Endoscope imaging control method and endoscope imaging control system

Abstract

The invention relates to an endoscope imaging manipulation method and an endoscope imaging manipulation system. The method comprises: capturing an endoscopic image of the laryngeal and / or pharyngeal tissue structure during examination of the laryngeal and / or pharyngeal using an endoscope inserted through the nose or mouth of the patient; feeding the captured endoscopic image to an instance of artificial intelligence trained to identify a suspicious region of the laryngeal and / or pharyngeal tissue structure, the suspicious region showing a different sign from healthy laryngeal and / or pharyngeal tissue; superimposing a marker on the captured endoscopic image, the marker indicating an area indicated as suspicious by the instance of artificial intelligence; and displaying the superimposed captured endoscopic image on a monitor.

Description

Technical Field

[0001] The invention relates to an endoscope imaging manipulation method and an endoscope imaging manipulation system. Background Art

[0002] Laryngeal and pharyngeal biopsy is a routine diagnostic procedure in ENT (ear, nose, and throat) medicine to evaluate and examine the larynx and other parts of the upper aerodigestive tract. The procedure is performed using a flexible or rigid endoscope inserted through the patient's nose or mouth. This allows visualization of the pharynx and larynx and identification of any suspicious areas or changes in the mucosal layers that may indicate inflammation or cancer.

[0003] Upon identification of a suspicious area in the pharynx or larynx, a biopsy from the suspicious area may be performed by removing a small tissue sample for examination. Historically, such biopsies have been performed under general anesthesia using a microscope. However, this approach involves significant costs, including operating room staff and the use of general anesthesia, as well as the potential for complications due to the use of general anesthesia, particularly in patients with comorbidities. Additionally, increased waiting lists due to the pandemic may result in delays in biopsy procedures, which may impact diagnostic accuracy and patient outcomes.

[0004] The primary goal of the above surgeries is to preserve anatomy, thereby ensuring good patient outcomes, including preservation of voice and function. These goals are achieved using less invasive biopsy techniques.

[0005] In this regard, an alternative method for performing a biopsy of the larynx is to use local anesthesia in an office setting, usually performed immediately after a diagnostic laryngoscopy. This variation of the technique involves introducing a flexible endoscope, which may be reusable or single-use, through the nose. Once a suspicious area is identified, a small clamp is inserted through the channel of the endoscope to obtain a tissue sample. The procedure takes only a few minutes and has been shown to be just as safe and efficient as traditional methods, while improving patient outcomes and diagnostic accuracy and reducing overall costs to the hospital.

[0006] However, this technique may present certain challenges during execution. When the patient is awake during surgery, anatomical structures may move and physiological reflexes may occur, which will potentially affect the success of the biopsy and require multiple attempts to obtain an adequate tissue sample. In addition, the patient may experience discomfort during the procedure.

[0007] These situations run counter to guidance recommendations, which recommend preserving as much tissue as possible to maintain voice and function.

[0008] One of the major challenges in performing a biopsy under local anesthesia is the precise identification of the core of the lesion. Determining the most likely anatomical location to yield the most accurate histological results is critical to the success of the procedure.

[0009] To achieve this, training and experience are necessary. However, trained and experienced physicians may be in short supply, especially in rural areas. In conclusion, in the areas and cases mentioned above, the traditional method of biopsy under general anesthesia will continue to be used, with its limitations explained previously.

[0010] Given this, there is a need for solutions that enable more ENT physicians to perform laryngeal biopsies under local anesthesia. Summary of the invention

[0011] This need is met by an endoscopic imaging manipulation method comprising:

[0012] capturing endoscopic images of tissue structures of the larynx and / or pharynx during examination of the larynx and / or pharynx using an endoscope inserted through the patient's nose or mouth,

[0013] feeding the captured endoscopic images to an instance of an artificial intelligence that is trained to identify suspicious areas of laryngeal and / or pharyngeal tissue structure that show signs of being different from healthy laryngeal and / or pharyngeal tissue,

[0014] superimposing markers on the captured endoscopic image, the markers indicating areas indicated as suspicious by the instance of artificial intelligence, and

[0015] The superimposed captured endoscopic image is displayed on a monitor.

[0016] Many clinical studies have shown correlations between different vascular patterns and various disease states. Changes in vascular morphology and density have been shown to reflect the severity of the disease. At this time, four different vascular classifications based on NBI have been widely adopted in the field of otolaryngology. These classifications help visually identify the type of lesions and diseases. The methods and systems of the present disclosure are directed to enhancing the diagnostic capabilities of otolaryngologists by combining examples of artificial intelligence (AI) and machine learning (ML) techniques with NBI to guide the targeted biopsy process, thereby improving patient treatment outcomes.

[0017] Such an overlay, for example to be displayed within a specific GUI (Graphical User Interface), will identify lesions and margin delineation with a higher accuracy, since small blood vessels that may be part of the lesion are usually not visible to the human eye. If missed, recurrence may result. Due to the use of NBI images, ENT physicians can be assisted by the above overlay even if they have little experience (e.g., lack of cases in less populated areas).

[0018] In an embodiment, the endoscopic image is an endoscopic NBI image. Narrow-band imaging (NBI) is a medical imaging technique that uses light of a specific wavelength to enhance the visualization of blood vessels, mucosa, and other structures. This technique is widely used in the ENT (ear, nose, and throat) field to improve the visualization of blood vessels.

[0019] In an embodiment, an example of artificial intelligence may be a convolutional neural network (CNN) having a classifier that has been trained by at least one of supervised learning and unsupervised learning of multiple endoscopic images of laryngeal and / or pharyngeal structures to classify suspicious areas of laryngeal and / or pharyngeal tissue in captured endoscopic images. Thus, using preset training data, images and videos and the learning capabilities of the system, the system may be trained to classify suspicious areas in NBI images of the larynx.

[0020] The term unsupervised learning generally describes a process in which only images of healthy laryngeal and / or pharyngeal tissue are used to train the neural network. Any area that does not conform to what the neural network considers to be healthy will be marked as suspicious. Supervised training takes the training a step further. Diseased areas are pre-labeled and classified in the training data so that the neural network is trained to not only identify suspicious areas, but also provide an estimate of what kind of lesion is shown in the NBI image.

[0021] In an embodiment, an instance of an artificial intelligence may have been trained to identify differences from healthy laryngeal and / or pharyngeal tissue as suspicious due to one or more of lesions, vascular morphology, vascular density, vascular pattern, and mucosal surface structure. Thus, the disclosed algorithm utilizes image analysis techniques to help otolaryngologists identify vascular patterns, density, morphology, and structure on mucosal surfaces. By providing such information, the algorithm is intended to support the physician's diagnostic capabilities and help identify potential pathologies.

[0022] In further embodiments, the overlay may be created as a color-coded or brightness-coded heat map. In particular, in further embodiments, the color coding or brightness coding of the heat map may be based on at least one of the density and pattern of blood vessels in the suspicious area.

[0023] Color-coded (heat) map overlays based on the density and pattern of blood vessels enhance targeted biopsies and reduce the likelihood of false results or unnecessary invasive procedures. Thicker and irregular areas can be represented by darker colors, while regular and normal blood vessels will be represented by lighter colors (e.g., greener). The color scheme can be determined according to the established classification.

[0024] Physicians will be able to view a heat map overlay on a specified lesion. The AI ​​instance will guide the physician to perform a biopsy in the darkest area of ​​the map, which represents the core of the lesion, where histology results are likely to indicate the highest degree of disease.

[0025] In an embodiment, the artificial intelligence instance may be configured to learn from new endoscopic images captured during subsequent laryngeal and / or pharyngeal examinations after initial training.

[0026] Upon request, the captured endoscopic image can be displayed on a monitor without an overlay, allowing the physician to examine suspicious areas that are not blocked by the overlay. Thus, the AI ​​instance can be activated or deactivated during surgery, depending on the physician's needs at any given time during surgery.

[0027] The object is also achieved by an endoscopic imaging manipulation system, which includes: a video endoscope suitable for being introduced through a patient's mouth or nose for laryngeal and / or pharyngeal examination; a light source capable of providing white light for white light imaging (WLI) and narrow band illumination for narrow band imaging (NBI), the light source being connected to the video endoscope or integrated into the video endoscope; an image analyzer connected to the video endoscope for receiving endoscopic images from the video endoscope, the image analyzer having an instance of artificial intelligence, the instance of artificial intelligence being trained to identify suspicious areas of laryngeal and / or pharyngeal tissue structures, the suspicious areas showing signs different from healthy laryngeal and / or pharyngeal tissues, the image analyzer being further configured to superimpose a mark on the captured endoscopic image, the mark indicating the area indicated as suspicious by the instance of the artificial intelligence; and a monitor connected to the image analyzer for displaying the endoscopic image provided by the image analyzer.

[0028] The system achieves the same features and advantages as the endoscopic imaging manipulation method.

[0029] In an embodiment, the image analyzer may be configured to apply an instance of artificial intelligence to the captured endoscopic image based on one of an imaging mode recognition signal or an image characteristic in the captured image that indicates NBI when NBI is applied.

[0030] In another embodiment, the image analyzer is configured to apply an instance of artificial intelligence to the captured endoscopic image upon request by a practitioner.

[0031] In an embodiment, an instance of artificial intelligence may be a convolutional neural network (CNN) having a classifier that has been trained by at least one of supervised learning and unsupervised learning on multiple endoscopic images of laryngeal and / or pharyngeal structures to classify suspicious areas of laryngeal and / or pharyngeal tissue in captured endoscopic images.

[0032] In another embodiment, an instance of the artificial intelligence has been trained to identify differences from healthy laryngeal and / or pharyngeal tissue as suspicious due to one or more of lesions, vascular morphology, vascular density, vascular pattern, and mucosal surface structure.

[0033] In an embodiment, the image analyzer may be configured to create the overlay as a color-coded or brightness-coded heat map, wherein in particular the image analyzer may be configured to base the color coding or brightness coding of the heat map on at least one of a density and a pattern of blood vessels in the suspicious area.

[0034] The GUI can be configured with user-customized colors, shapes, and other options to display the estimated shape, size, classification, and confidence level of the identified lesions. A lesion map depicting the parameters of the mucosal changes can also be displayed, possibly accompanied by written information about the size and type, with an associated accuracy rating. The GUI can also display patient information, imaging modes such as WLI (white light imaging) and NBI, and the activation status of the system, particularly the AI ​​instance.

[0035] Additional information that may be displayed using the GUI may be automatic recommendations for the appropriate number of biopsies to ensure optimal sampling and comprehensive mapping of all suspicious areas, based on the heat map area, shape, and size, or automatic recommendations and suggestions for compatible accessories to prevent potential damage to the endoscope due to use of incorrect instruments, based on identification of the endoscope connected to the system via an ID chip embedded in the endoscope. The GUI may also be recorded for later study or analysis.

[0036] The artificial intelligence instance can be configured to learn from new endoscopic images captured during subsequent laryngeal and / or pharyngeal examinations after the initial training has been completed.

[0037] To continue learning, the image analyzer can be configured to receive feedback from the practitioner performing the laryngeal and / or pharyngeal examination regarding: unmarked suspicious areas or lack of suspicious areas in the laryngeal and / or pharyngeal tissue structures, and / or whether the classification of one or more suspicious areas of the laryngeal and / or pharyngeal tissue structures is correct.

[0038] In an embodiment, the system can reconstruct the information into a 2D lesion map, providing a clearer representation of the anatomical region and location in question. This will help doctors better identify and map suspicious areas.

[0039] Other features of the invention will become apparent from the description of embodiments according to the invention as well as from the claims and the included drawings.Embodiments according to the invention may achieve individual features or a combination of several features. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The invention is described below based on exemplary embodiments without limiting the general intent of the invention, wherein for the disclosure of all details according to the invention, reference is explicitly made to the drawings which are not described in more detail in the text. The drawings show:

[0041] Figure 1 is a schematic diagram of the steps of the imaging manipulation method,

[0042] Figure 2 is a schematic diagram of an embodiment of an example of artificial intelligence,

[0043] Figure 3 is an example of a GUI showing the results of the present method,

[0044] Figure 4 training data representing instances of artificial intelligence, and

[0045] Figure 5 is a schematic diagram of a system according to the present disclosure.

[0046] In the drawings, elements of the same or similar type or respectively corresponding parts are provided with the same reference numerals to avoid the need to reintroduce the item.

[0047] Reference numerals list

[0048] 10GUI

[0049] 12 Heatmap Overlays

[0050] 14 Indication of suspicious areas

[0051] 16 Recommendations for the number of biopsy specimens

[0052] 18 Recommended clamps to use

[0053] 20Tissues of the larynx and / or pharynx

[0054] 30 Endoscope

[0055] 32 Controller

[0056] 34 Image Analyzer with AI Examples

[0057] 36 Light Source

[0058] 38 Display

[0059] 40 Feedback terminal DETAILED DESCRIPTION

[0060] Figure 1 A schematic representation of the steps of an imaging manipulation method is shown, which can be performed using a system, an exemplary embodiment of which is shown in Figure 5. In step S10, narrow band imaging (NBI) images of the patient's larynx are captured using a video endoscope, which may have an image sensor located at its distal end, or may be of a different type with an image sensor at the proximal end, or a separate camera head attached to a purely optical endoscope. In step S20, these images are sent to an image processor, which runs an instance of an artificial intelligence that has been trained with NBI images of the larynx to identify suspicious areas in the mucosal surface or laryngeal tissue. In step S30, the instance of the artificial intelligence processes the NBI images and identifies suspicious areas based on its training. The identification of suspicious areas can be accomplished by classifying these areas as suspicious, or in more complex cases, by classifying different types of suspicious areas based on the underlying disease and / or severity of the lesion.

[0061] The instance of artificial intelligence provides the suspicious region and the classification as output, so that in step S40, the image analyzer can superimpose a mark indicating the suspicious region found by the instance of artificial intelligence on the NBI image. The superimposed image is then displayed to the person performing the inspection or biopsy on the monitor in step S50.

[0062] Figure 2 A schematic diagram of an embodiment of an artificial intelligence that can be used to identify suspicious areas in NBI images of the larynx is shown. This embodiment of the embodiment of the artificial intelligence is configured as a convolutional neural network (CNN) with a classifier on the output side. The NBI image is fed into the CNN on the input side and feature extraction is performed by processing within the convolutional base ("convolution") of the CNN and pooling ("pooling") according to the known principles of the operation of convolutional neural networks. The convolution and placement phases can be repeated once or several times. As a result of the feature extraction, features of the NBI image that can indicate suspicious areas are identified. These features are then set into the classification layer of the CNN, which can be a fully connected layer or several layers, and processed towards the output layer, where each node represents a classification that the CNN has been trained on. In the simplest case, there may be only a few nodes in the output layer, which represent the presence or absence of suspicious areas. In more complex embodiments, the output layer may have several nodes representing different diseases, lesion types, severity, etc.

[0063] The explanation of the principles using convolutional neural networks should not be considered limiting, as other examples of artificial intelligence and machine learning can also be trained and used for this purpose, such as support vector machines, learning vector quantization or random forest models.

[0064] Figure 3An embodiment of a GUI 10 representing the results of the present method is shown. The GUI 10 would be displayed on a monitor to a physician performing an endoscopy and possibly a biopsy of a patient's larynx, and is divided into several partial screens displayed adjacent to each other. The large partial screen on the right displays an NBI image of the larynx currently being captured by the endoscope, wherein healthy laryngeal and / or pharyngeal tissue 20 is shown in an unchanged manner. Since the NBI image has been processed by an instance of artificial intelligence trained for this purpose, areas of disease or lesions have been classified as suspicious and are indicated as overlays 12 in the form of a heat map within the image. The overlays 12 have different colors to indicate the confidence level of the classification, resulting in the center of the suspicious area being displayed in a darker shade than its surrounding areas. As shown in FIG. Figure 3 As shown, additional boundary lines are drawn around suspicious areas.

[0065] In the left portion of the GUI 10, several widgets are displayed that provide information about suspicious areas and information about recommendations based on examples of artificial intelligence regarding findings in suspicious areas. The top widget may contain patient information as well as the duration of the procedure and an indication of the progress of the examination in the form of a progress bar, above which are displayed symbols representing anatomical milestones that have been passed during the examination, such as the mouth (diamond symbol), tonsils (circular symbol), and larynx (shell symbol). These symbols are highlighted because the endoscope tip has passed or reached these milestones. Later milestones on the way to the lungs have not yet been reached and remain gray.

[0066] Below the first insert, the second insert shows the outline of the suspicious area relative to its location inside the larynx. From the second insert, it can be seen that the suspicious area is located in the upper right part of the larynx. Its relative size with respect to the larynx can also be derived from this graphical representation.

[0067] The third plug-in contains two pieces of information, namely a number ("2") which is a suggestion for the number of biopsies to be taken from a particular suspicious area on the display, and a second piece of information which is a suggestion for the type of clamp to be used for the biopsy. The clamp can be introduced through a channel in the endoscope so that the physician performing the examination and biopsy can see the operation of the clamp as she performs the biopsy. The physician may have the ability to cause the image analyzer to stop displaying the overlay 12 so that the lesion or diseased tissue to be biopsied can be observed, or to change the transparency of the overlay. In another embodiment, only the outline of the suspicious area will be retained in the image. The representation 14 on the left side may be retained, including various colors or shades of the heat map, so that the physician will have both a clear view of the area she is operating on and a heat map representing the classification, which may indicate the most promising location for obtaining a tissue sample by a specific coloring or shading or by a unique mark at the location of the suggested location. If these marks are visible but not obtrusive, they can be retained in the large image when the rest of the heat map is made transparent or removed from the image for the purpose of the biopsy.

[0068] After the imaging has been switched to white light imaging (WLI), the marking of the suspicious area that has been identified in the NBI image by the instance of artificial intelligence can remain in the display, where some lateral movement and magnification due to the positional shift of the endoscope is detected in the image. This is particularly helpful in the actual act of obtaining a biopsy sample from the suspicious area.

[0069] Figure 4 An exemplary representation of training data for an artificial intelligence instance is shown. The top training data is an unclassified NBI training image containing only healthy tissue. Such training data is required for so-called unsupervised training of artificial intelligence. The training will cause the instance of the artificial intelligence to accept any laryngeal and / or pharyngeal tissue and services that conform to these training data as non-suspicious. If there is a deviation from the structure that has been trained as normal, the affected area will automatically appear and be classified as a suspicious area.

[0070] An alternative case of so-called supervised learning is based on freely classified training images, some of which show only healthy tissue, while others contain lesions or diseased areas, where such affected areas are simply indicated and classified as suspicious, or more finely classified with respect to their type and severity. Supervised training leads to an optimization such that specific nodes in the output layer of the CNN are trained to indicate various types and severities, respectively.

[0071] Figure 4The third type of input image indicated in is a new endoscopic NBI image captured during an inspection and biopsy performed using a pre-trained artificial intelligence instance, and is used to continuously update the training of the artificial intelligence instance in both unsupervised and supervised learning. In the case of unsupervised learning, the physician performing the inspection may indicate that the current view of the larynx has no suspicious areas. In this case, the NBI image can be used to further establish the basic facts of the NBI training images for unsupervised learning. Alternatively, in the presence of a suspicious area, the physician can confirm the finding and classification of the suspicious area, or change a finding that he has found to be incorrect, so that this image including its correct classification can be used for further supervised training of the artificial intelligence instance.

[0072] Figure 5 is a schematic diagram of a system according to the present disclosure. The system includes an endoscope 30, which can be a flexible endoscope having optics and an image sensor at its distal end, or a rigid endoscope having internal relay optics and an image sensor in its handle or in a separate camera attached to the endoscope. The image sensor is connected to a controller 32 of the system for controlling imaging, illumination, and other parameters of the endoscopic examination. The controller 32 may include a light source 36 capable of producing white light for WLI and replacing narrow-band illumination for NBI. The light source may also be implemented in the endoscope 30 and controlled by the controller 32. The controller 32 also includes an image analyzer 34 that runs an instance of artificial intelligence that is trained to identify suspicious areas in the laryngeal and / or pharyngeal tissue according to the previous description. The image analyzer 34 is also configured to provide an overlay to the NBI image that visualizes or highlights suspicious areas detected by the instance of artificial intelligence. For example, with respect to Figure 3 An example of such a stack is shown and described.

[0073] The superimposed image is displayed using a monitor or display 38, for example, as shown in FIG. Figure 3 The form of GUI 10 shown may include additional information, such as information regarding the size, type and / or severity of the suspicious area, or information regarding recommendations on the number and / or locations of biopsy samples to be obtained or information regarding the type of clamp to be used.

[0074] In addition, the system may include a feedback terminal 40 that provides feedback to the physician performing the laryngeal examination regarding the artificial intelligence instance's findings regarding suspicious areas, such as confirming or changing such findings. Such feedback may be used for further training of the artificial intelligence instance, such as regarding Figure 4As explained. Such a feedback terminal 14 can be a separate device, such as a computer, or integrated into a display, as indicated by the dotted line surrounding the display 38 and the feedback terminal 40, in which case the display has a touch screen, and by touching the touch screen, the GUI 10 is displayed in a defined area of ​​the screen for feedback. For example, touching an insertion area of ​​the GUI 10 displaying information about the type or severity of a lesion or other suspicious area can open a dialogue to correct the automatically generated information with the physician's findings. Instead of input from the touch screen, correction can also be performed via voice recognition or other suitable means.

[0075] All named features, including those features obtained from the drawings alone, as well as individual features disclosed in combination with other features, are considered to be important to the invention individually and in combination. Embodiments according to the invention may be realized by a single feature or a combination of several features. Features combined with the words "especially" or "particularly" will be considered as preferred embodiments.

Claims

1. An endoscopic imaging manipulation method, wherein: The endoscopic imaging manipulation method comprises: capturing endoscopic images of tissue structures of the larynx and / or pharynx during examination of the larynx and / or pharynx using an endoscope inserted through the patient's nose or mouth, feeding the captured endoscopic images to an instance of an artificial intelligence that is trained to identify suspicious areas of laryngeal and / or pharyngeal tissue structure that show signs of being different from healthy laryngeal and / or pharyngeal tissue, superimposing a marker on the captured endoscopic image, the marker indicating an area indicated as suspicious by the instance of the artificial intelligence, and The captured endoscopic image is displayed superimposed on a monitor.

2. The endoscopic imaging manipulation method according to claim 1, wherein: The endoscopic image is an endoscopic narrow-band imaging image.

3. The endoscopic imaging manipulation method according to claim 1, wherein: An example of the artificial intelligence is a convolutional neural network with a classifier that has been trained by at least one of supervised learning and unsupervised learning on multiple endoscopic images of laryngeal and / or pharyngeal structures to classify suspicious areas of laryngeal and / or pharyngeal tissue in the captured endoscopic images.

4. The endoscopic imaging manipulation method according to claim 1, wherein: The instance of the artificial intelligence has been trained to identify differences from healthy laryngeal and / or pharyngeal tissue as suspicious due to one or more of lesions, vascular morphology, vascular density, vascular pattern, and structure of the mucosal surface.

5. The endoscopic imaging manipulation method according to claim 1, wherein: An overlay superimposed on the captured endoscopic image is created as a color-coded or brightness-coded heat map.

6. The endoscopic imaging manipulation method according to claim 5, wherein: The color coding or brightness coding of the heat map is based on at least one of the density and pattern of blood vessels in the suspicious area.

7. The endoscopic imaging manipulation method according to claim 3, wherein: The instance of the artificial intelligence is configured to learn from new endoscopic images captured during subsequent laryngeal and / or pharyngeal examinations after initial training.

8. The endoscopic imaging manipulation method according to claim 1, wherein: Upon request, the captured endoscopic image is displayed on a monitor in a non-overlay manner.

9. An endoscope imaging manipulation system, wherein: The endoscopic imaging manipulation system includes: a video endoscope, which is suitable for being introduced through a patient's mouth or nose for laryngeal and / or pharyngeal examination; a light source, which is capable of providing white light and narrow-band illumination for white light imaging and narrow-band imaging, and is connected to or integrated into the video endoscope; an image analyzer, which is connected to the video endoscope for receiving endoscopic images from the video endoscope, and has an instance of artificial intelligence, which is trained to identify suspicious areas of laryngeal and / or pharyngeal tissue structures, and the suspicious areas show signs that are different from healthy laryngeal and / or pharyngeal tissues, and the image analyzer is also configured to superimpose a mark on the captured endoscopic image, and the mark indicates the area indicated as suspicious by the instance of the artificial intelligence; and a monitor, which is connected to the image analyzer for displaying the endoscopic image provided by the image analyzer.

10. The endoscopic imaging manipulation system according to claim 9, wherein: The image analyzer is configured to apply the instance of the artificial intelligence to the captured endoscopic image based on one of an imaging mode recognition signal or an image characteristic in the captured endoscopic image that indicates narrow-band imaging when narrow-band imaging is applied.

11. The endoscopic imaging manipulation system according to claim 9, wherein: The image analyzer is configured to apply the instance of the artificial intelligence to the captured endoscopic image upon request by a practitioner.

12. The endoscopic imaging manipulation system according to claim 9, wherein: An example of the artificial intelligence is a convolutional neural network with a classifier that has been trained by at least one of supervised learning and unsupervised learning on multiple endoscopic images of laryngeal and / or pharyngeal structures to classify suspicious areas of laryngeal and / or pharyngeal tissue in the captured endoscopic images.

13. The endoscopic imaging manipulation system according to claim 9, wherein: The instance of the artificial intelligence has been trained to identify differences from healthy laryngeal and / or pharyngeal tissue as suspicious due to one or more of lesions, vascular morphology, vascular density, vascular pattern, and structure of the mucosal surface.

14. The endoscopic imaging manipulation system according to claim 9, wherein: The image analyzer is configured to create an overlay superimposed on the captured endoscopic image as a color-coded or brightness-coded heat map.

15. The endoscopic imaging manipulation system according to claim 14, wherein: The image analyzer is configured to base color coding or brightness coding of the heat map on at least one of density and pattern of blood vessels in the suspicious area.

16. The endoscopic imaging manipulation system according to claim 9, wherein: The instance of the artificial intelligence is configured to learn from new endoscopic images captured during subsequent laryngeal and / or pharyngeal examinations after initial training.

17. The endoscopic imaging manipulation system according to claim 16, wherein: The image analyzer is configured to receive feedback from a practitioner performing the laryngeal and / or pharyngeal examination regarding unmarked suspicious regions or the lack thereof in laryngeal and / or pharyngeal tissue structures, and / or whether the classification of one or more suspicious regions of laryngeal and / or pharyngeal tissue structures is correct.

Images