Quantitative measurement of 219Rn radioactivity in exhaled breath from patients with bone metastasis of castration-resistant prostate cancer treated with 223RaCl2

Background The α-emitting radionuclide radium-223 (223Ra) is widely used for the treatment of bone metastasis in patients with castration-resistant prostate cancer. However, 223Ra decays into radon-219 (219Rn) which is a noble-gas isotope, and 219Rn may escape from patients treated with 223Ra via their respiration. In this study, we quantified the amount of 219Rn contained in the breath of patients treated with 223Ra to estimate its effect on the internal exposure dose of caregivers. Methods A total of 12 breath samples were collected using a breath collection bag from a total of six patients treated with 223RaCl2. Approximately 300 mL of exhaled breath was collected in a breath bag at 1 min and at 5 min after the start of 223RaCl2 administration. The contents of each bag were measured using an HPGe detector, and the amount of 219Rn was quantified based on the detection of the γ peak of 211Bi, which is a descendant nuclide of 219Rn, persisting in the breath bag. The effective dose to caregivers arising from the inhalation of 219Rn was estimated by referring to the scenario for the calculation of release criteria established for 131I therapy in Japan. Results A small peak for the 351-keV γ ray of 211Bi originating from the exhalation of 219Rn was observed. Using the observed γ peak of 211Bi, the average amounts of 219Rn per unit breath volume at 1 min and 5 min after the start of 223RaCl2 administration were calculated as 90 ± 56 Bq/mL and 28 ± 9 Bq/mL, respectively. The effective dose of 219Rn to caregivers was estimated to be 3.5 μSv per injection. Conclusions The amount of 219Rn in the exhaled breath of patients treated with 223RaCl2 was quantitatively calculated using breath collection bags. The internal radiation exposure of caregivers from 219Rn in the exhaled breath of patients treated with 223RaCl2 is relatively small.


Background
Recently, radium-223 ( 223 Ra) dichloride ( 223 RaCl 2 ; Xofigo®) has been widely used for the treatment of bone metastases in patients with castration-resistant prostate cancer (CRPC) [1,2]. This radionuclide emits α-particles within a short range in tissues and with a high linear energy transfer, making it suitable for the treatment of cancer cells with minimal side effects in the surrounding tissues. As shown in the decay chain of 223 Ra (Fig. 1), some descendant nuclides of 223 Ra also emit α-particles and contribute to its killing effect on cancer cells. However, 223 Ra decays into radon-219 ( 219 Rn) which is an isotope of a noble-gas element, and this 219 Rn may escape from patients treated with 223 Ra. Although the escape of 219 Rn from patients is probably minor because of the very short half-life of 219 Rn (3.96 s), the amount of 219 Rn escaping from patients should be quantified to allow the internal radiation exposure experienced by medical staff and caregivers to be estimated. Yamamoto et al. mentioned that 219 Rn may escape via the respiration of patients and reported that the concentration of the descendant nuclides of 219 Rn in a room increased during radionuclide therapy with 223 Ra, compared with that without the patients [3]. However, the amount of 219 Rn exhaled through the respiration of patients has not been quantified in a hospital setting. In this study, the breath of patients treated with 223 Ra was collected using a breath collection bag and the 219 Rn radioactivity was quantified by measuring the γ rays of 211 Bi, which is a descendant nuclide of 219 Rn, persisting in the breath bag. The purpose of this study was to evaluate the level of 219 Rn radioactivity in the exhaled breath of patients treated with 223 Ra to enable a precise estimation of the internal radiation exposure dose experienced by caregivers.

Measurement of expired gas from patients
The study protocol was approved by the institutional review board of Osaka University Hospital. First, we investigated whether the breath of the patients contains radionuclides emitting α-particles using an α survey meter (Hitachi TCS-232b, Tokyo, Japan). A patient (body weight 55.0 kg) received 223 RaCl 2 of 3.0 MBq (55 kBq/kg) for the treatment of bone metastases from CRPC. After informed consent was obtained from the patient, the α survey meter was placed closely to the mouth of the patient from the start of 223 RaCl 2 administration, and α-particles in the breath of the patient were directly measured until 3 min after the start of administration. Additionally, the count rate at 5 min after the start of administration was also measured.
Then, a total of 12 breath samples were collected using breath collection bags (Otsuka Pharmaceutical, Tokyo, Japan) from a total of six patients (71.3 ± 5.4 years old, body weight 59.5 ± 11.7 kg) received 223 RaCl 2 (3.4 ± 0.8 MBq). The administration of 223 RaCl 2 to the patients was conducted slowly over a period of approximately 1 min.
After informed consent was obtained from each patient, approximately 300 mL of exhaled breath was collected into a breath bag at 1 min and at 5 min after the start of 223 RaCl 2 administration. A breath sampling time of 1 min after the start of administration was selected because the amount of 219 Rn contained in a patient's breath is expected to reach a peak value at this time because of the slow administration of 223 RaCl 2 (see also Fig. 3 which is the result of the direct measurement of the breath of the patient by the α survey meter). A sampling time of 5 min is expected to correspond to the time at which the patient is likely to leave the injection room. After sampling, the breath sample was removed from the bag using a syringe (volume 50 mL, minimum volume scale 1 mL) to enable the bag to be folded into a small, fixed geometric shape (see also Fig. 2) for measurement using a high-purity germanium (HPGe) detector (Canberra BE-2020, Connecticut, USA). The volume of the breath collected in the bag Fig. 2 Photographs of a the breath bag with the glass-fiber filter after the removal of the breath sample and b the folded breath bag for the measurement with an HPGe detector was measured using the volume scale printed on the syringe (the uncertainty for the measured volume of the breath sample was estimated to be several milliliters). Because 219 Rn and its descendant nuclides may escape from the bag during the use of the syringe, the bag was allowed to stand for more than 80 s until almost all the 219 Rn atoms in the breath bag had decayed into non-volatile descendant nuclides prior to the removal of the breath sample. Then, the breath sample was removed through a glass-fiber filter (ADVANTEC GB-100R, Tokyo, Japan) attached to the syringe to catch the descendant nuclides of 219 Rn. After the removal of the breath sample, the breath bag, together with the glass-fiber filter, was subjected to γ-ray spectrometry using an HPGe detector for 1800 s at almost zero distance. The amount of 219 Rn in the breath sample was estimated as follows. Based on the observed γ peak of 211 Bi (351 keV), the radioactivity of 211 Pb in transient radioactive equilibrium with 211 Bi was calculated [4]. Then, from the 211 Pb radioactivity at the end of the breath collection, the amount of 219 Rn contained in the breath sample was estimated assuming that the number of 211 Pb atoms at the end of the breath collection would be equivalent to that of 219 Rn because of the very short half-life of 219 Rn. The measurement efficiency for the 351-keV γ peak of 211 Bi at the same distance with the measurement of the breath bag was evaluated based on the measurement of a self-produced 223 Ra point source, the radioactivity of which was determined using a 133 Ba standard source. The measurement efficiency was determined to be 2.63 ± 0.13%. The uncertainty for the efficiency is including the uncertainty of the radioactivity of the 133 Ba standard source (4.8%) and counting errors for the measurements of the 133 Ba standard source and the 223 Ra point source (less than 1%).

Estimation of radiation exposure experienced by medical staff
The descendant nuclides of 219 Rn exhaled through the breath of patients may float through the air by attaching to dust or aerosol particles. These nuclides can contribute to medical staff and caregivers being exposed to radiation internally. Therefore, we also investigated whether the inhalation of descendant nuclides of 219 Rn can be blocked by the use of a mask. Medical staff wore an N95 mask (3M 1870+, Tokyo, Japan) during the administration of 223 Ra to the patients. After administration, the outside and inside (face side) of the mask were measured using an α survey meter (Hitachi TCS-232b, Tokyo, Japan). A total of five masks worn by medical staff who stood close to the patients during 223 RaCl 2 therapy were measured during two injections (three medical staff for one injection [administration dose 3.9 MBq] and two staff for the other [administration dose: 3.0 MBq]). The measurements using the survey meter were conducted while placing the survey meter in close contact with the surface of each mask, and the count rate was recorded at 2 min after the start of measurement with a time constant of 30 s.

Estimation of radiation exposure experienced by caregivers
The radiation exposure experienced by caregivers was calculated based on documentation regarding the calculation of release criteria established for 131 I therapy by the Ministry of Health and Welfare in Japan [5]. The blood concentration of 223 Ra immediately after injection (C Ra ) was estimated using the following formula: where A injection is the injection dose of 223 Ra and V blood the blood volume in the human body. In this calculation, we used A injection = 3.4 × 10 3 kBq which is the average injection dose in this study and V blood = 5.0 × 10 3 mL, and C Ra was estimated to be 0.68 kBq/mL. The area under the curve (AUC) for the blood concentration of 223 Ra was obtained using the clinical trial data (phase I) for 223 RaCl 2 published in Japan by the Pharmaceutical and Medical Devices Agency (PMDA) (https://www.pmda.go.jp). Duration (T), which corresponds to the AUC up to 72 h post injection (0.674 kBq·h/mL) with the assumption that the blood concentration level immediately after the injection of 223 Ra is maintained, was calculated using the following equation: and T was estimated to be 0.99 h. The concentration of 219 Rn in the exhaled breath was assumed to be proportional to the blood concentration of 223 Ra. The total exhaled amount of 219 Rn was calculated by multiplying the radioactivity in the exhaled breath immediately after the injection by a factor of 0.99 h, assuming that the respiration rate was 12 breaths per minute and the volume of breath was 500 mL per single respiration. The effective dose coefficient for the inhalation of 219 Rn was calculated as 7.9 × 10 −9 mSv/Bq, with reference to the 222 Rn data (6.5 × 10 −6 mSv/Bq) [6] corrected by the effective half-life. Regarding the exposure scenario for caregivers, the volume of the room containing the patient was estimated to be 30 m 3 , air changes were regarded as once per hour, the daily respiration rate of the caregivers was estimated to be 20 m 3 , and the caregiver was assumed to always be present in the same room as the patient based on the release criteria for 131 I in Japan [5]. To estimate the internal exposure dose experienced by caregivers, an exposure factor of 0.5 was applied [7].

Results
The time activity curve obtained in the direct measurement of α-particles in the breath of the patient by the α survey meter is shown in Fig. 3. The count rate of α-particles exceeded the measurable range (more than 100 kcpm) at several tens of seconds after the start of administration and then gradually decreased to 20 kcpm until 5 min after the start of administration, showing that radionuclides emitting α-particle were exhaled through the respiration.
A typical γ-ray spectra obtained for the measurement of a breath bag is shown in Fig. 4 a. A small peak for the 351-keV γ ray of 211 Bi was observed. On the other hand, no γ peak of 269 keV from 223 Ra was observed. The 351-keV peak disappeared approximately 4 h after the collection of the breath sample, as shown in Fig. 4 b. This finding unambiguously shows that the breath samples from patients treated with 223 RaCl 2 did not contain 223 Ra.
The amount of 219 Rn contained in the breath samples as determined using γ-ray spectrometry is summarized in Table 1, together with the administration dose of 223 Ra  Rn per unit volume of breath is shown in Fig. 5. The values were calculated as 90 ± 56 Bq/mL and 28 ± 9 Bq/mL (mean ± standard deviation) at 1 min and 5 min after the start of administration, respectively.
The results of the mask measurements using the α survey meter are summarized in Table 2. Small count rates of less than 100 cpm for α-particles were detected on the outside of the masks. On the other hand, no α-particles were detected on the inside of the masks. The count rate on the outside of the masks varied widely depending on the medical staff, and the measurement for one mask showed no detection of α-particles even on the outside of the mask.
The cumulative amount of exhaled 219 Rn from the patient via exhaled breath until 72 h after the start of administration (A Rn,cumulative ) was calculated as follows: where A Rn, breath is the amount of 219 Rn contained in the unit volume of exhaled breath, V breath the volume of breath per single respiration (500 mL/respiration), N respiration the respiration rate per min (12 respiration/min) and T the duration with the assumption that the blood concentration level immediately after the injection of 223 Ra is maintained (0.99 h). Using 90 ± 56 Bq/mL as A Rn,breath , which is the amount of 219 Rn contained in the breath at 1 min after the start of administration, A Rn,cumulative was calculated as 32 MBq. By referring to the release criteria for patients treated with 131 I, the amount of 219 Rn inhaled by caregivers (A Rn,inhaled ) was estimated using the following formula: where V daily is the daily respiration rate of the caregivers (20 m 3 /d), V room the volume of the room containing the patient (30 m 3 ) and N air change the air changes per hour (1/h); A Rn,inhaled was calculated as 0.89 MBq. The effective dose of 219 Rn to caregivers (D eff ) was estimated as follows: where E is the effective dose coefficient for the inhalation of 219 Rn (7.9 × 10 −9 mSv/Bq) and F the exposure factor (0.5 [7]); D eff was estimated to be 3.5 μSv per injection.
Discussion This study showed that the amount of 219 Rn contained in the breath of patients can be quantitatively measured using an HPGe detector based on the detection of the γ peak of the descendant nuclide 211 Bi and a breath collection bag. In addition, the internal radiation exposure of caregivers from 219 Rn was estimated to be as small as 3.5 μSv per injection. The high count rate of α-particles observed in the direct measurement of the breath of the patient by the α survey meter would be attributed to 219 Rn, not to the descendant nuclides of 219 Rn, exhaled through the respiration because the average radioactivity of 211 Bi measured using the HPGe detector was in the order of 10 Bq and the count rate of α-particles from 211 Bi in the measurement with α survey meter would be quite low compared with that from 219 Rn (in the order of 10 kBq). Although the high count rate of α-particles from 219 Rn exhaled through the respiration was observed using the α survey meter, the quantitative measurement of the descendant nuclide of 219 Rn using the HPGe detector in this study clearly shows that the effective dose of 219 Rn to caregivers is small. The amount of 219 Rn contained in the breath of a patient at 1 min after the start of administration varied widely, as is shown in Table 1.
The reason for this variation can be attributed to the peak shift in the exhalation of 219 Rn from the breath of patients caused by the manual administration of 223 RaCl 2 .
Since the effective dose coefficient for inhaled 219 Rn has not been reported, the value for 219 Rn was estimated based on the effective dose coefficient of 222 Rn (6.5 × 10 −6 mSv/Bq) [6]. α-Particles emitted from the descendant nuclides are thought to be the main contributors to the effective dose coefficient because their radiation weighting factor is very high compared with those of electrons and γ rays. Since the descendant nuclides of 219 Rn would decay before clearance because of their short half-lives, analogous to those of 222 Rn [8], and the total α-particle energy from the descendant nuclides of 219 Rn would be roughly the same as that of 222 Rn, the effective dose coefficient of 219 Rn can be estimated by correcting the effective dose coefficient of 222 Rn with its effective half-life. Using the biological half-life of 55 min for Rn [9] (effective half-lives of 3.95 s for 219 Rn and 54.5 min for 222 Rn), the effective dose coefficient of 219 Rn was estimated to be 7.9 × 10 −9 mSv/Bq (= 6.5 × 10 −6 mSv/Bq × 3.95 s/54.5 min).
For the mask measurements, the observed α-particles were considered to have originated from 211 Bi, which was in transient radioactive equilibrium with a 211 Pb parent (half-life 36.1 min) because the half-life of 219 Rn is very short. The N95 mask was capable of catching the descendant nuclides of 219 Rn, that had attached to dust or aerosol particles, on the surface of the mask, but it was not thought to be capable of blocking the inhalation of 219 Rn. If 219 Rn had passed through the N95 mask, α-particles from the descendant nuclides of 219 Rn that had decayed during and after passing through the mask would have been detectable on the inside of the mask. However, α-particles were not detected on the inside of the masks in the present study. Therefore, 219 Rn is considered to have not reached the mouths of the medical staff. The reason for the variation in the count rate observed for measurements made for the outside of the masks, as shown in Table 2, is probably due to the change in the amounts of 219 Rn and its descendant nuclides reaching the medical staff via the airflow in the injection room. The radioactivity of 211 Bi caught by the mask, as calculated based on the measured count rate of αparticles, is expected to correspond to less than 10 Bq at the measurement start time using the instrument efficiency of 35.1%/2π written in the calibration certificate of the survey meter and an α branch of 99.724% for 211 Bi. Although the radioactivity of the descendant nuclides of 219 Rn caught by the mask was relatively small, wearing the N95 mask appeared to protect the internal exposure of the medical staff.
The effective dose from the inhalation of 219 Rn experienced by caregivers was estimated to be 3.5 μSv per patient injection. This value is significantly lower than the dose of 5 mSv per episode for caregivers recommended by the International Commission on Radiological Protection (ICRP) [10,11]; therefore, the internal radiation exposure from 219 Rn should not affect the release criteria for patients who have been treated with 223 RaCl 2 in Japan.

Conclusions
The amount of 219 Rn contained in the breath of patients treated with 223 Ra was determined using γ-ray spectrometry with an HPGe detector and breath collection bags. Although a small peak for γ rays of 211 Bi originating from 219 Rn contained in the breath was observed, the estimated internal radiation exposure to caregivers via the inhalation of 219 Rn was small and should not affect the release criteria of patients who have been treated with 223 RaCl 2 .