PT. SIEMENS INDONESIA (c) Daniel Kartawiguna, 2006 [email protected]1 1 Always Thinking Ahead. Computed Tomography Scanner (CT-Scan) (Pemindai Tomografi Komputer) Oleh: Daniel Kartawiguna, ST., MM., M.Acc. PT. SIEMENS INDONESIA – Medical Solutions [email protected]DISAJIKAN DALAM RANGKA PELATIHAN CT SCAN DOSEN ATEM JAKARTA, 19 JUNI 2006 2 Always Thinking Ahead. MATERI PELATIHAN I. Dasar Tomografi Komputer II. Sistem Tomografi Komputer Spiral III. Sistem Tomografi Komputer Multi Irisan (MSCT) IV.Pemeliharaan dan Perbaikkan V. Aplikasi Klinis MSCT
64
Embed
1 Computed Tomography Scanner (CT-Scan) Pemindai … · PT. SIEMENS INDONESIA (c) Daniel Kartawiguna, 2006 [email protected] 3 5 Always Thinking Ahead. POKOK BAHASAN 1.Tomografi
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Perkembangan kualitas gambar CTDari tahun 1972 - 2000
SIRETOM (1974)
SOMATOM Plus 4 UFC(1996)
12
Always Thinking Ahead.
12
Always Thinking Ahead.
Ringkasan Sejarah PerkembanganTomografi Komputer1917 J.H. Radon: Transformasi Radon, gambar dari
obyek yang tidak diketahui dapat digambarkan dari proyeksinya.
1963 A.M. Cormack: mengembangkan teknik untukmenentukan distribusi penyerapan tubuh manusia.
1972 G.N. Hounsfield dan J. Ambrose: menghasilkan gambar CT pertama kali untuk keperluan klinis.
1974 60 unit CT terpasang untuk pemeriksaan kepala1975 First whole body scanner in clinical use1979 Hounsfield dan Cormack dianugerahi hadiah Nobel1989 Spiral CT1998 Multislice CT2000 > 30000 clinical CT installations
• One could solve more than 260.000 unknown µ by measuring the attenuation in about 1.400 readingswith about 700 detector channels.Disadventage: calculation can be started only after a full
revolution, spiral scan wouldn`t be time effective
or
• We could simply add the projections by smearing them back upon each other across the image.
Overview over a CT systemA modern CT can be divided into three parts:
System Control
Runs the Operating System software (VMS, UNIX or Win NT.)
Runs the Application software to provide a user interface (SOMARIS).
Supports Scan Control by distribution of scan parameters and instructions.
Scan Control
Usually microprocessor based. Directly controls the scanner hardware via firmware routines stored in PROM. It is responsible for the safety of the patient, operator and system hardware. It distributes parameters from the host to the other functional groups in the system. It serves as the interface for all communication between “system control” and the gantry.
Image Processing
This is a specialized array of computers that are responsible for the reception of data from “scan control” and the subsequent reconstruction of an image.
The signals from the detector have a very wide range. To cover the whole dynamic range of input
signals, a
„Programmable Gain Amplifier“ (PGA) is used (also called FPA, floating point amplifier).
PGA: The PPA is an amplifier which selects its gain automatically. The selected gain can be 1, 8 or 64. The gain used is indicated by the two bits called `PGA Bits´.
PGA Decoding:In order to calculate the actual attenuation, the PGA bits are decoded in the SMI. This is done in the preprocessing step „PGA decoding“ (also called „FPA decoding“).
Amplification 64:If the signal from the detector was very small (i.e., high absorption in the scanfield), the amplifier will have used a factor of 64. The resulting data in the SMI will be the 14 bit from theADC, preceeded by many zeroes, in other words, a rather small numerical value.
Amplification 8:If the amplification was 8, the signal was larger. In the SMI, the 14 bit are shifted 3 bitto the left, equaling a multiplication by 8, or a larger numerical value.
Amplifiaction 1: If the signal from the detector was large (e.g., only air in the scanfield), thePGA will have used an amplification of 1. This will result in a large number in the SMI, because the 14 bit are shifted 6 bit to the left, equaling a multiplication by 64 or a rather large numerical value.
In the DMS, ADCs are used that can´t measure negative voltages. This would falsify the measurement, if very small detector signals ( = high absorptions ) have to be measured. To avoid this, in the DMS an offset voltage is added to the signal, the signal is measured, and in the SMI the offset signal subtracted again, leaving the true value only.
Offsets are channel specific: Because the analogue offset may be slightlydifferent for each ADC, or, to be precise, even for every integrator board channel, the actual offset has to be measured prior to the scan for every channel.
Offset measurement:With each scan start, a measurement is started without X-ray and the data are stored in the image processor as offset data.
Each detector has a different sensitivity. This can vary with the time and must be compensated.
The channel specific sensitivity differences are compensated by calibration. Technically, that means taking an air scan and then subtracting the channel values obtained in air from the normalized channel values.
Then pre- processing step calibration requires the base calibration tables, that are measured during the last calibration.
During the last tune up, the differences of each combination of kV, mA, slice width etc. were measured, so that only the base calibration is required on a daily basis.
The other calibration tables for different settings of kV, mA etc are calculated from the difference tables and the base calibration table.
Tubes generate polychromatic radiation, i.e.different wavelength are contained in the spectrum.
Just as with visible light, the higher energies or shorter wavelength can penetrate the objects better than the softer part of the spectrum.
Beam hardening causes in homogenous objects (e.g. a water phantom) an inhomogeneity. That means, the CT values in the center are different from the outer values.
The correction is done by taking data of a reference phantom (mostly a 20cm water phantom) and the correction data are used during pre- processing step „beam hardening“ for the correction of the scan data.
Just like „Calibration“, the pre-processing step ´Channel Correction´compensates for sensitivity differences of the detector. The difference is that the Channel correction compensates for nonlinearities in the area of attenuated radiation, i.e., with an object is in the scan field.
Parameter:The parameters which determine the detected radiation energy are:
Tube VoltageSlice ThicknessObject Attenuation (Head or body)
Correction tables:The sensitivity compensation is done with values which are determined
• Gerakan translasi dan rotasi• Berkas sinar-X berbentuk pensil (pencil beam)• Geometri berkas sinar paralel.• FOV (field of view) 24 cm.• Menggunakan 2 buah detektor sehingga sekali scan dapat
menghasilkan 2 irisan.• 160 berkas paralel/proyeksi.• 180 proyeksi dengan interval 1 derajat.• Detektor tidak dapat mendeteksi perbedaan intensitas sinar-X
yang sangat besar, oleh karena itu kepala yang diperiksa harusdikelilingi oleh kantong berisi air.
• Kristal NaI yang digunakan sebagai detektor memiliki waktu“afterglow” yang nyata.
• Keuntungan: pengaruh hamburan radiasi pada detektorditiadakan karena berkas sinar-X yang berbentuk pensil.
• Menggunakan 30 linear array detector.• Kerugian: adanya pengaruh radiasi hamburan dan
meningkatnya intensitas kearah tepi dari berkas sinar-X yang berbentuk kipas. Hal ini diatasi dengan penambahanfilter dasi kupu-kupu pada jendela tabung sinar-X.
• Keuntungan: waktu scan lebih singkat, yaitu antara 18 hingga 30 detik/irisan.
• Konfigurasi rotasi/rotasi.• Berkas sinar-x berbentuk kipas (fan beam).• Menggunakan detektor array.• Waktu scan 1 detik.• Kekurangan: kemungkinan terjadinya ring artefact karena
• Tabung sinar-X berputar dan detektor diam.• Detektor tersusun melingkar berbentuk lingkaran. • Sekitar 8000 buah detektor diperlukan.• Waktu scan 1 detik.• Kerugian: harga mahal, dosis radiasi pada pasien lebih
tinggi.• Keuntungan: tidak terjadi ring artefact.• Masalah: jarak antara tabung sinar-X dan elemen detektor
tidak semuanya sama -> diatasi dengan kalibrasi dannormalisasi saat scan.
Electron Beam Computed Tomography Scanner, waktu scan 50msDigunakan untuk scan jantung.
74
Always Thinking Ahead.
74
Always Thinking Ahead.
EBT Electron Beam TechniqueIn a conventional CT scanner, the x-ray tube moves round the patient, the x-ray beam is attenuated by the patient and the differences of the emergent beam are registered by a bank of detectors. This information is digitised and converted into a cross-sectional image. The exposure time for each slice is restricted by the time it takes physically to move the x-ray tube.
The Imatron CT Scanner has no x-ray tube but an electron gun which produces a beam of electrons at 130kV which is accelerated along a tube. The beam of electrons is focused, by an electro-magnetic coil, onto a small focal spot on a tungsten ring. This target area is then moved along the ring.
The x-rays are generated by this deceleration process and collimators define them into a fan-beam which ‘sweeps’ the patient. The x-ray density differences are registered by a bank of solid-state detectors and the output is digitised by the Data Acquisition System. Data is stored in the bulk memoryand the transferred to disc storage which is then converted into a cross-sectional image. There are no moving parts therefore exposure times can be reduced to 50ms per slice. Up to 17 slices per second can be taken, enabling the CT Scanner Unit to image moving structures such as the heart.
• Teknologi Slip-Ring sekitar tahun 1990-an.• Akuisisi data dilakukan dengan meja yang bergerak
sementara tabung sinar-X berputar sehingga gerakantabung sinar-X membentuk pola spiral terhadap pasiensaat dilakukan akuisisi data.
• Diterapkan pada konfigurasi rancangan CT generasi ke-3 dan ke-4.
78
Always Thinking Ahead.
78
Always Thinking Ahead.
Teknologi Slip-ringAliran listrik disalurkan melalui sejumlah konduktorberbentuk cincin yang disusun secara paralel, sehingga tidak menggunakan kabel lagi.
Rotasi Gantry yang kontinuMerupakan syarat untuk CT-Scan spiral/helical
Increment > Slice ThicknessIncrement > Slice ThicknessSelaSela antaraantara 2 2 potonganpotonganImages are further apartImages are further apartJumlahJumlah gambargambar yang yang lebihlebih sedikitsedikit..
Scan the whole region of Scan the whole region of interest in one breath hold interest in one breath hold
Reconstruction of overlapping Reconstruction of overlapping images without additional doseimages without additional dose
Retrospective reconstruction Retrospective reconstruction of slices in arbitrary position of slices in arbitrary position within the scanned volumewithin the scanned volume
No gaps since radiation alwaysNo gaps since radiation alwaystransmits the whole volumetransmits the whole volume
Rekonstruksi secara tumpang tindihakan memberikan resolusi sumbu-z yang lebih baik.
CT Scan SpiralCT Scan Konvensional
Akuisisi data tanpa celah dlm satu kali tahan napas.Waktu scan yang cepat terhadap valume yang besar.
Mengoptimalkan penggunaan media kontras.Dapat dilakukan rekonstruksi dengan menentukannilai increment yang dikehendaki.
88
Always Thinking Ahead.
88
Always Thinking Ahead.
Kapan pemeriksaan Spiral CT dilakukan?
Pemeriksaan dengan kontras.Pemeriksaan seluruh bagian tubuh. Pemeriksaan anak dan pasien trauma, yang memerlukan pemeriksaan secara cepat.Pemeriksaan anatomi yang panjang.
Spiral selalu digunakan untuk melakukan3D postprocessing (contoh CTA) !
Generasi ke-7: Multi Detector Array CT• Tabung sinar-X memiliki kapasitas panas yang terbatas.
Hanya 1% dari energi yang dikonversi menjadi sinar-X.• Dengan detektor multi array maka apabila kolimator dibuka
lebih lebar akan diperoleh data proyeksi lebih banyak. Dengan demikian maka penggunaan energi sinar-X lebihefisien.
• Pada detektor array tunggal, apabila kolimator dibuka lebihlebar maka akan diperoleh irisan yang lebih tebal yang akan mengurai resolusi spatial.
• Masalah: cone beam artefact.• Keuntungan: meningkatkan waktu scan hingga 0,33 detik,
resolusi dalam arah sumbu-Z hingga < 0,4 mm, dan dosisradiasi lebih rendah.
90
Always Thinking Ahead.
90
Always Thinking Ahead.
Pembentukan Gambar - IrisanSinar-X dilewatkan pada kolimator sehinggahanya menembus bagian potongan aksial dariobyek, yang disebut dengan irisan atau "slice"
Imagine a line emanating from the top center of one’s head and extending through the body toward the feet. This is known as the “long axis” of the body (like a chicken on a rotisserie). When speaking of scanner geometry, the long axis of the patient is aligned to the “Z” axis of the gantry. It is shown as going into the gantry, the same direction as table feed.The picture above also shows that the X-ray source and detector system rotate around this same Z axis. The slice shown then represents a “transaxial” section, that is perpendicular to the long axis.
A voxel is an “intangible object”. The operator of the CT has some control over its size, yet never really thinks about. The width of a single detector element plays a primary role in setting the “face” dimensions (length and width) of a voxel. But a voxel is three dimensional, so it also has depth which runs in the direction of the “Z” plane. This dimension is determined by the thickness of the chosen slice - typically a whole number between 1 to 10 mm. This is selected by the operator. In theory, we can calculate the attenuation of the X-ray beam by a single voxel. This is carried out by making many measurements at different angular positions around the object. The “Image Processor” will take this information and process it further until it has a single numerical value that represents the density of a tiny piece of the object scanned. It then assigns a luminance value and position of where this data is to be deposited in an “image matrix”.
• We can change the appearance of the image by varying thewindow width and level.
• This can spread a small range of CT numbers over a larger range of grey scale values.
• This makes it easy to detect very small changes in CT number.
CT number flexibility
104
Always Thinking Ahead.
104
Always Thinking Ahead.
Pembentukan Gambar - “Matrix”
35 36 3934 3331 34 33 35 3231 80 85 9078
Nilai koefisien pelemahan radiasi diukur kemudiandikodekan dan ditransfer ke komputer.Oleh komputer akan ditampilkan dalam gambar 2 dimensi yang disebut dengan matriks.
This illustration implies the metamorphosis of a voxelinto a pixel. It starts out as a three dimensional volume of matter. We measure it’s ability to attenuate X-ray. This information then is sent to a special computer where it is processed along with other data and becomes a “number” that will specify a “light intensity”or luminance value on a CRT display. The area that is filled by this “light” is known as a pixel and it has only two dimensions - length and width. A group of pixels placed in a specific order, side by side, result in the image that we see it displayed on the monitor.
106
Always Thinking Ahead.
106
Always Thinking Ahead.
Pembentukan Gambar - A>D>A*
Nilai masing2 elemn matriks dikonversikan kedalamgambar hitam putih sesuai dengan grey scale-nya.
Here the process continues, hiding the magic of “image reconstruction”, but revealing the desired end result - an image.
This slide also illustrates the conversion process that takes place, beginning with the acquisition of data in the gantry, where it is an analog signal. The “data” measured can be very small in amplitude. In order to maintain it’s integrity, it is converted to a digital value before it is sent over to the image processor (SMI5). The SMI5 uses this digital data to calculate an image. But before an image can be displayed on a monitor, it is necessary to convert the data back into analog form.
The image processor writes all backprojected data into a 5122 matrix, which is the so called reconstructed matrix.
It is on most CT- system a 5122 matrix.
The pixel size, mentioned on previous page, is a function of the reconstructed matrix.
The imager, responsible for windowing and calculation of HU values, is supplied from the backprojector with the 5122 reconstructed matrix.
A linear interpolation from 5122 to 10242 pixels is done, before the digital data of the matrix are converted to analog and connected as RGB to the monitor.
Image Display - WindowingClose the window - I can’t see anything...
The human eye can only resolve about 40 shades of gray at best. The human body is made up of tissue containing mostly water, but also calcium, phosphorous, as well as other minerals. However, all soft tissue is not the same, even though it consists of the same basic building blocks. There is a difference in density between muscle and the liver, for example. The same goes for bone - you have a variety of bone density in your body from the cartilage in your nose to the calcium rich bone of your pelvis. All of these organs attenuate X-rays differently, so they appear to us as different gray levels in the reconstructed image. Remember, each shade or intensity value correlates to how effectively an object attenuates X-rays. Bright white means a very effective attenuater, such as dense bone. Black means the opposite or virtually no attenuation, such as air.
Our image display system is designed to display no more than 256 gray shades at one time, but the HU scale extends from -1000 to 3096 or 4096 values in total! You could say that each value corresponds to a unique intensity or “gray level”. When a doctor is looking for details in the liver, he may be looking for small density changes. In order to see these small changes, we use a device called “windowing”. Since the liver only represents a very small portion of the HU scale, we choose a “center value” that corresponds to about the mean value of that organ and “open” the window just enough to see the desired detail. The maximum number of gray shades will not exceed 256 regardless of the window setting!
116
Always Thinking Ahead.
116
Always Thinking Ahead.
Pengaruh Windowing pada gambarNilai densitas CT didefinisikan berkisar antara -1000 hingga +3000, tetapi mata manusia paling baik hanya dapatmembedakan 30 - 40 tingkat terang gelap saja.
Dengan demikian pengatruan pilihan windwosharus disesuaikan dengan struktur yang akandilihat.
Secara normal bilangan CT yang dapat diukur berkisar -1024 sampai +3071, tetapi dengan SOMATOM Plus 4 dapatdiperluas (x10) dari -10240 sampai +30710 untuk dapatmenampilkan benda logam. Sehingga dimungkinkan untukmenampilkan bilangan CT yang nyata tidak tergantung dimanadan bagaimana window diposisikan.
Post operative femuralhead replacement – nilaibilangan CT = 6000 HU
122
Always Thinking Ahead.
122
Always Thinking Ahead.
Extended CT ScaleWhat’s the real value?
Besides just looking at an image to gain information, doctors will often request a quantitative measurement of an area of interest on the image. This can be performed by drawing an “ROI” (region of interest) over the object to be measured. The computer software then evaluates all of the data confined to that area and reports several statistics about it. The key one is called the mean value and it is expressed as HU’s. This is a representation of the X-Ray absorption characteristic of that object expressed in a numerical value. Remember, we normally only have about 4071 values to represent all objects in the human body. Also realize that anything very dense is displayed as white and if the density of that object is beyond 3071 HU - it is still going to be displayed at the maximum brightness level. In addition, if a quantitative measurement is taken - the result will be false, because the scale ends at 3071! This is where the value of being able to multiply the normal CT scale by a factor of ten comes into play. We now have over 30,000 values to choose from making very high density objects visible and measurable.
This feature of expanding the Hounsfield Unit scale is also available on the Somatom AR product line.
Note:
Extended windowing is only useful if the data is there to begin with. Depending on scanning parameters chosen and limitation of the scanning equipment, it may not be possible to accurately represent extremely dense objects.