Use of Molecularly-Cloned Haematopoietic Growth Factors In

Blood Reviews xxx (xxxx) xxxx

Contents lists available at ScienceDirect

Blood Reviews

journal homepage: www.elsevier.com/locate/blre

Review Use of molecularly-cloned haematopoietic growth factors in persons exposed to acute high-dose, high-dose rate whole-body ionizing radiations ⁎ Robert Peter Galea, , James O. Armitageb a Centre for Haematology, Department of Immunology and Inﬂammation, Imperial College London, London, UK b Department of Medicine, Division of Hematology and Oncology, Univ. Nebraska, Omaha, NE, USA

ARTICLE INFO ABSTRACT

Keywords: Exposure to acute, high-dose, high dose-rate whole-body ionizing radiations damages the bone marrow resulting Radiation in rapid decreases in concentrations of blood cells, especially lymphocytes, granulocytes and platelets with Nuclear associated risks of infection and bleeding. In several experimental models including non-human primate ra- Filgrastim diation exposure models giving molecularly cloned haematopoietic growth factor including granulocyte/mac- Sargramostim rophage colony-stimulating factor (G/M-CSF; sargramostim) and granulocyte colony-stimulating factor (G-CSF; Accidents filgrastim and pegylated G-CSF [peg-filgrastim]) accelerates bone marrow recovery and increases survival. Based on these data these molecules are US FDA approved for treating victims of radiation and nuclear incidents, accident and events such as nuclear terrorism and are included in the US National Strategic Stockpile. We discuss the immediate medical response to these events including how to estimate radiation dose and uniformity and which interventions are appropriate in different radiation exposures settings. We also discuss similarities and differences between molecularly cloned haematopoietic growth factors.

1. Introduction and bone marrow fitness (ability to respond to damage) before exposure. Conditions affecting prior bone marrow fitness include age, Exposure to acute, high-dose, high dose-rate whole-body ionizing health and exposure to bone marrow damaging drugs, chemicals and radiations damages the bone marrow resulting in rapid decreases in ionizing radiations. concentrations of blood cells, especially lymphocytes, granulocytes and At acute, uniform whole-body single-fraction doses < 2 Gy at dose- platelets with associated risks of infection and bleeding. Bleeding risk rates < 5 cGy/min most exposed persons have only modest decreases in can be decreased by platelet-transfusions and possibly by giving mo- blood cell concentrations which are without important immediate lecularly-cloned thrombopoietin (elthrombopag). Infection risk can be medical consequences and usually need no medical intervention. In decreased by giving prophylactic antibiotics and anti-fungal and -viral contrast, at acute doses > 12 Gy, especially at high dose-rates, most drugs and by possibly by microbial decontamination and a protected exposed persons will die immediately from neuro-vascular effects. At environment. However, because granulocyte transfusions are in- intermediate doses (> 2- < 12 Gy) damage to the bone marrow and effective, the only ways to increase blood granulocyte concentrations is gastro-intestinal tract dominate the immediate clinical picture. by giving molecularly-cloned myeloid growth factors (granulocyte/ Survivors of these acute effects will face delayed effects of acute ra- macrophage- and granulocyte-colony stimulating factors; G/M- and G- diation exposure (DEARE) including damage to the lungs, kidneys and CSFs) or by performing a haematopoietic cell transplant. heart. Our focus is on immediate medical interventions in persons ex- The rate and severity of decreases in blood cell concentrations fol- posed to 2–12 Gy total-body or partial-body with bone marrow lowing exposure to ionizing radiations depends on several radiation- shielding radiations at moderate dose-rates. A scheme of dose-based related variables such as type of radiation (X- or γ-rays, α-orβ-parti- toxicities and interventions in response to these exposures is displayed cles, neutrons, high-speed electrons and protons), dose, physical-che- in Fig. 1. Detailed discussions of interventions for DEARE are reviewed mical form (i.e. gas, particle, liquid), field and uniformity (i.e. whole-, elsewhere. [1,2] subtotal- or partial-body), dose-rate, route of exposure (i.e. external The above dose-range and rate recommendations are based mainly versus internal or combined), fractionation and other source-term on immediate or fall-out situation exposures to gamma and photons (X- parameters. Host-related variables are also important such as age, sex ray). However, in many emergency exposure scenarios such as an

⁎ Corresponding author at: Centre for Haematology, Department of Immunology and Inﬂammation, Imperial College London, London SW7 2AZ, UK. E-mail address: [email protected] (R.P. Gale). https://doi.org/10.1016/j.blre.2020.100690

Please cite this article as: Robert Peter Gale and James O. Armitage, Blood Reviews, https://doi.org/10.1016/j.blre.2020.100690 R.P. Gale and J.O. Armitage Blood Reviews xxx (xxxx) xxxx

Exposure to neutrons can be determined by analyses of blood and urine samples for neutron capture (typically 24‑sodium, 42‑potassium and 82‑barium) by gamma-ray spectroscopy. [12] However, physical measurements are technically demanding and often not immediately available to help estimating dose in accidental exposure scenarios especially if there are large numbers of victims and disrupted infrastructure or when technically competent persons are unavailable, in- jured or dead. Biological dosimetry can be done on blood or bone marrow samples including analyses of rates of decline of lymphocytes, granulocytes and platelets, di-centric chromosomes, micro-nuclei, premature chromosome condensation and gamma H2AX focus and chromosome painting. [13] In a mass casualty event the temporal development of nausea can be used to identify persons exposed to a dose > 2 Gy but there are many Fig. 1. Diagram of acute, high-dose and dose-rate whole-body radiation-in- false-positives, especially in the context of concurrent exposures to duced sub-syndromes and potential interventions. GI, gastro-intestinal; CNS/ chemicals and combustion products. Terrorists could synchronously CV, central nervous system/cardio-vascular. expose people to a nausea-inducing agent to increase perception of the magnitude of a radiation-related event. Computational-based dose re- improvised nuclear device (IND), immediate exposures may also in- constructions use source/dispersion models but require substantial time clude neutron radiations of variable energies. If so, the above dose and effort, have wide confidence interval and is rarely victim-specific. ranges should be adjusted to account for the increased relative biolo- Point-estimates of dose using these approaches are often inaccurate gical effect (RBE) of neutrons compared with photons or γ-rays. Also, with wide confidence intervals and/or credibility limits with discordant our focus is on external exposures. However, in many accident or nu- estimates from different techniques. [14,15] Data available to haema- clear terrorism scenarios there may be internal exposures such as from tologists may be sufficiently accurate for triage but not for critical ingestion of drinking water intentionally contaminated with a radio- therapy decisions such as giving molecularly-cloned haematopoietic nuclide from a radiation dispersion device (RDD), from handling a ra- growth factors or doing a haematopoietic cell transplant. It is important diation exposure device (RED) or from fallout from an IND. Therapy of haematologists understand the inaccuracy and imprecision of point- internal exposures is reviewed elsewhere. [3,4] In acute exposure set- estimates of dose in emergency settings which drives evaluation of tings damage from alpha particles is typically external and rarely life- benefit-to-risk ratios of potential interventions. It is wise to consider threatening. point-estimates as a best dose-estimate and understand this estimate may Two interventions given to accelerate granulocyte recovery after change substantially as new data are acquired. It is also important for exposures of 2–12 Gy photon/gamma exposure (or 1–7 Gy photon/γ- haematologists consider the 95 or 99% confidence interval around the ray and neutron exposure) are using molecularly-cloned haemato- best dose-estimate. In many accident settings the lower boundary of the poietic growth factors and/or performing a haematopoietic cell trans- best dose-estimate may be compatible with no intervention whereas the plant. Elsewhere we discuss why few if any exposed persons are ap- upper boundary may be compatible with an intensive intervention. propriate candidates to receive a haematopoietic cell transplant. [5] Uncertainties in dose-estimates and the relationship between point-es- The fundamental limitation is the unfavourable benefit-to-risk ratio of a timates, confidence intervals and credibility limits are displayed in transplant in this setting and the observation that no otherwise survi- Fig. 2. vable dose of acute whole-body radiation causes irreversible bone Consequences of inaccuracies and imprecision in dose-estimates marrow damage. [6] Furthermore, the likely scenario for radiation vary. For some interventions such as giving prophylactic oral anti- exposure from an terrorist attack (see below) is that it will be ill-de- biotics, an inaccurate best dose-estimate may be inconsequential. This is fined, heterogenous and non-uniform in the immediate and short-term less so for parenteral drugs, intravenous antibiotics, RBC- and platelet- fallout area. For most exposed persons there will be a non-uniform dose transfusions and molecularly-cloned haematopoietic growth factors. distribution to the body with sparing of some bone marrow areas. There is little tolerance for an inaccurate or imprecise best dose-estimate Consequently, our focus is on using molecularly-cloned haematopoietic in the context of a contemplated haematopoietic cell transplant. growth factors combined with other medical interventions. Because the Another important variable affecting the therapy metric is dose- major risk of death is from infection caused predominately by a low uniformity. In most emergency exposure scenarios it is unlikely there concentrations of blood granulocytes, we focus on use of drugs accel- will be uniform whole-body exposures. At doses > 10 Gy, shielding of erating granulocyte recovery, namely, granulocyte/macrophage- (sar- as little as 2.5 to 5% of the bone marrow is needed for a molecularly- gramostim and molgramostim; G/M-CSF) and granulocyte-colony stimulating factors (filgrastim and peg-filgrastim; G-CSF). A comprehensive 3-part review with detailed discussions of many topics we cover is available as are other typescripts. [7–10]

2. Estimating dose and dose-uniformity

Eﬀective therapy of persons exposed to acute high-dose whole- or partial-body ionizing radiations requires accurate estimates of dose and dose-uniformity. There are several complementary approaches to estimating dose including physical, biological and computational dosimetry. Physical dosimetry relies on devices such as radiation exposure badges, electron paramagnetic resonance (EPR) analyses of dental en- amel and clothes made from organic ﬁbers such as cotton, electron paramagnetic resonance spectroscopy of phone display glass and opti- cally-stimulated luminescence in smartphone resistors. [11,12] Fig. 2. Uncertainty in dose-estimates.

2 R.P. Gale and J.O. Armitage Blood Reviews xxx (xxxx) xxxx cloned haematopoietic growth factor such as filgrastim to be effective and makes the benefit-to-risk ratio of a haematopoietic cell transplant less attractive. [16–18] One should not assume persons exposed to uniform high-dose whole-body radiation cannot recover and need a haematopoietic cell transplant. These persons can recover bone marrow function albeit slowly and perhaps incompletely (see below).

3. G-CSF and GM-CSF signaling pathways and functional consequences

Receptors for sargramostim and filgrastim belong to the cytokine receptor super-family. The G-CSF receptor (G-CSFR) is a homo-oligo- dimer whereas the GM-CSF receptor (G/M-CSFR) is a hetero-oligo- dimer with a shared β-chain with the interleukin-3 and -5 receptors (Fig. 3). The G-CSFR is expressed primarily on neutrophils and bone Fig. 3. G- and G/M-CSF receptors (Adopted from [45]). marrow precursor cells. GM-CSF binds to the α-chain of the G/M-CSFR. ff The G/M-CSF-R is more widely-expressed than the G-CSF-R. Di erences rapid granulocyte recovery is considered supporting data. There are few ff in receptor expression account for most of the functional di erences direct comparisons of G/M- and G-CSFs and none in a monkey radiation fi between sargramostim and lgrastim (Fig. 4). These data were recently model. Studies in monkeys using bone marrow suppressive, potentially reviewed. [19] lethal radiation doses suggest roughly equivalent effects. In a study in humans receiving an allogeneic haematopoietic cell transplant G/M- fi 4. Sargramostim and lgrastim production CSF alone or combined with G-CSF was better than G-CSF alone. [23] Data from comparative studies suggest comparable safety and efficacy. ff There are fundamental di erences in some properties between [24] However, data in monkeys suggest use of sargramostim is asso- fi sargramostim and lgrastim. Sargramostim is produced in a yeast ciated with a greater incidence of non-infection-related fevers com- (Saccharomyces cerevisiae) expression system with glycosylation at the pared with filgrastim. Most studies in humans indicate similar rates of fi O-terminus. This contrasts to lgrastim which is produced in E. coli and granulocyte recovery and of infection which is not surprising based on ff is not glycosylated. Because these proteins bind to di erent receptors a the broader spectrum of sargramostim activity and on probable off- ff direct comparison of consequences of these di erences is not relevant. target effects. Sargramostim and filgrastim seem comparable when used ff Some data indicate di erences in biological activity when GM-CSF is to collect blood cells for a haematopoietic cell transplant. However, in ff ff di erently processed. However, few data suggest these di erences re- one randomized study giving filgrastim or sargramostim followed by ff ffi fi sult in di erences in safety or e cacy of sargramostim or lgrastim in filgrastim seems better than sargramostim alone when given to collect fi most clinical settings. [20] Curiously, use of lgrastim in most hae- blood cells for an autotransplant. [25] It is unlikely this difference is matology and oncology settings is substantially greater than that of important in the context of accidental high-dose radiation exposures. ff sargramostim. Reasons for this di erence are complex (see below). One potentially important difference between sargramostim compared with filgrastim and peg-filgrastim in the context of a radiation ffi fi 5. E cacy of sargramostim and lgrastim accident relates to conditions for storage. Labeling instructions for both drugs recommend refrigeration. However, unpublished data indicate ffi There are several relevant data sources for evaluating the e cacy of sargramostim can be stored for several months without refrigeration interventions with these drugs including: (1) humans after high-dose without losing efficacy1. Whether this is so for filgrastim or peg-fil- anti-cancer drugs, radiation therapy or both; (2) experimental radiation grastim is unknown. models of acute radiation syndrome, typically in mice, mini-pigs, dogs Based on these data from clinical trials sargramostim, filgrastim and and monkeys; and (3) humans with sustained low concentrations of pegfilgrastim are FDA-approved in overlapping but slightly different neutrophils or granulocytes. Data from humans after accidental radia- indications. Sargramostim is approved in persons with acute myeloid ffi tion exposures are di cult to interpret critically (discussed below). It is leukaemia (AML), lymphomas and acute lymphoblastic leukaemia important to understand there are no randomized trials of sar- (ALL) receiving an autotransplant, recipients of allotransplants, trans- fi fi gramostim, lgrastim or peg- lgrastim compared with no intervention plant recipients with delayed bone marrow recovery and to facilitate or placebo in accident settings and no comparative studies in animals collection of blood cells for a haematopoietic cell transplant. (including sub-human primates) or humans after accidental high-dose Filgrastim and pegfilgrastim are approved for use in persons with radiation exposures. solid cancers receiving intensive anti-cancer therapy, after intensive therapy of AML, to facilitate collection of blood cells for a haemato- 6. Bone marrow failure after cancer therapy poietic cell transplant and in persons with severe congenital or idio- pathic neutropenias. These differences have little impact on our dis- fi Data on the use of sargramostim and lgrastim in the context of cussion of use of these drugs in the context of a radiation accident. bone marrow failure are summarized elsewhere. [21,22] Briefly put, these drugs accelerate granulocyte recovery and reduce the frequency 7. Animal models of radiation-induced bone marrow failure of infections correlated with low blood granulocyte concentrations. fi Sargramostim has a much wider target cell range compared with l- Considerable data indicate sargramostim and filgrastim accelerate fi grastim and peg lgrastim (Fig. 4). Few clinical studies were structured recovery of granulocytes and improve survival of animals exposed to or powered to determine whether disease-free or progression-free sur- high doses of ionizing radiations under experimental conditions in- vivals or survival were increased. There are no randomized trials di- cluding rodents, dogs mini-pigs and monkeys. [7–10] fi rectly comparing sargramostim (or sometime molgramostim), l- In most models the endpoint of these interventions is the mid-lethal grastim or peg-filgrastim in this setting. This limitation is important dose (LD50) at 30 or 60 days (LD50/30-LD50/60). A comprehensive when we consider FDA approval of these drugs for accidental radiation exposures is based on data from studies in non-human primates where the primary, relevant clinical endpoint is survival and where more 1 Partner Therapeutics Medical Information Group.

3 R.P. Gale and J.O. Armitage Blood Reviews xxx (xxxx) xxxx

Fig. 4. Sites of action of diverse cytokines and haematopoietic growth factors (Adopted from [24]). comparison of experimental conditions and endpoints in diverse animal incidence of documented infections. A discussion of mathematical models is available. [7–10,26] models comparing data from monkey radiation studies and data from Our focus in this review is on models of acute radiation exposure in accidental radiation exposure in humans are reviewed in references 28 non-human primates (Summarized in Table 1). There are several im- and 29. portant variables in this setting including dose, dose-rate, energy, whether the exposure was whole-body or a fraction of the bone marrow 8. Filgrastim and pegfilgrastim in monkeys was shielded (partial-body exposure), haematopoietic growth factor given, interval between exposure and haematopoietic growth factor use In a series of studies monkeys received 7.5 Gy whole-body radiation and extent of additional support interventions such as parenteral fluids, (6 MeV, linear photons delivered at 80 cGy/min) or 10 to 11 Gy whole- oral or systemic antibiotics given prophylactically and/or ther- body radiation with 2.5 or 5% bone marrow shielding designed to more apeutically and RBC- and/or platelet-transfusions. closely mimic what might happen in a nuclear terrorism attack. Knowledge of these variable are critical in interpreting safety and Animals then received filgrastim or peg-filgrastim at diverse doses and efficacy of post-exposure haematopoietic growth factors. For example, schedules starting on day-1 or day-3 post-exposure. All animals re- kinetics of neutrophil recovery in non-human primates receiving 10 Gy ceived intensive supportive care. These interventions accelerated whole-body radiation versus radiation shielding 2.5 or 5% of the bone granulocyte recovery however survival was increased compared with marrow are shown in Fig. 5.[18] Shielding about 5% of the bone controls in only the 7.5 Gy studies [16–18,30,31]. marrow increased the LD50/60 dose from about 7.5 Gy to about 11 Gy, a These data indicate giving filgrastim or peg-filgrastim after a ra- 60% increase. These values reflect use of moderate supportive care diation dose > 10 Gy accelerates neutrophil recovery only when a small given because the shielded bone marrow is exposed to approximately amount of bone marrow is shielded. These studies were not powered to 0.5 Gy photon radiation. [27] determine whether there was a survival benefit. This observation is It is also evident that with 2.5 or 5% bone marrow shielding there is important in the context of radiation accidents or nuclear terrorism bone marrow recovery with no medical intervention save supportive where dose-uniformity is often difficult or impossible to determine and care measures. This is like our experience in humans receiving even likely to be non-uniform. Also, these interventions are most effective higher radiation doses. [6]Effects of sargramostim on neutrophil re- when given immediately post-exposure but there may be some efficacy covery and survival at doses > 10 Gy with or without bone marrow after intervals ≥3 days. This is another critical issue as it is unlikely shielding are unstudied. It is important to consider that improvement in these drugs could be given soon after a mass casualty setting for several some surrogate endpoints do not always correlate with clinical benefit. reasons including delays accurately estimating radiation-dose and in- For example, in the monkey partial shielding radiation model filgrastim frastructure disruption. Lastly, one study in monkeys with no post-ex- accelerated neutrophil recovery and decreased incidence of febrile posure antibiotics or blood transfusions reported no survival benefitof neutropenia but did not significantly improve survival nor decrease giving filgrastim to monkeys exposed to 7 Gy whole-body radiation

4 ..Gl n ..Armitage J.O. and Gale R.P.

Table 1 Monkey studies of molecularly-cloned haematopoietic growth factors.

Intervention, dose and interval N (intervention) N (control) Field/dose Supporta Day-60 survival vs. Days to neutrophils ≥ 1x10E+9/L vs. Days to platelets ≥ 1x10E+9/L vs. Infection rate vs. Study post-radiation Control (P-value) Control (P-value) Control (P-value) Control (P-value)

GM-CSF Sargramostim 48 h 36 36 TBI 6.55 Gyc Moderate 78% vs. 42% Median Day 18 vs. 20 (P = .0001) Median Day 16 vs. 18 (P = .0008) 32% vs. 63% [33] 7 μg/kg/day (P = .0018) (P < .0001) 18 18 7.13 Gyc 61% vs. 17% Median Day 17 vs. 19 (P = .0206) Median Day 16 vs. 20 (P = .0002) 37% vs. 84% (P = .0076) (P < .0001) Sargramostim 48 hb 88 88 TBI 7.13 Gyc Moderate 28% vs. 10% Mean Day 18.8 vs. 21.2 (P ≤ .001) Mean Day 16.6 vs. 17.4 (P ≤ .05) 68% vs. 82% (NS) [34] 7 μg/kg/day (P = .0032) 72 h 44 44 25% vs. 14% (NS) Mean Day 18.7 vs. 21.2 (P ≤ .001) Mean Day 17.5 vs. 17.4 (NS) 80% vs. 82% (NS) 96 h 44 32% vs. 14% (NS) Mean Day 18.8 vs. 21.2 (P ≤ .001) Mean Day 17.6 vs. 17.4 (NS) 55% vs. 82% (NS) 120 h 44 16% vs. 14% (NS) Mean Day 19.0 vs. 21.2 (P ≤ .001) Mean Day 17.4 vs. 17.4 (NS) 84% vs. 82% (NS)

G-CSF Filgrastim 24 h 24 22 TBI 7.5 Gye Intensive 79% vs. 41% Median Day 20 vs. 24.5 (P < .0001) Median Day 21 vs. 24 (P = .062) 58% vs. 86% (NR) [30] 10 μg/kg/day (P < .004) Filgrastim 48 h 40 40 TBI 7.5 Gye Intensive 52% vs. 50% Median Day 19 vs. 23 (P < .0001) Median Day 20 vs. 21.5 (P = .0289) 67.5% vs. 67.5% [46] 10 μg/kg/day (P = .82) (NR) Filgrastim 24 h 26 36 TBI 7 Gyd Moderate 31% vs. 36% NR NR 73% vs. 61% (NR) [32] 10 μg/kg/day (P = .4389) Pegﬁlgrastim 24, 192 h 23 23 TBI 7.5 Gye Intensive 91% vs. 48% Median Day 20 vs. 35 (P < .0001) Median Day 20 vs. 30 (P < .0001) 65% vs. 80% [31] μ

5 300 g/kg (P = .0014) (P = .29) Filgrastim 24 h 7 15 PBI/BM5 10 Gye Intensive 71% vs. 72% (NR) Mean Day 9.5 vs. 24.2 (NR) Mean Day 15.8 vs. 17.6 (NR) 0% vs. 29% (NR) [16] 10 μg/kg/day 72 h 8 87% vs. 72% (NR) Mean Day 12.4 vs. 24.2 (NR) Mean Day 16.5 vs. 17.6 (NR) 0% vs. 29% (NR) 24 h 5 22 11 Gye 40% vs. 59% (NR) Mean Day 10.0 vs. 23.2 (NR) Mean Day 12.8 vs. 19.8 (NR) 0% vs. 57% (NR) 72 h 8 75% vs. 59% (NR) Mean Day 12.1 vs. 23.2 (NR) Mean Day 14.7 vs. 19.8 (NR) 50% vs. 57% (NR) 120 h 8 62% vs. 59% (NR) Mean Day 13.0 vs. 23.2 (NR) Mean Day 14.8 vs. 19.8 (NR) 0% vs. 57% (NR) Filgrastim 24 h 7 15 PBI/BM5 10 Gye Intensive 71% vs. 73% (NR) NR NR NR [17] 10 μg/kg/day 72 h 8 87% vs. 73% (NR) NR NR NR 24 h 5 21 11 Gye 40% vs. 62% (NR) NR NR NR 72 h 8 75% vs. 62% (NR) NR NR NR 120 h 8 62% vs. 62% (NR) NR NR NR Filgrastim 24 h 8 12 PBI/BM2.5 10 Gye Intensive 50% vs. 42% (NR) Mean Day 20.5 vs. 27.0 (NR) Mean Day 29.6 vs. 27.0 (NR) 34% vs. 50% (NR) [18] 10 μg/kg/day 72 h 8 50% vs. 42% (NR) Mean Day 22.2 vs. 27.0 (NR) Mean Day 28.6 vs. 27.0 (NR) 75% vs. 50% (NR) Pegﬁlgrastim 24, 192 h 8 50% vs. 42% (NR) Mean Day 17.1 vs. 27.0 (NR) Mean Day 23.4 vs. 27.0 (NR) 50% vs. 50% (NR) 10 μg/kg/day 72, 240 h 8 50% vs. 42% (NR) Mean Day 20.3 vs. 27.0 (NR) Mean Day 22.2 vs. 27.0 (NR) 80% vs. 50% (NR)

Abbreviations: NR, not reported; NS, not signiﬁcant; PBI/BM2.5, PBI/BM 5, partial-body radiation with 2.5% or 5% bone marrow shielding; TBI, total-body irradiation. a Supportive care with prophylactic and symptomatic interventions in all studies. Moderate supportive care regimens provided prophylactic anti-emetics, antibiotics and anti-ulcer drugs to all monkeys. Based on clinical considerations these symptomatic interventions may also have been given: analgesics, parenteral ﬂuids, anti-emetics and oral nutrition. In addition to the moderate supportive care regimens, intensive supportive care regimens include the following symptomatic interventions: anti-diarrheals, −pyretics and antibiotics based on individual blood cultures and whole blood transfusions. b

Study-design included additional cohorts of monkeys treated at 48 h given prophylactic azithromycin with or without sargramostim. Adding azithromycin had no impact on any eﬃcacy endpoint. Survival data above Blood Reviewsxxx(xxxx)xxxx is pooled analysis of 176 monkeys in the intervention and control arms treated 48 h post-radiation. Day of recovery and infection rate data does not include cohort of monkeys given azithromycin. − c Dose rate 0.5 Gy min 1. − d Dose rate 0.6 Gy min 1. − e Dose rate 0.8 Gy min 1. R.P. Gale and J.O. Armitage Blood Reviews xxx (xxxx) xxxx

Fig. 5. Neutrophil concentrations in monkeys exposed to 10 Gy TBI (total body radiation) or 10 Gy PBI (partial body radiation) shielding 2.5 or 5% of the BM (bone marrow). [18]. from a 60Co gamma source. [32] reasons: (1) few subjects; (2) uncertainties regarding radiation dose, dose-rate and uniformity; (3) diverse intervals from exposure to starting therapy; (4) concomitant interventions; (5) confounding non-radiation 9. Sargramostim in monkeys injuries; and, most importantly, (6) no controls. However, most in- vestigators believe giving these drugs is useful in persons exposed to Non-human primates were exposed to acute whole-body ionizing – radiation using a 60Co gamma source at a 50–60% lethal-dose at day 60 2 12 Gy acute, whole-body radiation by accelerating granulocyte recovery, decreasing infections and possibly increasing survival. We used (LD50–60/60)or70–80% lethal-dose (LD70–80/60). [33] Prophylactic antibiotics were given without adjustment for blood culture data. No sargramostim to treat victims of a radiation accident in Brazil exposed to estimated doses of 2 to 7.3 Gy from a 131Cs. Our impression was that blood transfusions were given. The primary endpoint was day 60 sur- ff vival. Monkeys received daily subcutaneous sargramostim (7 μg/kg/d) this intervention was e ective but we lacked controls. [36] These data are shown in the Fig. 6. or control beginning 48 h post-radiation. Sargramostim significantly increased day 60 survival for the LD50–60/60 dose to 78% (95% confidence interval [CI], 61, 90%) versus 42% (26, 59%; P = .018) in 11. FDA animal rule controls. Day 60 survival for LD70–80/60 dose was 61% (36, 83%) versus 17% (4, 41%; P = .0076). Neutrophil and platelet recoveries were The FDA Animal Rule applies to development and testing of drugs accelerated and documented infections decreased. In another study with moderate supportive care sargramostim was given after delays of 48, 72, 96 and 120 h after whole-body radiation exposure of 7.13 Gy. A survival benefit was reported after the 48 h delayed dosing. Additional efficacy was shown in all cohorts of monkeys receiving sargramostim up to 120 h post-irradiation. Survival was improved by 18% at day 60 (P = .0032) compared to controls when sargramostim was administered at 48 h post-irradiation. [34] These data are summarized in a recent typescript. [35] Effects of recombinant human G/M-CSF (rhG/M-CSF) on radiation- induced bone marrow suppression was tested in a monkey model of high-dose, non-uniform exposure. rhG/M-CSF treatment for 7 days after 8 Gy non-uniform exposure accelerated bone marrow recovery. Monkeys received rhG/M-CSF continuously through an Alzet mini-os- motic pump implanted sub-cutaneously on day 3 post-exposure. Blood granulocyte and platelet concentrations recovered 4 and 7 days earlier compared with controls as did G/M progenitor cells. [10]

10. Radiation accidents

There are reviews of the role of molecularly-cloned haematopoietic growth factors, most of sargramostim and ﬁlgrastim in radiation acci- Fig. 6. Granulocyte concentrations in subjects treated with G/M-CSF after the dent. [9] These data are diﬃcult to interpret critically for several Goiania radiation accident. [36].

6 R.P. Gale and J.O. Armitage Blood Reviews xxx (xxxx) xxxx and biologicals to reduce or prevent serious/life-threatening conditions cloned haematopoietic growth factors are rate of granulocyte or neu- caused by exposure to lethal or permanently disabling chemical, bio- trophil recovery and survival. In the usual clinical setting rate of logical, radiological, or radioactive substances where human efficacy granulocyte recovery is (artificially) important because it is used to trials are not feasible nor ethical. [37] According to this regulation the determine when people can be released from isolation and or dis- FDA can rely on evidence from animal studies to provide substantial charged from hospital. However, in the context of a radiological event evidence of product effectiveness if: (1) there a reasonably well-un- rate of granulocyte recovery is a surrogate endpoint; survival is the key derstood mechanism for the toxicity of the agent and its amelioration or issue. prevention by the product; (2) the effect is shown in either > 1 animal As indicated, although there are no trials in humans comparing rates species expected to react with a response predictive for humans or one of granulocyte recovery between sargramostim, filgrastim and peg-fil- well-characterized animal species model (adequately evaluated for its grastim, these seem similar when tested in comparable clinical settings. responsiveness in humans) for predicting the response in humans; (3) However, at least in monkeys, there seems a survival endpoint ad- the animal study-endpoint is clearly related to the desired benefitin vantage of sargramostim over filgrastim and peg-filgrastim. This con- humans; and (4) data or information on the pharmacokinetics and clusion should be considered tentative because these studies used dif- pharmaco–dynamics of the product or other relevant data or informa- ferent experimental conditions. Effects of these drugs on a survival tion in animals or humans is sufficiently well understood to allow se- endpoint in humans is even more difficult to critically evaluate. [41]In lection of an effective dose in humans, and it is, therefore, reasonable to evaluating this issue one should consider that deaths from infection expect the effectiveness of the product in animals to be a reliable in- during cancer therapy, save for leukaemias, are rare. Consequently, dicator of its effectiveness in humans. most trials are not powered to detect a survival benefit. Moreover, in In the context of approval for drugs to be used following a radi- leukaemias most deaths from infection occur in persons with residual or ological or nuclear accident or incident the FDA position is to accept a resistant leukaemia rather than from delayed bone marrow recovery, a significant survival benefit in a non-human primate model of high-dose setting where molecularly-cloned haematopoietic growth factors are whole-body radiation. Acceleration of neutrophil recovery and reduced not expected to be effective. However, in some studies in which sar- infections are considered supportive surrogate endpoints but in- gramostim and filgrastim accelerated neutrophil recovery only sar- sufficient to warrant approval. [37] There is no stated FDA position on gramostim reduced severe infections. [42,43] non-human primates exposed to partial-body radiation. However, the partial-body radiation models with minimal bone marrow shielding 13.3. Effective interval post-exposure provides the ability to ask a different question relevant to the FDA and drug approval, e.g., what is the effect of G- and G/M-CSF on DEARE There seems a broader window of efficacy after high-dose radiation [16,17,27,38]. exposure for sargramostim compared with filgrastim when tested in monkeys. For example, filgrastim is most effective when given within 12. Drugs approved for acute radiation syndrome 1 day after radiation exposure with intensive supportive care [30,31]. In contrast, sargramostim seems effective up to 4 days after radiation It is impossible to test safety and efficacy of sargramostim, fil- exposure for some but not all endpoints. [33,34] However, the window grastim, peg-filgrastim or related drugs in the context of a radiation of efficacy depends on radiation dose and dose-uniformity and these accident. However, these drugs met requirements of the FDA Animal drugs are not tested or compared in comparable experimental monkey Rule and are approved for this setting and are part of the US Strategic models. Delays in giving these drugs is likely to depend on numbers of National Stockpile [39] and approved for use in victims of acute, high- exposed persons, speed of dose estimates, accessibility to victims, ac- dose whole-body radiation exposures based on data from animal studies cessibility to drugs, extent of infrastructure disruption, availability of including non-human primates [40] medical personnel and other variables. Given these complex con- founded variables it is difficult to recommend either drug as preferable. 13. Differences 13.4. Supportive care Although sargramostim, filgrastim and peg-filgrastim are effective in accelerating granulocyte recovery and reducing deaths in selected Another possible difference between sargramostim and filgrastim is monkey models (rhesus macaque) of acute radiation injury there are the role of additional interventions. Filgrastim appears to be most ef- several potential differences including: (1) target cell(s); (2) haemato- fective in the context of intensive support with parenteral fluids, blood poietic lineages affected; (3) rates of granulocyte or neutrophil re- transfusions and individualized antibiotic therapy [30] and was in- covery; (4) impact of delayed dosing after radiation-exposure; and (5) effective in increasing survival after a 7 Gy exposure with moderate impact of additional supportive interventions such as intravenous supportive care [32]. However, it was not tested without intensive fluids, antibiotics and RBC- and/or platelet -transfusions. We discuss supportive care at higher radiation doses with or without partial body these potential differences below. However, it should be recalled these shielding. Whether filgrastim or peg-filgrastim are effective when given data are derived from experimental models and that there is no direct with the type of moderate supportive care given to monkeys in ex- randomized comparison of sargramostim and filgrastim in animal periments with sargramostim and which might be expected after a models or humans after a radiation accident. large-scale radiation or nuclear accident or incident is untested. In contrast, sargramostim has only been tested in the setting of moderate 13.1. Sites of activity supportive care. Whether a survival benefit of sargramostim would be increased when combined with the intensive supportive care used in Sites of action of sargramostim and filgrastim are shown in Fig. 4. most experiments with filgrastim is also unknown. [33] These data Sargramostim acts on a less mature target cell (multi-potent progenitor suggest when intensive supportive interventions like individualized cell) and on a wider range of haematopoietic cell lineages than fil- antibiotic therapy based on blood culture data and blood transfusions grastim. Whether these differences are clinically-important is discussed are limited or unavailable there may be an advantage of sargramostim below. over filgrastim or peg-filgrastim but this is unproved.

13.2. Endpoints 14. Conclusions

The two most used endpoint to evaluate eﬃcacy of molecularly- Persons exposed to acute high-dose and dose-rate whole-body

7 R.P. Gale and J.O. Armitage Blood Reviews xxx (xxxx) xxxx ionizing radiations of 2 to 12 Gy are likely to benefit from receiving Declaration of Competing Interest molecularly-cloned haematopoietic growth factors such as sar- fi fi gramostim, lgrastim and peg- lgrastim which accelerate bone marrow RPG is on the Advisory Board of StemRad LLC. JOA is consultant to recovery and often improve survival in diverse experimental models Oncology Analytics, Partner Therapeutics, Samus Therapeutics, including non-human primates. There are considerable data using these Ascentage and Union Pacific Railroad. drugs after high-dose anti-cancer therapy and/or high-dose radiation therapy including, in some instances, acute whole-body ionizing ra- Acknowledgements diations. There is also considerable experience using these drugs after radiation and nuclear accidents with the impression they accelerate Profs. Vijay Singh (Uniformed Services Univ. of the Health Sciences) bone marrow recovery and improve survival. However, controls are and Thomas MacVittie (Univ. Maryland) kindly reviewed the typescript ffi lacking and claims of safety and e cacy in humans after radiation or and provided helpful comments and access to unpublished data. Brent nuclear accidents or incidents are unproved. Nevertheless, these drugs Gardner (Partner Therapeutics) gave us access to unpublished data and are FDA-approved based on the FDA Animal Rule and are included in helped assemble the Table. RPG acknowledges support from the the US Strategic National Stockpile. National Institute of Health Research (NIHR) Biomedical Research fi fi ff Sargramostim, lgrastim and peg- lgrastim have several di erences Centre funding scheme. based on mechanism of action and on- and off-target effects. Some data ff from non-human primate accident models suggest di erences in op- References timal interval after radiation-exposure and intensity of supportive care ff but direct comparisons are lacking. Sargramostim is e ective with [1] Unthank JL, Miller SJ, Quickery AK, Ferguson EL, Wang M, Sampson, et al. Delayed moderate supportive care but has not been tested with no or with in- effects of acute radiation exposure in a murine model of the H-ARS multiple organ tensive supportive care. Filgrastim is effective with intensive supportive injury consequent to < 10gy total body irradiation. Health Phys 2015;109:511–21. [2] Medhora M, Gao F, Gasperetti T, Narayan J, Khan AH, Jacobs ER, et al. Delayed care but not with moderate or no supportive care. [32] The lyophilized effects of acute radiation exposure (DEARE) in juvenile and old rats: mitigation by form of sargramostim may offer advantage in settings of infra-structure lisinopril. Health Phys 2019;118:529–45. disruption but peg-filgrastim may have advantages when access to [3] Bhattacharyya MH, Breitenstein BD, Metivier H, Muggenburg BA, Stradling GN, Volf V, et al. Guidebook for the treatment of accidental internal radionuclide con- medical personnel is limited. Both drugs can be self-given sub- tamination of workers. Radiat Prot Dosimetry 1992;41:1–49. cutaneously. Absent direct comparisons these conclusions should be [4] Kazzi Z, Buzzell J, Bertelli L, Christensen D. Emergency department management of considered tentative. The major activity and concern of physicians, patients internally contaminated with radioactive material. Emerg Med Clin N Am – heath authorities and government should be to prevent radiological and 2015;33:179 96. [5] Gale RP, Armitage JO, Hashmi SK. Emergency response to radiological and nuclear nuclear accidents and incidents including terrorism where these drugs accidents and incidents. BJH 2019. (doi.111/bjh.16138). might be used. Regardless of the potential efficacy of these drugs there [6] Baranov AE, Selidovkin GD, Butturini A, Gale RP. Hematopoietic recovery after – is no effective medical response to a nuclear war. Physicians need to acute high-dose total body radiation. Blood. 1994;83:596 9. ff [7] Singh VK, Garcia M, Seed TM. Review of radiation countermeasures focusing on make every e ort to prevent such an event. [44] injury-specific medicinals and regulatory approval status: part II. Countermeasures for limited indications, internalized radionuclides, emesis, late effects, and agents ffi 15. Future considerations demonstrating e cacy in large animals with or without FDA IND status. Int J Radiat Biol 2017;93:870–84. [8] Singh VK, Hanon BK, Santiago PT, Seed TM. A review of radiation countermeasures Our hope is radiation accidents will be prevented by increasingly focusing on injury-specific medicinals and regulatory approval status: part III. safe and effective human and computer-based intervention sand nu- Countermeasures under early stages of development along with ‘standard of care’ medicinal and procedures not requiring regulatory approval for use. Int J Radiat clear terrorism thwarted. Prevention is always better than cure. And Biol 2017;93:885–906. [Review]. physicians must get (re)-involved in efforts to prevent a nuclear war or [9] Singh VK, Seed TM. A review of radiation countermeasures focusing on injury- even the intended or unintended use of a nuclear weapon or improvised specific medicinals and regulatory approval status: part I. radiation sub-syndromes, animal models and FDA-approved countermeasures. Int J Radiat Biol nuclear weapon (IND). Politicians and policy-makers need to under- 2017;93:851–69. stand limitations in our ability to respond to any more than a small [10] Monroy RL, Skelly RR, Taylor P, Dubois A, Donahue RE, MacVittie TJ. Recovery radiation or nuclear incident, accident or event. from severe hematopoietic suppression using recombinant human granuloctye- macrophage colony stimulating factor. Ex Hematol 1988;16:334–8. [11] Sholom S, McKeever SW. Integrated circuits from mobile phones as possible Practice points emergency OSL/TI dosimeters. Radiat Prot Dosimetry 2016;170:398–440. [12] Muramatsu Y, Noda Y, Yonehara H, Ishigure N, Yoshida S, Yukawa M, et al. • Haematologists, oncologists, radio-biologists and first responders Determination of radionuclides produced by neutrons in heavily exposed workers of the JCO criticality accident Tokai-mura for estimating an individual’s neutron flu- need to understand how to estimate radiation dose and uniformity; ence. J Radiat Res 2001;42(Suppl):S117–28. • They need to understand the unavoidable imprecision involved in [13] National Council on Radiation Protection and Measurements. Report No. 163 – these estimates and wide confidence intervals; Radiation Dose Reconstruction: Principles and Practices. https://www.remm.nlm. fi gov/ars_wbd.htm; 2009. • They need to make a best estimate of potential bene ts and harms [14] Wojcik A, Oestereicher U, Barrios, Vraj A, Terzoudi G, Ainsbury EA, et al. The with an emphasis on doing no harm; RENEE operational basis: complement of established biodosimetric assays Int J • Haematopoietic cell transplants have a very limited role in the re- Radiat Biol 2017;93:15–19. [15] Ainsbury EA, Samaga D, Dellamonaca S, Marrale M, Bassinet C, Burbridge C, et al. sponse to radiation and nuclear incidents, accidents and events; Uncertainty on radiation doses estimated by biological and retrospective physical • Use of molecularity-cloned haematopoietic growth factors in per- methods. Radiat Prof Dosimetry 2017;178:382–404. sons exposed to acute, high-dose and dose-rate whole- partial-body [16] MacVittie TJ, Bennett AW, Farese AM, Taylor-Howel C, Smith CP, Gibbs AM, et al. ff The effect of radiation dose and variation in Neupogen initiation schedule on the radiation is generally safe and probably e ective based on data from mitigation of myelosuppression during the concomitant GI-ARS and H-ARS in a animal models. nonhuman primate model of high-dose exposure with marrow sparing. Health Phys 2015;109:427–39. Research agenda [17] MacVittie TJ, Farese AM, Parker GA, et al. The time course of radiation-induced lung injury in a nonhuman primate model of partial-body irradiation with minimal bone marrow sparing: clinical and radiographic evidence and the effect of • Develop radio-protective drugs which can be given post-exposure; Neupogen administration. Health Phys 2019;116:366–82. • Develop drugs to address delayed radiation effects to the lungs, [18] Farese AM, Bennett AW, Gibbs AM, Hankey KG, Prado K, Jackson 3rd. W. et al. efficacy of neulasta or neupogen on H-ARS and GI-ARS mortality and hematopoietic kidneys and heart. recovery in nonhuman primates after 10 Gy irradiation with 2.5% bone-marrow sparing. Health Phys 2019;116:339–53. [19] Singh VK, Newman VL, Seed TM. Colony-stimulating factors for treatment of the

8 R.P. Gale and J.O. Armitage Blood Reviews xxx (xxxx) xxxx

hematopoietic component of the acute radiation syndrome (H-ARS): a review. 2014;89:868–73. Cytokine. 2015;71:22–37. [33] Clayton NP, Charpentier EJJ, Emmons GT. Sargramostim significantly improved the [20] Stull DM, Bilmes R, Kim H, Fichtl R. Comparison of sargramostim and filgrastim in mortality rate at day 60 in a non-human primate model of hematopoietic acute the treatment of chemotherapy-induced neutropenia. Am J Health Syst Pharm radiation syndrome with minimal supportive care when administered 48 h after 2005;62:83–7. total body irradiation [poster]. Presented at: European radiation research society [21] Smith TJ, Khatcheressian J, Lyman GH, Ozer H, Armitage JO, Balducci L, et al. 2006 (ERRS) meeting; September 4-8, 2016; Amsterdam, Netherlands. update of recommendations for the use of white blood cell growth factors: an evi- [34] Zhong Y, Pouliot M, Downey AM, et al. Efficacy of delayed administration of dence-based clinical practice guideline. J Clin Oncol 2006;24:3187–205. Sargramostim up to 120 hours post exposure in a nonhuman primate total body [22] Klastersky J, de Naurois J, Rolston K, Rappoport B, Manchmeyer G, Aapro M, et al. radiation model. Int J Radiat Biol 2019. [Still in press]. Management of febrile neutropaenia: ESMO clinical practice guidelines. Ann Oncol [35] Singh VK, Seed TM. An update on sargramostim for treatment of acute radiation 2016;27:v111–8. syndrome. Drugs Today 2018;54:679–93. [23] Wan L, Zhang Y, Lai Y, Jiang M, Song Y, Zhou J, et al. Effect of granulocyte-mac- [36] Butturini A, De Souza PC, Gale RP, Cordiero JM, Lopes DM, Neto C, et al. Use of rophage colony-stimulating factor on prevention and treatment of invasive fungal recombinant granulocyte-macrophage colony stimulating factor in the Brazil ra- disease in recipients of allogeneic stem-cell transplantation: a prospective multi- diation accident. Lancet. 1988;332:471–5. center randomized phase IV trial. J Clin Oncol 2015;33:3999–4006. [37] https://www.fda.gov/emergency-preparedness-and-response/mcm-regulatory- [24] Mehta HM, Malandra M, Corey SJ. G-CSF and GM-CSF in neutropenia. J Immunol science/animal-rule-information. 2015;195:1341–9. [38] Cohen E, Hankey K, Farese AM, Parker GA, Jones JW, Kane A, et al. Radiation [25] Weaver CH, Schulman KA. Mobilization of peripheral blood stem cells following nephropathy in a nonhuman primate model of partial-body irradiation with myelosuppressive chemotherapy: a randomized comparison of filgrastim. Bone minimal bone marrow sparing. Part 1: acute and chronic kidney injury and the Marrow Transplant. 2001;Suppl. 2:S23–9. influence of neupogen. Health Phys 2019;116:401–8. [26] Williams JP, Brown SL, Georges GE, Hauer-Jensen M, Hill RP, Huser AK, et al. [39] https://www.phe.gov/about/sns/Pages/about.aspx. Animal models for medical countermeasures to radiation exposure. Radiat Res [40] https://www.fda.gov/drugs/nda-and-bla-approvals/animal-rule-approvals. 2010;173:557–78. [41] Dale DC, Crawford J, Klippel Z, Reiner M, Osslund T, Fan E, et al. A systematic [27] MacVittie TJ, Bennett A, Booth C, Garafolo M, Tudor G, Ward A, et al. The pro- literature review of the efficacy, effectiveness, and safety of filgrastim. Support Care longed gastrointestinal syndrome in rhesus macaques. The relationship between Cancer 2018;26:7–20. gastrointestinal hematopoietic and delayed multiorgan sequelae following acute [42] Rowe JM, Andersen JW, Mazza JJ, Bennett JM, Paietta E, Hayes FA, et al. A ran- radiation exposure. Health Phys 2012;103:427–53. domized placebo-controlled phase III study of granulocyte-macrophage colony [28] Graessle DH, Dörr H, Bennett A, Shapiro A, Farese AM, MacVittie TJ, et al. -stimulating factor in adult patients (> 55 to 70 years of age) with acute myelo- Comparing the hematopoetic syndrome time course in the NHP animal model to genous leukemia: a study of the eastern cooperative oncology group (E1490). radiation accident cases from the database search. Health Phys 2015;109:493–501. Blood. 1995;86:457–62. [29] Dörr H, Lamkowski A, Graessle D, Bennett A, Shapiro A, Farese AM, et al. Linking [43] Rowe JM, Rubin A, Mazza JJ, Bennett HM, Paietta E, Anderson JW, et al. Incidence the human response to unplanned radiation and treatment to the nonhuman pri- of infections in adult patients (> 55 years) with acute myeloid Leukemia treated mate response to controlled radiation and treatment. Health Phys with yeast-derived GM-CSF (Sargramostim): Results of a double-blind prospective 2014;106:129–34. study by the eastern cooperative oncology group. In: Hiddemann W, editor. Acute [30] Farese AM, Cohen MV, Katz BP, Smith CP, Gibbs A, Cohen DM, et al. Filgrastim Leukemias V. haematology and blood transfusion / Hämatologie und improves survival in lethally irradiated nonhuman primates. Radiat Res Bluttransfusion. 37. Berlin, Heidelberg: Springer; 1996. 2013;179:89–100. [44] Gale RP, Nathan DG. Physicians and the threat of nuclear war. ASCO POST July [31] Hankey KG, Farese AM, Blaauw EC, Gibbs AM, Smith CP, Katz BP, et al. 2019;25:82–3. Pegfilgrastim improves survival of lethally irradiated nonhuman primates. Radiat [45] Vainchenker W, Leroy E, Gilles L, Marty C, Plo I, Constantinescu SN, et al. JAK Res 2015;183:643–55. inhibitors for the treatment of myeloproliferative neoplasms and other disorders. [32] Gluzman-Poltorak Z, Vainstein V, Basile LA. Recombinant interleukin-12, but not F1000Res. 2018;7:82. granulocyte-colony stimulating factor, improves survival in lethally irradiated [46] Farese AM, Brown CR, Smith CP, Gibbs AM, Katz BB, Johnson CS, et al. The ability nonhuman primates in the absence of supportive care: evidence for the develop- of filgrastim to mitigate mortality following LD50/60 total body irradiation is ad- ment of a frontline radiation medical countermeasure. Am J Hematol ministration time-dependent. Health Phys 2014;106:39–47.