Eighteen patients (F: 3, M: 15) (8: colon Ca, 2: HCC, others: 8) were included in the current study and assigned consent form was obtained from each participant. A range of 111-222 MBq 99mTc-MAA was intra-arterially injected in the targeted liver at the interventional radiology department of Istanbul University-Cerrahpasa. After injection, whole-body scan (WBS) was acquired with adjusted settings includes peak-window at 140 keV (15% width), and one down-scatter window (15% width). A dual-headed gamma camera (Symbia™ T Series SPECT/CT) was used for imaging. According to this protocol, two planar images were created for each patient, as seen in Fig. 1. Region of interest ROI was delineated on the lungs and liver organs separately and the derived counts were used to calculate lung shunt fraction. The LSF was calculated from the peak-window image by:
$$\textrm{LF}=\frac{\sqrt{La\times Lp}}{\left(\sqrt{La\times Lp}+\sqrt{Lva\times Lvp}\right)}$$
(1)
La) lungs anterior counts. Lp) lungs posterior counts. Lva) liver anterior counts. Lvp) liver posterior counts.
Then, scatter corrected LSF was obtained and symbolized as LSFwb (SC) after applying scatter correction over the planar images by dual-energy windows (±15%) as follows:
$$\textrm{Ctrue}=\textrm{Cmain}-\left[\frac{Clow}{Wlow}\right]\ \frac{Wmain}{2}$$
(2)
Ctrue: organ’s scatter corrected counts, Cmain: counts frmom peak window image. Clow: counts from down-scatter window image, Wlow: scatter-window fraction. Wmain: peak-window fraction.
SPECT/CT scan was instantly conducted after scintigraphy encompassing lungs and liver regions. The windows settings were similar to those used in planar imaging and 64 projections were acquired with 25 seconds/projection. Scatter correction was carried out similarly by dual-energy window (15%). Iteration method of ordered subsets expectation maximization (OSEM) was applied for reconstruction with 10 iterations and 8 subsets, pursued by low pass filtering (Gaussian 9 mm) for noise suppression. The reconstruction was repeated to generate three types of SPECT images: first, SPECT images with attenuation and pixel-wise scatter correction; coded as (AC-SC), second: SPECT images with no attenuation and no scatter correction; coded as (NoAC-NoSC), and third: SPECT image with only scatter correction; coded as (NoAC-SC).
The segmentation of liver and lungs was performed over the CT images (Fig. 2) by freehand drawing contours. The injected healthy liver and tumor regions were segmented on the SPECT images via iso-contour tool (Fig. 3) based on an intensity threshold by region growing. The generated volume and counts were displayed on a desktop computer terminal for each organ and compartment. The IHL volume was compared between different SPECT-image types as seen in Fig. 4, to be next used in the suggested activity and lung dose calculation.
An experienced 90Y dosimetrist and physician were cooperated in the registration and segmentation process. The lung shunt fractions were calculated from (AC-SC), and (NoAC-SC) SPECT images by the below equation:
$${\textrm{LSF}}_{\textrm{spect}}:\textrm{Lc}/\left(\textrm{Lc}+\textrm{LVc}\right)$$
(3)
LSF spect: lung shunt fraction from SPECT images, Lc: lungscount, LVc: liver counts.
The lung shunt fractions were calculated by different modules involving: a- LSFwb (St): from the standard whole-body scan, b- LSFwb (SC): from scatter corrected whole-body scan; c- LSFspect (AC-SC): from SPECT/CT with attenuation and scatter correction; and d- LSFspect (NoAC-SC): from SPECT with only scatter correction.
Lung dose was calculated by the following equation:
$$\textrm{D}\ \left(\textrm{Gy}\right)=\textrm{A}\ \left(\textrm{GBq}\right)\times 49.33/\textrm{M}\ \left(\textrm{Kg}\right)$$
(4)
D: dose, A: activity, M: mass.
The absorbed dose to tumor and injected healthy liver was estimated using the following approaches:
-
1)
AC-SC (SPECT/CT): lung shunt fraction, tumor and target quantification were all made by SPECT/CT images with AC and SC.
-
2)
NoAC-SC (SPECT): lung shunt fractions, tumor and target quantification were all obtained from scatter corrected SPECT images.
-
3)
AC-SC + LSFwb (St) lung shunt fractions were computed from the standard WBS, while the tumor and target quantification was made from SPECT/CT (AC-SC).
-
4)
NoAC-NoSC+LSFwb (SC) lung shunt fractions were calculated from scatter corrected whole-body scans, while the tumor and target quantification were made from SPECT images (NoAC-NoSC).
-
5)
NoAC-NoSC+LSFwb (St): lung shunt fractions were calculated from the standard WBS, while the tumor and target quantification was from NoAC-NoSC SPECT images.
MIRD scheme was used in the dose calculation to tumor and non-tumor partitions as follows [8]:
$$\textrm{Liver}\ \textrm{uptake}=\left(1- SF\right)\left[\frac{m_{liver}}{\left({m}_{tumour}\times TLR\right)+{m}_{liver}}\right]$$
(5)
Activity to be administered for a certain absorbed dose:
$$\textrm{Activityadmin}\ \left(\textrm{mCi}\right)=\frac{{\textrm{dose}}_{\textrm{liver}}\ \left(\textrm{rad}\right)\times {\textrm{m}}_{\textrm{liver}}\ \left(\textrm{gm}\right)}{\textrm{184,000}\times \textrm{liver}\ \textrm{fractional}\ \textrm{uptake}\ }$$
(6)
$$\textrm{Tumor}\ \textrm{uptake}=\left(1-\textrm{SF}\right)\left[\frac{\textrm{TLR}\times {\textrm{m}}_{\textrm{tumor}}}{\left({\textrm{m}}_{\textrm{tumor}}\times \textrm{TLR}\right)+{\textrm{m}}_{\textrm{liver}}}\right]$$
(7)
$${\textrm{Dose}}_{\textrm{tumor}}\left(\textrm{rad}\right)=\frac{{\textrm{Activity}}_{\textrm{total}}\left(\textrm{mCi}\right)\times 184000\times {\textrm{UPTAKE}}_{\textrm{tumor}}}{{\textrm{m}}_{\textrm{tumor}}\left(\textrm{g}\right)}$$
(8)
Statistical analysis
The statistical analysis was performed via IBM/SPSS statistics 20 software. Two-sample t-test was used to compare the paired calculation methods in LSF, suggested 90Y activity and absorbed dose to tumor and IHL.
