Hematopoietic Toxicity

As previously mentioned (see the previous section on Radiobiology), the hematopoietic system is particularly sensitive to TBI, and lymphopenia is often seen with doses of 0.5 Gy and can be seen with doses of 0.3 Gy. Lymphopenia is typically followed by neutropenia, thrombocytopenia, and finally anemia. Soon after a TBI dose of 4 to 6 Gy has been given, lymphocytosis can be seen, but it typically is followed by neutropenia within 1 week. Three to 4 weeks after TBI, neutrophils fall to their minimum¹¹³ (Fig. 18-3). Regeneration of the HSC compartment depends on the total dose used because higher doses cause more rapid myelosuppression of greater duration. Administration of hematopoietic growth factors after TBI have the theoretical potential to alter hematopoietic system reconstitution, although reports in the setting of allogeneic HSCT have demonstrated an increased risk of GVHD and compromised survival¹¹⁴ and, therefore, routine use is controversial.^115,116 Of note, hematopoietic growth factors have only been used in the period following TBI, given the concerns raised by a trial in lung cancer, where growth factors increased pulmonary toxicity and thrombocytopenia when given concurrently with chemoradiation therapy.¹¹⁷

Figure 18-3 Representative hematologic response to total body irradiation, given as a single fraction of 200 cGy on day 0.

Data from Andrews GA: Radiation accidents and their management. Radiat Res Suppl 7:390-397, 1967; Figure 1, with permission from Radiation Research Society.

Oral Toxicity

As noted before, the salivary glands frequently are affected by TBI. Although acute parotitis is typically self-limited and can be managed with anti-inflammatory medicines, long-term salivary gland dysfunction can result in xerostomia, which may lead to dental caries. In a study of children who underwent allogeneic HSCT, the risk of developing impaired salivary function was 22% in those who received TBI as part of conditioning versus 1% in those who did not.¹¹⁸ Studies have shown that salivary flow can improve up to 1 year after the completion of TBI.¹¹⁹ Fractionating TBI was shown to reduce salivary dysfunction by 54% in one study.¹²⁰

Myeloablative conditioning regimens with and without TBI have been associated with abnormalities in tooth development in children.^120–122 In one series, myeloablative conditioning regimens using chemotherapy alone were associated with significantly higher rates of tooth developmental abnormalities than those involving TBI, although rates of salivary gland dysfunction were highest amongst the patients treated with single-fraction TBI.¹²³ Because of the increased risk of oral pathology associated with TBI, careful pretransplant evaluation by a dental specialist is recommended to minimize the risk of serious morbidity.¹²⁴ Pilocarpine has been noted to help relieve symptoms of xerostomia in patients treated with TBI.¹²⁵

Pulmonary Toxicity

The major dose-limiting toxicity of TBI is pneumonopathy (restrictive or obstructive lung disease), which can manifest early as pneumonitis and/or later as pulmonary fibrosis. In the setting of HSCT, radiation pneumonopathy can be difficult to distinguish from other causes of lung pathology; moreover, lung damage is likely multifactorial, with the risk of acute lung complications estimated to be 30% to 60%, depending on factors such as infection, conditioning regimen, GVHD, age, and diagnosis.¹²⁶ Likewise, late pneumonopathy occurs in 10% to 26% of patients and is associated with underlying lung dysfunction, type of conditioning regimen, acute and chronic GVHD and prophylaxis, donor and recipient age and immunocompatibility, stage of disease, and genetic predisposition.¹²⁷ TBI has been shown to be a risk factor for idiopathic pneumonia syndrome,¹²⁸ as well as for diffuse alveolar hemorrhage.¹²⁹ Although rates of pneumonopathy in patients receiving TBI vary widely (10% to 84%),¹³⁰ some series have reported pneumonitis in up to 20% of HSCT patients who never received TBI.¹³¹ In the modern era, with appropriate TBI techniques, the risk of pneumonopathy in patients treated with TBI may not be increased at all.¹³² Nevertheless, the significance of the problem is clear, given that mortality related to interstitial pneumonitis in patients treated with TBI can be 60% to 80%.^131,133,134

Several TBI-specific factors (e.g., total dose, fractionation, dose rate, and use of lung shielding) have been shown to have a significant bearing on the development of pulmonary complications. The total dose used during TBI has frequently been cited as a major factor influencing lung complications.^130,135,136 In two prospective RCTs using 12 Gy versus 15.75 Gy, higher rates of mortality were noted within the first 6 months in patients treated with 15.75 Gy, although pulmonary complications were not specifically cited as the excess cause of deaths.^{137–139,140} In a retrospective dosimetric study, a mean lung dose of more than 9.4 Gy was found to be an independent predictor for lethal pulmonary complications in patients receiving TBI to a total dose of 10 Gy in three daily fractions, at 0.055 Gy/minute using parallel opposed lateral fields.¹⁴¹ Two RCTs have demonstrated that fractionated TBI can reduce pneumonitis compared with single-fraction TBI, although only one study showed differences that were statistically significant.^89,142,143 A retrospective study found no difference in pneumonitis rates when comparing a single fraction of 6 Gy and three daily fractions of 3.33 Gy, suggesting that total doses of less than 10 Gy may not require fractionation to prevent toxicity, although no randomized data support this.¹⁴⁴ The necessity of hyperfractionation to prevent lung toxicity is unclear: A comparison of two prospective single-arm trials at the same institution revealed that conventional fractionation given with anteroposterior fields and lung blocks to a total dose of 12 Gy in daily 3-Gy fractions may not be any different than hyperfractionated TBI given twice daily with 1.7 Gy per fraction to a total dose of 10.2 Gy over 3 days, using parallel opposed lateral fields and no blocks.¹⁴⁵

Sampath and colleagues¹⁴⁶ recently reviewed 26 studies involving 1096 patients to create a dose-response model for predicting the risk of pneumonitis from TBI while taking other factors into consideration. Although unable to estimate the risk of pneumonitis for hyperfractionated regimens, they were able to determine the effect of fractionation and cyclophosphamide and busulfan on the risk of developing pneumonitis, in a dose-response model, as seen in Figure 18-4.

Figure 18-4 Incidence of pneumonitis based on dose-response model accounting for total body irradiation fractionation, as well as chemotherapy effects. Bu, busulfan; Cy and Cyc, cyclophosphamide; fx, fraction(s).

Data from Sampath S, Schultheiss TE, Wong J: Dose response and factors related to interstitial pneumonitis after bone marrow transplant. Int J Radiat Oncol Biol Phys 63(3):876-884, 2005; Figure 2,

Pneumonitis rates were not significantly different in a trial that randomized patients to high- or low-dose-rate TBI.⁸⁸ However, there is an abundance of retrospective clinical data suggesting that lowering the dose rate (<0.025 to 0.09 Gy/minute) does decrease the likelihood of pulmonary complications,^{130,136,140,147,148} especially if TBI is delivered as a single fraction.¹⁴⁹ If TBI is fractionated, some report that a low dose rate (<0.069 Gy/minute) is not necessary,¹³⁴ whereas others found a beneficial effect.^150,151 Studies of patients receiving fractionated TBI with lung shielding have demonstrated a reduction in pneumonopathy.^152,153 however, one RCT found no difference in pneumonopathy rates if the shielding allowed a lung dose of either 6 or 8 Gy in a single fraction.¹⁵⁴

Pulmonary function tests (PFTs) (e.g., spirometry and diffusion capacity) are often helpful in the assessment of patients with pulmonary symptoms or radiographic abnormalities. Studies of pulmonary function tests in patients treated with HSCT have demonstrated a deleterious effect on spirometry and diffusion capacity, which often resolves in the absence of other complicating factors^155–157 and is related to the TBI dose.¹⁵⁸ A retrospective study found that lung shielding had a small but significantly beneficial effect on pulmonary function tests 1 year after HSCT, especially in patients with abnormal function before HSCT.¹⁵⁹ In one study, busulfan and not TBI, was associated with a negative effect on PFTs.¹⁶⁰ There is no evidence that PFTs improve pulmonary outcome after HSCT in adulthood, and for this reason they are not recommended.¹⁶¹ However, some groups recommend baseline PFTs as part of long-term follow-up care for children treated with TBI.¹⁶² Counseling regarding smoking cessation is of critical importance for all patients, especially those who are at increased risk of developing lung injury. In the event of acute pneumonitis, high-dose steroids (30 to 60 mg of prednisone/day) typically alleviate symptoms within 24 to 48 hours.

Cardiovascular Toxicity

Cardiovascular toxicity in adults as a result of HSCT has been relatively rare, given the stringent selection criteria for patients treated with this aggressive therapeutic modality. Nevertheless, in adult patients who survived autologous or allogeneic HSCT, cardiac events were responsible for 2.4% and 3% of deaths, respectively; this finding represents a greater than expected occurrence.^163,164 Most of the recent literature has identified no association between TBI and the development of cardiovascular disease in adults.^165,166 In several detailed prospective analyses using plasma cardiac troponin and brain natriuretic peptide levels, electrocardiography and echocardiography revealed no evidence of cardiac dysfunction in previously healthy individuals treated with TBI.^167,168 However, a prospective study of children who underwent allogeneic HSCT found a 12% and 26% cumulative incidence of abnormalities in ejection fraction (<30%) on echocardiography before HSCT and 5 years after HSCT, respectively. This study revealed that TBI was associated with abnormalities in cardiac function on univariate analysis but not multivariate analysis; the 5-year cumulative incidence of cardiac abnormality was 26% or 2% in children treated with TBI, with or without prior anthracycline therapy, respectively.¹⁶⁹ Recent data from large studies of survivors of childhood cancer suggest that the risk of cardiac mortality is significantly elevated for children who receive heart doses of 5 Gy, although it should be noted that anthracycline doses of over 360 mg/m² similarly increased the risk.¹⁷⁰ Importantly, the rate of death from recurrence or progression of cancer exceeds the rate of death from cardiac disease by an approximate factor of 7.¹⁷¹ The reason for excess late cardiovascular toxicity in TBI in children is probably because of the association of TBI with the premature development of cardiovascular disease risk factors (e.g., hypertension, dyslipidemia, diabetes) that result in deleterious consequences rather than direct radiation cardiotoxicity.^172,173 Therefore it appears prudent to screen patients who have undergone HSCT for cardiovascular disease for risk factors, whether or not treatment involved TBI, to minimize late morbidity and mortality.

Hepatic Toxicity

Hepatotoxicity from TBI manifests primarily as veno-occlusive disease (VOD), also known as sinusoidal obstructive syndrome (SOS), of the liver. This clinicopathologic phenomenon was first described in 1977 by Shulman and colleagues,¹⁷⁴ who noted the onset of weight gain from ascites, painful hepatomegaly, and jaundice from centrilobular liver acinus necrosis 1 to 4 weeks after HSCT. Overall, up to 70% of patients who undergo HSCT can be affected by VOD. TBI, along with many other risk factors, has been implicated in the development of VOD.¹⁷⁵ An RCT and a meta-analysis found that patients treated with busulfan instead of TBI were significantly more likely to develop VOD.^176,177 Two RCTs have concluded that fractionated TBI reduces the incidence of VOD compared with single-dose TBI.^142,143 Another RCT found no difference in the rate of VOD when the dose rate was either 0.06 Gy/minute or 0.15 Gy/minute,⁸⁸ although a retrospective study of single-dose TBI found that dose rates of 0.07 Gy/minute were associated with less VOD than dose rates of 0.18 to 1.2 Gy/minute.¹⁷⁸ A TBI dose greater than 13.2 Gy has been reported to be associated with higher rates of VOD on univariate analysis,¹⁷⁹ although a dose greater than 12 Gy was not associated with VOD in another retrospective study.¹⁸⁰ Lawton and associates¹⁸¹ reported a nonstatistically significant 10% decrease in the rate of fatal VOD in patients treated with TBI as part of HSCT when a 10% attenuation liver block was employed. Ursodeoxycholic acid was effective in preventing VOD in an RCT of patients undergoing TBI followed by HSCT,¹⁸² although another RCT did not support this finding.¹⁸³ Reduced-intensity conditioning regimens may also prevent VOD. Treatment of VOD can include the fibrinolytic antithrombotic agent defibrotide; decompressing the sinusoids by a transjugular intrahepatic portosystemic shunt and liver transplantation is another, more invasive option for the management of severe disease.¹⁸⁴

Visual Toxicity

Cataracts are one of the most common complications of TBI. Patients may present with painless vision loss and may be noted to have opacification of the lens on examination. In one series of patients treated with TBI in one or two fractions, severe visual impairment was noted in about half of patients.¹⁸⁵ This problem has been noted to arise in a large proportion of patients treated with TBI, depending on the total dose, use of fractionation, and the dose rate. When considering risk factors associated with HSCT, steroid use,^186–188 prior cranial irradiation,^189,190 and the development of GVHD¹⁸⁸ have been shown to predispose to cataractogenesis, whereas heparin use appears to be protective.¹⁹¹ Delivering TBI in a single fraction is the single biggest risk factor for developing a cataract after HSCT.^191–196 High-dose-rate (>0.035 to 0.048 Gy/minute) TBI also appears to increase the risk of cataract formation.^{189,191,194,197} A prospective study that randomized patients to high- or low-dose-rate TBI found that the incidence of cataract 5 years after treatment was 12% and 34% in the low- and high-dose-rate arms, respectively, with 13% and 39% of the cataracts, respectively, occurring in patients who received fractionated or single-dose TBI.¹⁹⁸ Kal¹⁹⁹ and van Kempen-Harteveld²⁰⁰ recently reviewed the subject of TBI-induced cataracts and concluded that a biologically equivalent dose of 40 Gy yields a 10% chance of developing cataracts, using linear-quadratic modeling that included corrections for dose rate, with an α/β of 0.65 for late effects on the lens. On the basis of this, they suggest considering lens shielding for single-fraction TBI regimens,¹⁹⁹ although this is controversial in the setting of malignant disease.²⁰⁰ Given the frequency of cataracts occurring after TBI, patients should be monitored for the development of this complication.²⁰¹ Management of cataracts that impair vision or degrade quality of life may involve phacoemulsification and extraction; recent data suggest that these procedures are safe, with an adverse event rate of 0.1% for experienced surgeons,²⁰² and effective, with a 90% chance of 20/40 vision postoperatively.²⁰³

Renal Toxicity

Kidney dysfunction occurs in approximately 17% of survivors of HSCT²⁰⁴ and can manifest in a number of ways, with the most to least frequent syndromes being idiopathic chronic kidney disease, nephrotic syndrome, thrombotic microangiopathy (thrombotic thrombocytopenic purpura and hemolytic uremic syndrome), and acute renal failure.²⁰⁵ The syndrome most associated with TBI is thrombotic microangiopathy, which can manifest as nephritis, hypertension, proteinuria, and/or anemia 6 to 12 months after HSCT. HSCT-related risk factors for nephropathy can include GVHD, infections with cytomegalovirus or BK virus, nephrotoxic medications such as cytotoxic chemotherapeutic agents (cytarabine, cyclophosphamide, ifosfamide, cisplatin, retinoic acid, carmustine, actinomycin D, melphalan), antibiotics (acyclovir, ganciclovir, foscarnet, vancomycin, amphotericin, aminoglycosides), and immunosuppressants (cyclosporine, tacrolimus, methotrexate). The larger and more contemporary studies of chronic kidney disease after HSCT have not demonstrated an association with TBI.^206–210

Total dose has been implicated as the most important factor in predicting renal morbidity from TBI; a retrospective study found that GVHD and high-dose TBI (13.5 Gy) were associated with elevated serum creatinine levels.²¹¹ A prospective evaluation of renal function using radioisotopes found that early nephropathy was associated with age of less than 40 years, use of kidney blocks (possibly related to nephrotoxic contrast media given during simulation), and nephrotoxic drug use, whereas late nephropathy was associated with nephrotoxic drug use but not TBI dose.²¹² The benefit of kidney shielding has been assessed in two retrospective studies, both of which demonstrated significant improvement in long-term kidney function, with no evidence of dysfunction when the hyperfractionated dose was limited to 9.8 to 10 Gy.^213,214 Two recent dose-effect modeling studies have demonstrated that nephropathy is unlikely after a biologically equivalent total dose of 16 Gy (calculated using a linear quadratic model, with corrections for dose rate and an α/β of 2.5 Gy) and that fractionating TBI and delivery of a low dose rate (<0.10 Gy/minute) prevent kidney dysfunction.^199,215 Monitoring blood chemistry and counts, as well as blood pressure and urine studies, is advised given the prevalence of kidney disease after HSCT.

Although no therapies have been proven to treat HSCT-related nephropathy, medication therapies (antihypertensives, corticosteroids), plasma exchange, hemodialysis, and renal transplant are possible options for management.²⁰⁵ The angiotensin-converting enzyme inhibitor captopril prevents chronic nephropathy in patients with diabetes²¹⁶ and may reduce the risk of radiation nephropathy from TBI in preclinical studies.^217,218 A small RCT in which captopril was given to mitigate chronic renal failure in patients undergoing 12 Gy of fractionated TBI (with a kidney dose of 9.8 Gy) before HSCT revealed less nephropathy; however, the results were not statistically significant.²¹⁹

Endocrine Toxicity

Hypothyroidism is the most common endocrinopathy after HSCT.²²⁰ Most cases are overt and primary in nature, but subclinical and autoimmune thyroid dysfunction have also been noted. Hypothyroidism has been reported in up to 90% of patients after TBI delivered with a single dose of 10 Gy²²¹ and in up to 15% of patients treated with hyperfractionated TBI to 15 Gy.²²² These findings demonstrate the benefit of fractionation in avoiding this complication. One of the largest studies with a lengthy follow-up of children treated with or without TBI before HSCT revealed that TBI was not a risk factor for developing hypothyroidism when compared with equivalent conditioning with busulfan. In this study, the risk of hypothyroidism was higher in children undergoing HSCT before 10 years of age, and the risk continued to increase until 28 years after HSCT.²²³ A recent study demonstrated that reduced-intensity conditioning with lower doses of radiation during TBI was not associated with lower rates of hypothyroidism,²²⁴ suggesting that there is no dose threshold effect. Nevertheless, this complication should be monitored with assessment of thyroid hormones and can be easily managed with thyroid hormone replacement and monitoring.

Gonadal and reproductive endocrine functions can be altered by HSCT and TBI, as reviewed by Chemaitilly and Sklar.²²⁰ In males, Leydig cell function is typically preserved, given the dose of radiation delivered during full-dose TBI.²²⁵ However, in patients who have received prior testicular radiation or who will have a testicular boost during conditioning, Leydig cell function may be threatened. Germ cells are typically thought to be exquisitely sensitive to radiation. A recent study demonstrated that fertility is significantly reduced among boys treated with a radiation dose of more than 7.5 Gy to the gonads, although it should be kept in mind that other antineoplastics, such as alkylating agents, also profoundly reduce fertility.²²⁶ Although there have been case reports of men being treated with TBI and later having children, most men are rendered sterile by HSCT with full-dose TBI; most will experience azoospermia with reduced-dose TBI, but some may also be rendered sterile. With this in mind, male patients should be counseled regarding gamete cryopreservation before initiation of therapy. Likewise, postpubertal women are likely to experience ovarian failure as a result of intensive conditioning, including TBI, and should be counseled about this before treatment. However, about half of prepubertal girls who are treated with fractionated TBI will experience normal reproductive development.²²⁷ As reviewed by Schmidt and colleagues,²²⁸ several fertility-preserving options may be available to women who wish to bear children after HSCT.

Hypothalamic-pituitary function appears to be unaffected by TBI of adults to doses of 12 Gy.²²⁹ However, if additional brain radiotherapy has been or will be delivered, this complication could be encountered. Growth dysfunction may be caused by TBI through a variety of mechanisms, including growth hormone deficiency and skeletal dysplasia; other factors, including GVHD, liver dysfunction, and busulfan-based conditioning, may also contribute. Younger patients are more likely to be affected by growth impairment. In addition, children treated with single-fraction TBI (as opposed to fractionated TBI) are more likely to experience growth impairment.²³⁰ Growth hormone replacement therapy has been recommended to correct this deficiency,²³¹ although RCTs have not shown proven benefit²³² and treatment is associated with several significant toxicities.

Bone Toxicity

Bone health is another concern in patients that have undergone HSCT. Although TBI has been shown to be a risk factor for avascular necrosis of bone,²³³ a prospective study found no association between bone metabolism and conditioning with TBI; corticosteroid use is probably the major factor in HSCT causing bone health problems.²³⁴ Monitoring bone health with growth assessments, biochemical hormone assessments, and dual-energy x-ray absorptiometry (DEXA) scans is appropriate, and consideration of counseling, weight-bearing exercise, and use of calcium and vitamin D supplementation or antiresorptive agents (bisphosphonates) should be given in patients with evidence of abnormalities.

Nervous System Toxicity

Mild to moderate neurocognitive impairment has been noted in up to 60% of adults undergoing HSCT with TBI.²³⁵ However, these findings have not been consistent, because several studies have documented no impairment at all.^236,237 TBI-based conditioning (compared with non–TBI-based conditioning) has not been associated with neurocognitive impairment in adults who received full-dose²³⁸ or reduced-intensity conditioning.²³⁹ Children treated with TBI may experience mild neuropsychologic effects.²⁴⁰ The effects are more prominent in young children,²⁴¹ especially those under 3 years of age.²⁴² Similar to adults, no difference in neurocognitive function has been associated with TBI and non–TBI-based conditioning regimens in older children.²⁴³ However, in young children, TBI-based regimens appear to exert more of a negative effect than non–TBI-based regimens.²⁴⁴ The magnitude of the deficit has been described as statistically, but not clinically, significant (i.e., a deficit of 3 points in IQ) in recent investigations.²⁴⁵

Beyond generally mild cognitive effects of TBI on the nervous system, other types of toxicity are rare. Myelopathy is uncommon, but it has been reported to occur with TBI in conjunction with involved radiation therapy fields, even at cumulative doses considered to be “tolerable” (e.g., 45 Gy).²⁴⁶ Similarly, severe and fatal neurologic toxicity was reported in a series of children treated with TBI and was generally associated with prior whole brain radiotherapy to 18 Gy; therefore the cumulative radiation dose in this approach must be considered carefully.²⁴⁷ Other neurologic toxicity has not been associated with TBI in adults²⁴⁸ or children.²⁴⁹

Risk of Secondary Malignant Tumors

Beyond organ-specific toxicities, the risk of developing secondary malignant neoplasms is increased in patients who have undergone HSCT. Typically, three groups of secondary malignant neoplasms after HSCT are described, and they follow a distinct pattern of development²⁵⁰ (Fig. 18-5): myelodysplastic disease (MDS) and acute myelogenous leukemia (AML), posttransplant lymphoproliferative disorder (PTLD), and solid tumors.²⁵¹ There are multiple risk factors that may predispose to development of a secondary malignant tumor after HSCT, including, but not limited to, genetic aberrations, therapy given before HSCT, conditioning regimen, graft source and processing, posttransplantation immunosuppression, and GVHD³⁸; only the association of TBI with secondary malignant neoplasms will be discussed here.

Figure 18-5 Relative risk and chronology of second malignant tumors after allogeneic hematopoietic stem cell transplant (HSCT). GVHD, graft-versus-host disease; HD, Hodgkin’s disease; NHL, non-Hodgkin’s lymphoma.

Data from Ades L, Guardiola P, Socie G: Second malignancies after allogeneic hematopoietic stem cell transplantation. New insight and current problems. Blood Rev 16(2):135-146, 2002; Figure 1,

Secondary MDS or AML after HSCT has been reported to occur at rates between 1.1% at 20 months and 24.3% at 43 months. The World Health Organization classifies secondary MDS and AML as alkylating-agent and/or radiation-related disease, typically occurring 4 to 7 years after treatment, or topoisomerase-II inhibitor–related disease, typically occurring 6 months to 5 years after treatment.²⁵² Several studies have attempted to quantify the risk of developing secondary MDS or AML after undergoing TBI, with conflicting results. Studies from research groups from Minnesota,²⁵³ Newcastle,²⁵⁴ the City of Hope,²⁵⁵ and France²⁵⁶ have demonstrated no increase in rates of MDS or AML after myeloablative HSCT using TBI. Other studies from Nebraska,²⁵⁷ the European Group for Bone Marrow Transplantation (EBMT),²⁵⁸ Paris,²⁵⁹ Barcelona,²⁶⁰ and France²⁶¹ have found weak associations of borderline or no statistical significance. One study found an association of secondary MDS or AML with TBI on multivariate analysis when TBI was used with etoposide and cyclophosphamide.²⁶² A recent, detailed study of this subject found that among 4000 patients treated with autologous HSCT for lymphoma, 57 developed MDS or AML, with a cumulative incidence rate of 3.7%, 7 years after HSCT. A case-control analysis revealed a weak association of TBI and the subsequent development of MDS or AML that was not statistically significant. In a small subgroup analysis, a TBI dose of 13.2 Gy was associated with MDS or AML, but the authors cautioned that this group of patients may have shared other non–TBI-related factors that could have contributed to leukemogenesis. Older age also appeared to increase the risk.²⁶³ The dose threshold effect was not supported by a more recent study that found no differences in the rate of MDS or AML after TBI, whether 12 or 14 Gy of TBI was given.²⁶⁴

The association of TBI with the subsequent development of a PTLD has been described in several studies. Because it is thought to result from immune system dysfunction, PTLD occurs primarily in patients following allogeneic, not autologous, HSCT. The cumulative incidence of PTLD at 12 years varied by the number of risk factors the patient had, and ranged from 0.2% to 8.1%.²⁶⁵ Small early studies from Seattle²⁵¹ and the CIBMTR²⁶⁶ suggested an association between TBI use during HSCT and subsequent development of a PTLD. However, later research from Minnesota,²⁵³ Vancouver,²⁶⁷ and Nebraska²⁶⁸ suggested no association between TBI and PTLD. A follow-up study from the CIBMTR, the largest and most comprehensive study available, involving 26,000 patients who underwent allogeneic HSCT over a 30-year period, clearly demonstrated no association between PTLD and TBI, in contrast to their previous findings.²⁶⁵

The risk of solid tumors after HSCT involving TBI has been studied extensively.^269–279 The largest and most recently updated analysis from the CIMBTR suggests that among patients who undergo allogeneic HSCT, the cumulative incidence of developing a solid malignant tumor (including carcinomas in situ that are not of the skin) is 1%, 2.2%, and 3.3% at 10, 15, and 20 years after transplant, respectively, based on the competing risk analysis method.²⁸⁰ The number of secondary solid tumors was double what would have been expected in this population; significantly higher rates of oral cavity and pharyngeal, liver, central nervous system, thyroid, bone, and soft tissue cancers and melanomas were observed. TBI was found to be a risk factor for developing a secondary solid tumor, with the risk increased in those who survived more than 5 years from the HSCT and in those less than 30 years old at the time of HSCT. Although an earlier study found that TBI doses of less than 10 Gy were associated with fewer secondary solid tumors,²⁶⁹ this dose association was not maintained in the subsequent analysis.²⁸⁰ There was no difference in the rate of secondary solid tumors in patients treated with fractionated or single-dose TBI. There was no increased risk of squamous cell carcinomas in patients treated with TBI and no increased risk of secondary solid tumors in patients treated with TBI after age 30. Chronic GVHD and male sex were associated with the development of secondary solid tumors and squamous cell carcinoma, respectively.²⁸⁰

Given the increased risk of secondary malignant neoplasms after TBI, careful surveillance of long-term survivors is a desirable (but not yet proven effective) strategy for minimizing secondary morbidity and mortality. Guidelines published by the American Cancer Society,²⁸¹ the National Comprehensive Cancer Network (NCCN) (www.nccn.org), and the Children’s Oncology Group¹⁶² provide important references regarding assessing patients at higher risk of developing a secondary solid tumor.

Comparison of Techniques of Total Body Irradiation

A report from 1980 highlighted several approaches that have been explored in TBI.²⁸⁵ The common factor in the different ways of delivering TBI is the goal, which is to deliver the prescribed dose of ionizing radiation to the entire body in a uniform manner (±10% of the prescription dose). An opposed-parallel anterior-posterior (AP/PA) horizontal field setup was developed at Memorial Sloan-Kettering Cancer Center (MSKCC) many years ago.²⁸⁶ In this technique, the patient is positioned upright several meters from the source with the support of a custom-designed stand. In this technique, high-energy photons are used to increase dose homogeneity throughout the patient. As mentioned previously, high-energy photons will reduce the dose in superficial structures (skin and subcutaneous tissues), and therefore a beam spoiler is used to produce electrons that will increase the surface dose. The beam spoilers are affixed to the stand, which can also accommodate blocks for shielding, imaging equipment for field and patient position verification, and placement of dosimeters (Fig. 18-6). When using shielding, compensatory electron boosts can be delivered to superficial tissues such as those of the chest wall (Fig. 18-7).

Figure 18-6 Patient in position for total body irradiation using anteroposterior beam arrangement, with lung blocks in place.

Figure 18-7 CT axial section demonstrating electron-boost dose distribution.

An alternative to the AP/PA field arrangement is a lateral field arrangement, pioneered by Khan.²⁸⁷ In this approach, the patient is able to recline in a comfortable position during treatment, typically with the knees drawn toward the chest and the arms by the side. As with the AP/PA technique, the patient is typically positioned several meters from the radiation source. In this technique, the arms can provide shielding to the lungs from the lateral beams. However, as the lateral body thickness is quite disparate over the head-to-toe length of a patient, compensators (usually fixed to the gantry head) will often be required to reduce the dose to the head, neck, and legs. Positioning with this technique is critical; therefore performing and confirming several measurements as well as determining symmetry using in-room light fields and lasers are essential.

Although the AP/PA and lateral field arrangements are the most commonly employed TBI techniques, other methods exist. Children undergoing TBI may require conscious sedation during radiation therapy, making the previously described techniques impossible. In this situation, TBI can be delivered with the child lying on or near the floor. When a patient is in this position, the gantry is pointed down. The child’s smaller body size often will facilitate this arrangement; however, in larger children it may be necessary to match two fields. An alternative to this would be translating the patient or radiation source during treatment.²⁸⁶ Logistic considerations largely prohibit this type of approach.

In recent years, several novel approaches to delivering radiation therapy to the entire body have been described. One of these approaches involves rotational intensity-modulated radiation therapy (delivered by a TomoTherapy unit). This approach appears dosimetrically advantageous because it could allow targeting and radiation dose escalation in the bone marrow, and for this reason, it is often referred to as total marrow irradiation (TMI).²⁸⁸ However, few clinical data exist to support its use. Moreover, sparing the circulating hematopoietic cells from radiation therapy might defeat the purpose of delivering radiation to the entire body. Another alternative approach of delivering systemic radiation therapy is using radionuclides conjugated to monoclonal antibodies, commonly referred to as radioimmunotherapy, instead of external beam radiotherapy. The feasibility of this concept has been demonstrated by several groups, with encouraging results in allogeneic^289,290 and autologous HSCT.^291,292 The uncertainty in biodistribution and clearance of this form of radiotherapy is the major drawback.

Because there are no clinical data to suggest that one technique is advantageous over others, dosimetric benefits have largely dominated the choice of which TBI method to use. Because the AP/PA TBI technique is employed most frequently and is preferred under most conditions, the remaining discussion will focus on its application.

Simulation and Planning for AP/PA Total Body Irradiation

Patients undergoing TBI should be simulated in the treatment position, which is often most easily reproduced in the treatment room. For AP/PA TBI, height must be measured to ensure that the field will adequately encompass the entire patient; exceptionally tall patients might need to sit on the seat of the treatment stand. Other measurements of the AP thickness at the suprasternal notch, chest, abdomen, and pelvis are performed for dosimetric calculations.

While the patient is in the treatment position, radiographs of the thoracic cavity can be obtained to create lung shielding or other blocks. Careful attention should be paid to the size of the blocks so that they are produced in a manner proportional to the treatment field size and distance from the source; this often requires a magnification correction factor. For example, lung shielding is typically designed so that block edges are 1 to 2 cm from the bony thoracic periphery, diaphragm, and vertebral bodies. When shielding is used, some institutions choose to perform compensatory electron boosts matched to the blocks. In these situations, CT can be helpful for planning purposes (Fig. 18-7), although the obvious limitation is that CT cannot be performed when the patient is in the treatment position. Alternative imaging modalities (e.g., ultrasound) performed in the treatment position may be helpful in specific situations (e.g., chest wall boosts in patients with large, pendulous breasts).

Beam Energy, Angling, and Production

A wide variety of beam energies have been used for TBI. Most radiation oncology departments now use linear accelerators to generate ionizing radiation, although it should be noted that ⁶⁰Co generators, which typically produce lower-energy photons than linear accelerators, were the source for many of the early clinical studies involving TBI. As noted above, higher beam energies will achieve a more uniform dose distribution. Moreover, high-energy photons produced by linear accelerators may provide doses lower than desired to superficial structures, so beam attenuation may be needed. The thickness of the patient along the beam axis is the dominant factor dictating the choice of beam energy; typically, lateral field arrangements require higher beam energies than AP/PA field arrangements; however, this ultimately depends on patient size, beam spoilers, and source-axis distance. The source-axis distance can be up to 5 meters; it will often depend on the size of the treatment delivery room relative to the size of the radiation field (dictated primarily by the size of the patient) and may restrict the application of some specific geometric approaches.

The central axis of the beam should be incident at the umbilicus; this usually requires around 5 degrees of gantry angling. The collimator is usually opened to its maximum size (40 × 40 cm) and is rotated to 45 degrees, maximizing the height of the field. The total dose can be prescribed to any anatomic point, but it is often prescribed in the midplane of the pelvis or at the umbilicus.

Compensation

As noted above, higher beam energies will reduce the dose delivered to the superficial structures. For this reason, beam spoilers are often used. Typically an acrylic screen 1 to 2 cm thick is used to produce electrons. These electrons should increase the surface dose to at least 90% of the prescribed dose. If required, shielding is affixed to the plastic holder on the beam spoiler and placed using reference marks on the patient. For example, at MSKCC, a half-value layer block is employed for lung shielding. Electronic portal imaging can be used to confirm accurate placement of shields. Electron boost fields matched to the lung shields can also be set up using reference marks on the patient.

Calibration and Quality Assurance

Several groups recommend calibrating the treatment delivery unit for TBI using a dosimeter in a large water tank at the distance that will be used for treatment, in the treatment stand, if possible. Given the extended distance, large field size, and compensators used in this technique, a significant amount of scatter radiation will be produced and must be accounted for during calibration. For this reason, the typical percent depth dose (PDD) or tissue maximum ratio (TMR) measurements used for external beam radiation therapy should not be used for TBI; these measurements must be made specifically for the TBI setup. Moreover, variations in dose based on patient size and dose heterogeneity throughout the body should be anticipated and verified during the calibration process. In vivo dosimetry should be performed at several locations (i.e., along the central beam axis, under shielding blocks, and at distal extremities) to verify accurate dose delivery. Regular quality assurance checks should be undertaken according to AAPM guidelines.

Necessity of Using Total Body Irradiation in HSCT Conditioning

Although the earliest studies of HSCT for malignant hematopoietic diseases involved both chemotherapy and radiotherapy (see earlier section on history), in the 1970s several regimens using only chemotherapy (without TBI) were developed, largely for logistic reasons, because facilities adequate for performing TBI were not widely available.²⁹³ Soon thereafter, the availability of conditioning regimens for HSCT without TBI (primarily involving busulfan and cyclophosphamide),²⁹⁴ gave rise to several important clinical studies that investigated the necessity of TBI in HSCT in various hematopoietic diseases.

Acute Myelogenous Leukemia (AML)

There have been two RCTs comparing TBI-based conditioning regimens with non–TBI-based conditioning regimens for AML. In 1987, the Groupe d’Etude des Greffes de Moelle Osseuse (GEGMO) initiated a prospective multicenter RCT to evaluate whether TBI could be replaced by busulfan, in an effort to limit toxicity in allogeneic HSCT.²⁹⁵ A similar single-institution study was initiated at the University of Minnesota, for patients receiving autologous HSCT, but accrual was poor and the study was terminated early.²⁹⁶ As noted in web-only Table 18-3, available on the Expert Consult website, superior rates of overall survival, disease-free survival, and survival in general and fewer relapses were noted with the TBI conditioning regimen in both trials, with only the French trial demonstrating a statistically significant difference.^295,297 An updated combined analysis of three trials of allogeneic HSCT for AML that randomized patients to conditioning with or without TBI found marginally significant improvements in overall survival and disease-free survival rates. In patients treated with TBI, the estimated 10-year survival rate was 63%, and in those not treated with TBI, the rate was 51% (p = .068); TBI-treated patients had a 10-year disease-free survival rate of 57%, and untreated patients had a rate of 47% (p = .051). On multivariate analysis, TBI was associated with significantly less hair loss and no difference in the rates of GVHD, cataracts, avascular necrosis, or pulmonary complications.²⁹⁸ A retrospective analysis by the CIBMTR revealed that TBI-based conditioning yielded fewer relapses (especially in extramedullary and central nervous system sites) and less hepatic VOD than busulfan-based conditioning.²⁹⁹

Chronic Myelogenous Leukemia (CML)

Two RCTs have compared the efficacy and toxicity of allogeneic HSCT in patients with CML, preceded by conditioning with either TBI and cyclophosphamide or busulfan and cyclophosphamide^300–302 (see web-only Table 18-3, available on the Expert Consult website). Both studies showed no significant difference in rates of overall survival, event-free survival, or transplant-related mortality. The French study found more relapses in patients treated with TBI (especially fractionated TBI), but this was not seen in the study from Seattle, which used fractionated TBI exclusively. The Seattle trial did show significantly longer episodes of fever, more blood cultures revealing bacteria or fungi, more hospitalizations in the 100 days after transplantation, and a higher incidence of grade 2 to 4 GVHD in patients treated with TBI. An updated, combined analysis of three studies that randomized patients with CML to different conditioning regimens before allogeneic HSCT confirmed no significant difference in rates of survival (65% and 63%) and disease-free survival (52% and 42%) for patients treated with busulfan and cyclophosphamide or cyclophosphamide and TBI, respectively. There were no differences in the 5-year cumulative incidence of chronic GVHD (37% and 39%). However, cataracts were more common among patients treated with TBI.²⁹⁸ Retrospective data also suggest that busulfan-cyclophosphamide–conditioning regimens before allogeneic HSCT for CML may be favored because of lower relapse rates,³⁰³ although higher rates of GVHD, hepatotoxicity, and hemorrhagic cystitis have been reported with chemotherapy alone.³⁰⁴

Acute Lymphoid Leukemia (ALL)

The Pediatric Blood and Marrow Transplant Consortium conducted the only RCT comparing the effect of pre-allogeneic HSCT conditioning using busulfan or TBI, in children with ALL. A central nervous system boost of 6 Gy was given to patients with a history of prior central nervous system disease but no prior central nervous system irradiation if randomized to TBI; if randomized to busulfan, patients with a history of central nervous system disease received 18 Gy of central nervous system radiation therapy before receiving busulfan. The study showed no difference in overall survival rates; however, longer event-free survival times were noted for patients treated with TBI. Patients 6 years old or younger, as well as those in their first complete remission, appeared to experience the most benefit from the TBI regimen.³⁰⁵ A retrospective study by Davies³⁰⁶ using the CIBMTR data of 627 children with ALL who underwent HSCT after conditioning with cyclophosphamide and TBI (CY/TBI) or busulfan and cyclophosphamide (Bu/CY) found that the risk of relapse was not significantly different in the two cohorts. However, the risk of treatment-related mortality was significantly greater in the Bu/CY group. The use of CY/TBI was associated with superior rates of overall survival and leukemia-free survival. Based on these data, the authors concluded that less treatment-related mortality in the CY/TBI group led to greater rates of overall survival and leukemia-free survival. A multivariate analysis of adults with ALL who underwent HSCT preceded by conditioning with fractionated TBI to 12 Gy (lung dose of 9 Gy) or busulfan found that shorter times of event-free survival and disease relapse were more likely among patients who did not receive TBI on multivariate analysis. There were no differences in the rates of transplant-related mortality, hepatic VOD, or GVHD.³⁰⁷

Multiple Myeloma

The Intergroup Francophone du Myelome conducted a trial (9502) that randomized 399 patients with newly diagnosed multiple myeloma to high-dose therapy in preparation for autologous HSCT, after three cycles of vincristine-Adriamycin-dexamethasone chemotherapy. High-dose therapy consisted of melphalan 200 mg/m² (HDM200) or melphalan 140 mg/m² and 8 Gy of TBI delivered in four daily fractions without lung shielding (HDM140). Overall survival rates were longer in the HDM200 group (p = .05); however, event-free survival rates were not different in the two groups. The authors speculated that the improved overall survival rate was attributed to the better salvage regimens available after initial therapy with HDM200 in the protocol and suggested that the new standard conditioning regimen for autologous HSCT for MM should be without TBI.³⁰⁸

Mixed Diseases

Two prospective trials of conditioning have been carried out in groups of patients with leukemias and lymphomas. Because these trials included a heterogeneous group of diseases, results should be interpreted with caution. Nevertheless, in 1988, the Nordic Bone Marrow Transplantation Group initiated a multicenter RCT of 167 patients (27 children and 140 adults); 68 patients had AML, 38 had ALL, 57 had CML and 4 had lymphoma before allogeneic HSCT.¹⁷⁶ With 7 years of follow-up, the authors found a survival benefit in the TBI group among patients with “advanced” disease, with less toxicity compared with busulfan regimens; no differences were identified in the subset analysis of the different diseases that were treated.¹¹² The Southwest Oncology Group’s RCT (SWOG 8612) also investigated the role of TBI before allogeneic HSCT for ALL (n = 48), AML (n = 40), and CML (n = 34) patients who were not in their first remission. This regimen substituted etoposide for cyclophosphamide in the TBI group only, making a meaningful comparison difficult. Although rates of overall survival and disease-free survival appeared to be superior for “good-risk” patients in the TBI group, the differences were not statistically significant.³⁰⁹

In the setting of allogeneic HSCT, a meta-analysis of the RCTs by Hartman and colleagues¹⁷⁷ summarized how TBI influences the outcome of conditioning regimens before allogeneic HSCT. As illustrated in Figure 18-8, overall survival and disease-free survival rates are slightly better with TBI-based regimens, although the results were not statistically significant. As shown in Figures 18-9 and 18-10, rates of GVHD were roughly equivalent in the studies, whereas hepatic VOD was clearly more common with busulfan-based regimens. The occurrence of interstitial pneumonitis also appeared to be similar in both groups. Another recent meta-analysis of 18 prospective and retrospective studies comparing TBI-based conditioning regimens with those not involving TBI (busulfan-based regimens) found that TBI improved disease-free survival rates in AML and ALL but not in CML. Growth and development problems, interstitial pneumonitis, and cataracts were more common with TBI-based regimens, but transplant-related mortality, VOD, and hemorrhagic cystitis were more common with busulfan-based regimens.³¹⁰

Figure 18-8 Odds ratio and 95% confidence intervals of survival and disease-free survival in a meta-analysis of total body irradiation–based conditioning versus non–total body irradiation–based conditioning in five randomized controlled trials (RCTs).

From Hartman AR, Williams S, Dillon JJ: Survival, disease-free survival and adverse effects of conditioning for allogeneic bone marrow transplantation with busulfan/cyclophosphamide vs total body irradiation. A meta-analysis. Bone Marrow Transplant 22(5):439-443, 1998; Figure 1,

Figure 18-9 Odds ratio and 95% confidence intervals of acute and chronic graft-versus-host disease (GVHD) in a meta-analysis of total body irradiation–based conditioning versus non–total body irradiation–based conditioning in five and four randomized controlled trials (RCTs), respectively.

From Hartman AR, Williams S, Dillon JJ: Survival, disease-free survival and adverse effects of conditioning for allogeneic bone marrow transplantation with busulfan/cyclophosphamide vs total body irradiation. A meta-analysis. Bone Marrow Transplant 22(5):439-443, 1998; Figure 2,

Figure 18-10 Odds ratio and 95% confidence intervals of hepatic veno-occlusive disease and interstitial pneumonitis in a meta-analysis of total body irradiation–based conditioning versus non–total body irradiation–based conditioning in 4 and 3 randomized controlled trials (RCTs), respectively.

From Hartman AR, Williams S, Dillon JJ: Survival, disease-free survival and adverse effects of conditioning for allogeneic bone marrow transplantation with busulfan/cyclophosphamide vs total body irradiation. A meta-analysis. Bone Marrow Transplant 22(5):439-443, 1998; Figure 3,

There is much less data regarding the role of TBI before autologous HSCT. A non–TBI-based conditioning regimen is preferred for multiple myeloma (MM); however, no other trials have randomized patients to a TBI or non-TBI conditioning regimen before autologous HSCT, and, therefore, institutional preference primarily dictates choice of the technique in these situations.

Dose of Total Body Irradiation

The appropriate radiation dose to use during conditioning with TBI has been investigated by several groups. In 1985, researchers in Seattle conducted an RCT of two TBI regimens as part of the HSCT for AML. Seventy-one patients were randomized to receive 12 Gy in six daily fractions (2 Gy/fraction/day) or 15.75 Gy in seven daily fractions (2.25 Gy/fraction/day). TBI was delivered using opposed ⁶⁰Co sources, at 0.06 to 0.07 Gy/minute. The use of shielding was not described. Higher-dose TBI (15.75 Gy) was associated with higher rates of mortality in patients who did not relapse, although the authors note that this was limited to the first 6 months after transplant and to higher rates of grade 2 to 4 acute GVHD (p = .02). Patients receiving higher-dose TBI were less likely to receive adequate immunosuppression. Nonetheless, patients treated with lower-dose TBI (12 Gy) appeared to be more likely to relapse, although this finding was only marginally significant (p <.06).¹³⁷ A report of the long-term results (minimum follow-up of 7.5 years) of the trial revealed no difference in overall survival rates between the two TBI groups.¹³⁹

In a similar trial, also from Seattle, 116 patients with CML were randomized to receive either 12 Gy or 15.75 Gy, as described above. Higher-dose TBI (15.75 Gy) was associated with a superior relapse-free survival rate and with more grade 2 to 4 acute GVHD, although this was not statistically significant (p = 0.15). Initially, survival rates appeared to be superior in patients receiving lower-dose TBI (12 Gy). However, further follow-up revealed no significant difference in survival rates. Similar to the experience with AML, the authors concluded that higher-dose TBI was associated with greater relapse-free survival but at the cost of more acute GVHD, which likely led to transplant-related mortality that offset the survival benefit.¹³⁸

Retrospective data suggest a benefit to higher-dose (>12 Gy) TBI in certain clinical scenarios. The highest doses of TBI reported in the literature are 20 Gy before autologous HSCT,³¹¹ but at least one study has reported lower relapse rates and less transplant-related mortality in a subset of patients with ALL who received a TBI dose of more than 13 Gy.³¹² Although modeling studies suggest that higher biologically effective doses are associated with lower relapse rates and longer overall survival times,³¹³ the finding of excess toxicity with higher doses in clinical dose-escalation studies does warrant caution.^{314,315,316,317} For these reasons, fractionated TBI doses used before myeloablative HSCT typically range between 12 and 15 Gy.

Reduced-intensity conditioning regimens have been used more often over the last decade for patients who might not be able to tolerate a myeloablative HSCT (because of advanced age or comorbidities) but could benefit from the graft-versus-tumor effects of allogeneic HSCT. The role of TBI as part of reduced-intensity regimens is ill defined because prospective RCTs have not been completed. Likewise, the optimal dose of TBI as part of reduced-intensity autologous HSCT has not been defined in an RCT. Several notable studies of allogeneic HSCT have employed TBI doses of 2 to 8 Gy, often in conjunction with chemotherapy.^318–333 Web-only Table 18-4, available on the Expert Consult website, lists some representative allogeneic HSCT regimens using low-dose TBI.

Fractionation and the Total Body Irradiation Dose Rate

Laboratory-based radiobiologic data (see the previous section on Radiobiology) and clinical observations prompted trials investigating fractionation and dose-rate delivery in TBI. Thomas and colleagues were the first to conduct an RCT of single-dose or fractionated TBI in patients with AML. In this study, 53 patients with AML were randomized to receive TBI in a single 10-Gy dose or in a dose of 12 Gy in six daily fractions after receiving cyclophosphamide. TBI was delivered using opposed ⁶⁰Co sources, at a rate of 80 R/minute in air at midpoint (~0.06 Gy/minute). Disease-free survival times were significantly longer in the patients who were treated with fractionated TBI. There was no apparent difference in acute toxicity in the two treatment groups.⁸⁹ A follow-up report from the same institution described similar results using a slightly lower dose of single-fraction TBI (9.2 Gy using opposed ⁶⁰Co sources at 0.04 Gy/minute). Amongst the patients randomized, those who received fractionated TBI survived significantly longer than those receiving single-fraction TBI. Patients receiving single-fraction TBI experienced twice as many leukemic recurrences compared with those receiving fractionated TBI, although this was not statistically significant (p = .09). VOD of the liver was significantly more common in patients receiving single-fraction TBI (p = .02).¹⁴²

Ozsahin and colleagues⁸⁸ conducted a study investigating the effect of dose rate in patients undergoing single-dose or fractionated TBI as part of autologous or allogeneic HSCT for a variety of hematologic and lymphoid cancers. One-hundred fifty-seven patients receiving single-dose or multiple-fraction TBI were randomized to a low dose rate (0.06 Gy/minute or 0.03 Gy/minute, respectively) or a high dose rate (0.15 Gy/minute or 0.06 Gy/minute, respectively). Radiotherapy was delivered with ⁶⁰Co or 6-MV photons. Single-dose TBI was delivered to a total dose of 10 Gy with the lung dose limited to 8 Gy. Fractionated TBI was delivered to a total dose of 12 Gy in six fractions over 3 days, with the lung dose limited to 9 Gy. A variety of chemotherapies were used. The authors found no significant differences in rates of overall survival or relapse-free survival, pneumonitis, VOD, or GVHD. Cataracts were significantly more common amongst patients receiving single-fraction, high-dose-rate TBI. In a follow-up report on this study, in which the median follow-up time was 50 months, it was demonstrated that TBI in a single fraction or at a high dose rate was associated with the development of cataracts.¹⁹⁸

Girinsky and colleagues¹⁴³ conducted an RCT of 160 patients older than 15 years who were undergoing autologous or allogeneic HSCT for a variety of hematologic and lymphoid cancers. Patients were treated with a single fraction of TBI (STBI) with 10 Gy over 4 hours (average dose rate of 0.045 Gy/minute; instantaneous dose rate of 0.125 Gy/minute) or with hyperfractionated TBI (HTBI) with 14.85 Gy in 11 fractions over 5 days at a dose rate of 0.25 Gy/minute. A linear accelerator was used to deliver 18-MV photons with lung shielding; patients receiving STBI received a lung dose of 8 Gy, and HTBI patients received a lung dose of 9 Gy. After TBI, cyclophosphamide or melphalan was given, followed by HSCT. With a median follow-up of 8 years in 147 assessable patients, rates of overall survival and cause-specific survival were higher in the HTBI group; however, the differences were not statistically significant. A multivariate analysis that controlled for prognostic variables revealed greater cause-specific survival rates but not greater overall survival rates in patients treated with HTBI (p = .04). There was no significant difference in the rates of lethal pneumonitis, hepatic VOD, GVHD, graft failure, infection, or organ failure. However, the rate of hepatic VOD was significantly higher in the STBI group.

The necessity of hyperfractionation versus conventional daily fractionation has never been tested in a prospective trial. Investigators from the M.D. Anderson Cancer Center compared the outcome of patients in different prospective HSCT protocols and found that fewer patients in the conventionally fractionated group had recurrences of disease. These results should be interpreted with caution because this study was not randomized and other factors not related to TBI may have influenced the results.¹⁴⁵ Investigators in Seattle have conducted phase I/II studies of twice- and thrice-daily hyperfractionation protocols that did not suggest a significant benefit with hyperfractionation versus conventional fractionation when they were compared with historical results.^316,334

Organ Shielding during Total Body Irradiation

The necessity of shielding to prevent normal tissue toxicity has been investigated in several studies; because pneumonopathy has been considered the dose-limiting toxicity of TBI, the studies have primarily been concerned with lung shielding. Labar and associates¹⁵² conducted a trial in 64 patients with AML, ALL, and CML who were undergoing TBI and cyclophosphamide conditioning. Patients were treated with 12 Gy of TBI in three fractions given once daily and were randomized to TBI with or without lung shielding after stratifying by age group (<20, 21 to 30, 31 to 40, >40 years), sex, underlying disease, GVHD prophylaxis, and donor-recipient sex combination. TBI was delivered using a ⁶⁰Co source, and blocks were used to limit the estimated lung dose to 9 Gy in the group of patients treated with lung shields. There was no difference in rates of leukemia-free survival or relapse. Although interstitial pneumonitis was more common in the group of patients treated without lung blocks (12%) compared with those treated with lung blocks (4%), this difference was not statistically significant.

Girinsky and colleagues¹⁵⁴ conducted a similar study in a group of 85 patients with various lymphoid or hematopoietic cancers. Each patient received TBI with a dose of 10 Gy given in a single fraction over 4 hours (average dose rate of 0.04 Gy/minute; instantaneous dose rate of 0.125 Gy/minute). Patients were randomly assigned to receive a 6- or 8-Gy lung dose. The authors observed no differences in rates of overall survival or lung complications. However, chronic GVHD and infections were significantly less common in the low lung dose group. Interestingly, there were significantly more leukemia relapses in the low lung dose group. The authors speculated that the additional lung shielding may have fostered survival of leukemic cells, which led to higher rates of relapse.

Retrospective data support the benefit of shielding. Weshler and colleagues¹⁵³ reported a series of 44 consecutive patients who received 12 Gy of TBI; the first 23 patients were treated with opposed lateral fields and had no lung shielding in place, whereas the subsequent 21 patients had half-value layer transmission blocking after 6 Gy. Pneumonitis occurred in 26% of patients without lung shielding, whereas no pneumonitis was noted in those who had lung shielding. Investigators from Wisconsin have demonstrated sparing of renal and hepatic dysfunction with shielding.^181,213

Boosting in Total Body Irradiation

There is scant high-level evidence for integrating a boost into the radiotherapy plan involving TBI. The EBMT conducted an RCT to assess the benefit of splenic boosting before allogeneic HSCT in patients with CML in their first complete remission. Two-hundred twenty-nine participants were enrolled and stratified based on age (25 years) and T-cell depletion. Patients were randomized to receive a 10-Gy boost to the spleen (given in one, two or three fractions) within 14 days of HSCT. Overall, there was no difference in the groups. However, in subgroup analysis, it was determined that patients who did not undergo T-cell depletion and who had more than 3% circulating basophils before transplant demonstrated a statistically significant improvement in overall survival rates.³³⁵ At MSKCC, the testes of males with leukemia undergoing myeloablative HSCT are boosted with 4 Gy in one fraction, based on the results of a retrospective series that revealed significantly lower rates of testicular relapse with this treatment strategy.³³⁶ These findings have never been validated in a prospective study, but the other institutions have incorporated the findings in their radiotherapy strategies, as discussed by Quaranta and colleagues.³³⁷ In other situations, investigators have reported success in delivering an involved field boost to the central nervous system (in leukemia)³³⁸ or bulky nodal disease (in lymphoma),³³⁹ although the efficacy of these approaches has never been validated in prospective trials.

Total Body Irradiation in Nonmalignant Diseases

Because radiotherapy is most commonly employed in the treatment of malignant diseases dealt with in other sections of this book, special attention will be paid to the role of TBI as part of HSCT for nonmalignant diseases. Although there are reports of using TBI as part of conditioning before HSCT for nonmalignant acquired diseases such as severe aplastic anemia, paroxysmal nocturnal hemoglobinuria, systemic amyloidosis, and the acquired immunodeficiency syndrome caused by the human immunodeficiency virus, in general TBI is avoided in these situations.

Although no RCT data exist to support the use of TBI in the conditioning regimen of patients undergoing autologous HSCT for autoimmune conditions, several pilot studies have prompted the execution of phase III studies incorporating TBI. The basis for the role of TBI lies in preclinical studies of a rat model of autoimmune arthritis treated with pseudo-autologous HSCT. When the highest tolerable doses of TBI, cyclophosphamide, or busulfan were given, TBI was the most effective single conditioning agent.³⁴⁰ Saccardi and associates³⁴¹ summarized the results of phase I and II trials carried out by the EBMT using HSCT for multiple sclerosis (MS), which revealed stabilization or improvement of neurologic symptoms in 63% of patients, who generally had severe or progressive disease. Collectively, TBI was part of the conditioning regimen in 16 of the 178 patients. Statistical analysis revealed that busulfan, and not TBI, was associated with transplant-related mortality. Interestingly, Nash and colleagues³⁴² studied 26 patients in a pilot study using high-dose myeloablative conditioning before autologous HSCT, including 8 Gy of TBI, and found that the 3-year progression-free survival rate was 73%, with a 91% overall survival rate, with minimal toxicity. Concerns about the use of TBI in HSCT for MS were brought up by Samijin and associates³⁴³ as well as Burt and colleagues,³⁴⁴ who speculated that radiotherapy may have contributed to axonal injury or caused dysfunction of neural precursor cells, although direct evidence of this was lacking. For these reasons, the ongoing RCT to compare autologous HSCT (ASCT) with mitoxantrone drug therapy for severe MS will not involve TBI (www.astims.org/). Similar concerns regarding the toxicity of radiotherapy in patients with inflammatory bowel disease have prompted investigators to omit TBI from ongoing RCTs of autologous HSCT in Crohn’s disease.³⁴⁵

As in studies of MS and Crohn’s disease, autologous HSCT has been explored recently for patients with systemic sclerosis (SSc, scleroderma) in several pilot studies. Binks and colleagues³⁴⁶ reported the results of a phase I/II trial of autologous HSCT for poor-prognosis systemic sclerosis, using a variety of conditioning regimens, some of which included TBI. In their initial report, two of nine patients who received 8 Gy of TBI died of interstitial pneumonitis 1 and 3 months after TBI. This prompted investigators to minimize the TBI dose during further investigations of HSCT in this disease.³⁴⁷ After two of the first eight patients in a study by Nash and associates³⁴⁸ died of pneumonitis, these investigators shielded the lungs with five half-value layers after patients received the first 2-Gy fraction. With shielding in place, no further TBI-related pulmonary mortality has occurred; however, two patients developed myelodysplastic syndrome and a former smoker was diagnosed with stage 1 lung cancer.³⁴⁹ Autologous HSCT for systemic sclerosis is currently being studied in two RCTs; in the United States, the Scleroderma Cyclophosphamide or Transplantation (SCOT) trial, NCT00114530, will incorporate 8 Gy of TBI, whereas in the Autologous Stem Cell Transplantation International Scleroderma (ASTIS) trial in Europe (www.astistrial.com), TBI will not be used.

Several inherited diseases have been successfully treated with HSCT using TBI for conditioning; however, this approach is not desirable for children, who frequently need HSCT at a very young age. HSCT for thalassemia has been carried out in several Italian centers and generally does not involve TBI. However, reduced-intensity regimens involving low doses of TBI have been used in a few small series of patients undergoing HSCT for thalassemia and sickle cell disease with favorable results.^350,351 Immunodeficiency syndromes are often treated with HSCT, but TBI is not commonly employed during conditioning, probably because these syndromes are commonly caused by impaired DNA damage repair by nonhomologous end-joining. Phagocytic disorders have been treated with HSCT using myeloablative doses of TBI.³⁵² HSCT is considered the only curative therapy for osteopetrosis, and the first successful case involved TBI.³⁵³ Likewise, hundreds of patients have been successfully treated for lysosomal or peroxisomal storage diseases (i.e., mucopolysaccharidoses, mucolipidoses, leukodystrophies, glycoprotein and other disorders) with HSCT, often using TBI, although this approach is undesirable in younger children.^354–356 Recent studies have focused on non–TBI-based approaches in these diseases.³⁵⁷ Fanconi’s anemia is a disease commonly treated with allogeneic HSCT. In favorable patients with an HLA-matched sibling donor, TBI can usually be avoided. However, in patients with evidence of MDS or AML, use of TBI should be considered. TBI is typically delivered in a single fraction, given the underlying defects in DNA repair and cell-cycle checkpoint regulation in patients with Fanconi’s anemia. Prospective dose-escalation studies have found no benefit in doses beyond 4.5 Gy.³⁵⁸