Radiological imaging method

Abstract

The invention relates to a radiological imaging method, said method comprising a vertical scanning of a standing patient (20) by two radiological sources orthogonal to each other in two imaging directions along a vertical scanning direction (Z), wherein said radiological method comprises at least one operating mode in which: a frontal scout view is obtained to identify a specific skeletal location (21) within said frontal scout view, the driving current intensity and voltage intensity of said frontal radiological source are simultaneously adjusted (11), preferably automatically synchronized, as a function of the patient thickness and of said identified specific skeletal location (21) along said vertical scanning direction (Z), to improve the compromise between: reducing the overall radiation dose received by the patient (20) during said vertical scanning, and, for the frontal image, increasing the local image contrast of said identified specific skeletal location (21) at different imaging locations along said vertical scanning direction (Z).

Description

TECHNICAL FIELD

[0001] The present application relates to a radiological imaging method, with the aim of reducing the dose of radiation received by the patient, while still wishing to obtain a high quality image of the patient's body. BACKGROUND

[0002] The scanning stereoradiographic system exhibits an interesting ability to take simultaneously frontal and lateral views for 3D reconstruction of skeletal anatomical sites such as the spine or the pelvis, with a dose reduction of up to 50 and even up to 100 compared to a CT (Computed Tomography) scan, and a dose reduction of up to 10 for a single view image compared to a classic CR (Computed Radiography) or DR (Direct Radiography) system.

[0003] However, this system does not have AEC (Automatic Exposure Control) and the parameters of the patient scan can therefore only be chosen manually by the operator according to the visual assessment that he can make of the patient's anatomy. The operator can only choose between three possible anatomical sizes: small, medium or large. The spectral characteristics (kV and filter) and the intensity of the X-ray beam are adjusted according to the protocol chosen (whole body, whole spine, pelvis...) and the size of the patient (small, medium or large) with reference to a simple table. This system is interesting in helping the operator to choose the parameters, but the main drawback is linked to the only visual choice of the operator.

[0004] This drawback is also well known with standard 2D radiography and some automatic exposure control (AEC) was developed to achieve an automatic exposure duration. This AEC can stop the exposure once the dose objective is reached in a dosimeter unit, usually a radiotransparent ionization chamber placed between the patient and the detector (film, CR, DR). This AEC has also some drawbacks. The first one is again related to the choice of the spectrum (kV, filter) by the operator, usually chosen among 3 possible cross choices of protocol and patient size. The second one of this AEC is related to the type of detector used to acquire the image. As long as a film is used, this AEC can provide correct results because the effect of overexposure or underexposure providing too bright or too dark image density is clearly visible on the film. So, the dose objective of the AEC can be set to obtain a standard expected density of a film. But, when computed radiography (CR) and direct radiography (DR) captors started to become widespread 2D detectors used in radiology, other problems appeared because the operator was no more able to detect overexposure or underexposure by simply looking at the light or dark image density because these detectors and their automatic image processing can provide almost the same type of image density whatever the dose, the only difference being the noise in the image. The operator has some difficulties to set correctly the dose objective according to the noise in the image and the use of different types or vendor's CR or DR detector models is also a reason of very difficult good results.

[0005] IEC 62494-1 proposes to use an exposure index (EI) which depends on the noise and thus on the signal to noise ratio (SNR) of the image to define the dose objective of the AEC as an exposure index target (EIT). IEC 62494-1 also proposes to define a deviation index (DI) as the ratio of the achieved exposure index to the exposure index target, expressed in decibel DI = 10*log 10 (EI / EIT).

[0006] Considering a scanning radiography system, the known method for AEC of standard 2D radiography using a dosimeter unit placed between the patient and the detector is hardly compatible for different reasons. The first reason is related to the huge difficulty or even impossibility to change the exposure time which is only related to the scanning speed and the size of the scanning area, where for 2D systems the AEC dosimeter unit does not move relative to the patient and therefore this static measurement for 2D systems can stop the exposure immediately after reaching the exposure target. The second reason is the field of view of such dosimeter for scanning radiography systems which depends only on a very small portion of the relevant ROI (Region of Interest) for the diagnostic of the patient body, usually only one line of the total image. Using such a small portion of the relevant ROI has almost no chance to provide the a priori information which would be needed to handle the shooting parameters to obtain an exposure index close to the exposure index target in the relevant ROI and to obtain a satisfactory signal to noise ratio (SNR).

[0007] According to a first prior art outside the field of radiological vertical scanning imaging, a computed tomography imaging method based on a rotating radiation source emitting a very high radiation dose is known in patent application US 2011 / 0026668, the emission being made in a helical path along a horizontal scanning direction of a patient lying on a bed. The radiation source is driven by a current intensity and a voltage intensity. The current intensity is related to the radiation dose emitted by the radiation source per unit of time, for example one second. The voltage intensity is related to the energy of each photon emitted by the radiation source. In order to make such a horizontal scanning for a given specific lying patient, the voltage intensity is adjusted to adapt the emitted radiation dose along the horizontal scanning direction, or the current intensity is adjusted to adapt the emitted radiation dose along the horizontal scanning direction, the radiation dose being in any case very high for the patient thickness along the horizontal direction and at least fifty times the vertical scanning of a standing patient, thus improving the contrast of the overall image with the patient thickness along the horizontal direction.

[0008] In one embodiment, the voltage intensity adjustment is made only as a function of the angular position of the radiation source along the lying patient and not as a function of the horizontal position of the radiation source along the lying patient, the current intensity adjustment being made as a function of the horizontal position of the radiation source along the lying patient.

[0009] In another embodiment, the voltage intensity adjustment is made as a function of the angular position of the radiation source along the lying patient and as a function of the horizontal position of the radiation source along the lying patient, but the current intensity adjustment is not made as a function of the horizontal position of the radiation source along the lying patient.

[0010] But in this first prior art embodiment, the voltage intensity and the current intensity are not adjusted simultaneously as a function of the horizontal position of the radiation source along the lying patient. SUMMARY

[0011] The object of the present application is to at least partially alleviate the above-mentioned drawbacks.

[0012] More particularly, the present application aims at improving the trade-off between:

[0013] reducing the radiation dose received by the patient,

[0014] and improving the image quality of the patient's body or of the patient's organs.

[0015] The present application proposes to solve the problem of reducing the emitted radiation dose in a vertical scan of a standing patient, which is much lower than in computed tomography, whatever it is, while still reducing this emitted radiation dose, while looking for a different improvement of the image contrast along the vertical direction, according to the patient thickness variation along said vertical scan direction, by adjusting the amount of emitted radiation particles and by adjusting the intrinsic energy of each emitted radiation particle.

[0016] Thereby, not only the emitted radiation dose is reduced and kept minimized at low thickness zones of the patient's body while maintaining still sufficient emitted radiation dose at high thickness zones of the patient's body, but also the intrinsic energy of each emitted particle is adapted to the thickness value in the patient's imaged zone, so that not only the overall emitted radiation dose is improved but also the local image contrast.

[0017] This trade-off between reduced emitted radiation dose and enhanced image contrast can be different at each height along said vertical scan direction, or at least generally varies with the patient body thickness along said vertical scan direction.

[0018] Moreover, this local image contrast is enhanced:

[0019] This depends not only globally on the overall thickness of the patient, as in the first prior art,

[0020] but also on the local thickness guided on the specific bone positioning identified in the patient in the probe view.

[0021] This object is achieved by a radiological imaging method, said method comprising two mutually orthogonal imaging directions, a frontal radiation source and a lateral radiation source, vertically sliding to perform a vertical scan along a vertical scan direction of a standing patient, wherein said radiological method comprises at least one operating mode in which: a frontal scout view is obtained by a preliminary vertical scan of a standing patient along said vertical scan direction by said frontal radiation source, said frontal scout view is processed to identify a specific skeletal positioning within said frontal scout view, a drive current intensity of at least said frontal radiation source is adjusted along said vertical scan direction as a function of a patient thickness and said identified specific skeletal positioning along said vertical scan direction, a drive voltage intensity of said frontal radiation source is adjusted along said vertical scan direction as a function of a patient thickness and said identified specific skeletal positioning along said vertical scan direction, said drive current intensity and said drive voltage intensity of said frontal radiation source being adjusted simultaneously, preferably automatically synchronized, to improve a trade-off between: reducing an overall radiation dose received by the patient during said vertical scan, and increasing a local image contrast of said identified specific skeletal positioning at different imaging positions along said vertical scan direction for a frontal image. Drive voltage intensity is a synonym for drive tension intensity.

[0022] This object is also achieved by a radiological imaging method, said method comprising two mutually orthogonal imaging directions, a frontal radiation source and a lateral radiation source, vertically sliding to perform a vertical scan along a vertical scan direction of a standing patient, wherein said radiological method comprises at least one operating mode in which: a lateral scout view is obtained by a preliminary vertical scan of a standing patient along said vertical scan direction by said lateral radiation source, said lateral scout view is processed to identify a specific skeletal positioning within said lateral scout view, a drive current intensity of at least said lateral radiation source is adjusted along said vertical scan direction as a function of a patient thickness and said identified specific skeletal positioning along said vertical scan direction, a drive voltage intensity of said lateral radiation source is adjusted along said vertical scan direction as a function of a patient thickness and said identified specific skeletal positioning along said vertical scan direction, said drive current intensity and said drive voltage intensity of said lateral radiation source being adjusted simultaneously, preferably automatically synchronized, to improve a trade-off between: reducing an overall radiation dose received by the patient during said vertical scan, and increasing a local image contrast of said identified specific skeletal positioning at different imaging positions along said vertical scan direction for a lateral image.

[0023] This object is also achieved by a radiological imaging method, said method comprising two mutually orthogonal imaging directions, a frontal radiation source and a lateral radiation source, vertically sliding to perform a vertical scan of a standing patient along a vertical scan direction, wherein said radiological method comprises at least one operating mode in which: frontal and lateral scout views are obtained by performing a preliminary vertical scan of a standing patient along said vertical scan direction by said frontal and lateral radiation sources, said frontal and lateral scout views are processed to identify a specific skeletal location within said frontal and lateral scout views, the driving current intensity of at least said frontal and lateral radiation sources is adjusted along said vertical scan direction as a function of the patient thickness and of said identified specific skeletal location along said vertical scan direction, the driving voltage intensity of said frontal and lateral radiation sources is adjusted along said vertical scan direction as a function of the patient thickness and of said identified specific skeletal location along said vertical scan direction, said driving current intensity and voltage intensity adjustment of said frontal radiation source and said driving current intensity and voltage intensity adjustment of said lateral radiation source are both performed simultaneously, preferably automatically synchronized, to improve the trade-off between: reducing the overall radiation dose received by the patient during said vertical scan, and increasing the local image contrast of said identified specific skeletal location at different imaging locations along said vertical scan direction for frontal and lateral images.

[0024] Synchronized means:

[0025] For frontal or lateral images, the driving current intensity and voltage intensity adjustment are fully synchronized together, i.e. in phase with each other,

[0026] The driving current intensity of the frontal images is synchronized with the driving current intensity of the lateral images, possibly with a time shift (one is ahead or behind the other in vertical direction by some rows in order to reduce cross-scattering effects),

[0027] The driving voltage intensity of the frontal images is synchronized with the driving voltage intensity of the lateral images, possibly with a time shift (one is ahead or behind the other in vertical direction by some rows in order to reduce cross-scattering effects).

[0028] The preferred embodiments comprise one or more of the following features, which can be separate or together with any of the preceding objects of the invention, in partial or total combination.

[0029] Preferably, said identified specific skeletal location comprises the patient spine, preferably the patient spinal column.

[0030] Indeed, the patient spine is the specific skeletal location of most interest for detailed analysis within the patient; it is therefore used for driving voltage intensity and current intensity adjustment.

[0031] Alternatively, the specific skeletal positioning can also be the pelvis or the arm or leg of a standing patient along the vertical scan direction, depending on the region of interest within the patient body part being imaged.

[0032] Preferably, the driving current and voltage intensity adjustment of the frontal radiation source is also performed to reach a certain signal-to-noise ratio value, which is constant and common for most of the imaging positions along the vertical scan direction, preferably for all the imaging positions along the vertical scan direction, but which can take two different values for frontal and lateral images, respectively, for the frontal and / or lateral images.

[0033] Preferably, for each of the frontal and / or lateral images, the signal-to-noise ratio value is constant and predetermined for each different patient organ to be imaged.

[0034] Preferably, for the frontal image of the patient spine, the standard signal-to-noise ratio value corresponds to a number of X-ray photons received per detector pixel comprised between 50 and 70, the radiological imaging method operator preferably having the possibility to deviate from this standard value by at least + or - 20%, more preferably at least + or - 50% by manual command, and / or for the lateral image of the patient spine, the standard signal-to-noise ratio value corresponds to a number of X-ray photons received per detector pixel comprised between 20 and 40, the radiological imaging method operator preferably having the possibility to deviate from this standard value by at least + or - 20%, more preferably at least + or - 50% by manual command.

[0035] Thus, with a constant and optimized signal-to-noise ratio along or even all along the vertical scan direction, the local image contrast of the identified specific skeletal positioning at different imaging positions along the vertical scan direction is greatly improved, since this is the region of interest in the frontal and / or lateral images in practice.

[0036] Preferably, the frontal and / or lateral images are normalized, preferably by homogenization of the regions located only outside the patient body contour, after having undergone at least a first step of increasing the local image contrast of the identified specific skeletal positioning at different imaging positions along the vertical scan direction, in order to eliminate image artifacts from the driving current and voltage intensity adjustment.

[0037] In fact, due to these driving current intensity and voltage intensity adjustments, there are some artifacts in the frontal and / or lateral images, which superimpose some clear and dark gray level alternating waves on the blank part of the image (just outside the patient's body) or on very thin parts of the patient, making these images less satisfactory for interpretation by the operator of the radiological imaging method or at least requiring some training on his or her side.

[0038] Preferably, the frontal and / or lateral images, after normalization, undergo a contrast enhancement step.

[0039] Thus, on the one hand, the image artifacts resulting from these driving current intensity and voltage intensity adjustments are eliminated, while the contrast enhancement improved by these same driving current intensity and voltage intensity adjustments is not only preserved but also fully exploited.

[0040] Preferably, the identified specific bone positioning excludes metal parts, if any, such as metal prostheses of the patient's body skeleton parts or metal protections placed on the patient's body, for example, before carrying out the radiological imaging method.

[0041] In fact, these foreign (to the patient's body) objects introduced into or on the patient's body, being metal and thus stopping more radiation and X-rays than the rest of the patient's body, can cause some non-optimization of the emitted dose, which can lead to some overexposure or some underexposure of the emitted radiation at the height corresponding to these foreign objects. In the following modes, first the driving voltage intensity is constant, second the driving voltage intensity and the driving current intensity are both constant, the consequences can be even worse if the metal outliers are not excluded, because more or all parameters are chosen for the maximum thickness, thus the emitted radiation dose is higher or much higher than the required dose, which is very harmful to the patient.

[0042] Preferably, the current intensity and voltage intensity are adjusted: for larger patient thicknesses, the current intensity and the voltage intensity are simultaneously increased, for smaller patient thicknesses, the current intensity and the voltage intensity are simultaneously decreased, the current intensity variation rate is slower than the voltage intensity variation rate.

[0043] Thus, for large patient thicknesses, a larger number of higher energy radiation particles more easily pass through the patient's body, despite its large thickness, without having to increase the emitted radiation dose too much, all emitted signals being avoided from being absorbed by the large thickness patient's body due to the larger number of emitted radiation particles and the higher energy of each emitted radiation particle.

[0044] For small patient thicknesses, a larger number of lower energy radiation particles is absorbed by the patient's body, although its thickness is small, without having to increase the emitted radiation dose too much, since the smaller number of emitted radiation particles and the lower energy of each emitted radiation particle, all emitted signals are prevented from passing through the small thickness patient body.

[0045] Preferably, the current intensity adjustment is maximized, thereby also maximizing the vertical scan speed to a constant value.

[0046] Thus, for a given emitted radiation dose, and in turn for a given radiation dose received by the standing patient during the vertical scan, both are kept at the same level, the total vertical scan time is significantly reduced, with the advantage of reducing the likelihood of the standing patient moving and the impact of the patient moving, thereby reducing the blurring risk and the deformation risk of the frontal and lateral images to some extent, thereby still enhancing the signal-to-noise ratio of these frontal and lateral images.

[0047] Preferably, the operating mode can be manually turned on or off by a radiological imaging method operator.

[0048] Thus, this very advantageous way of operating a radiological imaging device is available, whereas if and when the operator of this radiological imaging device wants to remove this way, the way can be cancelled in order to operate this radiological imaging device, for example, completely manually. The radiological imaging method according to the advantageous embodiment of the present application presents 3 operating modes: full manual mode, AEC mode without adjustment, AEC mode with adjustment.

[0049] Preferably, the operating mode is dedicated to vertical scans of large and / or obese patients.

[0050] Preferably, the operating mode is dedicated to vertical scans of child patients.

[0051] The radiological imaging method according to the present application is even more focused on patients whose thickness can be particularly lower or particularly higher than the average build. This shows the very patient-specific ability of the radiological imaging method according to the present application. Of course, the radiological imaging method according to the present application is also very effective for patients of standard build.

[0052] Preferably, the current intensity adjustment rate does not exceed a predetermined threshold of 5 mA per millisecond, preferably a predetermined threshold of 2 mA per millisecond, more preferably a predetermined threshold of 1 mA per millisecond.

[0053] Thus, the radiological imaging method according to the present application can also be carried out with a relatively simple and inexpensive radiation source whose current intensity driving capability is relatively slow.

[0054] Preferably, said current intensity regulation ranges at least from 20 mA to 300 mA, and preferably from 10 mA to 400 mA.

[0055] Thus, the radiological imaging method according to the application can also be performed with a relatively simple and inexpensive radiation source having a relatively limited range of current intensity driving capabilities.

[0056] Preferably, said voltage intensity regulation ranges at least from 60 kV to 100 kV, and preferably from 50 kV to 120 kV.

[0057] Thus, the radiological imaging method according to the application can also be performed with a relatively simple and inexpensive radiation source having a relatively limited range of voltage intensity driving capabilities, while at the same time taking full advantage of the available range of voltage intensity driving capabilities.

[0058] Preferably, said vertical scan speed value ranges at least from 8 cm / s to 20 cm / s, and preferably from 4 cm / s to 30 cm / s.

[0059] Thus, the radiological imaging method according to the application can also be performed with a relatively simple and inexpensive radiation source having a relatively limited range of vertical scan speed capabilities, while at the same time taking full advantage of the available range of vertical scan speed capabilities.

[0060] Preferably, each of said frontal and lateral images is obtained by performing a preliminary vertical scan of the standing patient along the vertical scan direction with a reduced overall radiation dose compared to each of said frontal and lateral images, before obtaining each of said frontal and lateral images.

[0061] Thus, the regulation of the driving current intensity and driving voltage intensity, and possibly the vertical scan speed, can be determined only prior to performing the vertical scan according to the thickness profile and the specific skeletal positioning of the standing patient body along the vertical scan direction, which will result in the production of effective frontal and lateral images of the standing patient body with a limited and sufficient radiation dose sufficient to obtain high quality frontal and lateral images. The scout views can be produced at the cost of a rather limited overexposure to the emitted radiation.

[0062] Preferably, said reduced overall radiation is less than 10% of said overall radiation dose, and preferably less than 5% of said overall radiation dose.

[0063] Thus, the benefit is twofold: not only is the overexposure (+10% or +5%) during the production of the scout views very limited, but also the trade-off between optimizing the overall radiation dose received and enhancing the image contrast is very efficient.

[0064] Preferably, the pixels in the probe view are aggregated together to form an imaging zone, preferably in N x N pixel zones, more preferably in at least 10 x 10 pixel zones.

[0065] Thus, the image quality and image contrast of the probe view is enhanced, despite the very low level of emitted radiation dose used to produce this probe view.

[0066] Preferably, the image or the imaging zone is processed to identify salient points, which in turn are used to compute the thickness profile and identify the specific skeletal positioning of the standing patient along the vertical scan direction.

[0067] Thus, the thickness profile is computed and the specific skeletal positioning of the standing patient along the vertical scan direction is identified from the probe view, despite the very low level of emitted radiation dose, more easily and more efficiently.

[0068] Preferably, the image or the imaging zone is processed by a neural network to compute the thickness profile and identify the specific skeletal positioning of the standing patient along the vertical scan direction.

[0069] Thus, the specific skeletal positioning of the standing patient along the vertical scan direction is identified from the probe view, despite the very low level of emitted radiation dose, more easily and more efficiently.

[0070] Preferably, the 2 radiation sources are vertically sliding so as to vertically scan the pelvis or the spine or the whole body of the standing patient along a vertical scan direction.

[0071] Preferably, 2 radiation detectors are respectively associated with the 2 radiation sources, the 2 radiation detectors being 2 photon counting detectors (PCD), each photon counting detector being associated with an automatic image processing function that automatically balances image density to enhance image contrast, whatever the radiation dose received on the sensitive surface of the radiation detector.

[0072] Thus, it is more difficult for the radiological imaging method operator to correctly manually assess overexposure or underexposure of the radiation signal emitted by the radiation source. Moreover, photon counting detectors have an improved linearity and signal-to-noise ratio compared to gas detectors.

[0073] Preferably, 2 radiation detectors are respectively associated with the 2 radiation sources, the 2 radiation detectors being 2 multi-energy counting detectors, preferably 2 energy-resolved photon counting detectors (ERPCD).

[0074] Preferably, the radiation is X-ray.

[0075] A standing patient or standing posture patient is a patient in a weight-bearing posture, as opposed to a lying patient or a patient in a lying posture in computed tomography. Another patient weight-bearing posture that can replace the patient standing posture can be a patient seated posture.

[0076] Further features and advantages of the present application will appear from the following description of embodiments of the application, given by way of non-limiting examples only, with reference to the annexed drawings. BRIEF DESCRIPTION OF DRAWINGS

[0077] Figure 1 An example is shown of a part of a radiological imaging method according to an embodiment of the application, which handles the computation of the driving current intensity adjustment of the radiation source.

[0078] Figure 2 An example is shown of a specific bone localization by salient point detection, here the complete body of a patient containing the spine extending to the pelvis and legs.

[0079] Figure 3 An example is shown of a corresponding parameter, which is the median signal patch value as a function of vertical position in mm along the height of the patient.

[0080] Figure 4 An example is shown of a corresponding thickness profile in mm as a function of vertical position in mm along the height of the patient.

[0081] Figure 5 An example is shown of a radiological imaging method according to an embodiment of the application, which handles the computation of the driving current intensity and voltage intensity adjustment of the radiation source.

[0082] Figure 6 An example is shown of a voltage adjustment in kV as a function of patient equivalent thickness in cm, for frontal and lateral images respectively.

[0083] Figure 7 An example is shown of a current adjustment in mA as a function of patient equivalent thickness in cm, for frontal and lateral images respectively.

[0084] Figure 8 An example is shown of the number of X-ray photons per detector pixel obtained (indicative of the signal-to-noise ratio) as a function of patient equivalent thickness in cm, for frontal and lateral images respectively.

[0085] Figure 9 An example is shown of the radiation dose received by the patient in μGy (microgray) as a function of patient equivalent thickness in cm, for frontal and lateral images respectively.

[0086] Figure 10An example of a filtered frontal scout view after the salient point detection step but before the salient point filtering step is shown.

[0087] Figure 11 An example of a filtered frontal scout view after the salient point detection step and after the salient point filtering step is shown.

[0088] Figure 12 An example of a filtered lateral scout view after the salient point detection step but before the salient point filtering step is shown.

[0089] Figure 13 An example of a filtered lateral scout view after the salient point detection step and after the salient point filtering step is shown.

[0090] Figure 14 An example of a filtered frontal scout view after the deep neural network detection step is shown.

[0091] Figure 15 An example of a filtered lateral scout view after the deep neural network detection step is shown.

[0092] Figure 16 An example of a lateral image with constant driving current intensity and constant driving voltage intensity in constant current and voltage mode is shown.

[0093] Figure 17 An example of a lateral image with adjusted driving current intensity and constant driving voltage intensity in adjusted current and constant voltage mode is shown.

[0094] Figure 18 An example of a lateral image with adjusted driving current intensity and adjusted driving voltage intensity in dual adjusted current and voltage mode is shown.

[0095] Figure 19 An example of a final lateral image shown to a radiologist is shown. DETAILED DESCRIPTION

[0096] The present invention aims at providing a solution to provide a scanning stereoradiography system with an AEC system, this AEC being compliant with IEC 62494-1. This AEC system is designed for the scanning stereoradiography systems described in applications PCT / IB2016 / 000273 and PCT / IB2017 / 000986, incorporated by reference and owned by the same applicant, EOS-Imaging.

[0097] In a preferred embodiment, the two detectors of this scanning stereoradiography system are multi-energy counting detectors, also called energy-resolved photon counting detectors (ERPCD) with at least 2 energy bins. In another embodiment, the two detectors of this scanning stereoradiography system are single-energy counting detectors, also called photon counting detectors (PCD).

[0098] The use of photon counting detectors in multi-energy or single-energy form is advantageous compared to gas detectors for mainly two reasons. The first reason is that the signal of an ERPCD or PCD is linear with the incident flux and directly equal to the number of detected photons, while the signal of a gas detector is strongly non-linear and this non-linearity is quite complex to model for accurate correction of it. In ERPCD and PCD, there is still a non-linear behavior at high flux, called pile-up effect, but this pile-up effect can be well modeled and corrected by image calibration software. The second reason is that ERPCD and PCD have a very stable behavior and sensitivity, without the need for new calibration in a few months, and are not sensitive to room temperature variations, while a gas detector is far less stable and needs daily calibration, and the behavior and sensitivity of said gas detector can also vary in a few minutes according to room temperature variations. The stability of PCD and ERPCD and the photon counting function enable to directly use the count signal of the incident photons in the detector to evaluate the exposure index and the signal-to-noise ratio. The signal-to-noise ratio is directly equal to the root mean square of the signal. Other types of energy integrating detectors, such as gas detectors, do not have this advantage and need an accurate calibration to evaluate the SNR and thus the exposure index.

[0099] The radiological imaging method according to an embodiment of the application is based on the use of scout views under single-energy (ERPCD or PCD). Since one of the objectives of this scanning radiography system is the imaging of bones dedicated to orthopedics, the scout views are analyzed in this case to accurately find the axial skeleton or bones of the selected protocol. But for some other applications, a soft tissue organ protocol can be selected, for example the lung, in which case the scout views are analyzed to find the organ.

[0100] The relevant ROI for diagnosis defined according to IEC 62494-1 is defined by the union of a set of circular sub-ROIs, also called "patches" of about the size of a vertebra (diameter 4-5 cm), placed on a set of feature detection points or landmarks on the scout view according to the protocol-specific bone or organ search. This search of feature points can be embodied in two different ways: a specific salient point search algorithm or the use of a trained pose detection deep neural network.

[0101] The patient's equivalent thickness is then evaluated in each patch, and a certain selection rule scheme provides a vertical vector of equivalent thickness according to the patient's Z (vertical) position. When using mono-energetic probe views, the equivalent thickness is evaluated in a single material PMMA [Poly(methyl methacrylate)] equivalent.

[0102] The feature thickness is then processed using the vertical equivalent thickness vector, which is a safe detection of the most likely maximum thickness. The scanned parameters are then processed using the feature thickness and the equivalent vector thickness to obtain an exposure index in each patch as close as possible to the exposure target.

[0103] According to the operator's choice, the exposure parameters can be generated in several different modes:

[0104] The first mode, called "constant exposure mode", simply provides the best constant kV, mA, filter and scan speed for the scan and will provide a constant flux exposure control according to the definition of IEC 60601-2-44;

[0105] The second mode, called "elastic dose", can be used in two options. The two options will process the scan speed, the selected filter and the exposure time adjustment vector along the vertical axis and will provide a Z-axis exposure control according to the definition of IEC 60601-2-44.

[0106] The "elastic dose" can be used in a first option in which the voltage or tension (kV) is fixed, the current (mA) is adjusted along the Z axis, and in a second option in which both the voltage or tension (kV) and the current (mA) are adjusted along the vertical (Z) axis. The operating mode according to the application is the second option of this second mode called "elastic dose". The second option of this second mode called "elastic dose" is also called the automatic mode of the scan radiographic system.

[0107] The general principle of this automatic mode according to embodiments of the application is different from standard automatic exposure control (AEC) for 2D radiology detectors, such as CR (Computed Radiography) or DR (Digital Radiography). Standard radiology AEC uses real-time dosimetry at the back of the patient with a 2D detector entering horizontal to stop X-ray emission as soon as a target level is reached; it is based on ionization chamber cells with typical square shape with side size of 5-10 cm to adjust the exposure time. Typically, standard radiology AEC devices offer the operator a choice to use one of several different cells, for example in the middle, left or right of the 2D detector; thus, the exposure time is only well adjusted on the corresponding part of the cell area of the patient, not optimally on the whole area of the 2D detector. Moreover, standard radiology AEC cannot optimize all X-ray shooting parameters; voltage (kV) is usually manually selected based on a choice of modality and protocol and a filter is selected directly or indirectly.

[0108] The automatic mode is based on a very low dose scout view (also called preview or scanogram) that is required. AEC provides a constant exposure mode with constant flux exposure control and Z-axis exposure control respectively and an adjustment mode called "elastic dose". The constant flux exposure control system determines the optimal constant X-ray flux to be used over the whole scan sequence and the Z-axis exposure control system adjusts the incident X-ray flux along the Z-axis. The Z-axis is vertical when the patient is in standing or sitting position.

[0109] Figure 1 An example is shown of a part of a radiological imaging method according to embodiments of the application, which handles the calculation of the driving current intensity adjustment of the radiation source.

[0110] The radiological imaging method according to embodiments of the application comprises a method for handling the current intensity and voltage intensity adjustment of the radiation source along a vertical scan direction. Figure 1 and 5 A functional block diagram of an example implementation using a salient point is presented.

[0111] The following successive steps are performed:

[0112] In step 1, a scout view of a standing patient is acquired at reduced radiation dose. The scout view is necessary using the automatic mode and is acquired with a 0.5 mm thick copper filter and very low dose. The patient dose ratio between this scout view and the main photograph is less than 10% for long axis and local protocols. This step 1 of the method is by acquiring a scout view at reduced dose by vertical scanning along the patient.

[0113] In step 2, post-processing is performed on the scout view, wherein the pixels in the scout view are preferably clustered together in N x N pixel bins, for example in bins of at least 20 x 20 pixels, to obtain an imaged area with calibration and averaging binning. Here, the scout view is acquired at such a low level of dose that a large binning of 20 x 20 is applied to filter enough noise and obtain a higher level of confidence on the estimated thickness. This step 2 is a post-processing of the scout view image, containing uniformity correction using the gain calibration of the detector, and using a filter to improve the signal-to-noise ratio without introducing any bias to obtain a high level of confidence on the estimated thickness, for example a binning of 20 x 20 is very suitable.

[0114] In step 3, the salient points are detected as will be explained in more detail with reference to Figures 10 to 13 The salient point detection algorithm developed for the detection of validated axial skeleton is applied to the binned image. The detected salient points are then selected differently in the frontal and lateral images according to the patient orientation. The selected salient points are mainly along the axial skeleton consisting of lower limbs, pelvis, spine, neck and head. Metal prosthesis or metal protection detection is also used to remove and exclude the corresponding selected salient points when metal detection is performed in step 3, while in step 8 only outliers are selected with respect to thickness. This step 3 is the detection of specific skeletal locations and rejection of metal parts in the scout view. As an example, we use a description of the method using salient point detection, but another implementation using a specific neural network can also be used. The salient point detection method used consists in searching in a range of typical vertebra size, i.e. diameter of about 5 cm, to find a set of local maximum attenuation points in the scout view. The detected salient points are then selected differently in the frontal and lateral images according to the patient orientation. The selected salient points are mainly along the axial skeleton consisting of lower limbs, pelvis, spine, neck and head. There is only one selected salient point at one Z height (one Z position along the vertical scan direction), while there can be several different detected salient points at one Z height. Figures 10 to 13 The salient point detection and selection on the patient frontal and lateral scout views are presented respectively. Figures 10 to 13 The left part of Figures 10 to 13 presents the detected salient points, and Figure 14 The right part of 15 presents the selected salient points. Figures 10 to 13 An example of skeleton detection using a neural network method is presented, containing a limited number of landmarks, compared to Figure 14 and 15The limited number of landmarks presented correspond to the cervical vertebrae C2 and C7, the thoracic vertebrae T9, the sacrum, and the left and right femoral head, 1 / 3 diaphysis, trochlear center, distal tibia and proximal tibia of the lower limbs. The method of selection of the salient points also enables the removal of points detected on metallic parts, including for example metal prostheses or metal protectors for the breasts, gonads, ovaries or other sensitive parts of the body. Metallic parts can also be removed using a neural network method.

[0115] In step 4, from the salient points, a signal profile in mm as a function of the vertical position is obtained. This step 4 is processing the signal profile along the vertical scan of the patient. The median of the signal of the scout view image is processed in each circular patch centered on the selected salient point and associated with the height z of the corresponding salient point in the vertical scan reference. The patch size represents an area approximately equivalent to the size of a vertebra with a diameter of about 5 cm. The resulting signal profile function of z is sparse and the interpolation and extrapolation of this sparse signal profile provides a complete sampling of the signal vertical profile along the vertical height z over the entire height of the vertical scan required for the patient, including some bottom and top extensions of the scan region selected by the operator.

[0116] In step 5, from the previous signal profile, the patient thickness profile is calculated. Indeed, this step 5 is processing the thickness profile along the vertical scan of the patient. The signal profile function of z is processed using a calibration quadratic polynomial function to provide the corresponding PMMA thickness equivalent value according to equation 1:

[0117]

[0118] with mAs = mA 探查视图 * PMMA of maximum thickness = 600 mm.

[0119] The coefficients a, b and c of the polynomial are processed using a calibration.

[0120] In step 6, optionally, new upper and / or lower scan limits can be manually selected by the radiological imaging method operator. The operator can use the selection tools on the scout view image in the interface software to slightly decrease or increase the scan height.

[0121] In step 7, optionally, the patient thickness profile is cropped according to the previous new upper and / or lower scan limits. The processed thickness profile function of z is then cropped according to the operator's selection of the upper and lower limits of the vertical scan along the patient.

[0122] In step 8, the characteristic patient thickness is calculated from the previous thickness profile after having filtered out outliers such as metal prostheses or metal protections in step 3. This step 8 is the processing of the characteristic thickness. Considering only the maximum of the thickness profile determined previously in step 5 does not represent the patient thickness when the patient is sitting or legs on support or arms crossed. To define a representative thickness of the patient, the derivative of the thickness profile is processed to use the mean and standard deviation to define statistical parameters to remove outliers and then the characteristic thickness is defined as the maximum thickness without considering outliers.

[0123] In step 9, the image acquisition parameters are retrieved from the reference table according to the specific bone positioning (e.g. the exact position of the patient spine along the vertical scan direction) and the patient thickness variation along the vertical scan direction (i.e. along the patient height). Indeed, the processed characteristic PMMA equivalent thickness and the selected protocol are used to obtain the image acquisition parameters (kV only for Figure 1 are not used in the Figure 5 as they are the driving voltage intensity regulation, filter, scan speed). Step 9 is the selection in the image acquisition parameters table according to the processed characteristic thickness and the selected protocol. The parameter set includes for example the signal target and the reference kV. The reference kV, mA and speed can be used to acquire the patient image with the operator selection of the regulation disabled, but with the AEC maintained, in which case the constant values of scan speed and current will be processed according to the signal target and the characteristic thickness. The reference voltage (kV) can also be used to acquire the patient image with a constant voltage (kV) current regulation enabled. In addition, in the case of regulation enabled, only the signal target of the frontal and / or lateral view will be used in the following steps 10 or 11. As an example, the signal target for a spine examination is different for the frontal image and the lateral image and is typically about 60 photons / pixel for the frontal image and about 30 photons / pixel for the lateral image respectively. The radiologist can adjust these values within a limited allowed range.

[0124] In step 10, depending on the selected operating mode, a constant current intensity value along the vertical direction is selected, or preferably a variable adjustment of the current intensity, as a function of the precise location of the patient's spine along the vertical direction and the patient's thickness variation along the vertical scanning direction, to achieve a constant and common signal-to-noise ratio along the vertical scanning direction, i.e., a constant target X-ray photon number per detector pixel along the vertical scanning direction, which preferably has different values ​​for frontal images (e.g., 60) and side images (e.g., 30). The final step in automatic mode is to process the current (mA) adjustment vector of the images scanned along the patient's body parts to fit the exposure index target as closely as possible in each patch of the ROI. A calibration matrix is ​​used to establish the relationship between the measured signal values ​​as a function of a set of voltage (kV) values ​​and the PMMA equivalent thickness. The algorithm used to determine the current (mA) performs 2D interpolation in this matrix. When describing the dual-drive current intensity and drive voltage intensity adjustment, it is possible to... Figure 5 A complementary and detailed explanation of adjusting the intensity of the drive current was found there.

[0125] Figure 2 An example of specific bone localization detected by protrusion points is shown, in this case, the patient's complete body including the spine extending to the pelvis and legs. Figure 2 A typical example of a selected protrusion point is presented in a side view.

[0126] A continuous circle runs vertically along the patient's spine. This continuous protrusion SP is found on the lateral image of the patient's body. This lateral image of the patient's body is plotted with respect to the height Z (height along the vertical scan direction) as a function of the patient's width, the patient's thickness th perpendicular to the plane of the image, and both height and width w are expressed in mm.

[0127] Figure 3 An example of the corresponding parameter P is shown, which is the median signal patch value that varies with the vertical position Z in millimeters along the patient's height. Figure 3 The distribution of the corresponding full-sample signal is presented along the vertical scan of the patient. The selected prominence is... Figure 2 The side view is indicated by dots, which represent circular patches with a diameter of approximately 5 cm.

[0128] Figure 4 An example is shown of how the corresponding thickness distribution, expressed in millimeters, varies with the vertical position Z, expressed in millimeters, along the patient's height. Figure 4 Presented in Figure 2 The characteristic thickness of the side distribution presented in the middle was measured.

[0129] Patient thickness distribution plotted on Figure 4The patient thickness th in mm is represented as a function of the vertical position Z also in mm.

[0130] Figure 5 An example of a radiological imaging method according to an embodiment of the application is shown, which processes the calculation of the driving current intensity and voltage intensity regulation of the radiation source.

[0131] The same principle can be used for "elastic dose" under the voltage (kV) regulation option, to process in a first step a voltage (kV) regulation vector that increases the contrast according to the thickness of the scanned body part of the patient, and then in a second step a current (mA) regulation to fit as close as possible the exposure index target in each patch of the ROI. Voltage (kV) regulation and current (mA) regulation are performed simultaneously in step 11, steps 1 to 10 being similar to the previous reference Figure 1 The corresponding steps 1 to 10 are described.

[0132] In this step 11, according to the selected operating mode, the variable regulation of the voltage intensity and the variable regulation of the current intensity along the vertical direction are selected as a function of the exact position of the patient spine along the vertical scanning direction and of the patient thickness variation along the vertical scanning direction, to obtain an equivalent patient thickness variation along the vertical scanning direction (bone attenuation more radiation, so the bone corresponds to a thickness greater than its actual thickness compared to soft tissue), to reach a constant and common signal-to-noise ratio along the vertical scanning direction, i.e. a constant target number of X-ray photons per detector pixel along the vertical scanning direction, the constant target number of X-ray photons per detector pixel along the vertical scanning direction preferably having different values for frontal images (e.g. 60) and lateral images (e.g. 30).

[0133] Step 11 processes the current intensity and voltage intensity regulation of the radiation source along the vertical scanning direction. The thickness profile processed in step 5 is used to process the voltage regulation profile using specific relationships of the voltage function of the thickness, such relationships being the same or different for frontal and lateral images, depending on the anatomical site. Figure 6Voltage functions of thickness are presented, for example, for a full spine protocol. These voltage functions of thickness are determined for optimizing contrast of the skeletal regions of the body and minimizing dose. Contrast is known to be better at low voltages (kV), but using low voltages (kV) does not efficiently use the useful dose that is provided in the detector for a certain signal compared to the dose received at thick parts of the body. The selection is guided by a trade-off between better contrast leading to lower voltages (kV) and more efficient dose usage leading to higher voltages (kV). The processed voltage profile is used to process the optimal scan speed and current profile. The thickness profile, the voltage profile, and a 2D table representing the expected signal as a function of voltage and thickness for a reference current and scan speed are used to process the current profile needed at the reference scan speed to reach the signal target. The current profile can then be adjusted in a way that is proportional to the specific scan speed. For example, the current profile needs to be doubled to double the scan speed with respect to the reference scan speed. The maximum current profile allowed for a given scan speed can be inferred by the ratio of the maximum power to the voltage profile in case the source has a maximum output power. The optimal scan speed can be found looking at the higher allowed current profile.

[0134] The last step of the automatic mode is to process the voltage (kV) adjustment vector and the current (mA) adjustment vector of the images scanned along the patient body part to fit as close as possible the exposure index target in each patch of the ROI. The voltage (kV) adjustment vector and the current (mA) adjustment vector can be combined together into a dual voltage and intensity vector.

[0135] The calibration matrix is to establish the relationship between the measured signal values and the PMMA equivalent thickness as a function of a set of voltage (kV) values. The algorithm for determining the current (mA) does a 2D interpolation in this matrix.

[0136] The voltage (kV) and current (mA) can be processed as a scalar value for a corresponding characteristic thickness in constant AEC mode, or as one or both of two adjustment vectors in "elastic dose" mode to optimize the X-ray flux according to the estimated patient thickness along the vertical scan. The "elastic dose" mode enables a significant dose reduction for long axis protocols compared to the constant current (mA) mode based on the patient maximum thickness only.

[0137] The overall goal of the automatic mode is to reach a constant and repeatable signal level at the maximum thickness location of the patient scan area in constant mode and furthermore to reach a constant and repeatable signal level along the patient's scan axis skeleton and independent of the patient morphology and thickness in both modes.

[0138] Figure 6 and 7A standard relationship of simulated voltage (kV) and current (mA) as a function of equivalent PMMA thickness according to a full spine protocol in "elastic dose" mode is presented.

[0139] Figure 6 An example of voltage regulation in kilovolts as a function of patient equivalent thickness th in centimeters is shown for frontal and lateral images, respectively.

[0140] For the frontal curve F, the voltage increases regularly from about 60 kV to about 120 kV for an equivalent patient thickness ranging from about 10 cm to just below 30 cm, and then remains constant at about 120 kV for an equivalent patient thickness ranging from just below 30 cm to about 40 cm or about 50 cm.

[0141] For the lateral curve L, the voltage increases regularly from about 60 kV to about 120 kV for an equivalent patient thickness ranging from about 10 cm to just above 30 cm, and then remains constant at about 120 kV for an equivalent patient thickness ranging from just below 30 cm to about 40 cm or about 50 cm. The increase of the F curve is steeper than the increase of the L curve.

[0142] Figure 7 An example of current regulation I in milliampere as a function of patient equivalent thickness th in centimeters is shown for frontal and lateral images, respectively.

[0143] For the frontal curve F, the current increases regularly from about 30 mA to about 400 mA for an equivalent patient thickness ranging from about 10 cm to about 40 cm, and then remains constant at about 400 mA for an equivalent patient thickness ranging from about 40 cm to about 50 cm.

[0144] For the lateral curve L, the current increases regularly from about 20 mA to about 400 mA for an equivalent patient thickness ranging from about 10 cm to about 40 cm, and then remains constant at about 400 mA for an equivalent patient thickness ranging from about 40 cm to about 50 cm. The increase of the F curve is on average similar to the increase of the L curve.

[0145] Figure 8 An example of the number of X-ray photons per detector pixel S (representative of the signal-to-noise ratio) as a function of patient equivalent thickness in centimeters is shown for frontal and lateral images, respectively.

[0146] For the frontal curve F, the number of X-ray photons per detector pixel is rather constant at about 60 for an equivalent patient thickness ranging from about 10 cm to about 35 cm, and then decreases abruptly from about 60 to about 10 for an equivalent patient thickness ranging from about 35 cm to about 50 cm.

[0147] For the lateral curve L, the number of X-ray photons per detector pixel is fairly constant at about 27 for an equivalent patient thickness ranging from about 10 cm to about 40 cm, and then the number of photons abruptly decreases from about 27 to about 10 for an equivalent patient thickness ranging from about 40 cm to about 50 cm. The steep decrease in the F curve is similar to the steep decrease in the L curve, but extends to a larger range of equivalent patient thicknesses.

[0148] Figure 9 Examples of the radiation dose D received by the patient in μGy as a function of the patient equivalent thickness th in centimeters are shown for the frontal image and the lateral image, respectively.

[0149] For the frontal curve F, the radiation dose received by the patient first increases slowly from about zero to about 500 microGy for an equivalent patient thickness ranging from about 10 cm to about 30 cm, then abruptly increases from about 500 microGy to about 3000 microGy for an equivalent patient thickness ranging from about 30 cm to about 40 cm, and then remains approximately constant at about 3000 microGy for an equivalent patient thickness ranging from about 40 cm to about 50 cm.

[0150] For the lateral curve L, the radiation dose received by the patient first increases slowly from about zero to about 700 microGy for an equivalent patient thickness ranging from about 10 cm to about 33 cm, then abruptly increases from about 700 microGy to about 3500 microGy for an equivalent patient thickness ranging from about 33 cm to about 40 cm, and then remains approximately constant at about 3500 microGy for an equivalent patient thickness ranging from about 40 cm to about 50 cm. The L curve increases more slowly than the F curve in the first, slow-increase phase, but the L curve increases more rapidly than the F curve in the second, abrupt (or rapid) increase phase.

[0151] The "elastic dose" mode enables to obtain better results in exposure index accuracy than the constant mode, compared to the exposure index target per patch, especially on the long axis protocol. Moreover, the voltage (kV) adjustment and current (mA) adjustment options enable to obtain better accuracy in exposure index compared to the exposure index target than the current (mA) adjustment only option of the "elastic dose", because it enables a faster adjustment than mA only (limited to 1 mA / ms for 5 kV / ms, where ms is millisecond), thus a better tight fitting of the fast variations of the thickness of the patient body, and it also enables to overcome some generator current (mA) adjustment limitations of the X-ray tube, for example the current (mA) adjustment is currently usually limited to a minimum of 10 mA. Another improvement with the simultaneous use of the voltage (kV) adjustment is that the contrast for local thickness is also optimized to fit the exposure index target.

[0152] In all Figures 10 to 18 On the figure, the scales along the horizontal and vertical axes are in mm, after binning of the scout view image (which means that each 20x20 original square pixel area has been collected into one new pixel), for the scout view, the frontal and lateral images taken after the scout view are not such binned.

[0153] Figures 10 to 13 The frontal and lateral images are shown using a salient point detection step, before and after filtering, to focus on the patient spine (or to focus on the patient spine extending to one of his or her legs) and to exclude other bones, thus enhancing the image contrast of the area of interest (here the patient spine).

[0154] This filtering step aims at selecting only the salient points most likely to be on the axial skeleton, here on the spine, and on the leg extending the spine.

[0155] Figure 10 An example of filtered frontal scout view after the salient point detection step but before the salient point filtering step is shown. This patient body image is plotted with respect to the height Z (height along the vertical scan direction) as a function of the patient width, the patient thickness th being perpendicular to the plane of the figure, the height and the width w being expressed in mm.

[0156] Figure 11 An example of filtered frontal scout view after the salient point detection step and after the salient point filtering step is shown. This patient body image is plotted with respect to the height Z (height along the vertical scan direction) as a function of the patient width, the patient thickness th being perpendicular to the plane of the figure, the height and the width w being expressed in mm.

[0157] The 21 to 25 bones are found in patient 20. Patient 20 is plotted with respect to his or her height Z in mm, and with respect to his or her width also in mm.

[0158] The rules of the frontal image filtering step are as follows:

[0159] For each Z (vertical position) value, the salient point with the largest thickness is selected,

[0160] Even if a metal part corresponds to a salient point with the largest thickness (a metal part corresponding to a very steep change in attenuation or absorption in the vicinity of the patient's body), said metal part is excluded.

[0161] When filtering in the frontal image of patient 20:

[0162] Only a small part of the spine 21, the left leg 23 and the right leg 22 remain,

[0163] While the arms 24, the shoulders 25 and most of the right leg 22 have been filtered and are thus excluded.

[0164] Figure 12 An example of a filtered lateral scout view after the salient point detection step but before the salient point filtering step is shown. This patient body image is plotted with respect to height Z (height along the vertical scanning direction) as a function of patient width, the patient thickness th being perpendicular to the plane of the figure, both height and width w being expressed in mm.

[0165] Figure 13 An example of a filtered lateral scout view after the salient point detection step and after the salient point filtering step is shown. This patient body image is plotted with respect to height Z (height along the vertical scanning direction) as a function of patient width, the patient thickness th being perpendicular to the plane of the figure, both height and width w being expressed in mm.

[0166] The 31 to 35 bones are found in patient 30. Patient 30 is plotted with respect to his or her height Z in mm, and with respect to his or her width also in mm.

[0167] The rules of the lateral image filtering step are as follows:

[0168] For each Z (vertical position) value, the salient point closest to the patient's back is selected (as the patient is looking to the left, this corresponds here to the salient point on the right of the lateral image, but if the patient is looking to the right, it corresponds to the salient point on the left of the lateral image),

[0169] Even if the metal parts correspond to salient points with the largest thicknesses (metal parts corresponding to very steep attenuation or absorption variations in the vicinity of the patient's body), they are excluded,

[0170] Some isolated salient points that are too close to the patient's back, further from the spine than from the lateral image right side, are still excluded, such as some isolated salient points lost in soft tissue areas such as the buttocks or back muscle parts.

[0171] When filtering in the patient 30 lateral image:

[0172] Only the spine 31 and the left leg 32 are kept,

[0173] While the arms 33 and 34 and the upper and lower jaws 25 have been filtered and are therefore excluded.

[0174] The voltage and current intensity adjustments can be calibrated.

[0175] First, images with fixed current values can be obtained, with voltage adjustments made step by step, thus covering the voltage available range (50-140 kV).

[0176] Then, the obtained images are corrected by the calibration software which homogenizes the detector and corrects its non-linearity.

[0177] Then, the signal is measured for each voltage step value, these step values coming from the feedback measurement file of the radiation emitter post-generator.

[0178] Then, the evolution curve f(kV) is determined by interpolation, which gives the signal as a product of the evolution curve f(kV) with mA (and also for a given reference acquisition frequency of the detector), f can for example be expressed as the number of X-ray photons received by each pixel detector on the vertical axis divided by the current and voltage on the horizontal axis, and can also be a straight line when 40-120 photons per mA and 80-130 kV horizontally, and starts to gradually smooth when 10-40 photons per mA and 50-80 kV horizontally.

[0179] It is also possible to correct a specific image of the frontal or lateral:

[0180] The distribution of the driving voltage and current intensity adjustments is calculated from the scout view.

[0181] Then the image is acquired using the calculated adjustment distribution.

[0182] Then, the obtained images are corrected by the calibration software which homogenizes the detector and corrects its non-linearity.

[0183] The feedback measurement file of the radiation emission post-generator is then exploited to identify at each line j of the taken image the value of the voltage kV(j) and of the current mA(j) effectively sent by the generator at this line j.

[0184] The image is then normalized: the signal of each line j of the taken image is divided by the product [f(kV(j))*mA(j)].

[0185] Figure 14 An example of filtered frontal scout view after the deep neural network detection step is shown. This patient body image is plotted with respect to the height Z (height along the vertical scan direction) as a function of the patient width, the patient thickness th being perpendicular to the plane of the plot, both height and width w being expressed in mm.

[0186] A patient 40 is represented on a frontal image, the image having a landmark 41 plotted by a deep neural network.

[0187] Figure 15 An example of filtered lateral scout view after the deep neural network detection step is shown. This patient body image is plotted with respect to the height Z (height along the vertical scan direction) as a function of the patient width, the patient thickness th being perpendicular to the plane of the plot, both height and width w being expressed in mm.

[0188] A patient 50 is represented on a lateral image, the image having a landmark 51 plotted by a deep neural network.

[0189] Figure 16 An example of lateral image with constant drive current intensity and constant drive voltage intensity in constant current and voltage mode is shown. Figure 16 A lateral scan of a mannequin is presented with modulation disabled.

[0190] A patient 60 is represented on a lateral image. This patient body image is plotted with respect to the height Z (height along the vertical scan direction) as a function of the patient width, the patient thickness th being perpendicular to the plane of the plot, both height and width w being expressed in mm. There is a constant drive current intensity 61 along the vertical scan direction Z, for example about 350 mA. There is a constant drive voltage intensity 62 along the vertical scan direction Z, for example about 120 kV. The patient 60 lateral image has a medium quality.

[0191] Figure 17 An example of lateral image with modulated drive current intensity and constant drive voltage intensity in modulated current and constant voltage mode is shown. Figure 17 A lateral scan of a mannequin is presented with modulation enabled but with fixed voltage (kV).

[0192] A patient 70 is represented on a lateral image. This patient body image is plotted with respect to height Z (height along vertical scan direction) as a function of patient width, patient thickness th being perpendicular to the plane of the plot, both height and width w being expressed in mm. There is a driving current intensity adjustment 71 along the vertical scan direction Z, for example varying between about 10 mA and about 350 mA. There is a constant driving voltage intensity 72 along the vertical scan direction Z, for example about 120 kV. The lateral image of patient 70 has a significantly similar quality to the lateral image of patient 60, but is taken at a lower radiation dose and is adjusted to obtain the correct signal level at each position along the vertical scan direction.

[0193] Figure 18 An example of a final lateral image displayed to a radiologist is shown. Figure 18 A lateral scan of a phantom is presented with adjustment for voltage and current enabled.

[0194] A patient 80 is represented on a lateral image. This patient body image is plotted with respect to height Z (height along vertical scan direction) as a function of patient width, patient thickness th being perpendicular to the plane of the plot, both height and width w being expressed in mm. There is a driving current intensity adjustment 81 along the vertical scan direction Z, for example varying between about 10 mA and about 350 mA. There is also a driving voltage intensity adjustment 82 along the vertical scan direction Z, for example varying between about 50 kV and about 120 kV. The lateral image of patient 80 has a significantly better quality than the lateral image of patient 70, and a better quality than the lateral image of patient 60, and the lateral image of patient 80 is taken at a lower radiation dose than the lateral image of patient 60.

[0195] Compared to the image in Figure 16 The white or empty parts of the patient external image correspond to the adjustment applied by the adjustment profile. The image is then corrected by normalizing using the exception signal which is not attenuated to remove such adjustment of the white parts of the image to avoid compromising the radiologist analysis of the image. The demodulated image follows standard contrast enhancement processes before being presented to the radiologist. Figure 17 18 An example of a final lateral image displayed to a radiologist is shown.

[0196] A lateral scan image is presented to a radiologist, for example acquired using voltage and current adjustment including demodulation and contrast enhancement processing. The final lateral image quality is good. Figure 19 Figure 19

[0197] ​​The application has been described with reference to the preferred embodiments. However, many variations are possible within the scope of the application.

Claims

1. A radiological imaging method comprising: two radiation sources, one frontal and one lateral, with mutually orthogonal imaging directions, vertically sliding to perform a vertical scan along a vertical scan direction (Z) of a standing patient (20), wherein the radiological imaging method comprises at least one operating mode in which: a frontal scout view is obtained by performing (1) a preliminary vertical scan of the standing patient (20) along the vertical scan direction (Z) by the frontal radiation source, the frontal scout view is processed (2, 3) to identify a specific skeletal location (21) within the frontal scout view, a drive current intensity of at least the frontal radiation source is adjusted along the vertical scan direction (Z) as a function of patient thickness and the identified specific skeletal location (21) along the vertical scan direction (Z), a drive voltage intensity of the frontal radiation source is adjusted along the vertical scan direction (Z) as a function of patient thickness and the identified specific skeletal location (21) along the vertical scan direction (Z), the drive current intensity and voltage intensity adjustments of the frontal radiation source are performed simultaneously, and / or automatically synchronized, to improve a trade-off between: reducing the overall radiation dose received by the patient (20) during the vertical scan, and, for frontal images, increasing the local image contrast of the identified specific skeletal location (21) at different imaging locations along the vertical scan direction (Z).

2. A radiological imaging method comprising: two radiation sources, one frontal and one lateral, with mutually orthogonal imaging directions, vertically sliding to perform a vertical scan along a vertical scan direction (Z) of a standing patient (30), wherein the radiological imaging method comprises at least one operating mode in which: a lateral scout view is obtained by performing (1) a preliminary vertical scan of the standing patient (30) along the vertical scan direction (Z) by the lateral radiation source, the lateral scout view is processed (2, 3) to identify a specific skeletal location (31) within the lateral scout view, a drive current intensity of at least the lateral radiation source is adjusted along the vertical scan direction (Z) as a function of patient thickness and the identified specific skeletal location (31) along the vertical scan direction (Z), a drive voltage intensity of the lateral radiation source is adjusted along the vertical scan direction (Z) as a function of patient thickness and the identified specific skeletal location (31) along the vertical scan direction (Z), the drive current intensity and voltage intensity adjustments of the lateral radiation source are performed simultaneously, and / or automatically synchronized, to improve a trade-off between: reducing the overall radiation dose received by the patient (30) during the vertical scan, and, for lateral images, increasing the local image contrast of the identified specific skeletal location (31) at different imaging locations along the vertical scan direction (Z).

3. A radiological imaging method comprising: two radiation sources with mutually orthogonal imaging directions, a frontal radiation source and a lateral radiation source, vertically sliding to perform a vertical scan along a vertical scan direction (Z) on a standing patient (20, 30), wherein said radiological imaging method comprises at least one operating mode in which: by performing (1) a preliminary vertical scan along said vertical scan direction by said frontal and lateral radiation sources on a standing patient to obtain frontal and lateral scout views, processing (2, 3) said frontal and lateral scout views to identify specific skeletal locations (21, 31) within said frontal and lateral scout views, adjusting the drive current intensity of said frontal and lateral radiation sources along said vertical scan direction as a function of patient thickness and of said identified specific skeletal locations (21, 31) along said vertical scan direction (Z), adjusting the drive voltage intensity of said frontal and lateral radiation sources along said vertical scan direction as a function of patient thickness and of said identified specific skeletal locations (21, 31) along said vertical scan direction (Z), said drive current intensity and voltage intensity adjustments of said frontal radiation source and said drive current intensity and voltage intensity adjustments of said lateral radiation source are performed simultaneously and / or synchronously and automatically to improve a compromise between: reducing the overall radiation dose received by the patient (20, 30) during said vertical scan, and increasing the local image contrast of said identified specific skeletal locations (21, 31) at different imaging locations along said vertical scan direction (Z) for both frontal and lateral images.

4. Radiological imaging method according to any one of claims 1 to 3, characterized in that, said drive current intensity and voltage intensity adjustments of said frontal radiation source are also performed to reach a certain signal-to-noise ratio value which is constant and common for all said imaging locations along said vertical scan direction (Z) for both said frontal and / or said lateral images.

5. Radiological imaging method according to claim 4, characterized in that, said signal-to-noise ratio value takes two different values for said frontal and lateral images respectively.

6. The radiological imaging method of claim 4, wherein, said signal-to-noise ratio value is constant and predetermined for each different patient (20, 30) organ to be imaged for each of said frontal and / or lateral images.

7. The radiological imaging method according to claim 4, characterized in that: for frontal images of a patient's spine, a standard signal-to-noise ratio value corresponds to a number of X-ray photons received per detector pixel comprised between 50 and 70 and / or the radiological imaging method operator has the possibility to command manually a deviation from this standard value of at least + or - 20% or at least + or - 50%, and / or for lateral images of a patient's spine, a standard signal-to-noise ratio value corresponds to a number of X-ray photons received per detector pixel comprised between 20 and 40 and / or the radiological imaging method operator has the possibility to command manually a deviation from this standard value of at least + or - 20% or at least + or - 50%.

8. Radiological imaging method according to any one of claims 1 to 3, characterized in that, The frontal and / or lateral images are normalized after having undergone at least a first step of increasing the local image contrast at different imaging positions along the vertical scan direction (Z) of the identified specific skeletal positioning (21, 31) by homogenization of the areas located only outside the patient body contour in order to eliminate image artifacts from the driving current intensity and voltage intensity regulation.

9. Radiological imaging method according to claim 8, characterized in that, The frontal and / or lateral images, after normalization, undergo a contrast enhancement step.

10. Radiological imaging method according to any one of claims 1 to 3, wherein the identified specific skeletal positioning (21, 31) excludes metal parts.

11. Radiological imaging method according to any one of claims 1 to 3, characterized in that: - the current intensity and voltage intensity are regulated: - for larger patient thicknesses, the current intensity and voltage intensity are simultaneously increased, - for smaller patient thicknesses, the current intensity and voltage intensity are simultaneously decreased, - the current intensity variation rate is slower than the voltage intensity variation rate.

12. Radiological imaging method according to any one of claims 1 to 3, characterized in that, The driving current intensity regulation is maximized, so that the vertical scan speed is also maximized to a constant value.

13. Radiological imaging method according to any one of claims 1 to 3, characterized in that, The operating mode is manually turned on or off by the radiological imaging method operator.

14. Radiological imaging method according to any one of claims 1 to 3, characterized in that, The operating mode is not manually turned on or off by the radiological imaging method operator.

15. Radiological imaging method according to any one of claims 1 to 3, characterized in that, The driving current intensity regulation rate does not exceed a predetermined threshold of 5 mA per millisecond, or a predetermined threshold of 2 mA per millisecond, or a predetermined threshold of 1 mA per millisecond.

16. The radiological imaging method according to any one of claims 1 to 3, characterized in that, The driving current intensity regulation ranges at least from 20 mA to 300 mA, or from 10 mA to 400 mA.

17. The radiological imaging method according to any one of claims 1 to 3, characterized in that, The voltage intensity regulation ranges at least from 60 kV to 100 kV, or from 50 kV to 120 kV.

18. The radiological imaging method according to any one of claims 1 to 3, characterized in that, The vertical scan speed value ranges at least from 8 cm / s to 20 cm / s, or from 4 cm / s to 30 cm / s.

19. The radiological imaging method according to any one of claims 1 to 3, characterized in that, Before obtaining each of the frontal and lateral images, each of the frontal and / or lateral scout views is obtained by performing a preliminary vertical scan of the standing patient (20, 30) along the vertical scan direction (Z) with a reduced overall radiation dose compared to each of the frontal and lateral images.

20. The radiological imaging method of claim 19, wherein, The reduced overall radiation is less than 10% of the overall radiation dose, or less than 5% of the overall radiation dose.

21. The radiological imaging method according to any one of claims 1 to 3, characterized in that, Pixels in the scout view are grouped together in N x N pixel zones to form imaging zones.

22. The radiological imaging method of claim 19, wherein, Pixels in the scout view are grouped together in at least 10 x 10 pixel zones to form imaging zones.

23. The radiological imaging method according to any one of claims 1 to 3, characterized in that, The scout view is processed (3) to identify salient points which are in turn used to compute a thickness profile and to identify the specific skeletal positioning (21, 31) of the standing patient (20, 30) along the vertical scan direction (Z).

24. The radiological imaging method according to any one of claims 1 to 3, characterized in that, The scout view is processed by a neural network to compute a thickness profile and to identify the specific skeletal positioning (21, 31) of the standing patient (20, 30) along the vertical scan direction (Z).

25. The radiological imaging method according to any one of claims 1 to 3, characterized in that, The 2 radiation sources are vertically slidable in order to perform a vertical scan of the pelvis or spine or spinal column or whole body of the standing patient (20, 30) along the vertical scan direction (Z).

26. The radiological imaging method according to any one of claims 1 to 3, characterized in that, 2 radiation detectors are associated with the 2 radiation sources, the 2 radiation detectors are 2 photon counting detectors (PCD), each photon counting detector is associated with an automatic image processing function that automatically balances image density to enhance image contrast regardless of the radiation dose received on the sensitive surface of the radiation detector.

27. The radiological imaging method according to any of claims 1 to 3, characterized in that, 2 radiation detectors are associated with the 2 radiation sources, the 2 radiation detectors are 2 multi-energy counting detectors, or 2 energy resolving photon counting detectors (ERPCD).

Images