Invented by Gabriel Fine, Nathan Silberman, Individual

The market for systems for guiding medically invasive devices relative to other medical devices via image processing is experiencing significant growth and innovation. These systems play a crucial role in improving the accuracy and safety of medical procedures, particularly those involving minimally invasive surgeries. Image processing technology has revolutionized the field of medicine by providing real-time visualization and guidance during surgical procedures. It allows surgeons to navigate through complex anatomical structures with precision and accuracy, reducing the risk of complications and improving patient outcomes. This technology is especially valuable in procedures such as cardiac catheterization, endoscopy, and laparoscopy. One of the key drivers behind the growth of this market is the increasing demand for minimally invasive procedures. These procedures offer numerous benefits over traditional open surgeries, including reduced pain, shorter hospital stays, and faster recovery times. However, they also present unique challenges for surgeons, as they require precise navigation through small incisions or natural orifices. Image-guided systems provide surgeons with the necessary tools to overcome these challenges and perform procedures with greater confidence and accuracy. Another factor contributing to the market’s growth is the continuous advancements in imaging technology. The development of high-resolution imaging modalities, such as magnetic resonance imaging (MRI), computed tomography (CT), and ultrasound, has significantly improved the quality and clarity of medical images. These advancements enable surgeons to visualize anatomical structures in real-time and make informed decisions during procedures. Furthermore, the integration of artificial intelligence (AI) and machine learning algorithms into image-guided systems has further enhanced their capabilities. These algorithms can analyze medical images, identify relevant anatomical structures, and provide real-time feedback to surgeons. This technology not only improves the accuracy of device placement but also reduces the risk of complications and enhances patient safety. The market for systems for guiding medically invasive devices relative to other medical devices via image processing is highly competitive and fragmented. Numerous companies are actively involved in developing and commercializing these systems, each offering unique features and functionalities. Some of the key players in this market include Medtronic, Philips Healthcare, Siemens Healthineers, GE Healthcare, and Stryker Corporation. In conclusion, the market for systems for guiding medically invasive devices relative to other medical devices via image processing is witnessing significant growth and innovation. The increasing demand for minimally invasive procedures, advancements in imaging technology, and integration of AI and machine learning algorithms are driving this growth. These systems are revolutionizing the field of medicine by improving the accuracy and safety of medical procedures, ultimately benefiting both patients and healthcare providers.

The Individual invention works as follows

A system and method of guiding invasive devices relative to other devices is disclosed. Imaging devices can produce images of an invasive medical device inside a body. Annotated images with orientation and position data can be used to train a trained model of the invasive device. The trained model can be applied by an imaging computer system to unannotated pictures of the medical device in order to determine the current position and orientation of the medical device. “At least one of visual position and orientation information representing current orientation of the medical device in relation to another medical device can be outputted.

Background for System for guiding Medically Invasive Devices relative to Other Medical Devices via Image Processing

The successful completion of invasive medical devices is dependent on the physical manipulation of the tools. The correct positioning and movement of the devices is essential to (a) complete the procedure in a timely manner, (b) avoid harm to the patients, and (c), limit radiation exposure to both the patient and the operator. Two-dimensional (2D) imaging technologies are available to help medical practitioners. Imaging technologies that provide 2D views in real-time have been developed for medical practitioners to assist them. Fluoroscopy, ultrasound and other imaging technologies provide guidance and orientation to practitioners in 2D space. These 2D imaging technologies can also provide 2D views on a graphic display of the invasive instruments themselves.

The document relates in general to medical vision tools that provide orientation information to medical devices within two-dimensional image projections. These can be used by practitioners for image-guided procedure, such as procedures performed by interventionalists. Interventional radiologists (cardiologists, gastroenterologists etc.) and surgeons are among the professionals who use medical vision tools. When performing a medical treatment, doctors often rely on technology. Tracking systems can provide information on the positioning of medical instruments in relation to other instruments, patients and/or coordinate reference systems. Medical practitioners can use tracking systems to determine the position of a medical instrument when it is out of their line of vision or to confirm the alignment of the device. “A tracking system can also be used to aid in presurgical planning.

A system and method of guiding invasive devices relative to other devices is disclosed. Imaging devices can produce images of an invasive medical device inside a body. Annotated images with orientation and position data can be used to train a trained model of the invasive device. The trained model can be applied by an imaging computer system to unannotated pictures of the medical device in order to determine the current position and orientation of the medical device. “At least one of visual position and orientation information representing current orientation of the medical device in relation to another medical device can be outputted.

In one implementation, an image-generating device is used to create one or more 2D images of the invasive device within the patient. The trained model was generated by one or several machine learning algorithms that were trained on annotated 2D images of the invasive device with orientation and position information. The imaging computer system is programmed: to receive the one- or two-dimensional images of an invasive device from the imaging device; to access the trained models for the device from the database; to determine the orientation and current position of an invasive device by applying the trained model to one or both of the 2D pictures; to use the trained model in determining orientation and current position from unannotated 2D photos of the device; to have a database that stores a model trained from annotated 2D photographs of the device an The imaging computer system receives the one 2D image of the medical device generated by the imaging device. It then accesses the trained model from the database to determine the current orientation of the medical device in the patient.

Such an application can include optionally one or more features. These features can be combined to create any permutation. The system may also include a computer system for training the model of the invasive device. The training computer system may be programmed to receive the annotated 2D image of the medical device that includes orientation and positioning information. The annotated 2D image can depict the medical device inside a patient. A practitioner can manually annotate a portion (or all) of the 2D images with orientation and position information. At least a part of the 2D annotated images can be computer-generated images simulating the use of the invasive device in patients as imaged with the imaging device. One or more of the machine learning algorithms may include a deep learning algorithm that is supervised. The model trained can include a model with a short-term long memory.

The trained model can be tailored to the specific combination of the invasive device and imaging device. Other trained models may be used with other combinations (i) of the invasive device or another invasive device and (ii), the imaging device, or other imaging devices. The database can include one or more other trained models. The imaging device may be an x ray imaging device, and the two-dimensional images can include one or both x ray images. The imaging device may include an ultrasound and the two-dimensional images can be ultrasound images. The imaging device may include a computerized Tomography (?CT?) The imaging device can include a computerized tomography (?CT?) Magnetic resonance imaging (?MRI?) can be included in the imaging device. The imaging device can include a magnetic resonance imaging (?MRI?) The imaging device may include a nuclear imager and the two-dimensional images can include nuclear images. Magnetic resonance imaging (?MRI?) The one or two 2D images can be one or several MRI images. The current orientation of an invasive medical device can include its current roll, pitch and yaw. The trained model can be determined by using the cosine distance-loss function. The current roll and pitch can be discretized, and the trained model determined by sigmoid cross entropy. The current position may include information on (i) anterior, posterior, cranial, caudal, and left, right positions. A single image of one or more 2D pictures can determine the current orientation and location of the invasive device. A sequence of images taken from one or more 2D pictures can determine the current orientation and location of the invasive device. A single image of one or more 2D pictures and a reference picture for the invasive device can determine the current orientation and location of the medical device.

The disclosed systems and devices may be used in conjunction with any of these imaging modalities, with or without contrast administration (e.g. angiography).

Certain Implementations may offer one or more benefits.” It is possible to improve the imaging of medical devices, allowing medical practitioners to better visualize their position and orientation before and/or while performing medical procedures. These improved and additional information can help practitioners perform medical procedures more safely and quickly. Three-dimensional (3D?) knowledge, for example, can help a practitioner to complete many operations more quickly and safely. For example, having three-dimensional (?3D?) 2D imaging can be insufficient to represent the device’s orientation and position, increasing the time practitioners take to complete the procedure. The amount of radiation a patient receives is directly proportional to how long the procedure lasts. Therefore, longer procedures, such as those that use 2D imaging, can result in more radiation. This can cause risks and cost to both patient and operator. By giving practitioners 3D information about a device’s orientation and position, they can perform operations more efficiently and quickly, thus reducing risks to both patient and practitioner.

In another instance, 3D data (e.g. position information, orientation) can be retrofitted onto imaging devices that provide only 2D images. By using machine-learning techniques to infer orientation and position information from 2D images, imaging devices, which are only capable of delivering 2D information, can be enhanced to provide 3D data without the need for additional hardware or components. This can improve the operation of medical imaging devices with minimal cost.

The accompanying drawings and description below detail one or more embodiments. Additional features and benefits will be evident from the drawings and claims.

DESCRIPTION DU DRAWINGS

FIG. “FIG.

FIG. “FIG.

FIG. “FIG. This creates the appearance of bands of varying grayscale intensities along the device.

FIG. FIG.

FIG. “FIG.

FIG. “FIG.

FIG. “FIG.

FIG. “FIG.

FIG. “FIG.

The “Like references symbols” in various drawings indicates like elements

Click here to view the patent on Google Patents.