it is composed of some parts that may also be spatially distributed.

Claims

1. A method for automatically analyzing 2D medical image data (MID) including attached objects (EO), in, The additional objects include external objects (EOs). The method includes the following steps: - Using the first modality (2), 2D medical image data (MID) is acquired from the examination area (ROI) of the patient (P). - Use different modalities (C) to acquire additional image data (EID) from the inspected area (ROI), wherein the additional image data (EID) includes external image data. - Perform automated image analysis applicable to the attached object (EO) based on the acquired 2D medical image data (MID) and the acquired additional image data (EID). The automatic image analysis includes: - A separate first analysis of the additional image data (EID), and -A subsequent second analysis of the medical image data (MID) based on the results of the first analysis. Its features are, -The first analysis includes: - Extract information about the type and / or location of the external object (EO) from the external image data (EID). - Identify the type (T) of the external object (EO), and / or locate the external object (EO) in the external image data (EID). - Select a dedicated Medical Image Analysis Application (MIAA) based on prior identification and / or localization to identify and / or locate external objects (EOs) in the medical image data (MID), and -The second analysis includes: - Use the Medical Image Analysis Application (MIAA) to identify and / or locate external objects (EOs) in the Medical Image Data (MID).

2. The method according to claim 1, wherein, The first analysis includes the steps of identifying and / or locating additional objects (EOs) in the additional image data (EID).

3. The method according to claim 1, wherein, - The extraction step includes: detecting the path (CS) of external objects (EOs) in the external image data (EID), and - The steps for selecting a dedicated medical image analysis application (MIAA) include: adjusting the analysis application to focus on the detection path (CS).

4. The method according to claim 3, wherein, Adjusting the analytics application includes considering the route (CS) of the external object (EO) through at least one of the following steps: - Suppress overlapping external object (EO) regions in medical images, and / or - Weight the segmentation results or heatmap regression results. - Process medical image sub-regions with the overlapping external objects (EOs) respectively.

5. The method according to any one of the preceding claims, wherein, The first analysis includes detecting the insertion location (IP) and / or type (T) of a portion of the external device.

6. The method according to claim 5, wherein, - The external device includes a port catheter having a cover comprising a pattern (PT) that varies slightly from the tip to the port device. - The first analysis includes at least one of the following: - The type (T) of the port conduit is identified based on the overlay pattern (PT). -The insertion position (IP) of the port conduit is located based on the overlay pattern (PT). - Estimate the insertion depth (ID) based on the coverage pattern (PT).

7. An analytical apparatus (40, 50), comprising: - A first input interface (41) is used to acquire 2D medical image data (MID) including additional objects (EO) from the examination area (ROI) of the patient (P) acquired through the first modality (2), wherein the additional objects include external objects (EO). - A second input interface (42) is used to acquire additional image data (EID) from the inspection area (ROI) using different modalities (C), wherein the additional image data (EID) includes external image data. - Analysis units (43, 53) are used to perform automatic image analysis applicable to the attached object (EO) based on the acquired 2D medical image data (MID) and the acquired additional image data (EID). The automatic image analysis includes: - A separate first analysis of the additional image data (EID), and -A subsequent second analysis of the medical image data (MID) based on the results of the first analysis. Its features are, -The first analysis includes: - Extract information about the type and / or location of the external object (EO) from the external image data (EID). - Identify the type (T) of the external object (EO), and / or locate the external object (EO) in the external image data (EID). - Select a dedicated Medical Image Analysis Application (MIAA) based on prior identification and / or localization to identify and / or locate external objects (EOs) in the medical image data (MID), and -The second analysis includes: - Use the Medical Image Analysis Application (MIAA) to identify and / or locate external objects (EOs) in the Medical Image Data (MID).

8. A medical imaging system (1), comprising: - Scanning unit (2), the scanning unit (2) being used to acquire measurement data (MD) from the examination area (ROI) of the patient (P), - Reconstruction unit (4a), said reconstruction unit (4a) being used to reconstruct image data (MID) based on acquired measurement data (MD), - The analytical apparatus (40) according to claim 7, - Additional modality (C), which is used to acquire additional image data (EID) from the inspected region (ROI).

9. A computer program product having a computer program that can be directly loaded into a memory device of a medical imaging system (1), the computer program having program segments for performing all steps of the method according to any one of claims 1 to 6 when the computer program is executed in the medical imaging system (1).

10. A computer-readable medium storing a program segment that can be read and executed by a computer unit so as to perform all steps of the method according to any one of claims 1 to 6 when the program segment is executed by the computer unit.

Images