The present invention relates to a method for determining a damage characteristic value of a kidney. The present invention further relates to a corresponding device for determining a damage characteristic value of a kidney.
Treatment-resistant hypertension describes a condition characterized by the inability to achieve a sufficient lowering of systolic blood pressure below 160 mmHg in spite of the, in some cases concurrent, administration of at least three antihypertensive agents, including diuretic. As indicated in W. C. Cushman et al., Journal of Clinical Hypertension, Vol. IV, No. VI, 2002, pp. 393 to 404, this disease affects even 8% of the optimally treated hypertensive patients, with an increasing trend; it therefore represents a very large patient cohort. The consequences of high blood pressure existing over many years may include: arteriosclerosis (atherosclerotic vascular disease), coronary heart disease (narrowing of the coronary blood vessels), myocardial infarction, cardiac insufficiency (chronic heart failure), atrial fibrillation, brain damage or stroke (high blood pressure being the primary risk factor for this), calcification or aneurysms of the aorta with risk of rupture and bleeding, kidney damage and renal failure, eye damage with diminution of visual acuity, etc.
Renal denervation, also known as renal sympathetic denervation, renal nerve ablation, renal denervation, or renal artery denervation, is a relatively new method for successfully lowering blood pressure in the long term. Denervation is generally understood to mean the complete or partial interruption of nerve tracts between an organ and the brain. Renal, from the Latin “ren”, kidney, means “relating to the kidney” or “appertaining to the kidney”. In renal denervation, an instrument known as an ablation catheter, more precisely a radiofrequency current ablation catheter, is typically introduced into the renal artery via an access site in the groin of a patient. A radiofrequency current is emitted, causing nerve fibers extending around the kidney vessel to be ablated, with the result that the corresponding nerve pathways, which in some instances conduct spurious signals from the kidney to the brain and back, are interrupted. The kidney is accordingly decoupled from the sympathetic nervous system.
By means of this method it is possible, according to H. Krum, Transcatheter Cardiovascular Therapeutics (TCT) Conference, 2009, Sep. 21-26, 2009, San Francisco, Calif., USA, to achieve a clinically significant reduction in blood pressure of about 30 mmHg on average in approx. 80% of patients. Contrasting with the high success rate of 80%, there remain 20% of patients in whom no reduction in blood pressure can be achieved through renal denervation therapy. The cause for when the treatment works for a patient, i.e. the patient is what is called a “responder”, and when it does not, has not been conclusively explained to date. Currently, little to no attention is devoted to the difference between “responders” and “non-responders”, attempts being made instead to treat as many patients as possible. This serves on the one hand to help the medication-resistant patients and on the other hand a more extensive database is generated as a result. Current publications, such as in N. Goulding, A new cure for hypertension—Renal Denervation, The Boolean, 2011, put forward the hypothesis that the success of a renal denervation is dependent on what is called the damage state of the kidney. Currently, this damage state of the kidney is measured by, for example, measuring the protein content in urine and/or blood, the creatinine level in the blood being determined in addition. However, these measurements only describe the state of damage of the kidney indirectly. Precise knowledge about the renal factors on which the regulation of blood pressure is dependent is so important because the kidney constitutes the primary medium- and long-term regulatory organ, with the short-term regulation being effected primarily by way of baroreceptors. The long-term regulation controls the blood pressure by sending signals via the nerves to the brain, resulting in a change in the total blood volume, the heart rate, and the change in vessel diameters. However, in certain circumstances, or under certain preconditions, the kidney also sends out “spurious” signals which lead to a pathological increase in blood pressure, without this being necessary for the body. It could also be said that the diseased kidney is the source of a pathological afferent signal which stimulates the efferent sympathetic activity. In the case of renal denervation, such signals from the kidney are prevented by the corresponding nerve tracks being ablated and thus deactivated. As a result the long-term regulation, which in the pathological case is incorrect, can be switched off and the body regulates the blood pressure primarily by way of short-term regulation mechanisms, such as the baroreceptors.
The object of the present invention thus consists in disclosing a method which enables a damage characteristic value of a kidney to be determined. It is furthermore the object of the invention to describe a device by means of which a damage characteristic value of a kidney can be determined.
The invention achieves this object by means of a method for determining a damage characteristic value of a kidney having the features of the first independent claim and by means of a device for determining a damage characteristic value of a kidney having the features of the second independent claim. Advantageous embodiments are described in dependent claims.
A basic concept of the invention is a method for determining a damage characteristic value of a kidney, wherein images comprising the kidney and the renal artery that were obtained by a three-dimensional digital subtraction angiography scan that was carried out with administration of contrast agent at the proximal end of the renal artery and comprises a fill run and a mask run are input into the method, and wherein the method comprises the following method steps:
The invention is based on the consideration that a renal denervation is medically successful in particular when certain conditions are present which are responsible for the cited spurious signals. Toward that end the method presented enables a damage characteristic value of a kidney to be determined which, based on perfusion properties and resistance characteristics of the arterioles, describes a decreased perfusion of the kidney. Perfusion or blood flow is the term generally used to describe the supply of organs or organ parts with blood. The blood for supplying tissues with oxygen, nutrients and so forth is supplied via arteries, while the outflow, for carrying away metabolic products and carbon dioxide, is effected via veins. A renal denervation treatment generally appears to be promising when the damage characteristic value of a kidney is comparatively large.
The medical examination of blood vessels, i.e. in angiography, is frequently undertaken by means of the digital subtraction angiography method, DSA for short, using digital, X-ray based imaging systems. In said method, at least one image, called the mask or mask image, of a region that is to be examined, e.g. the kidneys, is acquired in the first instance in the so-called background phase or mask phase by means of an X-ray machine, which image can also reveal bone structures, inter alia. Whereas in the further course of the examination, the so-called uptake phase, arterial phase, subtraction phase or fill phase, a contrast agent, e.g. an iodine-containing substance or an isotonic saline solution, is supplied to the vessels, further images, called fill images, are acquired, possibly after a delay of about 3 to 6 seconds, the reaching of what is termed the steady state, at relatively low frame repetition rates, for example 0.5 to 6 frames per second. Thereafter, or at the same time, the mask image is subtracted in each case from said fill images by means of digital image processing. In particular, the logarithmized grayscale values of the images are subtracted. What is obtained as result in the form of DSA result images are continuously anatomy-free images showing only the vascular system or, as the case may be, the current location of the contrast agent. In addition to two-dimensional digital subtraction angiography acquisitions, three-dimensional digital subtraction angiography acquisitions, or 3D DSAs, are also known. In principle these are obtained like 2D DSAs, with spatial images of the region that is to be examined being used. The spatial images can be obtained with the aid of a C-arm X-ray machine. According to the prior art it is possible, by means of a suitable series of X-ray images acquired from different directions around the examination subject, to calculate a spatial image, also referred to as a C-arm CT by analogy with computed tomography.
In a basic concept of the invention, images obtained from a three-dimensional digital subtraction angiography scan performed in preparation for the method according to the invention and known per se are accordingly input, said images comprising at least one kidney and its associated renal artery as the region to be examined. In this case the three-dimensional digital subtraction angiography scan was performed with administration of contrast agent at the proximal end of the renal artery and comprises a fill run and a mask run. Alternatively, the contrast agent injection may also have been administered in the aorta, proximally to the side branches of the renal arteries, and the three-dimensional digital subtraction angiography images include both kidneys and both associated renal arteries. This embodiment has the advantage that both kidneys are acquired for the method steps of the method according to the invention.
In a first method step of the method according to the invention, a parenchymal blood volume (PBV) is determined. By parenchyma, in medicine, is understood a tissue, also called parenchymal tissue, which exercises a specific function and is in most cases enclosed by supporting connective tissue or, as the case may be, carrier or structural tissue. In relation to the kidney, parenchyma denotes the tissue formed by the specific function cells of the kidney, in this case the nephrons. The subtraction from images of the fill run and the mask run that were obtained in the three-dimensional digital subtraction angiography scan is input into the parenchymal blood volume. It is conceivable that the parenchymal blood volume is determined by the difference between the logarithmized grayscale values of the images from fill run and mask run:
PBV=fill run−mask run.
An arterial input function (AIF) is determined in addition. The arterial input function represents the average grayscale value, or average logarithmized grayscale value, of the renal artery or renal arteries. Furthermore, the parenchymal blood volume is normalized by means of the arterial input function. It is conceivable that the normalized parenchymal blood volume is visualized, e.g. in color-coded form, on a presentation means, e.g. a computer monitor. The normalization of the parenchymal blood volume in respect of the arterial input function (AIF) has the advantage that the representation and values of the normalized parenchymal blood volume are essentially independent of a contrast agent protocol, i.e. independent of injection site, density, dilution, flow rate, etc., and consequently is now essentially dependent only on the perfusion or, as the case may be, on the blood volume of the kidney.
In a second method step of the method according to the invention, the kidney or kidneys is/are segmented in the three-dimensional digital subtraction angiography images. Segmentation is a subsidiary discipline of digital image processing and of so-called machine vision. Many segmentation methods, i.e. for generating regions related by content by combining neighboring pixels, in the spatial domain by combining neighboring voxels, according to a specific homogeneity criterion, are known in the technical literature. Pixel-, edge- and region-oriented methods are examples. Also known are model-based methods in which a specific shape of the objects is taken as the starting point, as well as texture-based methods in which an internal homogeneous structure of the objects can also be taken into account. An average normalized parenchymal blood volume value, henceforward denoted by xA, is determined from the images of the segmented kidney or kidneys. A total value of the parenchymal blood volume of the kidney or kidneys, henceforward denoted by xB, is determined in addition, the normalized parenchymal blood volume values of the segmented kidney or kidneys being incorporated into the determination of the total value of the parenchymal blood volume.
In a third method step of the method according to the invention, an average normalized parenchymal blood volume normal value, henceforward denoted by xA,Norm, and/or a total normal value of a parenchymal blood volume, henceforward denoted by xB,Norm, is received. At least one damage characteristic value of the kidney is subsequently determined, the average normalized parenchymal blood volume value xA and the average normalized parenchymal blood volume normal value xA,Norm being input into a first damage characteristic value, henceforward denoted by yA, and/or the total value of the parenchymal blood volume xB and the total normal value of a parenchymal blood volume xB,Norm being input into a second damage characteristic value, henceforward denoted by yB.
In a fourth method step of the method according to the invention, at least one damage characteristic value of the kidney, yA and/or yB, is output. The value(s) can be output for example as a numeric value or in color-coded form on a presentation means, e.g. a computer monitor.
An advantageous embodiment provides that a first damage characteristic value of a kidney is the percentage deviation of the average normalized parenchymal blood volume value from the average normalized parenchymal blood volume normal value, and/or a second damage characteristic value of a kidney is the percentage deviation of the total value of the parenchymal blood volume from the total normal value of a parenchymal blood volume.
This means that in this embodiment the damage characteristic values, yA or yB, which can also be interpreted as degrees of damage to the kidney, are calculated as percentage deviations of the normal values from the corresponding determined values, i.e.
y
A=1−(xA/xA,Norm) or
y
B=1−(xB/xB,Norm).
Another advantageous embodiment provides that a histogram of the parenchymal blood volume is incorporated into the determination of the arterial input function.
The arterial input function, which represents the average grayscale value or, as the case may be, the average logarithmized grayscale value of the renal artery or renal arteries, is determined in this embodiment by means of a histogram analysis of the parenchymal blood volume.
In an advantageous embodiment variant, as an alternative to the determining of the arterial input function with the aid of a histogram analysis, the following method steps can be incorporated into the determination of the arterial input function:
Since a segmentation of organs and vessels is often available as a function for example in modern medical imaging devices, such as computed tomography, magnetic resonance or C-arm X-ray machines, it is advantageously possible in a first method substep to have recourse to an already segmented 3D dataset comprising the segmented kidney and the segmented renal artery. This is then received for example in a first method substep by the imaging device, i.e. for example transferred into a memory of a computing and control means.
In a second method substep, a nonrigid 3D/3D registration of the parenchymal blood volume with the segmented 3D dataset is carried out. By registration, image registration or image fusion in digital image processing is understood a process in which two or more images of the same or at least a similar scene are brought into the best possible correspondence with one another. During this process one image is typically defined as the reference image, the other images being called subject images. In order to achieve an optimal alignment of the subject images to the reference image, a compensating transformation is determined. The images to be registered may be different from one another if they have been acquired from different positions, at different times or by means of different imaging devices. Image registration methods are prevalent primarily in medical image processing, being used for bringing the images acquired by means of different imaging methods or modalities into alignment with one another. In such a process a distinction can be made between rigid or inflexible registration methods and nonrigid or flexible registration methods. Registration by means of elastic models lends itself as a suitable method where there are complex local distortions between two images. In this case the registration is generally accomplished iteratively through minimizing an energy functional. Registration algorithms for 3D images are digital image processing methods that are known per se.
In a third method substep, the renal artery is extracted from the parenchymal blood volume. Since the positions of the kidney and in particular of the renal artery are known from the segmented 3D dataset and these have been transferred into the parenchymal blood volume by means of the registration process, the renal artery can be extracted or determined from the parenchymal blood volume.
In a fourth method substep, the arterial input function is determined by averaging over the grayscale values of all pixels of the extracted renal artery of the parenchymal blood volume. By a pixel is understood in this case a spatial picture element or voxel.
In a development of the last-described embodiment variant, the segmented 3D dataset received in the first method substep, which comprises the segmented kidney and the segmented renal artery, and the nonrigid 3D/3D registration of the parenchymal blood volume with the segmented 3D dataset performed in the second method substep are incorporated with particular advantage into the segmentation of the kidney in the second method step. Since the positions of kidney and renal artery are known from the received segmented 3D dataset and these have been transferred into the parenchymal blood volume by means of the registration process, the kidney can be extracted or determined from the parenchymal blood volume. In other words, there are synergy effects to be leveraged here which for example enable faster data processing.
It has proved to be advantageous if an addition of all of the normalized parenchymal blood volume values of the segmented kidney is input into the determination of the total value of the parenchymal blood volume.
In the second method step, the total value of the parenchymal blood volume is determined following segmentation of the kidney and determination of the average normalized parenchymal blood volume value. This is advantageously calculated by an addition of all of the normalized parenchymal blood volume values of the segmented kidney.
With particular advantage the average normalized parenchymal blood volume normal value and/or the total normal value of a parenchymal blood volume are/is the average normalized parenchymal blood volume normal value and/or the total normal value of a parenchymal blood volume of a healthy examination subject.
As described, an average normalized parenchymal blood volume normal value xA,Norm and/or a total normal value of a parenchymal blood volume xB,Norm are/is received in the third method step and can be incorporated into the determination of the at least one damage characteristic value of the kidney. In a preferred embodiment variant, the first damage characteristic value of a kidney is the percentage deviation of the average normalized parenchymal blood volume value xA from the average normalized parenchymal blood volume normal value xA,Norm, and/or the second damage characteristic value of a kidney is the percentage deviation of the total value of the parenchymal blood volume xB from the total normal value of a parenchymal blood volume xB,Norm. With particular advantage one or both normal values xA,Norm or xB,Norm is or are corresponding values of healthy examination subjects, in particular of healthy living creatures of the same species. The normal values can be derived for example from measurement series or statistics in which the corresponding values, i.e. the average normalized parenchymal blood volume value xA and/or the total value of the parenchymal blood volume xB, have been determined for healthy examination subjects, such as human or animal patients.
In a further advantageous embodiment, the average normalized parenchymal blood volume normal value and/or the total normal value of a parenchymal blood volume are/is the average normalized parenchymal blood volume normal value and/or the total normal value of a parenchymal blood volume of an examination subject in the case of which a renal denervation treatment was not successful.
In this embodiment, the normal values are derived from examination subjects in the case of which a renal denervation treatment was not successful, i.e. from so-called “non-responders”. This choice has the advantage that an increasing amount of statistical material would be available which would increase the probability of a correct prediction.
It is also conceivable for the purpose of selecting the normal values to correlate further characteristic values of the examination subject, such as sex, age, weight, type and symptoms of the hypertension, with the current examination subject in order in this way to ensure a better correspondence and comparison possibility.
In another embodiment variant of the invention, a warning is issued if the first and/or second damage characteristic value of a kidney lie/lies within a predefinable damage deviation range.
In this case the at least one damage characteristic value of the kidney is compared with a received damage deviation range and a warning is issued, for example in the form of a text output on a presentation means, such as a computer monitor, if the at least one damage characteristic value of the kidney lies within the damage deviation range. The damage deviation range is produced for example from measurement series of a plurality of examination subjects that have been determined as so-called “non-responders”, i.e. in the case of which a renal denervation treatment was not successful, and whose damage characteristic values have been combined to form said damage deviation range. If a damage characteristic value of the kidney of a currently examined examination subject now lies within the damage deviation range, it is to be assumed that this examination subject also cannot be treated successfully by means of a renal denervation and a warning to this effect is issued. It is also conceivable that a probability value for a renal denervation treatment is indicated which is all the greater, the more the at least one damage characteristic value of the kidney deviates from the corresponding normal value of the kidney.
Furthermore, further improvements can be achieved if the method is executed automatically.
Automatically executed methods require fewer interventions on the part of a user and as a result are generally less susceptible to error and are processed more rapidly.
A further basic concept of the invention relates to a device for determining a damage characteristic value of a kidney. The device comprises a computing and control means having an input means and an output means, the computing and control means being configured to execute one of the above-described methods.
The computing and control means can be configured as a computer to which a monitor and a keyboard are connected as output means and input means, respectively. The capability enabling the method steps described in the features to be performed can consist for example in the loading, into the computing and control means, of a computer program which is executed for the purpose of carrying out the method steps.
The exemplary embodiments described in more detail hereinbelow represent preferred embodiment variants of the present invention.
Other advantageous developments will become apparent from the description taken in conjunction with the following figures, in which:
To sum up, further embodiments and advantages of the invention are described. The invention proposes a method which defines a value, in particular a damage characteristic value of a kidney, which can help a user of the method, e.g. a physician, to estimate the chances of success of a renal denervation treatment. The invention is based on the knowledge that a kidney having function values that lie within a “normal range” does not conduct any pathological signals to the brain and that the cause of a pathological increase in blood pressure does not lie with the long-term regulation by the kidney, but is determined by other factors. In this case a renal denervation would not lead to medical success. In the case of a kidney having function values that lie outside of the “normal range”, however, it is to be assumed that the heretofore described spurious signals do occur, i.e. that the pathological increase in blood pressure arises from spurious signals of the kidney. A renal denervation would be indicated in this case and should be performed. The probability that a renal denervation would be successful is to be estimated as significantly higher than in the first case.
The methodologies described offer the following advantages, amongst others:
Number | Date | Country | Kind |
---|---|---|---|
10 2012 217 724.8 | Sep 2012 | DE | national |
Filing Document | Filing Date | Country | Kind |
---|---|---|---|
PCT/EP2013/068828 | 9/11/2013 | WO | 00 |