Datortomogrāfijas dozu optimizācija, saglabājot diagnostisku attēla kvalitāti: fantoma pētījums
Notiek ielāde...
Datums
Autori
Journal Title
Journal ISSN
Volume Title
Publisher
Rīgas Stradiņa universitāte
Rīga Stradiņš University
Rīga Stradiņš University
Kopsavilkums
Datortomogrāfijas (DT) izmeklējumus pasaulē izmanto aizvien biežāk. Jonizējošā starojuma ekspozīcijas dēļ, DT izmeklējumi ir saistīti ar kanceroģenēzes risku, kas pieaug, palielinoties saņemtā starojuma dozai.
DT attēlu kvalitāte ir tieši atkarīga no saņemtā starojuma dozas. Atbilstoši aizsardzības pret jonizējošā starojuma iedarbību galvenajam principam ALARA (as low as reasonably achievable), izmeklējuma veikšanai ir jāizmanto tik zema starojuma deva, cik vien saprātīgi ir iespējams. Samazinot starojuma dozu, attēlā palielinās troksnis (attēls kļūst graudains), kas līdz noteiktam dozas pazeminājumam, neskatoties uz attēla graudainību, ļauj tajā saglabāt diagnostisku informāciju. Tādēļ bērnu radiologi jēdziena “perfekts attēls” vietā runā par jēdzienu “diagnostisks attēls”, un bērnu radioloģijā ir ierasts strādāt ar graudainiem jeb “trokšņainiem” attēliem. Attēlu graudainības kompensēšanai lieto dažādas tehnoloģijas. Viens no šādu tehnoloģiju veidiem ir t.s. iteraktīvās rekonstrukcijas tehnikas (IRT), kas ļauj optimizēt attēlu kvalitāti, virtuāli “nogludinot” ar zemāku jonizējošā starojuma dozu iegūtu graudainu attēlu. Tādējādi attēls kļūst tīkamāks radiologa acij un, atbilstoši optimizējot DT izmeklējuma skenēšanas protokolus, iespējams samazināt izmeklējumā pielietoto jonizējošā starojuma dozu.
Pētījuma mērķis bija atrast optimālāko IRT galvas DT izmeklējumiem pediatrijā, kas ļautu samazināt izmeklējumā pielietoto jonizējošā starojuma dozu, saglabājot diagnostisku attēla kvalitāti.
Lai to izdarītu, tika veikti 10 galvas DT izmeklējumi ar 64 slāņu DT iekārtu Philips Ingenuity, skenējot PIXY antropomorfo testa fantomu, katru reizi mainot ekspozīcijas parametrus (kV un mAs), kā rezultātā DT dozas indekss (CTDIvol - volume computed tomography dose index) katram izmeklējumam atšķīrās (2.3-35.4 mGy). Katrai izmeklējumu sērijai tika veiktas piecas rekonstrukcijas ar Philips IRT algoritmiem iDose 1-5 mīksto audu un kaulu logos. Katrā rekonstrukcijas sērijā tika izvēlēti trīs slāņi (katrai skenēšanas reizei vieni un tie paši), kuros sertificēts radiologs veica 15 radiodensitātes trokšņa (SD – Standard Deviation) mērījumus. Lai kvantitatīvi izvērtētu attēlu kvalitāti, tika aprēķinātas densitātes un mediānās trokšņa vērtības, un attiecīgā signāla-trokšņa attiecība (SNR - signal-noise ratio). Lai spriestu par labāko IRT pielietošanai praksē, ņēma vērā gan attēla kvalitāti, gan CTDIvol.
Rekonstruētajām attēlu sērijām galvas mīkstajiem audiem iDose 5 vienmēr nodrošināja vismazāko attēla troksni un augstāko SNR, salīdzinājumā ar pārējiem rekonstrukcijas algoritmiem pie tās pašas dozas vērtības, kas deva iemeslu secināt, ka salīdzināmos attēla kvalitātes līmeņos CTDIvol var samazināt līdz pat 40%. CTDIvol pārsniedzot 30 mGy, SNR samazinās un attēla troksnis pieaug.
Vērtējot rekonstruētās attēlu sērijas galvaskausa kauliem, neviena IRT nebija pārliecinoši pārāka par citām un attēla kvalitāte bija tieši saistīta ar izmeklējumā pielietoto starojuma dozu. CTD
Computed tomography (CT) imaging is being used more frequently worldwide. Due to ionising radiation exposure, CT examinations are associated with a risk of oncogenesis depending on the radiation dose received. Unfortunately, CT image quality depends on the dose of ionising radiation used for the examination. According to the main principle of radiation protection ALARA (as low as reasonably achievable), it is advised to use the lowest possible dose for the radiological examination. By lowering the ionising radiation dose used for the examination, image noise increases (image becomes grainy). Until a certain point of dose reduction, even though the image is “noisy”, it is possible to maintain diagnostic informativity of the image. The goal of examination in pediatric radiology is to acquire “diagnostic image” rather than “perfect image”, so in pediatric radiology it is common to work with “noisy” images. Different technologies in pediatric CT imaging are used to improve image quality, such as iterative recontruction techniques (IRT), which may compensate quality of “noisy” CT images obtained using lower radiation dose. IRT make the appearance of grainy images acquired with lower dose of ionising radiation look smooth, so the image becomes more pleasant for the radiologist to work with. By optimising CT scanning protocols and IRT use, it is possible to lower the dose used in examination. The aim of the study was to find the most optimal IRT to perform head CT examinations with lower radiation doses, maintaining diagnostic informativity of the image. In this research 10 head CT image series using a PIXY antropomorhic human training phantom were scanned with a 64-slice CT machine with dose affecting parameters - kV and mAs – changed every time. CT dose index (CTDIvol - volume computed tomography dose index) for each scan differed (2.3-35.4 mGy). For every scan, five reconstructions were performed using IRT algorithms (Philips iDose, from 1 to 5), for the bone and brain windows. One radiologist performed 15 measurements of radiodensity and image noise in 3 same exact slices of every reconstructed series. The median density and noise values, along with the corresponding signal-noise ratio (SNR - signal-noise ratio) were calculated to estimate the most optimal IRT providing the best image. Both image quality and dose (CTDIvol, mGy) used in examination were taken into account to determine the best IRT for use in practice. For the brain, reconstructions with iDose 5 always provided the least noise and the highest SNR in comparison with other reconstructions made with the same dose. At comparable image quality levels, CTDIvol can be reduced by 40%. When CTDIvol is over 30 mGy, SNR drops and noise slowly increases. For bones, no IRT was convincingly better than any other; image quality is closely correlated with ionising radiation dose used for the examination. When the CTDIvol is more than 20 mGy, there is a slight superiority for iDose 4. Therefore in
Computed tomography (CT) imaging is being used more frequently worldwide. Due to ionising radiation exposure, CT examinations are associated with a risk of oncogenesis depending on the radiation dose received. Unfortunately, CT image quality depends on the dose of ionising radiation used for the examination. According to the main principle of radiation protection ALARA (as low as reasonably achievable), it is advised to use the lowest possible dose for the radiological examination. By lowering the ionising radiation dose used for the examination, image noise increases (image becomes grainy). Until a certain point of dose reduction, even though the image is “noisy”, it is possible to maintain diagnostic informativity of the image. The goal of examination in pediatric radiology is to acquire “diagnostic image” rather than “perfect image”, so in pediatric radiology it is common to work with “noisy” images. Different technologies in pediatric CT imaging are used to improve image quality, such as iterative recontruction techniques (IRT), which may compensate quality of “noisy” CT images obtained using lower radiation dose. IRT make the appearance of grainy images acquired with lower dose of ionising radiation look smooth, so the image becomes more pleasant for the radiologist to work with. By optimising CT scanning protocols and IRT use, it is possible to lower the dose used in examination. The aim of the study was to find the most optimal IRT to perform head CT examinations with lower radiation doses, maintaining diagnostic informativity of the image. In this research 10 head CT image series using a PIXY antropomorhic human training phantom were scanned with a 64-slice CT machine with dose affecting parameters - kV and mAs – changed every time. CT dose index (CTDIvol - volume computed tomography dose index) for each scan differed (2.3-35.4 mGy). For every scan, five reconstructions were performed using IRT algorithms (Philips iDose, from 1 to 5), for the bone and brain windows. One radiologist performed 15 measurements of radiodensity and image noise in 3 same exact slices of every reconstructed series. The median density and noise values, along with the corresponding signal-noise ratio (SNR - signal-noise ratio) were calculated to estimate the most optimal IRT providing the best image. Both image quality and dose (CTDIvol, mGy) used in examination were taken into account to determine the best IRT for use in practice. For the brain, reconstructions with iDose 5 always provided the least noise and the highest SNR in comparison with other reconstructions made with the same dose. At comparable image quality levels, CTDIvol can be reduced by 40%. When CTDIvol is over 30 mGy, SNR drops and noise slowly increases. For bones, no IRT was convincingly better than any other; image quality is closely correlated with ionising radiation dose used for the examination. When the CTDIvol is more than 20 mGy, there is a slight superiority for iDose 4. Therefore in
Description
Pediatrija
Pediatrics
Veselības aprūpe
Health Care
Pediatrics
Veselības aprūpe
Health Care