Master of Science Thesis in Medical Physics

Image Based Method for Bone Marrow Dosimetry -in patients treated with 177Lu-DOTATAE

Julia Söderström

Supervisors: Peter Bernhardt Johanna Svensson

Department of Radiation Physics University of Gothenburg Gothenburg, Sweden June 2014 0

Abstract Background Peptide receptor radionuclide therapy with 177Lu-DOTATATE has shown to be an effective treatment of patients diagnosed with metastatic neuroendocrine tumors. Organs at risk in this treatment are primarily the kidneys and the bone marrow. Kidney dosimetry has already been investigated but less effort has been made to investigate the absorbed dose to the bone marrow. The absorbed dose limit to the bone marrow in radionuclide therapy is often set to 2 Gy. It is generally assumed that treatment with 177Lu-DOTATATE generates lower absorbed doses to bone marrow; however, hematological toxicity is often achieved during treatment. The aim of this study was to develop an image based method for bone marrow dosimetry. Material and methods This study included 51 patients with metastatic neuroendocrine tumors who were all treated with 177Lu-DOTATATE. Planar anterior and posterior scintigraphic images were acquired at 2, 24, 48 and 168 hours post injection. By using a segmentation tool in the in-house made software RONSO, the high uptake in the images was separated from the low uptake. Thereby the images were divided into two compartments; one containing the tissue related activity which area corresponded to the area of the tumors and organs (i.e. the high uptake), and one containing the blood related activity corresponding to the low uptake in the images. Absorbed dose to the bone marrow was calculated as sum of; the self-dose from charges particles of cumulated activity concentration in bone marrow, which was assumed to be equal to that in blood; the cross-dose from the high uptake solely contributed by the γ-radiation; and a crossdose from the remainder of the body. The hematological toxicity was graded according to CTCAE (Common Terminology for Adverse Events) 4.0 by NCI (National Cancer Institute, USA). The relation between the hematological toxicity and absorbed dose was then studied. Results The median of the absorbed dose to the bone marrow was 0.17 Gy (range 0.08 – 0.49 Gy) for the first therapy cycle; where an average amount of 7.5 GBq (range 4.2 – 8.3 GBq) 177LuDOTATATE was administered to the patients. For all the treatment cycles the median of absorbed dose was 0.51 Gy (range 0.16 – 1.35 Gy) where an average amount of 7.1 GBq per treatment cycle was administered. No correlation between the absorbed dose to the bone marrow and the hematological toxicity was found, though the hematological toxicity seemed to depend on the cross-dose from the high uptake. Conclusion The image based method developed in this work might be useful for bone marrow dosimetry. In treatment with 177Lu-DOTATATE the self-dose made the largest contribution to the absorbed dose to bone marrow. There was a considerable variation of absorbed dose between patients. The absorbed doses to bone marrow obtained from this image based method are in the same order of magnitude that has been reported from measurement of blood samples. The method requires to be validated for a more patient specific purpose, which can be done by blood dosimetry from blood sampling. Since no correlation between absorbed dose to bone marrow and hematological toxicity was found, the hematological toxicity might not be due to irradiation of bone marrow of this magnitude. Instead it might be due to irradiation of the spleen, since the spleen has been reported as the organ with the highest uptake of the radiopharmaceutical in treatment with 177Lu-DOTATATE. This could explain why the hematological toxicity seemed to depend on the absorbed dose from the compartment which included the high uptake.

1

Abbreviations 177

Lu BAC BRA BRAarea Bq CTCAE DOTA DOTATATE GFR Gy h.p.i. keV MDS MIRD NET nNUFTI PRRT Rmax Rmean ROI SPECT t ThI TRA TRAarea TBRA TTRA W

lutetium 177 Blood Activity Concentration Blood Related Activity area of the compartment containing BRA Becquerel Common Terminology for Adverse Events Dodecaneteteriaminepentaacetic acid the somatostatin analogue octretate bounded to the chelate DOTA Glomerular filtration rate Gray hours post injection kilo electron volt myelodysplastic syndrome Medical Internal Radiation Dose neuroendocrine tumor normalized Number of Uptake Foci versus Threshold Index Peptide Receptor Radionuclide Therapy maximum range mean range Region of Interest Single Photon Emission Computed Tomography thickness of the mean organ containing activity in TRA threshold index Tissue Related Activity area of the compartment containing TRA thickness of the compartment containing BRA thickness of the compartment containing TRA weight of patient

2

Table of content Abstract ................................................................................................................................................................... 1 Abbreviations .......................................................................................................................................................... 2 Introduction ............................................................................................................................................................. 4 Materials and methods ............................................................................................................................................ 7 Patients and image acquisition ............................................................................................................................ 7 Image analysis .................................................................................................................................................... 7 Whole body regions of interest (ROI) ............................................................................................................ 7 Segmentation .................................................................................................................................................. 7 Activity determination ........................................................................................................................................ 9 Calibration of the thicknesses of the compartments ..................................................................................... 10 Blood Activity Concentration....................................................................................................................... 11 Kinetics ........................................................................................................................................................ 11 Dosimetry ......................................................................................................................................................... 12 Self-dose ....................................................................................................................................................... 12 Cross-dose from TRA .................................................................................................................................. 13 Cross-dose from remainder .......................................................................................................................... 13 Total absorbed dose from all treatment cycles ............................................................................................. 14 Correlation with hematological toxicity ........................................................................................................... 14 Statistical Methods............................................................................................................................................ 14 Results ................................................................................................................................................................... 15 Image analysis .................................................................................................................................................. 15 Whole body region of interest (ROI) ............................................................................................................ 15 Segmentation ................................................................................................................................................ 15 Activity determination ...................................................................................................................................... 17 Calibration of thickness of the compartments .............................................................................................. 17 Kinetics ........................................................................................................................................................ 19 Dosimetry ......................................................................................................................................................... 20 Correlation with hematological toxicity ........................................................................................................... 23 Discussion ............................................................................................................................................................. 26 Conclusion ............................................................................................................................................................ 29 Future aspects ........................................................................................................................................................ 30 Acknowledgement ................................................................................................................................................ 31 References ............................................................................................................................................................. 32 Appendix I ............................................................................................................................................................ 34 Appendix II ........................................................................................................................................................... 35

3

Introduction 177

Lu-DOTATATE treatment is a form of peptide receptor radionuclide treatment (PRRT) which has been reported as an effective treatment in patients with metastatic neuroendocrine tumors (NET) (1). To avoid toxicity due to the absorbed dose to normal tissue, individual dosimetry has shown to be of great importance (2). Since hematological toxicity often is achieved in patients undergoing treatment with 177Lu-DOTATATE, bone marrow dosimetry is important. In this work an image based method for bone marrow dosimetry is examined. In Sweden almost 400 persons are diagnosed with NETs on a yearly basis. NETs are slow growing neoplasms that are originated from the neuroendocrine cell system. The neuroendocrine cell system is found throughout the entire body and acts as the link between the nervous system and the endocrine system. The majority of the NETs start in the gastrointestinal tract, most commonly in the small intestine (3). These tumors often overexpress neuroendocrine hormone receptors such as somatostatin, serotonin etc. (4), where an overproduction of serotonin can cause diarrhea, heart disease or flushes (5). The disease is often advanced at diagnosis and the treatment options will thereby be of a palliative intent. Surgery is the most common treatment of metastatic NETs. But when surgery is no longer an option, due to the metastasis, patients are often treated with interferon or somatostatin analogues for palliation and symptom relief. In some cases these treatments can also give an anti-proliferative effect (6, 7). The overexpression of somatostatin receptors of the NETs has been used to develop diagnostics with radionuclide bounded to somatostatin analogues (8) and later on peptide receptor radionuclide treatment (PRRT) (9, 10). The safety and effectiveness for PRRT of NETs was first reported by Kwekkeboom et al. (1), and also a positive impact of life has been reported (11). In PRRT with 177Lu-DOTATATE the radioactive isotope 177Lu of the lanthanide lutetium is bounded to the somatostatin analogue octreotate via the chelate Dodecaneteteriaminepentaacetic acid (DOTA) (12). 177Lu has a half-life of 6.7 days and decays solely by β--decay. When decaying it emits low-energy charged particles as β-particles, Auger electrons and conversion electrons with a mean energy of 147 keV per decay (13), a Rmean of 0.5 mm and a Rmax of 2 mm (14). Besides the emission of charge particles 177 Lu also emits γ-radiation, which is of great use for gamma camera imaging. The main γradiation has an energy of 208.4 keV (11.0 %) and 113 keV (6.4 %) (15). Organs at risk in 177Lu-DOTATATE treatment are the kidneys and the bone marrow (12). The highest absorbed dose has been found in the spleen, though its clinical relevance is uncertain. In some cases liver toxicity has been reported (16). Hematological toxicity has sometimes become relevant in clinical practice. Although kidney dosimetry has already been investigated less effort has been made in bone marrow dosimetry (2). In this work the focus will be on the absorbed dose to the bone marrow. The bone marrow is an organ that is found in the cavities of long bones, as seen in figure 1, and the trabecular bone of the vertebrae, sternum, ribs, and the flat bones of the cranium and pelvis. It consists of a sponge-like tissue framework called stroma; myeloid tissue in which the production of blood cells takes place; lymphatic tissue and numerous blood vessels and sinusoids. There are two types of bone marrow, red and yellow. The production of blood cells takes place in the red marrow; while the yellow marrow mostly contains fat tissue and do not produce blood cells (17). 4

Spongy bone with red marrow Compact bone Yellow marrow Blood vessels

FIGURE 1. Anatomy of long bone in adult. The long bone consists of spongy bone which contains the red marrow, and yellow marrow. Numerous of blood vessels are passing through the bone marrow (18).

Due to irradiation of the red marrow hematological toxicity can occur. This can be manifested as myelodysplastic syndrome (MDS) or leukemia. MDS occurs in less than < 1 % of the patients and are medical conditions related to an underproduction of myeloid blood cells and in some cases it can develop to anemia. The probability of developing leukemia is approximately 2 % with an absorbed dose of 2 Gy to the bone marrow (19). A maximum absorbed dose of 2 Gy is used at Sahlgrenska University Hospital. This absorbed dose limit is generally accepted to avoid underproduction of blood cells (20), but later reports has showed the limit could be up to 3 Gy by a more patient-specific dose calculation (21). The contributions to the absorbed dose to the red marrow due to PRRT with 177LuDOTATATE will be derived from the activity circulating in blood (self-dose), the accumulated activity in tissue and tumor with high uptake (cross-dose), and cross-dose from the blood in the remainder of the body. Since it has been shown that there is no significant activity uptake in bone marrow and that the activity concentration in bone marrow is identical to that in blood, the self-dose can be calculated solely from the activity concentration in blood (2). This absorbed dose will only arise from the charged particles emitted by 177Lu due to that less than 2 % of the γ-radiation are absorbed locally in small organs such as the bone marrow (22). The cross-dose to the red marrow will solely arise from the γ-radiation since maximum range of the charged particles is 2 mm in tissue. Preliminary results indicate that in approximately 10 % of the patients a toxicity of grade 3 was reported after one or more treatment cycles; where the hematological toxicity was graded according to CTCAE (Common Terminology for Adverse Events) 4.0 by NCI (National Cancer Institute, USA). As seen in figure 2 A and B, the hematological toxicity seemed to depend on the whole body residential time and the distribution of the radiopharmaceutical in the body. Also patients with a higher tumor burden tended to develop a severer degree of hematological toxicity, 1.30±0.68 for small tumor burden respectively 1.67±0.66 for large tumor burden. The tumor burden was visually estimated in the 24 hour planar image for each patient at the first treatment. It was graded from 1 (minor tumor burden) to 5 (major tumor burden) with accordance to the number of tumor uptakes, the intensity of the tumor uptakes and the estimated volume of the tumor uptakes. A tumor burden of 1 or 2 were grouped as 5

small tumor burden, 4 or 5 to large tumor burden, and a tumor burden of 3 represented a medium tumor burden. A.

B.

FIGURE 2. In an earlier study of these patients it was seen that the hematological toxicity tended to increase with increasing residence time (A) and the residential also time seemed to be longer for a larger tumor burden.

With this in mind the work in this study investigates the correlation between the distribution of the radiopharmaceutical and the hematological toxicity. This by dividing the signal in two compartments: one containing the blood activity and one containing the activity of the tumors and the organs related with a high uptake. The aim of this study was to develop an image based method for bone marrow dosimetry by quantification of tissue related activity and blood related activity in patients treated with 177Lu-DOTATATE, and examine the correlation between the absorbed dose contribution from these compartments and the hematological toxicity.

6

Materials and methods Hematological toxicity often occurs in patients treated with 177Lu-DOTATATE. Individual dosimetry is important to avoid toxicity due to the absorbed dose to normal tissue. In this work an image based method for bone marrow dosimetry was developed.

Patients and image acquisition This study includes 51 patients (20 females and 31 males), all diagnosed with metastatic neuroendocrine tumors, who all underwent 177Lu-DOTATATE treatment (1-6 cycles) at Sahlgrenska University Hospital in Gothenburg. An average amount of 7.5 GBq (range 4.2 – 8.3 GBq) 177Lu-DOTATATE was administered to the patients as a 30 minutes infusion. The data used in this work were obtained from anterior and posterior planar whole body scintigraphic images acquired at 2, 24, 48 and 168 hours post injection (h.p.i.). The gamma cameras used for imaging were Picker IRIX (Marconi, Phillips, Holland) and Millennium VG Hawkeye (General Electric Medical Systems, Milwaukee, WI, USA) with the sensitivity of 11.05 cps/MBq and 7.35 cps/MBq. Both cameras were equipped with medium energy parallel-hole collimators.

Image analysis Whole body regions of interest (ROI) Geometric mean images were derived from the anterior and posterior planar images. A whole body region of interest (ROI) was created for each geometric mean image by isosurface setting. To ensure that the whole body ROI covered the whole body and that it was too not noisy the image was first filtered by a Butterwort low pass filter at order 2 with a cutoff frequency of 0.03. Then the isosurface setting was applied to make the whole body ROI to fit the body contour, which was visually evaluated. The whole body ROI was then applied on the unfiltered geometric mean images. For those images which included an uribag, the signal from the uribag was set to zero by manually drawing a ROI around it and then changing its in-value. Then the whole body ROI was created as described above. Sometimes the whole body ROI excluded pixels around high uptake; this was due to the filtering of the images. To avoid this, the values of those pixels were set to the same value as of the high uptake after filtering and then the isosurface setting was applied again. Segmentation The whole body activity, AWB, was divided into two compartments; Tissue Related Activity (TRA) and Blood Related Activity (BRA). TRA was assumed to correspond to the activity in the tissue i.e. the signal from the high uptakes; meanwhile BRA was assumed to correspond to the rest of the signal. The sum of the activity in the two compartments corresponded to the whole body activity, equation1, (1) In order to separate the compartments from each other a segmentation tool in the in-house made software RONSO was used. Figure 3 shows the plot provided by the segmentation tool, where the number of segments was presented on the y-axis and the threshold value was presented on the x-axis. The threshold-value was presented with the highest value to the left and the lowest to the right and it represented the minimum signal in a segment. For the threshold value a threshold index, ThI, was calculated by 7

(2)

The number of segments on the y-axis was normalized to the maximum segments, this normalization was called nNUFTI (normalized Number of Uptake Foci versus ThI), where uptake foci were equal to segments.

FIGURE 3. The plot provided by the segmentation tool in RONSO. The number of segments is presented at the y-axis and the threshold-value at the x-axis. The red line represents the chosen threshold-value.

The segmentation tool created a ROI around the high uptake for a chosen threshold-value. The threshold value is represented by the red line in figure 3. Figure 4 shows how the threshold value influenced the created ROI, where the size of the multi-colored ROI increased with increasing ThI. For a specific threshold value the ROI assumed to correspond to the counts in the Tissue Related Activity, TRA. Thereby the rest of the signal in the whole body ROI assumed to correspond to the counts in the Blood Related Activity, BRA, which could be seen as the counts to the right of the red line in figure 3. The specific threshold value which separated TRA from BRA was individually adapted for each patient to a level where the noise started to appear. For this value the nNUFTI was noted. Then a mean of nNUFTI for all patients was taken to the value that separated TRA from BRA, this was then visually evaluated for each patient. To survey the impact of nNUFTI, the absorbed dose to bone marrow was calculated for different nNUFTI. To keep the area of TRA constant, TRA was only segmented in the 24 h.p.i. geometric mean images. The ROI of TRA was then applied on the 2 h.p.i., 48 h.p.i. and 168 h.p.i. geometric mean images and was manually translocated and rotated to fit the uptakes in the images. The counts in BRA were calculated as the difference of the counts in the whole body ROI and the counts in TRA, and the area of BRA as the difference of the area of the whole body minus the area of TRA. The impact on absorbed dose due to in which image the segmentation was made was then studied.

8

A.

B.

D.

C.

FIGURE 4. A. The 24 h.p.i. geometric mean image of patient 50 with whole body ROI. The green whole body ROI and the multi-colored ROI from segmentation with: B. ThI =0.92 and nNUFTI=0.02, C. ThI =0.97 and nNUFTI=0.10, D. ThI =0.99 and nNUFTI=0.90.

Activity determination Activity determination in the scintigraphic planar images was executed by using the conjugate view formula, first described by J.S. Fleming (23), √ (

)

(3)

where √(CAPCPA) was the geometric mean of the anterior and posterior count rate, t the thickness of the organ containing activity, T the body thickness, k the sensitivity of the gamma camera and µ the effective linear attenuation set to 0.1032 cm-1. The effective linear attenuation was earlier obtained by measurements of 177Lu in a petri dish with a diameter of 10 cm at different depths in a Plexiglas phantom. Figure 5 shows how the conjugate view method was applied on the geometric mean images in order to determine the activity in each compartment. The orange color corresponds to the distribution of the radiopharmaceutical in the body, where the small dots correspond to BRA and the larger orange area corresponds to TRA. The distribution of BRA was assumed to be homogenous inside the thickness TBRA and was calculated by √ (

)

(4)

9

TRA was assumed to be homogenously distributed inside a mean organ with thickness t and the body thickness of this compartment was TTRA. TRA was calculated by equation 5, where the last term was a correction of the blood related activity outside of t in TTRA. TRAarea were the area of TRA and BRAarea the area of BRA. BRA was assumed to be homogeneously distributed in the volume of TRA outside t. The thickness t was derived as mean of the thickness of the main organs that contributed to the signal in TRA i.e. kidneys, spleen and liver (17). (

√ (

)

) (5)

At the time 0 h.p.i. it was assumed that TRA was zero and that BRA was equal to the administered activity. The impact of t and µ on the absorbed dose to bone marrow was determined for four patients by calculating the absorbed dose for different t and µ.

FIGURE 5. Schematic image how the conjugate view method was applied on the geometric mean images in order to determine the activity in each compartment.

Calibration of the thicknesses of the compartments The thicknesses of both compartments, TTRA and TBRA, were calibrated to ensure that the calculated activity corresponded to the true activity. This was done for those patients who had not urinated before the 2 h.p.i. image acquisition, and thereby contained an activity equal to the administered activity at that moment. The thicknesses were calculated as an iterative process in MATLAB® (see Appendix I).

10

A thickness weight-dependent thickness was calculated, equation 6, where W was the weight of the patient, the whole body area, A, was determined as the mean of the area of the whole body ROIs of all images for each patient, and ρ the density (ρ = 1 g/cm3). A linear correlation between the weight-dependent and the calibrated thicknesses was made. TTRA and TBRA could thereby be determined for those patients who had urinated before the 2 h.p.i. image acquisition, by first calculating the weight-dependent thickness and then inserting it in the expressions of the linear correlations. This method was tested by inserting the values from patients used for calibration into the linear correlation into activity calculations, equation 4 and 5. Then the ratio between the calculated and administered activity was calculated. ( )

(6)

For those patients whose calculated activity was larger than the administered activity when using the linear correlation for calculation of TTRA and TBRA, the thicknesses were instead calculated by the iterative process in MATLAB. Blood Activity Concentration Since no significant uptake of 177Lu-DOTATATE in bone has been reported, the concentration of activity in bone marrow was assumed to be identical identical to that in blood, (2). The blood activity concentration, BAC, was calculated by

(7) where BRAvolume fraction was the fraction of the total body volume which corresponded to the volume of BRA. Kinetics The kinetics of TRA, BRA and BAC were studied in time-activity diagrams. For BAC biexponential function was fitted to the data points, by the curve fitting tool in MATLAB, by the form (8)

( )

which could be described in two elimination phases. Where A(t) was BAC at the time t, A0 the start value of each phase, and λ the decay constant. Since the elimination was divided into two phases, these phases had different effective half-lives. The effective half-life was determined by ( ) (9) for each phase. For TRA, the kinetic was divided into two or three phases; one uptake-phase and, one or two elimination phases. The elimination phases were described by a biexponential function as in equation 8 or by a single exponential curve, due to the number of data points. The uptake-phase was described by equation 10 in the interval 0 h.p.i. to the time of the highest uptake measured. ( )

(

)

(10)

11

Dosimetry The absorbed dose to the red marrow was calculated according to the MIRD scheme (24). The contribution to the absorbed dose to the bone marrow were obtained from the self-dose, the first term in equation 11; the cross-dose from TRA, the second term; and the cross-dose from the remainder of the body, the third term. ̃

̃

̃

(11)

where S (S-value) was the absorbed dose per unit cumulated activity and à the cumulated activity. The S-value was summarized over the energies, Ei, multiplied by the intensity of the energy, ni, and the specific absorbed fraction, Φ, from source to target at that energy. ∑

(

)

(12)

The cumulated activity was calculated by ̃

(13)

for each phase, where τ was the residential time and A0 the constant a from equation 8. For the elimination phases the residential time was calculated by

(14)

∫

where t1 and t2 was the integration limits, where t2 was set to infinity. For BAC t1 was set to zero, but for TRA it was set to the time of the highest uptake. The residence time of the uptake phase of TRA was determined by equation 15, where t1 was set to zero and t2 to the time of the maximum measured uptake. ∫(

)

(15)

The total accumulated activity was obtained as the sum of accumulated activity in phase in each compartment. Self-dose The small amount of locally absorbed γ-radiation, as mentioned in the introduction, resulted in a self-dose to the bone marrow solely arisen from the emitted charged particles, hence ϕBMBM=1 g-1 in equation 16. ̃( (16) ) Δ equaled nE in equation 12 and was 147 keV per decay, and the accumulated blood activity concentration, Ã(BAC), was calculated as the accumulated activity described earlier. The self-dose was calculated separately for each elimination phase in order to examine which elimination phase that made the largest contribution to the self-dose.

12

Cross-dose from TRA The cross-dose from TRA was contributed from the activity in kidneys, liver, spleen and tumors in all patients. As earlier mentioned, those organs had the highest uptake in treatment with 177Lu-DOTATATE (22). Hence, the S-value of TRA was calculated as the mean of the S-values of these organs, table I. The specific absorbed fraction, Φ, was calculated from the organs in TRA to the red marrow by interpolation of the photon energies in the tables in MIRD no. 5 (25). The cross-dose from TRA to the red marrow was the obtained by ̃

∑

(17)

(

)

(18)

TABLE I. S-values of the main organs included in TRA to the red marrow for 177Lu.

Source organ Kidneys Liver Spleen Mean:

Sred marrow  source organ (J/(Bqs kg) 5.8 · 10-17 2.5 · 10-17 2.7 · 10-17 3.7 · 10-17

Cross-dose from remainder The cross-dose from the reminder of the body was only contributed by the emitted γ-radiation. The cumulated activity in the remainder of the body was set to cumulated activity in BRA, the S-value was calculated as a mean of the rest of the organs that was not included in TRA, and also excluded bone, ovaries and testis. TABLE II. S-values of the remaining organs which were not included in TRA to the red marrow for 177Lu.

Source organ Adrenals Bladder contents Stomach contents Small intestine plus contents Upper large intestine contents Lower large intestine contents Lungs Muscle Pancreas Thyroid Mean:

Sred marrow  source organ (J/(Bqs kg) 5.6 · 10-17 3.4 · 10-17 2.5 · 10-17 6.5 · 10-17 5.6 · 10-17 7.6 · 10-17 2.9 · 10-17 3.1 · 10-17 4.3 · 10-17 1.8 · 10-17 4.5 · 10-17

13

Total absorbed dose from all treatment cycles Since most of the patients underwent more than one treatment cycle, the total absorbed dose to the bone marrow was of interest. In order to estimate the absorbed dose to bone marrow for all the treatment cycles the absorbed dose per unit administered activity from the first treatment cycle was multiplied with the total administered activity for all treatment cycles. For the first 15 patients the segmentation analysis was made for all treatment cycles in order to obtain the individual variation in absorbed dose per unit administered activity.

Correlation with hematological toxicity The hematological toxicity was earlier graded according to CTCAE (Common Terminology for Adverse Events) 4.0 by NCI (National Cancer Institute, USA) from grade 0 to 4; this for the decrease of leucocytes, erythrocytes and thrombocytes. The correlation between the hematological toxicity and the absorbed dose to bone marrow from one and all treatment cycles was studies in a box and whiskers plot. This was also performed for the self-dose and the cross-dose. Since no specific consideration was taken to the tumors when applying the svalues; the correlation between hematological toxicity and absorbed dose was studied solely for patient without bone metastasis.

Statistical Methods The statistical analyze was performed by a Jarque-Bera test to check for normal distribution. For the data that did not pass, results are given as a median and total range.

14

Results Image analysis Whole body region of interest (ROI) To be able to apply the segmentation analysis in the images a whole body ROI had to be created, since the segmentation only could be performed inside a whole body ROI. The whole body ROI was also needed in order to obtain the total count rate and thereby the count rate of BRA as the difference between the total count rate and the count rate in TRA. Figure 6 shows whole body ROIs for two patients with filtering (B and D) and without filtering (A and C) before creating the ROIs with isosurface setting at the 168 h.p.i. geometric mean images, where signal was the most noisy. A. .

B.

C.

D.

FIGURE 6. ROIs of the 168 h.p.i. geometric mean image. A. Whole body ROI of patient 1 with unfiltered image; B. Whole body ROI of patient 1 by first filtering the image; C. Whole body ROI of patient 50 with unfiltered image; D. Whole body ROI of patient 50 by first filtering the image.

Segmentation TRA contained kidneys, liver, spleen and tumor for all patients. In some cases it also contained bladder and intestine. In the segmentation analysis it was seen that a nNUFTI of 0.10 (SD 0.03) could separate the compartments from each other. Figure 6 shows the impact of different nNUFTI to the absorbed dose to bone marrow, data are presented as the relative deviation from absorbed dose calculated from nNUFI = 0.10. A different choice of nNUFTI would result in a change of absorbed dose to bone marrow of maximum 16 %. Smaller nNUFTI resulted in an increase of the absorbed dose; meanwhile a higher nNUFTI resulted in a decrease of the absorbed dose. A nNUFTI equaled to 0.15 – 0.25 would maximum decrease the absorbed dose with 10 %. The impact on the self-dose and cross-dose is shown in Appendix II. 15

Figure 7 shows the impact on absorbed dose due to in which image the segmentation of TRA and BRA was performed. Data are presented as the relative deviation from segmentation in the 24 h.p.i. images. Segmentation in the 168 h.p.i. images yielded a higher dose for patients 7, 9 and 28 than for segmentation in 24 or 48 h.p.i. images. For patients 5, 7 and 9 segmentation in the 2 h.p.i. and 48 h.p.i. yielded similar absorbed doses to that in the 24 h.p.i. image, also for segmentation in the 48 h.p.i. for patient 28. Segmentation in the 168 h.p.i. image yielded a lower absorbed dose for patient 5 but higher absorbed doses for patients 7, 9 and 28. The impacts on the self-dose and cross-dose are shown in Appendix II.

Relative deviation in absorbed dose to bone marrow (%)

20 15 10 5

Patient no. 5 Patient no. 7

0 -5

0.05

0.10

0.15

0.20

0.25

0.50

0.80

Patient no. 9 Patient no. 28

-10 -15 -20

nNUFTI

FIGURE 6. Impact of different nNUFTI to absorbed dose to bone marrow for patient 5, 7, 9 and 28. Data are presented as the relative deviation from absorbed dose to bone marrow calculated from nNUFTI =0.10.

Relative deviation in absorbed dose to bone marrow (%)

80 60 40 Patient no. 5 Patient no. 7

20

Patient no. 9 0 2

24

48

168

Patient no. 28

-20 -40

h.p.i.

FIGURE 7. Impact on absorbed dose to bone marrow from which image the segmentation was made; 2, 24, 48 or 168 h.p.i. for four different patients, data are given as the relative deviation from segmentation in 24 h.p.i. images.

16

Activity determination Calibration of thickness of the compartments The calibration of the thicknesses of the compartments gave the following linear correlation between the calculated weight dependent thickness T(W) and TTRA respectively TBRA. ( )

(20)

( )

(21)

Inserting these thicknesses in the calculation of TRA and BRA for the patients who had not urinated before the 2 h.p.i. image acquisition a relative standard deviation of 12 % was obtained. Figure 8 shows the ratio between the calculated activity with the calibrated thicknesses inserted and the administered activity for the non-urinated patients.

Acalculated/ Aadministered (%)

140 120 100 80 60 40 20 0 0

5

10

15

20

25

30

Non-urinated patient FIGURE 8. The ratio of the calculated activity and the administered activity for the non-urinated patients, i.e. the patient who had not urinated before the 2 h.p.i. image acquisition and thereby contained an activity identical to the administered activity.

Figure 9 shows the impact of the effective linear attenuation, µ, on the absorbed dose to bone marrow. Data are presented as the relative deviation from the absorbed dose to bone marrow calculated with µ = 0.1032 cm-1. A change of µ = 0.1032 cm-1 to µ = 0.0992 – 0.1072 cm-1 resulted in a maximum change of the absorbed dose by ± 11 %. As seen in figure 10 a change the value of the organ thickness, t, of TRA resulted in a change in absorbed dose of maximum ± 12 %. The data in the figure are presented as the relative deviation in absorbed dose calculated from t = 8 cm. The impact of t and µ on the self-dose and cross-dose are shown in Appendix II.

17

Relative deviation in absorbed dose to bone marrow (%)

15 10 5 Patient no .5 Patient no. 7

0 0.0992

0.1012

0.1032

0.1052

0.1072

-5

Patient no. 9 Patient no. 28

-10 -15

µ (cm-1)

FIGURE 9. Impact on the absorbed dose to bone marrow different effective linear attenuation coefficients, µ, t for four patients. Data are presented as the relative deviation from absorbed dose to bone marrow calculated from µ = 0.1032.

Relative deviation in absorbed dose to bone marrow (%)

12 10 8 6 4

Patient no. 5

2

Patient no. 7

0

Patient no. 9

-2

6

7

8

9

10

11

12

Patient no. 28

-4 -6 -8

t (cm)

FIGURE 10. Impact on absorbed dose to bone marrow from different organ thicknesses, t, of TRA. Data are presented as the relative deviation from absorbed dose to bone marrow calculated from t = 8 cm.

18

Kinetics Figure 11 shows the typical kinetics of Tissue Related Activity (TRA) and Blood Related Activity (BRA) normalized to their maximum value; this at 2 h.p.i. for TRA and 0 h.p.i. for BRA. As seen in the figure, BRA had a shorter residential time than TRA. 1 0.9 Normalized activity

0.8 0.7 0.6 0.5

TRA

0.4

BRA

0.3 0.2 0.1 0 0

24

48

72

96

120

144

168

h.p.i. FIGURE 11. Typical kinetics of TRA and BRA normalized to their maximum values.

Figure 12 shows a typical a blood activity concentration curve, with a biexponential fit, over time for one patient. This curve represents the blood activity concentration in patient 19 where the effective half-life was 1.28 h and 55.6 h for the first and second phase, respectively. The median effective half-life of the first phase was 1.22 h (interquartile range 0.90 – 1. 29 h) and 52.6 h (interquartile range 46.2 – 58.9 h), respectively. R2 of the biexponential fit was greater than 0.99 for all the patients.

Activity concentration (MBq/g)

0.30 0.25 0.20 0.15 0.10 0.05 0.00 0

24

48

72

96

120

144

168

h.p.i. FIGURE 12. Typical blood activity concentration curve, with a biexponential fit, in patient treated with 177 Lu-DOTATATE. This curve represents the blood activity concentration in patient 19 with effective half-life of 1.28 h and 55.6 h for the first respectively the second phase.

19

Dosimetry The median of the absorbed dose to the bone marrow from the first therapy cycle in patients were 0.17 Gy. A histogram over the absorbed doses is shown in figure 13 and shows the range of absorbed dose to bone marrow 0.08 – 0.49 Gy. The estimated total absorbed dose for all treatment cycles for all patients had a median of 0.51 Gy (range 0.16 – 1.35 Gy). 25

n = 49

Frequency (%)

20 15 10 5 0 0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

Absorbed dose to bone marrow (Gy) FIGURE 13. Total absorbed dose to bone marrow of 49 patients from the first therapy cycle.

Figure 14 shows the mean absorbed dose to the bone marrow per unit administered activity for all treatment cycles for the 15 first patients. The results indicate variation of absorbed dose per unit administered activity from 8.85 to 61.0 mGy/GBq with a mean of 22.2 mGy/GBq, there was also an individual variation. 70

D/Aadministered (mGy/GBq)

60 50 40 30 20 10 0 1

2

3

4

5

6

7

8

9

10 11 12 13 14 15

Patient no. FIGURE 14. The absorbed dose to bone marrow per unit administered activity for the 15 first patients. The results are given as the mean for all treatment cycles. Patient nos. 6 and 15 only received one treatment. Error bars indicate ± 1 SD.

20

There was a larger individual variation in cross-dose from TRA than in self-dose, as seen in figure 15. Data are presented as the relative deviation from the mean of the absorbed dose per administered activity for the cross-dose and the self-dose, respectively. 70

Relative deviation (%)

60 50 40 Cross-dose

30

Self-dose 20 10 0 1

2

3

4

5

6

7

8

9 10 11 12 13 14 15

Patient no. FIGURE 15. Relative deviation from the mean of the absorbed dose per administered activity for all treatment cycles of treatment with 177Lu-DOTATATE for the 15 first patients. Patients 6 and 15 only underwent one treatment cycle.

The self-dose made the largest contribution to the absorbed dose in bone marrow in all the patients, as seen in figure 16. As also shown in figure 15, was that the cross-dose from TRA to the bone marrow also was considerable.

Contribution bone marrow dose (%)

100 90 80 70 60 50 40 30 20 10 0 Cross-dose

Self-dose

Remainder

FIGURE 16.Relative contribution to absorbed dose in bone marrow from the cross-dose from TRA, self-dose from BRA, and remainder of the body. The limits of the boxes indicate the first quartile and the third quartile; the whiskers show the minimum value respectively the maximum value.

21

As seen in figure 17 the largest contribution to the self-dose was derived from the second elimination phase. 100

Contribution to self-dose (%)

90 80 70 60 50 40 30 20 10 0 1st phase

2nd phase

FIGURE 17.Relative contribution to self-dose from the first phase and the second phase. Results are given as the mean percental contribution to self-dose for the patients in the first treatment cycle for each phase. Error bars indicates ± 1 SD.

There was no correlation between the self-dose and cross-dose as seen in figure 18. The line in the diagram shows the linear regression of the data point with an R2 of 0.0069. 0.45 y = 0.1836x + 0.1134 R² = 0.0069

0.40

Self-dose (Gy)

0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 0.00

0.05

0.10

0.15

0.20

Cross-dose (Gy) FIGURE 18. The cross-dose plotted against the self-dose to the bone marrow for each patient from the first treatment cycle. The line indicated the linear regression with an R 1 of 0.0069.

22

Correlation with hematological toxicity

Abdorbed dose to bone marrow (Gy)

Figure 18 shows the hematological toxicity measured after the first treatment cycle patients plotted against the absorbed dose to the bone marrow. Patients with a hematological toxicity of grad 2 or 3 were seen as one group. 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 0

1

2-3

Hematological toxicity (CTCAE) FIGURE 18. The hematological toxicity for absorbed dose to bone marrow. The limits of the boxes indicate the first quartile and the third quartile; the whiskers correspond to the minimum respectively the maximum data point inside ± 1.5 times the interquartile range (the height of the box), and the asterisk corresponds to the a value outside this range

Absorbed dose to bone marrow (Gy)

The absorbed dose for all treatment cycles estimated from the absorbed dose to bone marrow per unit activity from the first treatment cycle plotted against the hematological toxicity after all treatment cycles is seen in figure 19. No correlation between the absorbed dose and hematological toxicity was found. 1.40 1.20 1.00 0.80 0.60 0.40 0.20 0.00 0

1

2

3

Hematological toxicity (CTCAE) FIGURE 19. Hematological toxicity after all treatment cycles and the absorbed dose to bone marrow for all treatment cycles estimated from the absorbed dose per administered activity from the first activity. The limits of the boxes indicate the first quartile and the third quartile; the whiskers correspond to the minimum respectively the maximum data point inside ± 1.5 times the interquartile range (the height of the box), and the asterisk corresponds to the a value outside this range.

23

No correlation was found between the estimated self-dose after all treatment cycles and the hematological toxicity, as seen in figure 20. 1.00 0.90

Self-dose (Gy)

0.80 0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00 0

1

2

3

Hematological toxicity (CTCAE) FIGURE 20. Hematological toxicity after all treatment cycles and self-dose for all treatment cycles estimated from the absorbed dose per administered activity from the first activity. The limits of the boxes indicate the first quartile and the third quartile; the whiskers correspond to the minimum respectively the maximum data point inside ± 1.5 times the interquartile range (the height of the box), and the asterisk corresponds to the a value outside this range.

Figure 21 shows the correlation between the hematological toxicity and the estimated crossdose for all treatment cycles. 0.45 0.40 Cross-dose (Gy)

0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 0

1

2

3

Hematological toxicity (CTCAE) FIGURE 21. Hematological toxicity after all treatment cycles and cross-dose for all treatment cycles estimated from the absorbed dose per administered activity from the first activity. The limits of the boxes indicate the first quartile and the third quartile; the whiskers correspond to the minimum respectively the maximum data point.

24

The hematological toxicity seemed to occur with a similar frequency for patients with and without bone metastasis, as seen in figure 22. 50 45

Frequency (%)

40 35 No bone metastasis (n=32)

30 25

Bone metastasis (n=16)

20 15 10 5 0 0

1

2

3

Hematological toxicity (CTCAE) FIGURE 22. Frequency of the occurrence of hematological toxicity for patients with and without bone metastasis.

Total abdorbed dose to bone marrow (Gy)

Figure 23 shows the hematological toxicity for the patients who had no bone metastasis plotted against the absorbed dose after all treatment cycles, calculated as earlier described. No correlation between the absorbed dose to the bone marrow and the hematological toxicity for patients without bone metastasis. 1.20 1.00 0.80 0.60 0.40 0.20 0.00 0

1

2-3

Hematological toxicity (CTCAE) FIGURE 23. Hematological toxicity plotted against the total absorbed dose to bone marrow for all treatments for patients without bone metastasis. The limits of the boxes indicate the first quartile and the third quartile; the whiskers show the minimum value respectively the maximum value.

25

Discussion The image based method developed in this work might be useful for bone marrow dosimetry in patients treated with 177Lu-DOTATATE since absorbed doses are of the same order of magnitude as earlier reported, considering the administered activity (2, 22). Due incorrect image acquisitions at the first treatment cycle, the segmentation analysis could not be applied to two of the 51 patients. The ROI that corresponded to TRA was made from the segmentation analysis of the image acquisitioned at 24 hours post injection. The self-dose and cross-dose from segmentation in the 24 h.p.i. images was similar to that in the 48 h.p.i. images, the differences was probably due to if TRA contained feces and/or urine. But since the organs were most visible and also the SPECT was acquired at this time, the segmentation was implemented in the 24 h.p.i. images. The self-dose calculated from segmentation in images 2 h.p.i. and 168 h.p.i. was considerable overestimated, while the cross-dose was underestimated. This due to that ROI of TRA did not cover the high uptakes. Sometimes when applying the ROI of TRA created from segmentation in the 24 h.p.i. image on the rest of the images it did not perfectly match the uptakes after translocation and rotation of the ROI, since the patient was not positioned in the exactly same way for all image acquisitions. This could have been solved with a more flexible ROI. A nNUFTI of 0.10 was implemented as the value that separated TRA from BRA. The ROI created from this value visually contained the high uptakes and some noise. It was supposed to contain some noise in order to keep all the segments with a high count rate in TRA and not in BRA. nNUFTI smaller than 0.10 did not cover all the high uptakes for most of the patients. Hence the self-dose and thereby absorbed dose to bone marrow would be overestimated as seen for segmentation with nNUFTI 0.05, since the high uptake would be included in BRA. A nNUFTI of 0.15, 0.20 or 0.25 yielded an absorbed dose to bone marrow equal to nNUFT1 = 0.10, which resulted in a maximum change of ± 9 %. A nNUFTI of 0.50 or 0.80 underestimated the absorbed dose for most of the patient since BAC became smaller; this resulted in a maximum deviation of -16 % of the absorbed dose to bone marrow. The variation of nNUFTI made a larger impact on the cross-dose than the self-dose. The iterative limits for calibration of the thickness of TRA and BRA was set in order for the fraction of the calculated activity and administered activity to be less than 1.000001 and greater than 0.99999. This could seem to be a small interval, but when increasing this interval in some cases there was no difference between the thickness of TRA and BRA. In which case yielded a larger relative standard deviation when inserting the thicknesses obtained from the linear correlation into the activity calculations. Before changing the iteration limits, the calibration gave no difference between TTRA and TBRA when BRA was much larger than TRA. For some patients when using the linear correlations for calculation of the compartment thicknesses the fraction between the total calculated activity and the administered activity was greater than one, in those cases the thicknesses was obtained in by the calculations used in calibration of the thicknesses for the non-urinated patients. By the same assumptions the calculated activity was probably underestimated in some patients. By using the same effective linear attenuation coefficient in the entire BRA, it will be overestimated since no consideration was taken of the attenuation in lung tissue. The attenuation in lung tissue is less than attenuation in soft tissue. Besides this, 10 percent of the total blood volume is located in the lungs (17). The lungs were hardly ever excluded from 26

BRA, and thereby a large part of the blood volume in BRA corresponds to the blood in lung. Besides this the blood flow is different in the lung compared with the rest of the body. An alternative would be to solely measure the measuring the signal from the lung to relate to the activity in blood. When changing the effective linear attenuation from µ = 0.1032 to µ = 0.0992 – 0.1072 cm-1 a maximum change of ± 10 % of absorbed dose to bone marrow and self-dose were obtained. There was larger impact of the cross-dose with µ. This could have been avoided by instead using scout images for attenuation correction. The organ thickness, t, in TRA was set to 8 cm which corresponded to the mean thickness of kidneys, liver and spleen (17). A change of t to a value in the range 6 – 12 cm resulted in a change of the absorbed dose to bone marrow of maximum ± 12 %. The change was mostly due to change of cross-dose. The uptake-phase of TRA was underestimated due to that the peak of the activity was probably not covered by the data points. But since the cross-dose did not contribute with the largest amount to the absorbed dose in bone marrow, this was probably of minor relevance. The mean of the S-values of the kidneys, liver and spleen was solely taken to describe the absorbed fraction to the red marrow per cumulated activity of TRA. TRA contained kidney, liver, spleen and tumor for all the patients. In some cases TRA also contained bladder and intestines and thereby feces and urine, but since the highest uptakes in organs were found in kidney, liver and spleen only a mean these S-values were applied. This was also performed by Forrer et al. (2). The S-value of the kidneys was higher than the S-values of the liver and spleen, but the cumulated activity in liver and spleen has been reported to be higher than in kidneys. Thereby the applying of the mean of the S-values resulted in an overestimation of the cross-dose. When TRA contained bladder and intestines the S-value of TRA was underestimated, since the S-values of these organs are higher than the liver and spleen. This could have been avoided by using a mean of the S-values of the organs included in TRA which was weighted for the fraction of the cumulated activity in and size of each organ of TRA. But since the cross-dose did not make the largest contribution to the absorbed dose to bone marrow, this probably did not influence the absorbed dose that much. No specific Svalue was applied for the tumors; thereby the cross-dose for those patients with bone metastasis was more uncertain. The S-value from the remainder of the body was calculated as a mean of the S-values of the organs not included in TRA. There were some differences in those values. Bone, testis and ovaries were excluded as organs of the remainder of the body. The equalizing of BRA to the administered activity at 0 h.p.i. was an assumption to capture the first elimination-phase. By this assumption no consideration was taken to that the activity was administered as an infusion during 30 minutes, and thereby the effective half-life of this phase of BAC was underestimated. Comparing the effective half-life of the first phase of BAC obtained in this work with Sandstrom et al.(22), who obtained the effective half-life in the first phase by blood dosimetry, it is seen that the effective half-life of the first phase in this work is slightly shorter; 1.22 h (interquartile range 0.90 – 1. 29 h) compared to 1.61 h (interquartile range 1.44 – 1. 83 h). For the second elimination phase the half-life in this study was slightly longer to that in Sandstrom et al.(22); 52.6 h (interquartile range 46.2 – 58.9 h) in this study compared to 49.5 h (interquartile range 45.1– 56.6 h). Compared to Sandstrom et. al. the last data point of BAC was measured in a later stage in this work; 168 h.p.i. in this work compared to 96 h.p.i. (22). This might explain why the second elimination phase was longer in this study.

27

The absorbed doses to bone marrow by using this images based method were of the same order of magnitude as when using blood dosimetry (2, 22), considered the administered activity. For the first therapy cycle the absorbed dose to the bone marrow was in the range 0.08 – 0.49 Gy with a median of 0.17 Gy, which was well below the absorbed dose limit of 2 Gy. But since most of the patients underwent more than 1 treatment cycles the total absorbed dose was of interest. The estimated median absorbed dose for all treatment cycles were 0.51 Gy (range 0.16 – 1.35 Gy), in this estimation no consideration was taken to the individual variation. The cross-dose had the largest individual variation. This was probably due to that it varied a lot with the choice of thickness and effective linear attenuation. There was also a variation of the absorbed dose between the patients. The largest contribution to the absorbed dose to the bone marrow was received by the self-dose. The cross-dose from TRA gave a considerable contribution in most of the patients. The contribution from the self-dose was slightly larger than earlier reported (2, 22). This due to that the last data point of BAC in this work was measured in a later stage and thereby the second elimination had a longer effective half-life which increases the absorbed dose. No correlation between the self-dose and crossdose from TRA was seen; which confirmed that there were no systematic errors in the segmentation analysis or in the calibration of the thickness. No correlation between hematological toxicity and the calculated absorbed dose to the bone marrow was seen, which has been reported earlier (2). This could depend on that the magnitude of the absorbed dose with a median of 0.17 Gy was low in comparison with the absorbed dose limit of 2 Gy, but nor between the estimated total absorbed dose (range 0.16 – 1.35 Gy) and the hematological toxicity a correlation was found. But the hematological toxicity seemed to depend on the cross-dose, though there was a large uncertainty in the determination of the cross-dose. For patients with bone metastasis the cross-dose was more uncertain, since no consideration was taken to the tumors when applying the S-value. When excluding those patients no correlation between the absorbed dose and hematological toxicity was found in patients without bone metastasis. There is a possible explanation to that the hematological toxicity did not arise from irradiation of bone marrow of this magnitude. Instead it might be due to irradiation of the spleen, since the highest uptake of the radiopharmaceutical in treatment with 177Lu-DOTATATE is in spleen (22). The spleen act as reservoir of the blood cells, and also breaks down the old blood cells by mechanical filtering. Besides this the spleen also plays a big part in the immune defense (26). This will also explain why the hematological toxicity seemed to depend on the cross-dose, since a part of the cross-dose was derived from the accumulated activity in spleen.

28

Conclusion In this work an image based method that might be useful for bone marrow dosimetry has been developed. The absorbed doses to the bone marrow obtained from this method were of the same order of magnitude as earlier reported from work where the blood dosimetry were obtained from blood samples, though this method needs to be validated by blood sampling on an individual basis. The median of the absorbed dose to the bone marrow from the first treatment cycle was 0.17 Gy (range 0.08 – 0.49 Gy) and 0.51 Gy (range 0.16 – 1.35 Gy) for all the treatment cycles. There was a considerable variation of absorbed dose between patients. No correlation between the absorbed dose to bone marrow and the hematological toxicity was found. There is a probability that the hematological toxicity did not occur from irradiation of the bone marrow. It might occur from irradiation of the spleen. This since the organ with the highest uptake of the radiopharmaceutical within 177Lu-DOTATATE treatment has been reported to be the spleen. This can explain why the preliminary results showed that the hematological toxicity seemed to depend on the whole body residence time and that it seemed to depend on the cross-dose in this work.

29

Future aspects The image based method presented in this work yielded absorbed doses to bone marrow in the same order of magnitude on a group level as earlier reported, even though this method requires a validation on an individual basis. The validation can be obtained from blood dosimetry from blood samples. Since the second elimination phase of the blood activity concentration made the largest contribution to the self-dose it is important to validate this phase. Validation can also be performed by measurements in phantoms or by simulating different activity distributions in SIMIND. Since no correlation with the absorbed dose to the bone marrow and hematological toxicity was found it would be interesting to investigate the correlation of the absorbed dose and the percental change in leucocytes, erythrocytes and thrombocytes. But the hematological toxicity seemed to depend on the cross-dose from the high uptakes. The hematological toxicity might not be caused by irradiation of the bone marrow. Maybe it was caused by irradiation of the spleen, since the spleen is reported as the organ with the highest uptake of the radiopharmaceutical and besides this serves as a reservoir of the blood cells. This can explain why the hematological toxicity seemed to depend on the cross-dose. Thereby it would be of great interest to study the correlation between the absorbed dose to the spleen and the hematological toxicity.

30

Acknowledgement First of all, I would like to express my gratitude to my supervisors Peter Bernhardt and Johanna Svensson for this opportunity. Thank you for all your helpful guidance, encouragement and feedback; it has been a real pleasure to work with you. A special thanks to Tobias Magnander, Emma Wikberg and Rebecca Herman for all the help with the work in software RONSO. Thanks to my friends and family for all your support. Especially thanks to my classmate and good friend Frida Svensson, for all the great discussions and helpful comments during this work. I would also like to thank you for all the good times and support during these five years.

31

References 1.

2.

3. 4. 5.

6. 7. 8.

9.

10.

11.

12.

13.

14.

Kwekkeboom DJ, de Herder WW, van Eijck CH, Kam BL, van Essen M, Teunissen JJ, et al. Peptide receptor radionuclide therapy in patients with gastroenteropancreatic neuroendocrine tumors. Seminars in nuclear medicine. 2010;40(2):78-88. Forrer F, Krenning EP, Kooij PP, Bernard BF, Konijnenberg M, Bakker WH, et al. Bone marrow dosimetry in peptide receptor radionuclide therapy with [177Lu-DOTA(0),Tyr(3)]octreotate. European journal of nuclear medicine and molecular imaging. 2009;36(7):1138-46. Oberg K. Biotherapies for GEP-NETs. Best practice & research Clinical gastroenterology. 2012;26(6):833-41. Rehfeld JF. The new biology of gastrointestinal hormones. Physiological reviews. 1998;78(4):1087-108. Vinik AI, Woltering EA, Warner RR, Caplin M, O'Dorisio TM, Wiseman GA, et al. NANETS consensus guidelines for the diagnosis of neuroendocrine tumor. Pancreas. 2010;39(6):713-34. Arnold R, Frank M. Control of growth in neuroendocrine gastroenteropancreatic tumours. Digestion. 1996;57 Suppl 1:69-71. Oberg K. Interferon-alpha versus somatostatin or the combination of both in gastro-enteropancreatic tumours. Digestion. 1996;57 Suppl 1:81-3. Krenning EP, Bakker WH, Breeman WA, Koper JW, Kooij PP, Ausema L, et al. Localisation of endocrine-related tumours with radioiodinated analogue of somatostatin. Lancet. 1989;1(8632):242-4. Valkema R, De Jong M, Bakker WH, Breeman WA, Kooij PP, Lugtenburg PJ, et al. Phase I study of peptide receptor radionuclide therapy with [InDTPA]octreotide: the Rotterdam experience. Seminars in nuclear medicine. 2002;32(2):110-22. Anthony LB, Woltering EA, Espenan GD, Cronin MD, Maloney TJ, McCarthy KE. Indium-111-pentetreotide prolongs survival in gastroenteropancreatic malignancies. Seminars in nuclear medicine. 2002;32(2):123-32. Khan S, Krenning EP, van Essen M, Kam BL, Teunissen JJ, Kwekkeboom DJ. Quality of life in 265 patients with gastroenteropancreatic or bronchial neuroendocrine tumors treated with [177Lu-DOTA0,Tyr3]octreotate. Journal of nuclear medicine : official publication, Society of Nuclear Medicine. 2011;52(9):1361-8. de Jong M, Krenning E. New advances in peptide receptor radionuclide therapy. Journal of nuclear medicine : official publication, Society of Nuclear Medicine. 2002;43(5):617-20. Larsson M, Bernhardt P, Svensson JB, Wangberg B, Ahlman H, ForssellAronsson E. Estimation of absorbed dose to the kidneys in patients after treatment with 177Lu-octreotate: comparison between methods based on planar scintigraphy. EJNMMI research. 2012;2(1):49. Cremonesi M, Botta F, Di Dia A, Ferrari M, Bodei L, De Cicco C, et al. Dosimetry for treatment with radiolabelled somatostatin analogues. A review. The quarterly journal of nuclear medicine and molecular imaging : official publication of the Italian Association of Nuclear Medicine (AIMN) [and] the International Association of Radiopharmacology (IAR), [and] Section of the So. 2010;54(1):37-51. 32

15. 16.

17.

18.

19.

20. 21.

22.

23.

24.

25.

26.

Web page http://nucleardata.nuclear.lu.se/toi/nuclide.asp?iZA=710177 [Cited 23rd of May 2014]. Kwekkeboom DJ, de Herder WW, Kam BL, van Eijck CH, van Essen M, Kooij PP, et al. Treatment with the radiolabeled somatostatin analog [177 Lu-DOTA 0,Tyr3]octreotate: toxicity, efficacy, and survival. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2008;26(13):2124-30. Basic anatomical and physiological data for use in radiological protection: reference values. A report of age- and gender-related differences in the anatomical and physiological characteristics of reference individuals. ICRP Publication 89. Annals of the ICRP. 2002;32(3-4):5-265. Web page http://www.1177.se/VastraGotaland/Tema/Kroppen/Rorelseapparaten/Skelett-och-leder/ [Cited 1st May 2014]. Coleman CN, Blakely WF, Fike JR, MacVittie TJ, Metting NF, Mitchell JB, et al. Molecular and cellular biology of moderate-dose (1-10 Gy) radiation and potential mechanisms of radiation protection: report of a workshop at Bethesda, Maryland, December 17-18, 2001. Radiation research. 2003;159(6):812-34. Nonstochastic effects of ionizing radiation. Annals of the ICRP. 1984;14(3):133. Dorn R, Kopp J, Vogt H, Heidenreich P, Carroll RG, Gulec SA. Dosimetryguided radioactive iodine treatment in patients with metastatic differentiated thyroid cancer: largest safe dose using a risk-adapted approach. Journal of nuclear medicine : official publication, Society of Nuclear Medicine. 2003;44(3):451-6. Sandstrom M, Garske-Roman U, Granberg D, Johansson S, Widstrom C, Eriksson B, et al. Individualized dosimetry of kidney and bone marrow in patients undergoing 177Lu-DOTA-octreotate treatment. Journal of nuclear medicine : official publication, Society of Nuclear Medicine. 2013;54(1):33-41. Fleming JS. A technique for the absolute measurement of activity using a gamma camera and computer. Physics in medicine and biology. 1979;24(1):17680. Howell RW, Wessels BW, Loevinger R, Watson EE, Bolch WE, Brill AB, et al. The MIRD perspective 1999. Medical Internal Radiation Dose Committee. Journal of nuclear medicine : official publication, Society of Nuclear Medicine. 1999;40(1):3s-10s. Snyder W, Ford M, Warner G. MIRD pamphlet no. 5, revised: Estimates of specific absorbed fractions for photon sources uniformly distributed in various organs of a heterogeneous phantom. . New York, NY: Society of Nuclear Medicine. 1978. Web page http://www.ne.se/lang/mj%C3%A4lte [Cited 23rd of May 2014].

33

Appendix I MATLAB® code for calibration of the thicknesses of each compartment. function [ thick ] = calibration_of_thickness( data_) % calibration_of_thickness is a function which calculates the thicknesses % of each compartment for those patient who had not urinated before the 2 % h.p.i image acquisition t=8; % organ thickness my=0.1032; % linear attenuation [x,y]=size(data_); thick=zeros(x,5); for l = 1:x % loop of all the calibration patients patient=data_(l,1); treatment=data_(l,2); IA=data_(l,3); k=data_(l,4); C_TRA=data_(l,5); C_BRA=data_(l,6); TRA_area=data_(l,7); BRA_area=data_(l,8); for T = t:0.01:30 % loop of the "base"-thickness if thick(l,1)>0 break end for T_delta=0:0.01:T % creates a loop for the "base"-thickness

T_BRA= T - T_delta; % thickness of BRA BRA = C_BRA * T_BRA * my/(k * 240 * (1-exp(- my * T_BRA))); T_TRA= T + T_delta; % thickness of BRA TRA = C_TRA * t * my * exp(my * T_TRA /2)/... (k * 240 *(exp(my*t/2)-exp(-my*t/2)))... -BRA*(T_TRA-t)*TRA_area/(T_BRA*BRA_area);

A = TRA + BRA; if A/IA0.999999 thick(l,1)=patient; thick(l,2)=treatment; thick(l,3)=T; thick(l,4)=T_delta; thick(l,5)=A/IA; break % breaks when optimal thicknesses are found end end end end

end end

34

Appendix II Figures 24 to 27 shows the impact of nNUFTI, in which image the segmentation was made, thickness of the organs included in TRA (t), and the effective linear attenuation (µ) on the cross-dose. Figures 28 to 31 shows the impact of nNUFTI, in which image the segmentation was made, thickness of the organs included in TRA (t), and the effective linear attenuation (µ) on the cross-dose.

Relative deviation cross-dose (%)

60 40 20 Patient no. 5 Patient no. 7

0 0.05

0.10

0.15

0.20

0.25

0.50

0.80

-20

Patient no. 9 Patient no. 28

-40 -60

nNUFTI

FIGURE 24. Impact of different nNUFTI on cross-dose for patient 5, 7, 9 and 28. Data are presented as the relative deviation from cross-dose calculated from nNUFTI = 0.10

Relative deviation cross-dose (%)

40 20 0 2

24

48

-20

168

Patient no. 5 Patient no. 7

-40

Patient no. 9

-60

Patient no. 28

-80 -100

h.p.i.

FIGURE 25. Impact on cross-dose from which image the segmentation was made; 2, 24, 48 or 168 h.p.i. for four different patients, data are given as the relative deviation from segmentation in 24 h.p.i. images.

35

Relative deviation cross-dose (%)

30 20 10 Patient no .5 Patient no. 7

0 0.0992

0.1012

0.1032

0.1052

0.1072

-10

Patient no. 9 Patient no. 28

-20 -30

µ (cm-1)

FIGURE 26. Impact on cross-dose from different effective linear attenuation coefficients for four patients. Data are presented as the relative deviation from crocs-dose calculated from µ = 0.1032.

Relative deviation cross-dose (%)

20 10 0

-10

6

7

8

9

10

11

12 Patient no. 5 Patient no. 7

-20

Patient no. 9

-30

Patient no. 28

-40 -50 -60

t (cm)

FIGURE 27. Impact on crocs-dose from different organ thicknesses, t, of TRA. Data are presented as the relative deviation from the crocs-dose calculated from t = 8 cm.

36

Relative deviation in self-dose (%)

50 40 30 20

Patient no. 5

10

Patient no. 7 Patient no. 9

0

-10

0.05

0.10

0.15

0.20

0.25

0.50

0.80

Patient no. 28

-20 -30

nNUFTI

FIGURE 28. Impact of different nNUFTI on self-dose for patient 5, 7, 9 and 28. Data are presented as the relative deviation from self-dose calculated from nNUFTI = 0.10

Relative deviation in self-dose (%)

120 100 80 Patient no. 5

60

Patient no. 7 40

Patient no. 9

20

Patient no. 28

0 2 -20

24

48

168

h.p.i.

FIGURE 29. Impact on self-dose from which image the segmentation was made; 2, 24, 48 or 168 h.p.i. for four different patients, data are given as the relative deviation from segmentation in 24 h.p.i. images.

37

Relative deviation in self-dose (%)

12 10 8 6 4 2 0 -2 -4 -6 -8 -10 -12

Patient no .5 Patient no. 7 0.0992

0.1012

0.1032

0.1052

0.1072

Patient no. 9 Patient no. 28

µ (cm-1)

FIGURE 30. Impact on self-dose from different effective linear attenuation coefficients for four patients. Data are presented as the relative deviation from the self-dose calculated from µ = 0.1032.

Relative deviation in self-dose (%)

15 10 5 Patient no. 5 Patient no. 7

0 6

7

8

9

-5

10

11

12

Patient no. 9 Patient no. 28

-10 -15

t (cm)

FIGURE 31. Impact on the self-dose from different organ thicknesses, t, of TRA. Data are presented as the relative deviation from the self-dose calculated from t = 8 cm.

38