Medical Isotope Production Without Highly Enriched Uranium (2009)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

4 Molybdenum-99/Technetium-99m Supply Reliability T he statement of task for this study calls on the National Academies to evaluate the “availability of medical isotopes for . . . future domestic use,” and also to “identify any reliability of supply issues that could arise as a result of conversions from highly enriched uranium (HEU) to low enriched uranium (LEU) production of medical isotopes” (see Sidebar 1.2). These supply reliability issues are addressed in this chapter, specifically with respect to molybdenum-99 (Mo-99) and technetium-99m (Tc-99m). Supply reliability is primarily an issue for reactor production of Mo-99. The downstream elements of the Mo-99/Tc-99m supply chain (i.e., tech­ netium generator production and Tc-99m distribution; see Chapter 3) are in relatively better shape with respect to reliability. Mo-99 supply reliabil- ity has been a concern in the United States since the late 1980s when the Cintichem Reactor was shut down (Chapter 3). These reliability concerns arise from the following three factors: 1. Increasing demand, both domestically and globally, for Mo-99; 2. Continued reliance on a small number of aging foreign reactors for Mo-99 production; and   There have been some recent problems with technetium generators, but these have been less frequent and have had a smaller impact on the medical isotope user community than the disruption of Mo-99 supplies.   Reliability is affected by both supply and demand; if demand grows without new produc- tion then reliability can suffer. 55 

56 MEDICAL ISOTOPE PRODUCTION WITHOUT HIGHLY ENRICHED URANIUM 3. Increasing difficulty of transporting Mo-99 across international borders, especially by air. Supply reliability is in comparatively better shape in other major world regions, notwithstanding the recent outages that are discussed elsewhere in this chapter. As noted in Chapter 3 and discussed in more detail elsewhere in this chapter, Europe is in the process of replacing two aging reactors; Australia recently commissioned a new reactor (Open Pool Australian Lightwater [OPAL]) and plans to bring a new Mo-99 production facility online in 2009; and Argentina recently upgraded its Mo-99 production facilities and has also begun planning to replace its aging reactor (RA-3). The discussion in this chapter is organized into three sections. The first provides an examination of general Mo-99 supply reliability issues indepen- dent of whether HEU or LEU targets are used to produce this isotope. The second section provides an examination of Mo-99 supply reliability issues that could arise as a result of conversion from HEU to LEU targets. The third and final section provides findings to address the study charge. MOLYBDENUM-99 SUPPLY RELIABILITY: GENERAL ISSUES The general supply reliability issues that will be discussed in this sec- tion arise roughly on two timescales: days to weeks (short timescales) and months to years (long timescales). Over short timescales, supply reliability problems are primarily the result of: 1. Planned or unplanned facility outages combined with limited ­excess capacity for Mo-99 production elsewhere. 2. Problems with transporting Mo-99 from production facilities to technetium generator producers. Such disruptions, although temporary, can lead to severe disruptions in diagnostic imaging procedures that can affect the continuity of patient care. All of the reactors that produce Mo-99 must be shut down periodically for refueling and other maintenance. Even the best run and maintained reactors will be shut down on a monthly or more frequent basis for a ­total of at least 50 days per year (see Table 3.2). The operational programs and planned shutdowns of reactors in Europe (Belgian Reactor II [BR2], High Flux Reactor [HFR], and Osiris) and South Africa (Safari-1) are coordinated so that there is available reactor capacity for medical isotope production. Also, the two European Mo-99 producers (Mallinckrodt and   However, as discussed elsewhere in this chapter, unplanned shutdowns can result in insuf- ficient reactor capacity. 

MOLYBDENUM-99/TECHNETIUM-99m SUPPLY RELIABILITY 57 SIDEBAR 4.1 Shutdown of the NRU Reactor On November 18, 2007, AECL shut down the NRU reactor for what was intended to be 5 days of routine maintenance. During the shutdown, inspectors from Canada’s nuclear regulatory agency (Canadian Nuclear Safety Commission [CNSC]) discovered that AECL had been operating the reactor without upgraded emergency backup power systems for the reactor’s cooling pumps. The CNSC ordered the reactor to remain shut down until installation of these backup systems was completed. The shutdown of the NRU caused shortages of Mo-99 in the United States and Canada, causing the cancellation of medical procedures and outcries from the medical community (see Ad Hoc Health Experts Working Group on Medi- cal Isotopes, 2008). The Canadian Parliament passed a bill that allowed AECL to resume operation of the reactor for 120 days despite any conditions under the Nuclear Safety and Control Act relating to the installation of the emergency backup power systems. NRU was restarted on December 16, 2007, with one of the two emer- gency backup systems for the cooling pumps installed; the second backup supply was installed shortly after the reactor began operating. MDS Nordion reported that it ­received the first batch of Mo-99 after the shutdown from AECL on December 19. Once it became clear that NRU would have an extended shutdown (i.e., for more than a few days), Mo-99 producers in Europe and South Africa increased production of Mo-99 and took steps to distribute this isotope around the world to help offset supply disruptions (AIPES, 2007). In spite of this increased production, they were not able to replace all of the lost production. Institut National des Radioéléments [IRE]) have agreements with multiple European reactors for target irradiation services (see Figure 3.1). Until recently, the Canadian reactor operator Atomic Energy of C ­ anada Limited (AECL) did not coordinate its reactor outage schedules with other reactor operators (Collier, 2008). However, after an extended shutdown in 2007 (discussed below and in Sidebar 4.1), the Canadian government announced that it was developing a new protocol for sharing information among reactor operators, isotope suppliers, and the medical establishment. Unplanned reactor shutdowns can severely disrupt Mo-99 supplies, and these disruptions can have serious impacts on the quality of patient care. Supply disruptions can lead to the reduction of Tc-99m that is avail- able for patient procedures. Some of these procedures can be resched- uled, but others, especially emergency procedures, cannot be postponed without poten­tially serious medical consequences. Such shutdowns have 

58 MEDICAL ISOTOPE PRODUCTION WITHOUT HIGHLY ENRICHED URANIUM SIDEBAR 4.2 A Selected Chronology of Events That Have Affected Mo-99 Supply to North America 1989 Cintichem Reactor, the only domestic supplier of Mo-99 to the United States, is permanently shut down. 1992 The U.S. Department of Energy (DOE) begins an effort to produce Mo-99 in its reactors (see Chapter 3). 1999 DOE ends its efforts to produce Mo-99 after a solicitation of private com- panies yields no interest (see Chapter 3). 2001 Mo-99 shipments to the United States by air are halted temporarily after the September 11 terrorist attacks. 2002 HFR is shut down for 42 days because of reactor operation safety concerns. 2005 Production of Tc-99m generators by Mallinckrodt is shut down in the United States on November 18 because of a product recall. Production is not restarted until April 2006. 2006 NRU reactor is shut down for approximately 6 days because of a technical problem. 2007 NRU reactor is shut down for 24 unplanned days by its regulator to a ­ ddress safety concerns. 2008 HFR is voluntarily shut down in August 2008 after a corrosion problem in the primary cooling system is discovered. The reactor is not scheduled to come back online until February 2009. IRE is shut down in August 2008 after I-131 was unexpectedly vented through a stack. The facility received approval to restart on November 4, 2008. A scheduled 5-day shutdown of NRU Reactor in December 2008 was extended for several additional days. Because HFR was also shut down at the time, there were supply shortages in the United States and Canada. resulted from worker strikes as well as reactor maintenance and reactor upgrades that could not be taken care of during planned outages. Several unplanned shutdowns have occurred during the past 20 years (Sidebar 4.2 provides selected examples); two major reactor shutdowns occurred while this ­National Academies study was in progress: • A November 2007 shutdown of the National Research Universal (NRU) reactor for scheduled maintenance was extended for almost a month after the regulator discovered that a safety upgrade had not been made (Sidebar 4.1). This outage was reported to have affected more than 50,000 patient procedures in the United States (Perkins et al., 2008), although the basis for this estimate is unclear. An unpublished survey by the Society for 

MOLYBDENUM-99/TECHNETIUM-99m SUPPLY RELIABILITY 59 Nuclear Medicine indicated that 84 percent of respondents’ facilities were affected by this shutdown, and about 40 percent of respondents’ facilities were operating at half capacity or below (the survey results are provided in Ad Hoc Health Experts Working Group on Medical Isotopes [2008]). • The HFR reactor was shut down in late August 2008 after small gas bubbles of unknown origin and composition were discovered in the primary cooling system. A subsequent investigation determined that the gas bubbles were the product of corrosion of an aluminum sleeve where it contacted concrete. A possible fix has been identified, but the operator now estimates that the reactor will not be restarted until February 16, 2009. At the time this shutdown occurred, the other four major production reactors (NRU in Canada; BR2 and Osiris in Europe; and Safari-1 in South Africa) were either shut down for maintenance or had scheduled shutdowns planned in the near future. Within a week of the August 2008 shutdown of HFR, IRE also shut down its isotope production facilities in Fleurus, Belgium, after 40 GBq (a little over 1 curie) of iodine-131 (I-131) gas was unexpectedly released to the air outside the plant. The facility regulator did not approve a restart until November 4, 2008. This “perfect storm” of coincidental shutdowns is having substantial global impacts on Mo-99 availability. These outages are expected to disrupt Tc-99m supplies in Europe for at least 4–6 weeks and are also having an impact on North American markets. However, the supply disruptions have been somewhat less than expected because Mallinckrodt has been able to produce Mo-99 at its Petten facility from HEU targets irradiated in the Osiris reactor in France. Nevertheless, the European Association of Nuclear Medicine recently characterized the iso- tope supply situation as “turning from a short term shortage to a ‘chronic disease.’” The American Society of Nuclear Medicine has established a task force to examine alternative means for isotope production within the United States (SNM, 2008).   Pitting corrosion of aluminum materials in contact with concrete is probably the single most serious materials aging problem in research reactors. Such corrosion is particularly likely to occur in heat-affected zones close to welds in reactors and in the lining of spent fuel pools.   For example, Mallinckrodt has informed its customers that they will get less Mo-99 than they have ordered.   According to European press reports, customers were expecting to receive only about 30 percent of their normal deliveries of Tc-99m but were instead receiving 65–70 percent. Tc-99m supplies may further ease when the BR2 reactor restarts in late October 2008, but addi­tional shortages are expected again in November 2008 when BR2 and Osiris shut down for scheduled maintenance. The committee has not independently verified these press reports.   EANM Press Release, December 2, 2008. 

60 MEDICAL ISOTOPE PRODUCTION WITHOUT HIGHLY ENRICHED URANIUM Difficulties in moving radioactive material across international borders can also affect Mo-99 supply reliability, especially when air transport is involved. Cross-border shipments of medical and industrial radioactive materials are regulated by individual countries, usually in accordance with the International Atomic Energy Agency’s (IAEA) International Regulations for the Safe Transport of Radioactive Material (IAEA, 2004). IAEA’s model regulations allow radioactive materials such as Mo-99 to be transported in commercial airliners. However, airline companies can refuse to carry these shipments, and individual airline pilots can refuse to carry shipments even if company policies allow it. The IAEA has reported (IAEA, 2004) that it is becoming increasing difficult for companies to ship radioactive materials by air. Mo-99 producers told the committee that although cross-border ship- ments are still manageable, they are becoming less reliable. A representative of MDS Nordion told the committee that it avoids the use of passenger aircraft for Mo-99 shipments to the United States. Instead, it uses char- ter ­ aircraft and trucks to ship Mo-99. A representative of the Australian N ­ uclear Science and Technology Organisation (ANSTO) reported to the committee that it encounters an “adverse” Mo-99 cross-border shipping event once every 3 weeks on average. Such events include shipments being laid off at airports or delayed in customs. These adverse events often ­occur without notice. They disrupt Mo-99 delivery schedules and may delay patient care. Medical isotope producers and technetium generator manufacturers are aware of these supply reliability issues and they cooperate with each other to minimize the impacts of disruptions. For example, producers have agreements in place (including the necessary U.S. Food and Drug Administration approvals; see Sidebar 8.1) to obtain alternative supplies of Mo-99 during temporary disruptions. Mo-99 producers will also ramp up production when possible to supply each others’ customers with Mo-99 or technetium generators. There is enough surge capacity at existing reactors to ­temporarily cover Mo-99 shortages caused by short-duration shutdowns of single reactors, but such surges cannot be maintained indefinitely because reactors need to be shut down periodically for routine maintenance and refueling. In fact, recent experience suggests that unplanned shutdowns that extend beyond about a week have the potential to cause severe supply disruptions, as demonstrated by the November 2007 shutdown of NRU that was discussed earlier in this chapter (see also Sidebar 4.2).   Of course, ground-based disruptions can also occur. For example, a fire in the Chunnel between France and Britain on September 11, 2008, disrupted Tc-99m supplies to Britain during early October. 

MOLYBDENUM-99/TECHNETIUM-99m SUPPLY RELIABILITY 61 Many of the steps taken by producers to increase the reliability of Mo-99 supplies have relatively low cost, and of course the additional business that comes with supplying a competitor’s customers adds to that producer’s profits. However, producers have also taken some relatively high-cost steps to increase reliability that might not be seen as necessary or prudent if cost were the only business consideration. As noted in Chapter 3, for example, Mallinckrodt has 10 hot cells for Mo-99 production at its Petten, Netherlands, facility, even though other producers typically operate with fewer hot cells. MDS Nordion decided to build two Maple reactors at AECL to irradiate targets for Mo-99 production, even though one reactor had more than enough capacity to meet its current production needs. The obsolescence of existing Mo-99 processing facilities, which consist of hot cells, the target processing equipment contained within them, and ancillary support facilities, is not a major concern for supply reliability. Some of these facilities have been operating for decades, and the commit- tee received no reports of disruptions owing to major equipment malfunc- tions. Many of the major components of the hot cell itself (e.g., windows, m ­ anipulators) can be repaired or replaced. The target processing equipment contained within the cells (Figure 2.7) can be replaced with off-the-shelf items or can be easily fabricated at relatively low cost. The greatest single threat to supply reliability is the approaching obso- lescence of the aging reactors that current large-scale producers utilize to irradiate HEU targets to obtain Mo-99 (see IAEA, 2008). The continued opera­tion of these aging reactors (Table 3.2) is a testament to their good d ­ esign and construction, and to the success of reactor safety and mainte- nance programs. However, unlike processing facilities, not all of the compo- nents of these reactors can be easily maintained or replaced. For example, buried or concrete-encased pipes and some structural components of the reactor are difficult to access; replacing them could be expensive and could require extended (months to years) reactor shutdowns. These components include, depending on the reactor design, structural elements of the ­reactor core, the reflector, the reactor containment vessel, and the reactor pool liner. The three reactors that are currently being used to irradiate targets for Mo-99 production in Europe (HFR, BR2, and Osiris) were commissioned in the 1960s (Table 3.2) and will be reaching the ends of their planned lives between about 2015 and 2020. Efforts are under way to construct two replacement reactors that could be used to produce medical isotopes. These reactors would presumably irradiate targets on a contract basis for current Mo-99 producers in Europe (Mallinckrodt and IRE) and any new producers that have nearby target processing facilities. Ground was broken in 2007 for construction of the Jules Horowitz Reactor in Cadarache, France (Iracane, 2007). This 100 MWt materials 

62 MEDICAL ISOTOPE PRODUCTION WITHOUT HIGHLY ENRICHED URANIUM test reactor will be used for nuclear fuel research and the production of medical isotopes. The reactor is being constructed with funding from the Commissariat à l’Énergie Atomique (French atomic energy commission), Electricité de France, several research institutes, and AREVA. The reactor is planned to be commissioned by 2014. Construction cost for this reactor is estimated to be about 500 million euros (ESFRI, 2006). Four organizations, including NRG and Mallinckrodt, are developing a business plan and conceptual design with research reactor builders for a new multipurpose reactor, named “Pallas,” to replace HFR (van der Schaaf et al., 2008). The site for this reactor has not yet been selected but, accord- ing to NRG staff, it is likely to be built at Petten. The primary applications of this new reactor will be nuclear research and isotope production. The target date for completion of this reactor is 2016. About 40 percent of the U.S. supply of Mo-99 currently comes from Europe (Chapter 3), and so these new reactors will likely contribute to an improved reliability of supply for the United States. This assumes, of course, that the current European reactors can continue to operate until these new reactors come online. However, the other 60 percent of U.S. supply is produced in a 51-year-old Canadian reactor (NRU). When it an- nounced its decision to discontinue work on the Maple reactors (Chapter 10), AECL also ­announced its intention to seek a 5-year license extension for NRU (from 2011 to 2016). A representative of the Canadian govern- ment told the committee that this upgrade would require expenditures of “hundreds of millions of dollars.”10 It is not clear to the committee whether such upgrades could be made without extended shutdowns of NRU. In the committee’s judgment, a particular concern for upgrading the NRU is the possible need to replace its aluminum reactor vessel, or ­ alandria.11 The original NRU calandria was replaced in the early 1970s c because of corrosion, and the reactor was shut down for over 2 years while this replacement was made. There is no other reactor on the Chalk River site that could be used to produce Mo-99 during an extended outage of NRU.   Three vendors have been invited to submit designs for this reactor: Korea Atomic Energy Research Institute (KAERI, South Korea), AREVA (France), and Investigaciones Aplicadas Sociedad del Estado (INVAP, Argentina). A final design has not yet been selected, nor has funding been committed for construction. The 2016 date was characterized to the committee by NRG staff as “optimistic.” 10  Sylvana Guindon, Natural Resources Canada, verbal communication with committee chair Chris Whipple and study director Kevin Crowley, June 20, 2008. 11  The calandria is a sealed drum-shaped vessel that contains the heavy water moderator. This vessel is penetrated by a series of horizontal fuel channels and vertical channels for control rods. 

MOLYBDENUM-99/TECHNETIUM-99m SUPPLY RELIABILITY 63 Finally, reliability of Mo-99 supply will depend on the continued availability of HEU12 until Mo-99 producers are able to convert to LEU. A ­ lthough the recently enacted Burr Amendment (Sidebar 1.3) has increased the short-term reliability of Mo-99 supply by ensuring continued access to HEU by producers in Belgium, Canada, France, Germany, and the Nether- lands, its impact on long-term supply reliability is unclear. Long-term access to HEU is likely to be driven by unforeseen events that are out of Mo-99 producers’ direct control. For example, the U.S. government could decide to restrict or eliminate exports of HEU in the future because of security concerns or in direct response to a terrorist attack. If that were to happen, the Burr Amendment will have decreased the reliability of supply if it has slowed conversion efforts by HEU-based producers, which appears to be the case for at least one producer, MDS Nordion (Chapter 10). LEU CONVERSION The conversion of Mo-99 production from HEU to LEU would ­increase reliability of Mo-99 supplies in one important respect: namely, it would remove longer-term uncertainties associated with the continued availability of HEU for Mo-99 production. However, Mo-99 production using LEU targets could utilize the same reactors and the same or similar process- ing facilities used for current HEU-based production. Consequently, the reliability-of-supply concerns described previously for current HEU-based production would also apply to LEU-based production. Additionally, con- version itself could lead to reliability of supply problems if not carried out in a technically sound manner. The technical aspects of conversion are discussed in some detail in Chapter 7. FINDINGS With respect to its charge to assess the availability of Mo-99 for future domestic use and identify any reliability-of-supply issues that could arise as a result of conversions from HEU- to LEU-based production, the com- mittee finds that: 1. Reliability of supply is primarily a problem for the reactor produc- tion of Mo-99. Recent Mo-99 disruptions have impacted the availability of 12  Continuing to make HEU available for Mo-99 production is a U.S. government policy decision, not a technical decision. From a purely technical perspective there is enough excess U.S.-controlled weapon grade HEU to supply Mo-99 production for a very long time at cur- rent rates of consumption. As discussed in Chapter 1, about 40–50 kg of HEU is used annually to support global production of Mo-99. There are hundreds of metric tons of HEU in the U.S. stockpile (http://nnsa.energy.gov/nuclear_nonproliferation/1978.htm). 

64 MEDICAL ISOTOPE PRODUCTION WITHOUT HIGHLY ENRICHED URANIUM this isotope for medical use and are affecting the continuity of patient care in the United States and elsewhere. 2. The supply of Mo-99 to the United States is fragile over a number of different timescales. This fragility occurs because: • Mo-99 is highly perishable owing to its short (66-hour) half-life. • It is produced in a small number of reactors, all of which are shut down periodically for planned and unplanned maintenance. There is limited excess capacity when a major reactor is shut down for ­extended periods (weeks) or more than one reactor is shut down simultaneously even for shorter periods. • It is produced in reactors that are about 40–50 years old and have uncertain additional remaining lifetimes. • It is not produced domestically. • It is produced with HEU, which could be restricted in the future. • There are long supply lines from some producers in Europe and South Africa to users in the United States. • There can be difficulties involved in moving radioactive mate­ rials across international borders, especially by air. As demonstrated by the 2007 NRU reactor outage and 2008 HFR outage, the sustained shutdown of reactors used by either MDS Nordion or Mallinckrodt would result in the substantial disruption of supplies to the United States and worldwide, as would the simultaneous shutdown of reactors used by both companies even for short periods. 3. AECL’s May 2008 announcement that it will discontinue develop- ment work on the Maple reactors is a blow to worldwide supply reliability and increases U.S. vulnerability to supply disruptions. 4. Reliability of Mo-99 supply is likely to become a serious problem for the United States in the early part of the next decade without new or refurbished reactors: The operating license for the NRU reactor expires in 2011 and substantial investment and refurbishment will apparently be required to obtain a license extension; moreover, the European replacement reactors (Jules Horowitz and Pallas) will not yet be operational. HFR and NRU can probably continue to meet incremental growth in Mo-99 demand if those reactors can remain operational, but continued operations are not assured through the next decade. There is enough surge capacity at existing reactors to cover shortages caused by the shutdown of a single reactor, but such surges can not be maintained indefinitely. 5. Conversion from HEU-based to LEU-based production of Mo-99 would improve supply reliability because it would remove uncertainties associated with the continued availability of HEU for Mo-99 production. 

MOLYBDENUM-99/TECHNETIUM-99m SUPPLY RELIABILITY 65 However, conversion would not address any of the other supply reliability concerns associated with current HEU-based production. Moreover, con- version itself could lead to reliability-of-supply problems if not carried out in a technically sound manner. 6. Although there are other potential foreign and domestic sources of Mo-99 supply (see Chapter 3), it will take some time (5–10 years and possibly longer) for substantial supplies from these producers to become available (see also Chapter 10). As discussed in Chapter 10, government assistance is likely to be required to improve U.S. supply reliability.