Advanced

Optimization of Acquisition Time for 177- Lu SPECT with Two Energy Windows

Mörnsjö Centofanti, Fanny (2019) MSFT01 20191
Medical Physics Programme
Abstract
Background & Aim: The imaging process following the treatment of 177Lu-DOTATATE is in some medical clinics demanding on both patients and the clinic. Usually a number of images are acquired, with different time intervals to the time of injection. To facilitate for the patient and create the opportu- nity to treat more patients it is desirable to simplify the imaging process. At Skåne University hospital (SUS) the number of imaging sessions has been reduced from four to one, acquired 96 hours p.i. Image acquisition is done with SPECT imaging, which is used for determination of absorbed dose. The main objective is a SPECT image of the kidneys, but there is also a great interest to map out the absorbed dose to tumours. In order to do so a two... (More)
Background & Aim: The imaging process following the treatment of 177Lu-DOTATATE is in some medical clinics demanding on both patients and the clinic. Usually a number of images are acquired, with different time intervals to the time of injection. To facilitate for the patient and create the opportu- nity to treat more patients it is desirable to simplify the imaging process. At Skåne University hospital (SUS) the number of imaging sessions has been reduced from four to one, acquired 96 hours p.i. Image acquisition is done with SPECT imaging, which is used for determination of absorbed dose. The main objective is a SPECT image of the kidneys, but there is also a great interest to map out the absorbed dose to tumours. In order to do so a two FOV SPECT image has to be acquired. Thus the total acquisition time becomes very long and patient movements has shown to be a great difficulty when matching the two FOV together.

At SUS image acquisition is made with one energy window around the 208 keV photopeak and by adding a second one around the 113 keV photopeak, the acquisition time can hopefully be reduced. The determination of absorbed dose is based on a mean value of the activity concentration inside a VOI de- fined from CT image. Thus the task is to lower the acquisition time but still maintain a mean value with low uncertainty. This results in the question formulation: For SPECT imaging, of patients treated with 177Lu-DOTATATE, how much can the acquisition time be reduced when collecting information from 113 keV photopeak and 208 keV photopeak?

Method: SPECT projection data of a cylindrical phantom, within it 6 spheres of different sizes, were simulated with Monte Carlo code SIMIND. Projection sets with activity in the spheres and activity in the background were simulated separately and added together in a later step. The reason for the separate simulation was to easily be able to modify the activity concentration in the spheres and background inde- pendent of each other. Different number of counts in the projection data can be a result of either different activity concentrations in source regions or due to different acquisition times. A quantity, named ”dis- integration concentration value”, was defined as the number of radioactive decay during the acquisition time, per unit volume, to be able to change the acquisition time in the projection data. The disintegration concentration in the spheres was changed by multiplying with so called Disintegration Concentration value (DC-value) [MBq s/ml], derived from activity concentration multiplied with an acquisition time. Thus a higher DC-value is derived from a longer acquisition time, therefore the spheres obtained a higher disintegration concentration. A set of 10 projection sets were created, all multiplied with different DC- values, i.e. different time per projection. The projection data of the sphere and background were added together and four different contrasts of disintegration concentration, between the spheres and the back- ground, were evaluated for all 10 DC-values. For every projection set 50 noise realisation were created. Reconstruction was done with OS-EM algorithm, using 6 subsets and 40 iterations. Projection data for 113 keV and 208 keV were simulated separately and data from the reconstructed images were later added together. Masks of the spheres was created to extract data from the reconstructed images. The data were evaluated by the recovery of the activity concentration and CV, where the standard deviation, of the 50 mean values from the 50 noise realisations, was divided with both the measured disintegration concentration and the true disintegration concentration, respectively.

Result & Conclusion: Image acquisition with two energy windows, one around the 113 keV photo- peak and one around 208 keV photopeak, results in time per projections according to DC-value=22.50 MBq s/ml. For tumours it indicates a time per projection of 22.5 seconds but for kidneys the time per projection would have to be 225 seconds, due to its lower activity concentration 96 hours p.i. compared to the activity concentration in tumours at that time. At SUS the time per projection is 45 seconds, meaning that for image acquisition of tumours the time per projection can be shortened by a factor 2. However for image acquisition of kidneys the time per projection needs to be prolonged by a factor 5 compared to today’s image protocol. (Less)
Popular Abstract (Swedish)
Ur ett samarbete mellan biokemi och strålningsfysik har den nuklearmedicinska grenen växt fram. Genom kemiska reaktioner kan man binda radioaktiva atomkärnor till ämnen som i människokroppen har en mål- sökande egenskap gentemot specifika delar i kroppen. På så sätt riktas ett läkemedel mot ett önskat område, exempelvis ett organ eller receptorer som som uttrycks på bland annant tumörceller.
När den radioaktiva atomkärnan sönderfaller och strålning emitteras skadas tumörvävnaden och tumören kan hämmas. Ur ett diagnostiskt syfte önskar man få en visuell bild över hur läkemedlet har fördelat sig i kroppen. Detta åstadkommer man med hjälp av en Single Photon Emission Computed Tomographic sys- tem, så kallad SPECT. Systemet använder sig av... (More)
Ur ett samarbete mellan biokemi och strålningsfysik har den nuklearmedicinska grenen växt fram. Genom kemiska reaktioner kan man binda radioaktiva atomkärnor till ämnen som i människokroppen har en mål- sökande egenskap gentemot specifika delar i kroppen. På så sätt riktas ett läkemedel mot ett önskat område, exempelvis ett organ eller receptorer som som uttrycks på bland annant tumörceller.
När den radioaktiva atomkärnan sönderfaller och strålning emitteras skadas tumörvävnaden och tumören kan hämmas. Ur ett diagnostiskt syfte önskar man få en visuell bild över hur läkemedlet har fördelat sig i kroppen. Detta åstadkommer man med hjälp av en Single Photon Emission Computed Tomographic sys- tem, så kallad SPECT. Systemet använder sig av den energi, i form av fotoner, som frisläpps vid sönderfallet av den radioaktiva atomkärnan för att skapa bilden. För att erhålla en tillräckligt bra bild för det avsedda ändamålet måste SPECT-systemet utsättas för en tillräcklig mängd fotoner. Det innebär att bildtagningen tar en viss tid, i de flesta fall omkring 30 minuter. Dagens bildtagning är mycket krävande för patienterna, då de måste ligga helt stilla under hela bildtagningen och i samband med att antalet patienter ökar sätter bildtagningstiden en gräns för hur många patienter sjukhuset kan ta emot.

I detta arbete är det en typ utav nuklearmedicinsk behandling som är i centrum, nämligen behandling med 177Lu-DOTATATE. Den radioaktiva atomkärnan förkortas 177Lu, som binds till den receptorsökande pep- tiden, TATE, med hjälp av ett kelat kallat DOTA. Läkemedlet används för behandling av neuroendokrina tumörer, vilket är ett samlingsnamn för tumörer som härstammar från neuroendrokrina celler. På dessa tumörer sitter somostatinreceptorer som peptiden TATE binder till. När 177Lu sönderfaller kommer strål- ningen, kallat betastrålning, ge upphov till den terapeutiska effekten med behandlingen och strålningen i form att fotoner användas som informationsbärare för SPECT-systemet att skapa en bild. Genom bilden kan man se hur läkemedlet har fördelat sig i kroppen och man vill kunna mäta ståldosen som njurar och tumörer erhåller. Framförallt är stråldosen till njurarna av intresse, då det sätter gränsen för hur många gånger be- handlingen kan ges till en patient.

Då 177Lu sönderfaller kan fotonerna erhålla olika energier. Vilka energierna fotonerna får beskrivs av en statistisk fördelning, där vissa fotonenergier är vanligare än andra. Vid dagens bildtagning, på Skånes Uni- versitetssjukhus (SUS), samlar man endast in information från en av 177Lu möjliga fotoner, nämligen den mest förekommande fotonen med energin 208 keV. Dock finns det även en annan relativt vanligt förkom- mande foton vid 177Lu-sönderfall, en foton med energin 113 keV. Genom att samla in information från båda dessa fotoner är förhoppningen att en fullgod bild kommer uppnås under en kortare bildtagningstid.

I den insamlade signalen finns, tillsammans med information till bilden, statiska fluktuationer, även kallat brus, tillsammans med den önskade signalen. Vid längre insamlingstider kommer den önskade signalen vara högre i förhållande till bruset, men i samband med att insamlingstiden förkortas kommer skillnaden mellan önskad signal och brus minska. Det innebär att vid för korta insamlingstider kommer den önskade signalen inte synas i bilden, då den drunknar i bruset.
Genom att immitera bildtagningsprocessen med en dator kan man undersöka hur ändringar av olika parametrar påverkar bruset och slutresultatet av bilden. I arbetet ändras parametern bildtagningstid, för att avgöra hur mycket den kan förkortas då man samlar in information från båda fotonrena för att skapa bilden. Målet är att på så kort bildtagning som möjligt erhålla en bild där stråldosen till njurar och tumörer kan mätas med stor säkerhet.

Resultaten i detta arbete visar att ur perspektivet av bildtagning av tumörer, kan bildtagningstiden halveras jämt emot dagens bildtagningstid. Då det framförallt är njurarna man är intresserad av är det den bild- tagningstid som njurarna kräver, för en säker uppskattning av stråldosen, som sätter gränsen för hur kort bildtagningen kan bli. Vid bildtagningen är koncentrationen av läkemedel generellt är lägre i njurarna i förhållande till tumörerna, vilket innebär att längre bildtagning krävs för att kunna uppskatta stråldosen till njurarna. Ur perspektivet av njurarna visar resultatet att dagens bildtagningstid behöver förlängas med en faktor 5 för att kunna mäta stråldosen med stor säkerhet. (Less)
Please use this url to cite or link to this publication:
author
Mörnsjö Centofanti, Fanny
supervisor
organization
course
MSFT01 20191
year
type
H2 - Master's Degree (Two Years)
subject
language
English
id
8999119
date added to LUP
2020-01-07 09:32:39
date last changed
2020-01-07 09:32:39
@misc{8999119,
  abstract     = {Background & Aim: The imaging process following the treatment of 177Lu-DOTATATE is in some medical clinics demanding on both patients and the clinic. Usually a number of images are acquired, with different time intervals to the time of injection. To facilitate for the patient and create the opportu- nity to treat more patients it is desirable to simplify the imaging process. At Skåne University hospital (SUS) the number of imaging sessions has been reduced from four to one, acquired 96 hours p.i. Image acquisition is done with SPECT imaging, which is used for determination of absorbed dose. The main objective is a SPECT image of the kidneys, but there is also a great interest to map out the absorbed dose to tumours. In order to do so a two FOV SPECT image has to be acquired. Thus the total acquisition time becomes very long and patient movements has shown to be a great difficulty when matching the two FOV together.

At SUS image acquisition is made with one energy window around the 208 keV photopeak and by adding a second one around the 113 keV photopeak, the acquisition time can hopefully be reduced. The determination of absorbed dose is based on a mean value of the activity concentration inside a VOI de- fined from CT image. Thus the task is to lower the acquisition time but still maintain a mean value with low uncertainty. This results in the question formulation: For SPECT imaging, of patients treated with 177Lu-DOTATATE, how much can the acquisition time be reduced when collecting information from 113 keV photopeak and 208 keV photopeak?

Method: SPECT projection data of a cylindrical phantom, within it 6 spheres of different sizes, were simulated with Monte Carlo code SIMIND. Projection sets with activity in the spheres and activity in the background were simulated separately and added together in a later step. The reason for the separate simulation was to easily be able to modify the activity concentration in the spheres and background inde- pendent of each other. Different number of counts in the projection data can be a result of either different activity concentrations in source regions or due to different acquisition times. A quantity, named ”dis- integration concentration value”, was defined as the number of radioactive decay during the acquisition time, per unit volume, to be able to change the acquisition time in the projection data. The disintegration concentration in the spheres was changed by multiplying with so called Disintegration Concentration value (DC-value) [MBq s/ml], derived from activity concentration multiplied with an acquisition time. Thus a higher DC-value is derived from a longer acquisition time, therefore the spheres obtained a higher disintegration concentration. A set of 10 projection sets were created, all multiplied with different DC- values, i.e. different time per projection. The projection data of the sphere and background were added together and four different contrasts of disintegration concentration, between the spheres and the back- ground, were evaluated for all 10 DC-values. For every projection set 50 noise realisation were created. Reconstruction was done with OS-EM algorithm, using 6 subsets and 40 iterations. Projection data for 113 keV and 208 keV were simulated separately and data from the reconstructed images were later added together. Masks of the spheres was created to extract data from the reconstructed images. The data were evaluated by the recovery of the activity concentration and CV, where the standard deviation, of the 50 mean values from the 50 noise realisations, was divided with both the measured disintegration concentration and the true disintegration concentration, respectively.

Result & Conclusion: Image acquisition with two energy windows, one around the 113 keV photo- peak and one around 208 keV photopeak, results in time per projections according to DC-value=22.50 MBq s/ml. For tumours it indicates a time per projection of 22.5 seconds but for kidneys the time per projection would have to be 225 seconds, due to its lower activity concentration 96 hours p.i. compared to the activity concentration in tumours at that time. At SUS the time per projection is 45 seconds, meaning that for image acquisition of tumours the time per projection can be shortened by a factor 2. However for image acquisition of kidneys the time per projection needs to be prolonged by a factor 5 compared to today’s image protocol.},
  author       = {Mörnsjö Centofanti, Fanny},
  language     = {eng},
  note         = {Student Paper},
  title        = {Optimization of Acquisition Time for 177- Lu SPECT with Two Energy Windows},
  year         = {2019},
}