The present invention relates to a system and a method for image processing and analysis according to the preamble of the

independent claims.

Neurodegenerative diseases drastically affect the neuronal system of humans, in particular the brain, leading to

deteriorated brain function. A large number of such diseases is known. Generally, brain cells and tissue are disintegrated as a consequence of such diseases, which is manifested in

neurological defects. In order to analyze and diagnose disease states of the human brain, systems and methods were developed for image analysis of anatomical brain structures based on medical images of a patient. Brain defects of a patient are traditionally analyzed and diagnosed by imaging techniques, e.g. magnetic resonance imaging, and downstream computational means or methods. The analysis ideally indicates a defective/disease state over a healthy/normal state of the patient's brain.

WO 2009/146703 A2 describes a system and a method for volumetric analysis of medical images. The system comprises image analyzing means and statistical analyzing means for providing a confidence interval of the volume or the shape of an anatomical structure. The normality of said anatomical structure is quantified. The system does not allow visualization of acquired data or

representation of the analyzed anatomical structure.

It is therefore an objective of the present invention to avoid disadvantages of the known systems and methods for image

processing and analysis, and in particular to provide improved tools in order to analyze and evaluate anatomical structures in disease states. This objective is solved with the systems and the method according to the independent claims.

According to the present invention a system for analysis, evaluation and visualization of at least one anatomical

structure in medical images is provided. In particular the system is suitable for analysis, evaluation or visualization of an anatomical brain structure. The medical images represent at least a part of a patient or a person. Said system comprises: a) an image analyzer for segmenting a region of at least one anatomical structure, in particular an anatomical brain structure, on the basis of said images;

b) a storage for storing datasets relating to the segmented

region of the anatomical structure;

c) a generator of (a) graphical representation ( s ) of the

segmented region of the anatomical structure, in particular the anatomical brain structure, on the basis of said

datasets relating to the segmented region, and

d) a visualizer, which is adapted to display a time series of the segmented region of anatomical structures by displaying graphical representations, which were generated on the basis of said datasets relating to the segmented regions of images, in particular of the entire brain, at at least two different points in time.

The system allows visualization of the disease progress of the segmented region of an anatomical structure. Thus, the

practitioner can assess and diagnose a disease state over time, based on the time series, which relates to graphical

representations at at least two points in time.

The visualizer may be designed for superposition of said

graphical representations with or without a representation of the anatomical, in particular brain, structure, in particular with or without a representation of the surface of the

anatomical, in particular brain, structure.

The segmented region may also be visualized in the context of the entire brain. The segmented region may be superimposed on any image, which preferentially is a medical or artificial image. An artificial image is to be understood to be an image, which is not necessarily directly related or taken from the patient. It can be any suitable reference image, which may be e.g. artificially created on the basis of a diagnostic image.

The visualizer may be designed for displaying said time series of segmented regions of anatomical structures by displaying individual graphical representations

side by side at the same time; or

superimposed; or

in a consecutive manner.

By offering displaying/visualizing alternatives to a user, the system can display the graphical representations in a user- friendly way so that the state of an anatomical structure is easily and conveniently evaluated.

The visualizer may be designed for displaying said time series of the segmented region of anatomical structures by providing a time zoomer. The time zoomer allows scrolling through a time line, wherein missing time points may be simulated on the basis of a linear or non-linear regression approximation. Thus, a prognosis on how the disease will develop can be made and transitional phases lacking data can be simulated and

visualized. The image analyzer may further be designed for calculating characteristics of the volume, the shape, and/or the surface area of said brain structure and/or for calculating the cortical thickness of any region of the brain cortex and/or for

determining such parameters of regions (s) of interest. Thereby, characteristics typically can be a length and/or width related to the afore-mentioned parameters. Characteristics can also be the volume and/or the thickness. Region of interest is to be understood as a predetermined region. In this way, defined parameters of a region of interest can be measured and

calculated to provide a basis for evaluation of the regions of interest .

The region of interest may be selected from the group: frontal lobe, parietal lobe, temporal lobe, occipital lobe, anterior cingulate, median cingulate, posterior cingulate cortex,

precuneus cortex, cuneus cortex, lingual cortex, fusiform gyrus, parahippocampal gyrus, inferior frontal gyrus, rolandic

operculum, inferio frontal gyrus, medial orbitalfrontal cortex, gyrus rectus, superior frontal gyrus, middle frontal gyrus, heschl gyrus, insula, supplementory motor area, precentral gyrus, postcentral gyrus, paracentral lobule, superior occipital gyrus, middle occipital gyrus, inferior occipital gyrus,

olfactory cortex, superior parietal gyrus, supramarginal gyrus, angular gyrus, superior temporal gyrus, middle temporal gyrus, inferior temporal gyrus.

In addition, the region of interest may be also selected from the group: hippocampus, amygdala, caudate nucleus, putamen, pons, thalamus, global pallidus, brainstem, cerebellum.

The system may be designed to allow for selection of: a) at least one reference image comprising a segmented reference representation, and

b) a registration for providing a mapping between the medical image/s and the reference image/s.

The system may be designed for normalizing said medical images, preferably prior to analysis of said medical images. Normalizing the medical image means fitting the medical image into a

template and/or adjustments and corrections of pixel intensities and size (length and width) , which will facilitate the

comparison of the sequence.

The system may be designed for providing a confidence interval of the characteristics of the shape and/or the volume of the at least one anatomical structure, in particular the at least one anatomical brain structure, as at least one numerical value, and/or a graphical representation in the analyzed medical image and/or in the reference image.

The system may be designed for analyzing, evaluating and

illustrating a variation over time of the characteristics of the shape and/or the surface area and/or the volume of at least one anatomical structure. In particular at least one anatomical brain structure and more particularly the cortical thickness may be analyzed and evaluated or illustrated. The syetem may further be designed for representing at least a part of a patient or person, based on individual data of the patient or person.

Thereby changes over time of the at least one anatomical

structure can be determined. In this way, disease evolution and/or changes over time can be visualized. A disease state can be thus related to a reference image and the deviation from normality determined. The at least one reference image may be a previously acquired medical image of the patient or person or an artificial image created e.g. on basis of such medical image. Thus, a direct comparison between a prior state and the actual state used to evaluate the disease state.

The system may include at least one database. The at least one database comprises more than one dataset relating to a graphical representation of each patient or person on the basis of medical images acquired at at least two different points in time.

The medical image may comprise at least a part of the entire brain of the patient or person.

The medical image may be a magnetic resonance image, a computer tomography or any suitable medical image or artificial images created on the basis thereof.

The image analyzer may be designed for analyzing anatomical structure ( s ) . The anatomical structure (s) is any segmentable brain structure, preferably any of the following neuro- anatomical structures: hippocampus, amygdala, caudate nucleus, putamen, pons, thalamus, global pallidus, brainstem and/or cerebellum.

Further regions of interest corresponding to analyzable

anatomical structures include: frontal lobe, parietal lobe, temporal lobe, occipital lobe, anterior cingulate, median cingulate, posterior cingulate cortex, precuneus cortex, cuneus cortex, lingual cortex, fusiform gyrus, parahippocampal gyrus, inferior frontal gyrus, rolandic operculum, inferio frontal gyrus, medial orbitalfrontal cortex, gyrus rectus, superior frontal gyrus, middle frontal gyrus, heschl gyrus, insula, supplementory motor area, precentral gyrus, postcentral gyrus, paracentral lobule, superior occipital gyrus, middle occipital gyrus, inferior occipital gyrus, olfactory cortex, superior parietal gyrus, supramarginal gyrus, angular gyrus, superior temporal gyrus, middle temporal gyrus, inferior temporal gyrus.

The system may be designed for aiding in the diagnoses of diseases, and/or initial stages of diseases, based on the deviation from normality, and/or the confidence interval of the analyzed volume, and/or shape of the at least one anatomical structure, diseases such as depression, Alzheimer's disease, epilepsy, post-traumatic stress disorder, schizophrenia, anxiety and/or Parkinson.

The system may be designed for being hosted on and/or being provided through at least one web server of medical images and graphical representations. In this way, the system does not need to be run on a local computer system but is accessed and run through a web server. Data storage and access is feasible from any remote computer system having access to the web server.

The visualizer may be designed for visualization of differences over time between said anatomical structures. Thus, differences can be illustrated by highlighting or usage of particular color schemes .

Another aspect of the invention relates to a system for

analysis, evaluation and visualization of at least one

anatomical structure, in particular an anatomical brain

structure, in a medical image. Said medical image represents at least a part of a patient or a person, preferably as described above. Said system comprises: a) an image analyzer for segmenting a region of at least one anatomical structure, in particular an anatomical brain structure, on the basis of said image, and

b) a storage for storing at least one dataset relating to the segmented region of the anatomical structure; and

c) a generator of (a) graphical representation ( s ) of the

segmented region of the anatomical structure, in particular the anatomical brain structure, on the basis of said dataset relating to the segmented region, and

d) a visualizer, which displays the graphical representation, which was generated on the basis of said dataset relating to the segmented regions of an image, in particular of the entire brain.

According to this aspect of the invention, the said visualizer is designed for web-browser-based visualization of said

anatomical structure/s.

The system allows easy visualization of the segmented region of an anatomical structure. Thus, the practitioner can assess and diagnose a disease state without the need of generating

graphical representations on site. There is hence no need for storage a calculation capacity at the practitioner's premises. "Web-browser visualization based" is to be understood such that the visualization is facilitated through a web-browser, which accesses data stored on the web-server.

The storage for storing at least one datasets may be designed to be hosted on a server and/or the generator of graphical

representations may be designed to be provided on a web server.

The visualizer may be designed for visualization on at least one client. Client includes local computers, notebooks or tablets. The visualizer may be designed for displaying a time series of segmented regions of anatomical structures by displaying

graphical representations, which were generated on the basis of datasets relating to segmented regions images recorded at at least two different points in time. A time series allows comparison of states of anatomical structures over time and changes or disease evolution can be identified.

The visualizer may be designed for visualization of differences over time between said anatomical structures via a web browser.

The visualizer may be designed for displaying of said time series of segmented regions of anatomical structures facilitated by displaying individual graphical representations

side by side at the same time; or

superimposed; or

in a consecutive manner.

By offering displaying/visualizing alternatives to a user, the system can display the graphical representations in a user- friendly way through a web-browser so that the state of an anatomical structure is easily and conveniently evaluated.

The invention further relates to a method for analysis,

evaluation and visualization of at least one anatomical

structure, in particular an anatomical brain structure, in a medical image. Said medical image represents at least a part of a patient or a person. The method comprises the steps:

a) recording and/or providing of images of at least one

anatomical structure at at least two points in time;

b) image analysis for segmenting a region of at least one

anatomical structure, in particular an anatomical brain structure, on the basis of said images; c) storage of datasets relating to the segmented region of the anatomical structure;

d) generation of (a) graphical representation ( s ) of the

segmented region of the anatomical structure, in particular the anatomical brain structure, on the basis of said datasets relating to the segmented region, and

e) visualization of a time series of the segmented region of anatomical structures by displaying graphical

representations, which were generated on the basis of said datasets relating to the segmented regions of images, in particular of the entire brain, at at least two different points in time.

The method allows visualizing of the disease progress of the segmented region of an anatomical structure. Thus, the

practitioner can assess and diagnose a disease state over time based on the time series, which relates to graphical

representations at at least two points in time.

A time series of said graphical representations of segmented regions of anatomical structure may be displayed

side by side at the same time, or

sumperimposed; or

in a consecutive manner.

By offering displaying/visualizing alternatives to a user, the graphical representations in a user-friendly way can be

displayed such that the state of an anatomical structure is easily and conveniently evaluated.

Characteristics of the shape, and/or the surface area and/or the volume of said brain structure and/or for determining regions (s) of interest, in particular the cortical thickness in the entire brain, may be calculated. In this way, defined parameters of a region of interest can be measured and calculated to provide a basis for evaluation of the regions of interest.

The method may comprise the steps:

a) providing at least one reference image comprising a

segmented reference representation, and

b) registering of reference image/s for providing a mapping

between the medical images and the reference image/s.

The at least one reference image may comprise more than one segmented reference representation.

The method may further comprise the step normalizing said

medical images. Normalizing the medical image means fitting the medical image into a template and/or adjustments and corrections of pixel intensities and size (length and width) . Normalizing said medical image is preferably performed prior to analyizing said medical image.

The method may provide a confidence interval of the

characteristics of the volume and/or the shape of the at least one anatomical structure, in particular the at least one

anatomical brain structure, as at least one numerical value, and/or a graphical representation in the analyzed medical image and/or in the reference image.

The method may comprise the steps

analysis, evaluation and illustration of a variation over time of the characteristics of the shape and/or the surface area and/or the volume of at least one anatomical structure, in particular of at least one anatomical brain structure, more particularly the cortical thickness, in said medical images acquired at different points in time, and representing at least a part of a patient or person, based on individual data of the patient or person.

Thereby, changes over time of the at least one anatomical structure are determined. In this way, disease evolution and/or changes over time can be visualized. A disease state can be, thus, related to a reference image and the deviation from normality determined.

Further, at least one previously acquired medical image of the patient or person may be taken as a reference image. Thus, a direct comparison between a prior state and the actual state used to evaluate the disease state.

At least one database may comprise more than one dataset

relating to a graphical representation of each patient or person on the basis of medical images acquired at at least two

different points in time.

The medical image may comprise at least a part of the entire brain of the patient or person.

The medical image may preferably be a magnetic resonance image, CT or any suitable medical image.

The the analyzed anatomical structure (s) may be any segmentable brain structure, preferably any of the following neuro- anatomical structures: hippocampus, amygdala, caudate nucleus, putamen, pons, thalamus, global pallidus, brainstem and/or cere ^¬ bellum.

Further the method may aid in the diagnoses of diseases, preferably such as depression, Alzheimer, epilepsy, post ^¬ traumatic stress disorder, schizophrenia, anxiety and/or Parkinson, and/or of initial stages of diseases. The diagnoses is based on the deviation from normality, and/or on the

confidence interval of the analyzed volume, and/or on the shape of the at least one anatomical structure.

Medical images and graphical representations may be hosted on and provided through at least one web server. In this way, the system does not need to be run on a local computer system but is accessed and run through a web server. Data storage and access is feasible from any computer system.

Differences over time between said anatomical structures may be visualized via a web browser. Thus, differences can be

illustrated by highlighting or usage of particular color

schemes .

The invention further relates to a method for analysis,

evaluation and visualization of at least one anatomical

structure, in particular an anatomical brain structure, in a medical image. The medical image represents at least a part of a patient or a person. The method comprising the steps:

a) recording and/or providing of an image of at least one

anatomical structure, and

b) image analysis for segmenting a region of at least one

anatomical structure, in particular an anatomical brain structure, on the basis of said images, and

c) storage of a dataset relating to the segmented region of the anatomical structure; and

d) generation of (a) graphical representation ( s ) of the

segmented region of the anatomical structure, in particular the anatomical brain structure, on the basis of said datasets relating to the segmented region, and e) visualization of the segmented region of the anatomical structure, in particular the anatomical brain structure, by displaying graphical representation ( s ) , which was generated on the basis of said dataset relating to the segmented regions of the image recorded or provided, in particular of the entire brain.

The visualization of said anatomical structures is hereby facmade in a web-browser-based manner. The method allows

visualization of the disease progress of the segmented region of an anatomical structure. Thus, the practitioner can assess and diagnose a disease state over time based on the time series, which relates to graphical representations at different points in time. Additionally, the method does not need to be run on a local computer system but is accessed and run through a web server. Data storage and access is feasible from any computer system.

The storage for storing at least one datasets may be maintained on a server and/or the generation of (a) graphical

representation ( s ) is run on a server. The visualization may be performed on at least one client. Thus, only the visualization is performed on a local computer through a web-browser but data storage and calculations are run on the server.

Differences between said anatomical structures may be visualized via the web browser.

The invention is described in more detail in the following figures .

Figure 1 : Hippocampus shown in the context of the brain; Figure 2 : Visualization of disease progress of the isolated hippocampus ;

Figure 3 : A sketch of computational elements;

Figure 4 : Flow chart of method steps.

Figure 1 shows the whole brain 1 of a patient based on a medical image. By way of example, the right 2 and the left hippocampus 3 are indicated (dotted black line) . Therein, the right 2 and the left hippocampus 3 is illustrated in the context of the whole brain 1 in order to visualize and display the spatial

arrangement of the hippocampi 2, 3 in the brain.

Figure 2 shows a comparison of data recorded at two examinations of a patient. A second medical image (t _4a ) was recorded four years after a first medical image (to _a ) · I ⁿ figure 2, a

superposition of the first and the second calculated border of hippocampus is shown. The borders of the hippocampi at to _a are shown as black dotted lines. The borders of the hippocampi at t _4a are shown as white dotted lines. The superposition of data derived from images recorded at t _0a and t _4a illustrates the decrease in size of the right and the left hippocampus due to Alzheimer disease.

Figure 3 shows computational/hardware elements of the system. A processor includes software instructions for image analysis and generation of graphical representations. The processor is located on a server. A storage stores and saves medical images and datasets derived from the medical images. The storage is also located on the server. The graphical representations are displayed on a client at the practitioner's site, which is separate from the server. Figure 4 shows several steps of the method. Images of an

anatomical structure are recorded at two points in time (a) . The recorded images are analyzed for a segmenting region of an anatomical structure on the basis of said images (b) . Datasets relating to the segmented region are created on basis of the images recorded at the two points in time. Afterwards, the datasets are stored in the storage (c) . Graphical

representations of the segmented region of the anatomical structure on the basis of the datasets are generated (d) .

Finally, a time series of segmented region of the anatomical structure is visualized by displaying graphical representations, which were generated on the basis of said datasets (e) . The datasets relate to the segmented region.