The clear patient benefits of PRRT with 177Lu-DOTATATE have inevitably increased the number of interventions and studies on this isotope and its theragnostic couples, such as 68 Ga-DOTATOC [12, 26, 39]. However, there are still some challenges regarding its use, such as lack of trained personnel or fully standardised procedures [12], which may result in high doses to the NM staff. In consequence, this study aims to analyse the occupational exposure during PRRT-Lu, using both passive and active dosimetry.

The use of lead aprons in NM departments is not as clear-cut as in interventional radiology, existing both supporting and opposing arguments for their use [40], such as the time extension that their heavy weight may cause, their diminishing attenuating ability with increasing photon energies, or the potential production of Bremsstrahlung radiation [41]. Nevertheless, the latter is less important in the case of 177Lu than other therapeutic pure beta emitters, such as yttrium-90 (90Y), due to the lower energy of its beta spectrum [22]. On top of that, the shielding material with the lowest atomic number (PMMA) is the first in contact with the nuclide, so that the Bremsstrahlung production is minimised and only the remaining gamma radiation from the decay is stopped by the apron. Moreover, these photons (113 and 208 keV) are less energetic than those emitted by other radioisotopes used in positron emission tomography (PET), such as 18F (511 keV) or 68 Ga (1077 keV), for which the use of the apron is not recommended due to the low attenuation capacity at these energies. In our hospital, the use of lead apron was introduced after the first few sessions. Therefore, with only 3 sessions performed without apron, a thorough hypothesis testing was not possible. Still, it was found that the doses and dose rates received with the apron were substantially lower than without, while the treatment time remained the same. Sghedoni et al. [22] compared the physician's personal equivalent doses over and under the lead apron during labelling and administration of 177Lu-DOTATOC and 90Y-DOTATOC. They obtained that radiation was almost completely stopped by the apron for 90Y, as it is a pure beta emitter and PMMA was used first, whereas the 177Lu-gamma emission was just partially attenuated, so they suggested that its use during 177Lu procedures should not be considered essential. However, they used a 0.25 mm lead equivalent apron, while those used in our study were 0.5 mm, so their gamma-stopping efficacy is higher. Manogue et al. also recommend the use of a lead apron during 177Lu-PSMA-617 therapy [42]. Accordingly, this study suggests that the use of a 0.5 mm lead equivalent apron during 177Lu-DOTATATE administration is recommended, as we experienced dose rates and whole-body doses reduction without prolongation of time.

It has been obtained that the dose rate behind the lead shielding remains at a constant value about 0.4 µSv/h. However, it should be noted that the natural background radiation plays an important role in this measurement. A dose rate value of 0.2 µSv/h was measured before the beginning of each session, which is consistent with the natural gamma dose rate value in the south of Galicia (Spain) [43]. Therefore, staying behind the lead screen can be associated with a dose rate of 0.2 µSv/h. In addition, the fact that dose rates decrease with time near the vial and increase near the patient is expected, as they follow the flux of the radiopharmaceutical. Dose rates measured immediately after the infusion of 177Lu-DOTATATE at 1 m from the chest of the patient reported in the literature were found to range (on average) from 16.2 to 34.9 µSv/h [14, 19, 44], and 48 µSv/h for 7.4 GBq of 177Lu-PSMA-617 [27]. Our results are in line with these studies, although within a larger range. This can be explained by the variability that certain patient-related factors may introduce, such as the body mass index (BMI) as described by Bellamy et al. [17], and by possible variations in timing and distance at the moment of measurement.

Ambient dose rates were similar for both protocols except at the change of the infusion rate, which is expectable as the timing is different. Protocol 2 takes less time than protocol 1, so although the dose rates are higher in the 5 min approach, total time is shorter, which can explain why dose rates and normalised doses received by each worker from both protocols showed no significant differences. Nevertheless, the second protocol has been adopted in our hospital from now on, as we obtained no significant differences in radiological protection, it is faster and is the actual protocol recommended by the EMA [45].

The administration step involves the highest dose rates and accounts for the majority of the cumulative dose, followed by the end of treatment, so there is room for optimisation in these steps if further dose reduction is required. Also, approaches to the patient or vial beyond the lead screen should be reduced as much as possible, as doses were not found to be negligible. Although the dose rates for both groups of workers are approximately the same, the nurses exhibit slightly higher values than the physicians, except for N4, who also presents lower values than the other nurses. Firstly, it should be noted that the administration of therapeutic lutetium is a longer and more complex process than the administration of a diagnostic radiopharmaceutical, as accidents are more likely to occur and therefore different sessions can be very different from each other. The value of the cumulative dose and dose rates also depends on the behaviour of the individual worker, with some considering it a priority to be closer to the patient or vial to ensure that the liquid does not spill, and others considering it more important to perform the procedure quickly and keep as much distance as possible, which can also vary between different sessions, even if they are carried out by the same person. In addition, patient behaviour can also introduce a lot of variability, since if the patient experiences discomfort, they will call the staff more often (so there will be more approaches) and exposure will increase. Therefore, it is most likely that the difference in dose rates between N4 and the other nurses is due to variability between sessions, as well as in the number of sessions performed and their behaviour during them. There is no significant difference in the level of training between nurses and they have approximately the same years of experience. On the other hand, physicians are responsible for vial and needle manipulation during administration and at the end of administration, so they receive higher doses on their hands compared to nurses due to the contribution of short-range beta emission. However, as nurses are in charge of patient care and monitoring, they make more approaches to the vial/patient (as can also be seen in Fig. 5) and thus spend more time close to radioactive sources. It is therefore to be expected that doses and dose rates obtained with the PED, which relate to whole-body doses and do not take account of betas, will be higher for this group of workers.

For most measurements there is good agreement between Hp(10)/A measured with PED and OSL dosemeters, as in these cases the dose measured with PED are below the LDL of 50µSv of OSLs, as well as OSLs. Only two sets showed higher Hp(10)/A values with OSL (sets P1#2 and N2#2). However, these sets were used during the same period of the measurement campaign, and they exhibited very high background doses due to long time between reset and readout and potentially due to transit doses. Besides, the correction with a single background dosemeter is imprecise, so there is a significant chance that the dose is due to fluctuations in background doses. Finally, it should be noted the high value of Hp(0.07) obtained in set P1#4. The fact that Hp(0.07) is so high, but that no value was obtained for Hp(10), indicates that this is a dose from a beta field, which is not measured by PEDs, so it is likely to have been slightly contaminated. It was verified that the physician’s monthly personal dosemeters, used during the same period as the measurement campaign and also placed at chest level, did not show high doses of Hp(10), which reinforces the option of contamination. On the other hand, the results obtained with PEDs are consistent with other studies that have measured Hp(10), also with electronic dosemeters. In the study by Sghedoni et al. [22] the physician’s Hp(10) was obtained during administration of 177Lu-DOTATATE in similar conditions to the present study (i.e. with electronic dosemeters, under apron). They resulted in 6.6 [1–24] µSv, which normalised to the injected activity (an average of 5.5 GBq to 3–4 patients) is 0.40 [0.06–1.45] µSv/GBq, in line with our results with PEDs. Other studies also show similar values (0.36 ± 0.16 µSv/GBq [23] and 0.24 ± 0.05 µSv/GBq [14]), although no information is given on the use of aprons. A study involving administration of 177Lu-PSMA-617 obtained 0.60 ± 0.05 µSv/GBq for physicians behind a lead shield [46]. Additionally, in a previous study on Hp(10) doses during manipulation of 68 Ga-DOTATOC [47] it was concluded that electronic dosemeters could underestimate the dose due to the non-linear response of the Geiger-Müller detector at the 68 Ga photon energies (1077 keV) and the pure beta field. However, the linear response of PED is straighter in the range of the 177Lu-photon energies (208 keV) [48]. In addition, the fact that both detectors are placed under the apron also introduces uncertainty, since as demonstrated in previous studies [49] the response of passive dosemeters is more influenced than that of active dosemeters in the Hp(10) reading, so it would be interesting to perform double dosimetry (detectors above and below the apron). Since no literature was found on whole-body doses from 177Lu administration with OSL dosimetry, further measurements would be needed. These results also highlight the need for caution when comparing doses from different dosimetry systems, especially when the LDLs are different or unknown.

Extremity dosimetry is a concern in NM practices, but according to a recent review by Kollaard et al. [26], only a few publications have addressed finger exposure with novel radiopharmaceuticals, with only three articles on 177Lu found at the time [13, 22, 24]. These studies presented Hp(0.07)/A values ranging from 10 to 66 µSv/GBq, which are similar to those presented in our study for both physicians and nurses excluding outliers (8–70 µSv/GBq), being 45.2 [20–70.3] µSv/GBq the highest maximum normalised dose obtained for physicians and 15.4 [8.5–33.1] µSv/GBq for nurses. As seen from Fig. 6, doses in the ND hand are qualitatively higher than in the D hand for both physicians and nurses. However, according to Table 6, physicians presented higher maximum normalised doses in the D than in the ND hand. This can be explained because the calculations in Table 6 are very conservative, as they have been obtained from the mean maximum values of each worker, therefore providing an estimate of the worst-case scenario. In this case, the values of set #2 of P2 have been considered outliers (Additional file 1:Table S5), and the thumb position of the D hand (A) of set #1 for the same physician, although not counted as an outlier, also shows a very high value compared to the rest (70 µSv/GBq versus a range of 19–25 µSv/GBq for P1, as seen in Table 5). Therefore, this value counts as the highest for this worker, and carries much weight when calculating the median with the P1 value, which shows a mean maximum value in the D hand of 20.1 µSv/GBq, giving a median of 45.2 µSv/GBq in the D hand (Table 6 and/or Additional file 1: Table S5), higher than the ND hand. However, according to the individual values for each position (Additional file 1: Table S5), without considering the maximum doses for each worker, the ND hand shows median values ranging from 11 to 47 µSv/GBq and the D hand from 10 to 21.3 µSv/GBq for physicians, and 3–11 µSv/GBq and 6–11 µSv/GBq for the ND and D hand, respectively, for nurses. Therefore, it is observed that the dose deposition is higher in the ND hand than on the D. In addition, this difference is more evident for physicians as they are responsible for manipulating the vial, which is generally done with the ND hand, so that they can handle the rest of the equipment (needles, plunger of the syringe, etc.) with the D hand. These results match with the fact that physicians received statistically significant higher doses than nurses in almost all locations (Additional file 1: Table S4). Nevertheless, although doses are qualitatively higher on the ND hand, especially for physicians, no statistically significant differences were found between both hands for either physicians or nurses. In addition, no significant differences were found between doses on the D and ND for any location (Additional file 1: Table S3), although the tip of the thumb and index finger show higher values in both hands for both physicians and nurses. These results are in contrast to those obtained after the manipulation of other radiopharmaceuticals, such as 68 Ga [47], 18F or 99mTc, as shown in the ORAMED study [25, 38], in which the ND is significantly more exposed than the D hand. This is explained due to the direct handling of the vial/syringe and the pure beta emission of these isotopes. With the gravity method used to infuse 177Lu-DOTATATE, the vial is always inside the PMMA shield, so the physician does not need to hold the vial directly except in rare occasions, which explains why there is less difference between the D and ND hands than with other radiopharmaceuticals. Also, because of the dual emission of 177Lu, in contrast to the pure beta emission of 68 Ga or 18F, the dominant hand is also exposed to the gamma radiation. In addition, the administration of these diagnostic radiopharmaceuticals is faster and entails few steps, taking no more than 2–3 min of vial or syringe handling. This is opposed to the administration of therapeutic 177Lu, which is a more complex and irregular administration, that can take up to 30–40 min and entails several steps in which workers should approach radioactive source, and therefore are more prone to accidents, such as extravasation of radioactive liquid from the vial, patient movements during administration, vomiting, etc. This aspect not only introduces greater variability between the results of different sessions and workers, which is why the doses received at each position show a wide interquartile range (Fig. 6), but also leads to a more homogeneous exposure between the D and ND hand compared to the administration of a diagnostic radiopharmaceutical.

It was found that the wrist and ring CND dosemeters underestimated the maximum dose values. Nevertheless, this underestimation is expected, as they are located on the wrist and at the base of the ring finger, so it does not mean that these dosemeters perform poorly. In fact, the values recorded with the rings resulted similar to the values recorded with the TLDs located at the same position (e/E) in almost all the dosemeter sets. This is also reflected in the value of the CFs (Table 7) which is similar for the ring and positions e/E. In the case of the wrist detectors, the CFs show larger differences, as expected, except for the nurse’s D hand. Thus, as set out by other studies, for measuring extremity doses in NM the use of ring over wrist dosemeter is recommended [25, 47].

The CFs, or dose ratios, were calculated at the base of the middle (d/D) and ring (e/E) fingers, as these are the most common positions for placing a ring dosemeter. The calculated median CFs suggest that physicians should correct the middle and ring doses by at least a factor of 5 and 6 respectively, and nurses by a factor of 3 and 4 respectively, preferably located on the ND hand. To the best of our knowledge, and according to Kollaard et al. [26] only one publication has reported on dose ratios [24], obtaining a value of 1.6, which is very different from our results. However, this value was calculated by averaging the dose ratios over the three measured fingers (thumb, index and middle), which may underestimate the maximum CFs. In addition, the estimated values are similar to those generally recommended in the literature for the assessment of maximum finger doses in nuclear medicine [50].

It should be noted that, compared to the ORAMED project [25, 38, 51], the position of a TLD at the base of the index finger was not taken into account in this study, a factor that could have potentially yielded more detailed insights into the dose distribution over the hands. However, it is noteworthy that in everyday clinical practice, the ring dosemeter is not exclusively worn on the index finger. In fact, through an ongoing measurement campaign within the SINFONIA project in various European hospitals, the base of the middle finger was one of the observed positions, also consistent with our hospital's practice. Subsequently, during the same campaign a comparison of the dose at the base of the middle and ring finger to the dose at the base of the index finger of the non-dominant hand was performed. It was obtained that the ratio index/ring finger showed a mean of 1.2 [0.6–2.2] and the ratio index/middle a mean of 1.1 [0.6–1.9]. This implies that the mean difference between these two monitored positions and the base of the index finger is approximately 10–20%, falling within the range of 0.60–2.2. Therefore, although it could have provided additional insights on the dose distribution, the absence of TLD data from the base of the index finger is relatively inconsequential due to minor dose variations among the base of these three fingers, underscored by the fact that the index finger is not the exclusive position where the ring dosemeters are worn in routine clinical practice.

Regarding doses to the eyes in terms of Hp(3)/A, a median of 2.02 [1.84–2.20] µSv/GBq and 1.76 [1.00–2.53] µSv/GBq was found for physicians and nurses, respectively. No significant differences were found between left and right eyes, although given the isotropic nature of gamma emission, it is expected. No studies on the estimation of Hp(3), especially from 177Lu management, were found to compare our results, so further measurements are needed to validate them, as there may be large inter-worker variability. Nevertheless, these first measurements indicate that there is no large concern to reach the annual eye dose limits.

Finally, annual dose estimates were made in terms of the maximum number of sessions and patients expected to be treated within safe dose limits, wearing a 0.5 mm lead equivalent apron. Outliers were also included in these calculations, as they inevitably represent the potential hazards faced by the staff in real treatments, such as cross-contamination or unexpected irradiation. With outliers, the most restrictive value was found to be a limit of 59 sessions/year for P2 due to skin dose, followed by 100 sessions/year for P1 due to eye dose and to 197 for P1 due to Hp(10)/A recorded with OSLs. Assuming that a patient requires 4 sessions of 7.4 GBq each, administered 8 weeks apart in the same year, the annual limit of patients per worker according to these values would be 15, 25 and 49 due to Hp(0.07), Hp(3) and Hp(10), respectively. Without outliers and OSL Hp(10) measurements, the session (patient) limit per worker would increase up to 965 (241), limited by the skin dose. In our hospital 21 sessions were performed in 2022 (12 by P1 and 9 by P2), involving 7 patients, so even accounting with outliers, dose limits were not reached due to 177Lu treatments. Nevertheless, the number of annual patients in other hospital has been found to be larger in the literature, from 11 to 25 patients per worker [4, 24], so the risk of overexposure should be considered in these cases. Besides, in out hospital 177Lu-DOTATATE is received in individual patient doses, so only administration is monitored (from pre to post activity verification), while in other hospitals also preparation and dispensing could be performed in-house. On top of that, the real situation would be more complex as these estimations only refer to 177Lu-DOTATATE treatments and usually one worker would perform several procedures with multiple radionuclides. In addition, as treatments with 177Lu-PSMA-617 are expected to start in the near future, exposure to 177Lu will increase and must be taken into account. It is therefore particularly important to ensure that staff dosimetry is adequate to ensure that ICRP limits are not exceeded.

In addition, it is worth noting that the measured activity of each vial averaged 7121 ± 105 MBq, ranging from 6808 to 7289 MBq, which is less than the intended activity of 7400 MBq/cycle. This difference can be explained because all radiopharmaceuticals, including Lutathera vials, are distributed regionally from a central radiopharmacy to hospitals. In the studied hospital, Lutathera vials are shipped from the factory to the radiopharmacy one day before treatment, then transported to the hospital on the treatment day without any handling at the radiopharmacy. Although the process is carefully planned so that the vial reaches the hospital with an activity level of 7400 MBq at the time determined by the treating physician, unforeseen circumstances and external factors beyond the control of the staff, such as delays in delivery or patient arrival, can lead to small variations in the vial's activity level, causing a 4–5% variance from the target of 7400 MBq.

This study has thoroughly analysed the doses received by workers administering 177Lu-DOTATATE and has shown that its results are consistent with the limited existing literature. However, it presents some limitations. Firstly, there is no information on post-administration dose rates (4, 6 or 24 h after administration), which would be of great interest as staff may also approach the patient during this period. However, based on other studies, these doses are rather low [14, 23]. It should be noted that the inpatient basis was chosen for internal organisation and logistic reasons, but it would be possible to perform the administration on an outpatient basis, as suggested by other authors [14, 52], under the requirement that the patient returns within 24 h of administration for the scan and given that it complies with RP measures. The relationship between the usual ring and wrist dosemeters and TLDs was also investigated, but the results would have been more robust if they had been worn on either the D or ND hand, rather than both. However, this reflects the reality of staff dose monitoring, as workers often wear the detectors in different positions. On the other hand, the effect of the lead apron was studied by comparing dose rates and whole-body doses received with and without the apron using the real-time dosemeter, but a more complete analysis would have required double dosimetry (one dosemeter above and one below the apron), as has been done in other studies [49]. However, this approach was not possible as only one electronic dosemeter was available for the physician and one for the nurse, as well as only one OSL in each set used. The number of sessions performed without apron was also limited, so as mentioned before, a thorough hypothesis testing was not possible. In addition, the study covers data from 2 physicians and 4 nurses, but a larger sample size could provide more statistical power and add information on those dosemeters for which few results were obtained, such as OSLs. Finally, some values were reported as outliers and attributed to possible cross-contamination, but it is not possible to be sure that this is indeed the reason for these high values. Therefore, the effect of cross-contamination should be investigated in the future and compared with the values obtained in this study.