Melphalan/TBI Is Not More Carcinogeneic Than

Home , Melphalan

Bone Marrow Transplantation (2000) 25, 365–370  2000 Macmillan Publishers Ltd All rights reserved 0268–3369/00 $15.00 www.nature.com/bmt Melphalan/TBI is not more carcinogeneic than cyclophosphamide/TBI for transplant conditioning: follow-up of 725 patients from a single centre over a period of 26 years

S Kulkarni1, R Powles1, J Treleaven1, S Singhal1,2, C Horton1, B Sirohi1, N Bhagawati1, D Tait1, R Saso1, S Killick1, R Pinkerton1, A Atra1, S Meller1 and J Mehta1,2

1Leukaemia and Myeloma Units, Royal Marsden Hospital, Sutton, UK

Summary: and TBI. Bone Marrow Transplantation (2000) 25, 365– 370. As there is concern regarding the high carcinogenic Keywords: allogeneic bone marrow transplantation; potential of melphalan (Mel), 725 patients with haema- second malignant neoplasms; conditioning therapy tological malignancies who received allogeneic (n = 714) or syngeneic (n = 11) transplants over the last 26 years were followed-up to evaluate if melphalan was more likely to result in secondary malignant neoplasms Bone marrow transplantation has been increasingly used for (SMNs) than cyclophosphamide (Cy). Three hundred the treatment of haematological and immunological dis- and ninety-five were treated with Cy/TBI and 330 with orders. Transplant conditioning regimens use alkylating Mel/TBI. Twelve patients developed non-haematological agents as well as total body irradiation. Both of these SMN. Median time to develop a SMN was 7 years modalities can be carcinogenic and there are growing con- (range 2–17 years). Age-adjusted rate was significantly cerns about transplant-induced malignancies. As the suc- higher than in the general population (observed 12 cess rates of BMT have improved over the years, the popu- expected 1.2, risk 10; P Ͻ 0.0001). The cumulative over- lation of long-term survivors who may be at increased risk all risk of developing a SMN at 2, 5, 10 and 15 years of second malignancies has grown. A number of studies has 1–4 post transplant was 0.4% (95% CI 0.1–2.6%), 1.7% estimated the risk of second malignancies post transplant (95% CI 0.6–4.4%), 6.4% (95% CI 2.8–10.8%) and and there are reports suggesting an increased risk of both 6.6% (95% CI 3.4–12.4%), respectively. Even though haematological and non-haematological malignancies in 5 age-adjusted rates were higher than the general popu- patients receiving allogeneic and autologous transplants. lation melphalan/TBI was not associated with higher A multi-centre study involving almost 20000 patients age-adjusted risk than Cy/TBI (increased risk 7.9 vs has recently been published analysing the risk of solid can- 6 11.4; P = NS). The cumulative overall risk of SMNs was cers after BMT. Melphalan is becoming increasingly popu- not different with CY/TBI or Mel/TBI (8/393 vs 4/363; lar for conditioning prior to BMT because it causes no blad- 7 10 year risk 4.4%, 95% CI 1.8–10.6 vs 8.4%, 95% CI 2.9– der toxicity and may have a superior anti-leukaemia effect. 22.9; P = NS). The risk was significantly higher with use However, there is concern regarding the high carcinogenic 8–10 of additional cranial or cranio-spinal irradiation (17.5% potential of melphalan in the non-transplant setting. This vs 2.7% at 10 years; P = 0.0241). Transplants for acute single centre analysis was undertaken to evaluate the risk lymphatic leukaemia resulted in a higher incidence of of second malignant neoplasms following allogeneic bone SMNs than did transplants for other diseases (ALL: marrow transplantation after a variety of conditioning regi- 17.4%, 95% CI 6.3–42.6%; other diseases: 3.4% (95% mens. Secondary leukaemia and myelodysplastic syn- 1.3–8.5%, P = 0.0469). The risk of SMN for patients dromes are excluded in this analysis. with chronic GVHD was 8.4% (95% CI 3.7–18.7%) as compared to 3.5% (95% CI 1–11.1%) for patients with- out chronic GVHD (P = NS). No factor was associated Materials and methods with independently increased risk in multivariate analysis. Use of melphalan and TBI for transplant condition- Seven hundred and twenty-five patients who received allog- ing does not appear to be associated with higher risk of eneic (n = 714) or syngeneic transplants (n = 11) were second malignant neoplasms than cyclophosphamide identified from the prospectively maintained leukaemia database at the Royal Marsden Hospital.11 Transplants were carried out over the period 1973–1997. Patients transplanted for aplastic anaemia or Fanconi’s anaemia were Correspondence: Prof RL Powles, Leukaemia and Myeloma Units, Royal excluded from the analysis because of their predisposition Marsden Hospital, Downs Road, Sutton, SM2 5PT, UK 2Current address: Division of Transplantation Medicine, South Carolina to cancer. Patient characteristics are shown in Table 1. Cancer Center and Richland Memorial Hospital, Columbia, SC, USA Transplants were for AML (n = 378), ALL (n = 164), CML Received 19 July 1999; accepted 14 October 1999 (n = 86), and other diseases (n = 97). Median patient age Second malignant neoplasms after allogeneic stem cell transplantation S Kulkarni et al 366 Table 1 Patient characteristics Statistical analysis

Cy/TBI Mel/TBI P value The follow-up period was 11 days to 15.3 years for patients receiving melphalan and 11 days to 26 years for those Total 395 330 receiving cyclophosphamide. One hundred and thirty-four Sex in the melphalan group and 185 in the cyclophosphamide Male 224 206 NS group were followed for more than 1 year. Female 171 124 NS For individual patients, the number of person-years at Median age at BMT (years) 22 30 NS risk was calculated from the date of transplant until last (range) (2–49) (3–57) date of follow-up or death. To calculate the expected num- Diagnosis ber of cancers in the study population, incidence rates were AML 210 168 0.001 16 ALL 82 82 obtained from England and Wales cancer registers. The CML 80 6 ratio of observed to expected number of cancers and 95% Others 23 74 conﬁdence intervals were calculated assuming a Poisson Type of graft distribution for the observed cancers. The cumulative over- Syngeneic 7 4 NS all risk of developing a SMN was calculated using the Matched sibling 318 318 Kaplan–Meier method.17 Impact of age at the time of graft Mismatched family member 37 3 Ͻ Ͼ Matched unrelated 33 5 ( 15 years or 15 years), sex, diagnosis (AML, ALL, CML or others), GVHD prophylaxis (cyclosporine vs Dose of TBI 9.5 Gy single fraction 131 95 0.001 cyclosporine and methotrexate), conditioning regimen 10.5 Gy single fraction 98 227 (Cyclo/TBI vs Mel/TBI), dose of TBI (9.5 Gy single frac- Other 166 8 tion or 10.5 Gy single fraction), acute GVHD (present or GVHD prophylaxis absent), chronic GVHD (present or absent) were calculated CyA 262 154 0.001 using Kaplan–Meier method and compared using log rank + CyA MTX 95 170 test. Multivariate analysis using the Cox regression model MTX alone 14 1 18 T depletion 18 1 was performed. None 7

Results

Twelve patients developed non-haematological SMN following allograft or syngeneic transplant. Table 2 shows the at the time of transplant was 26 years (range 9 months– details of SMN. Sites of second malignancy were skin = = 58 years). Donors were HLA-identical siblings (n = 636), (basal cell: n 2), oral cavity (squamous cell: n 2), brain = = mismatched family members (n = 40), matched unrelated (PNET: n 3), breast (adenocarcinoma: n 2), colon = = (n = 38) or syngeneic twins (n = 11). All the patients were (adenocarcinoma: n 1) and thyroid (medullary cell: n followed until 1 August 1998 or death. Syngeneic recipients 1). One patient developed large cell non-Hodgkin’s lym- did not receive GVHD prophylaxis. Cyclosporine alone phoma with no evidence of EB virus. The median time was used as prophylaxis in 416 patients, cyclosporine and elapsing between transplant and developing a SMN was 7 years (range 2–17 years). short-course methotrexate in 265, methotrexate alone in 15, The age-adjusted rate of SMNs was significantly higher and 19 received T cell depleted grafts. There were more than in the general population (observed 12, expected 1.2; patients who received cyclosporine alone as GVHD increased risk of 10, P Ͻ 0.0001). The risk was signifi- prophylaxis in the Cy/TBI arm (P = 0.001) while in cantly elevated for skin, brain, breast and thyroid cancer Mel/TBI arm use of 10.5 Gy as the TBI dose was higher = (Table 3). The overall risk of developing a SMN at 2, 5, (P 0.001). 10 and 15 years post transplant was 0.4% (95% CI 0.1– Details of the conditioning regimens have been pre- 12–15 2.6%), 1.7% (95% CI 0.6–4.4%), 6.4% (95% CI 2.8– viously published. In the cyclophosphamide group, the 10.8%) and 6.6% (95% CI 3.4–12.4%), respectively = dose of TBI was 9.5Gy as a single fraction (n 131), or (Figure 1). 10.5 Gy as a single fraction (n = 98) and in 166 either higher dose or fractionated TBI was used. In the melphalan group, 9.5 Gy single fraction TBI was used in 95, 10.5 Gy Comparison of melphalan and cyclophosphamide single fraction in 227 and eight received a higher dose or There appeared to be no difference in the cumulative risk fractionated TBI. of SMN with use of cyclophosphamide or melphalan with In addition to TBI, 128 patients also received cranial (n TBI for conditioning (8/395 vs 4/330; P = NS) (Figure 2). = = 113) or craniospinal (n 15) irradiation for CNS prophy- Also, TBI dose appeared to have no influence on the inci- laxis. dence of SMNs (9.5 vs 10.5 Gy, 4.5% vs 8.7%, P = NS). Cases of second malignant neoplasms (SMN) were ident- Both melphalan and cyclophosphamide were associated ified from the case files, patient interviews and information with a higher age-adjusted rate as compared to the general from referring physicians. Data concerning histological population (melphalan: O/E–4/0.53, risk: 7.6, P Ͻ 0.0001; findings, staging and treatment of SMN were obtained. cyclophosphamide: O/E–8/0.67, risk: 11.94; P Ͻ 0.0001)

Bone Marrow Transplantation Second malignant neoplasms after allogeneic stem cell transplantation S Kulkarni et al 367 Table 2 Patients with second malignant neoplasms

Diagnosis Type of graft Conditioning Site 2nd cancer Histology

AML allo/sib Cy/TBI abdomen NHL (large cell) AML allo/sib Cy/TBI brain PNET (supratent.) ALL allo/sib Mel/TBI tongue squamous cell ALL allo/sib Mel/TBI brain PNET (supratent.) ALL allo/sib Cy/TBI mouth squamous cell AML synge Cy/TBI skin basal cell AML allo/sib Cy/TBI colon adenocarcinoma ALL allo/sib Cy/TBI brain PNET (medulla) AML allo/sib Cy/TBI thyroid medullary cell AML allo/sib Mel/TBI breast adenocarcinoma AML allo/mm Cy/TBI skin basal cell ALL allo/sib Mel/TBI breast adenocarcinoma

Table 3 Comparison of observed and age-adjusted expected incidence of SMNs

Site Observed (O) Expected (E) Ratio (O:E) 95% CI P value

Total 12 1.2 10 1.41–38.4 Ͻ0.0001 Cyclo/TBI 8 0.67 11.94 1.3–26.7 Ͻ0.0001a Mel/TBI 4 0.53 7.6 1.1–42.3 Ͻ0.0001a Skin 2 0.15 13 0.24–60.3 Ͻ0.001 Brain 3 0.043 69.8 0.36–120 Ͻ0.0001 Oral 2 0.01 200 6.23–2892 Ͻ0.0001 Breast 2 0.5 4 0.24–27.1 0.003 Colon 1 0.09 11.1 0.111–55 0.05 Thyroid 1 0.01 100 0.73–365 Ͻ0.001 NHL 1 0.04 25 0.2–37.2 Ͻ0.01 aNo difference between Mel/TBI and Cyclo/TBI.

30 30 n = 725 O = 12 E = 12 M/TBI n = 330 O = 4 E = 3.6 Cy/TBI n = 395 O = E = 8.4 Chi-squared = 0.05 df = 1 P = 0.8231 20 20

Cy/TBI M/TBI 10 10 % Probability of secondary cancer % Probability of secondary cancer 0 0 0 5 10 15 20 25 0 5 10 15 20 25 Time since allograft (years) Time since allograft (years)

Figure 1 Overall risk of developing second malignant neoplasm at 2, 5, Figure 2 Use of melphalan or cyclophosphamide and TBI on incidence 10 and 15 years post transplant. of second malignant neoplasms. but there was no difference in the expected number of eases: 3.4%, 95% CI 1.3–8.5% P = 0.047, Figure 3). The SMNs with cyclophosphamide or melphalan. It was not risk of SMN was significantly higher in patients who possible to evaluate the risk in the two groups for different received cranial or cranio-spinal irradiation (17.5% vs 2.7% anatomical sites due to the small number of patients. at 10 years; P = 0.0241, Figure 4). Age of the patient above or below 15 years was not asso- = ciated with higher risk (7/532 vs 5/193; P NS) and there Acute and chronic GVHD was no influence of the sex of the patient on risk. Transplants for acute lymphatic leukaemia appeared to There was little influence of GVHD prophylaxis on inci- have a higher incidence of SMN than did transplants for dence of SMNs (CyA: 7/416, CyA + MTX: 3/265). Five other diseases (ALL: 17.4%, 95%CI 6.3–42.6%; other dis- hundred and forty-six patients developed acute GVHD after

Bone Marrow Transplantation Second malignant neoplasms after allogeneic stem cell transplantation S Kulkarni et al 368 30 chronic GVHD as compared to 3.5% (95% CI 1–11.1) in ALL n = 164 O = 4 E = 1.6 Other n = 561 O = 8 E = 10.4 patients not developing chronic GVHD. Chi-squared = 3.95 df = 1 P = 0.0469 Risk of second malignancy was not inﬂuenced by the type of immunosuppressive drugs used for the treatment of 20 ALL Other acute or chronic GVHD. In the multivariate analysis no predictor was associated with signiﬁcantly higher risk of second malignant neoplasms. 10 Treatment and outcome of SMNs All patients except one received treatment for the SMN.

% Probability of secondary cancer 0 Details of treatment and outcome are shown in Table 4. 0 5 10 15 20 25 The patient with abdominal NHL died before commencing Time since allograft (years) the therapy for the disease. Basal cell carcinoma was treated Figure 3 Impact of diagnosis (ALL vs Other) on incidence of second with local surgery in both patients and they are alive with- malignant neoplasms. out recurrence. The two patients with breast cancer received modiﬁed mastectomies, local radiotherapy and adjuvant chemotherapy. One died with disseminated disease and one 3 is alive with no evidence of cancer. Patients with oral P = 0.0241 malignancies were treated with surgery and local radio- Cr. RT therapy. Both died with advanced disease. The patient with medullary carcinoma of the thyroid was treated with radio- 2 No Cr. RT iodine and is alive. One patient with PNET of the medulla is alive after surgery and radiotherapy with the disease under control, but the two with supratentorial PNETs died with disease. The case with carcinoma of the colon died Probability 1 after the surgery. Five patients out of 12 are alive (with no evidence of disease). The SMN was the main cause of death in all the patients who died.

0 0 2000 4000 6000 8000 Discussion Days from BMT Use of high-dose chemo-radiotherapy for transplant con- Figure 4 Effect of cranial or cranio-spinal irradiation on the risk of ditioning is associated with various long-term compli- second malignant neoplasms. cations including cataracts, sterility, pulmonary dysfunction and endocrine dysfunction19,20 and second malignancy. transplant and seven developed second cancers, as com- Most of these complications are treatable, but transplant- pared to ﬁve out of 161 who did not develop acute GVHD induced SMNs may not be as amenable as are other compli- (P = NS). Patients who developed chronic GVHD did not cations. Curtis et al6 found that in a large population of show a higher incidence of SMNs (7/228 vs 5/497; P = transplanted patients, the risk for solid cancers was 2.7 0.22, Figure 5). The 10 year risk of developing a second times higher than in the general population. Risk increased cancer was 8.4% (95% CI 3.7–18.7) in patients with with time, and at 15 years it was as high as 6.7%. In our

30 Table 4 Treatment of second malignant neoplasms and status CGVHD n = 228 O = 7 E = 4.9 No CGVHD n = 497 O = 5 E = 7.1 Diagnosis Treatment Response Status Chi-squared = 1.53 df = 1 P = 0.2161 20 NHL (generalised) No — Died PNET (brain) S + RT No Died Squamous (tongue) S + RT No Died No CGVHD PNET (brain) S + RT No Died + 10 Squamous (mouth) S RT No Died CGVHD Basal cell (skin) RT Yes Alive Adenocarcinoma (gut) S No Died PNET (medulla) S + RT Yes Alive Medullary Ca (thyroid) Radioiodine Yes Alive

% Probability of secondary cancer 0 Adenoca. (breast) mast/RT/chemo Yes Alive 0 5 10 15 20 25 Basal cell (skin) RT Yes Alive Adenoca. (breast) mast/RT/chemo No Died Time since allograft (years)

Figure 5 Effect of chronic GVHD on incidence of second malignant RT = radiotherapy; S = surgery; Mast = mastectomy; Chemo = chemo- neoplasms. therapy.

Bone Marrow Transplantation Second malignant neoplasms after allogeneic stem cell transplantation S Kulkarni et al 369 study we observed a cumulative risk of 6.4% at 10 years established that risk is directly related to radiation dose.21,22 and 8.7% at 15 years. The age-adjusted incidence of SMNs It is very likely that in transplant patients the deciding fac- was 10 times higher than in the general population. The tor may not only be the conditioning therapy, but other number of patients at risk beyond 10 years was relatively factors including immunosuppression and immune dysfunc- small and obviously this limits the ability to estimate the tion may play equally important roles. Use of cranial or actual risk in long-term survivors. Although direct compari- cranio-spinal irradiation for CNS prophylaxis was associa- son with the IBMTR data may not be valid, it is interesting ted with higher incidence of SMN but this lost its signifi- to note that overall incidence of SMN was higher in our cance in multivariate analysis. The small number of patients cohort. This may partly be due to inclusion of basal cell with SMN in this study did not allow evaluation of the skin cancers or comparatively small numbers in the present effect of cranial prophylaxis in melphalan/TBI and study but probably more due to under-reporting of cases cyclophosphamide/TBI groups separately. from the many centres involved in the IBMTR study. The In a previous study, acute GVHD was associated with an SMNs in our series included basal cell carcinoma, oral increased risk of haematological SMN.24 Chronic GVHD is squamous cancers, adenocarcinoma breast, PNET, NHL a known risk factor for developing squamous cancers but and adenocarcinoma of colon. These are a few of the com- its impact on other cancers seems equivocal. In our patients, mon cancers reported in other series as well. Transplan- there was no increased risk with either acute or chronic tation may be associated with certain malignancies that do GVHD but this may be due to the small number of patients not have an established association with radiation. We had in this series. It is possible that deaths in the immediate one case of medullary carcinoma of the thyroid that has no post-transplant period reduced the number of patients at proven association with radiation, and Curtis et al6 noticed risk available for evaluation of impact of acute GVHD upon a higher incidence of melanoma in patients who had risk. It is also possible that the long-term immune dysfunc- received radiation, although this association has not been tion associated with chronic GVHD and its treatment are established. more carcinogenic. Curtis et al6 found no effect of immune The carcinogenic effect of radiation is time and age dysfunction on incidence of SMN but there was an dependent, with long latent periods and a higher incidence increased incidence of melanoma and squamous cell carci- in the younger population.21–23 Surprisingly, we found no noma. All our patients with squamous carcinoma had correlation with age and risk of SMN, possibly due to the chronic GVHD, and there was no difference relating to type relatively small number of patients younger than 15 years of GVHD prophylaxis. T cell depletion is known to of age at the time of transplant. Even though the age- increase the risk of SMN specifically lymphoproliferative adjusted incidence was 7 and 11 times higher with mel- disorders, but as only 19 patients received T cell-depleted phalan and cyclophosphamide, respectively, there was no marrow in our study it was not possible to evaluate risk in difference in incidence between cyclophosphamide/TBI this group. and melphalan/TBI. All anticancer alkylating agents are It was observed that acute lymphatic leukaemia was recognised as being carcinogenic, and this varies among associated with a higher risk for SMN than were other dis- the drugs in this group. Melphalan is likely to have more eases although this lost its significance in multivariate carcinogenic potential than does cyclophosphamide and this analysis. The higher risk observed in univariate analysis is especially true for haematological malignancies.8,9 In one may partly be explained by use of cranial radiation for CNS trial of chemotherapy for breast cancer, melphalan was prophylaxis in most patients with ALL. Primary disease is associated with an 11.2% risk as compared to cyclophos- known to influence the risk of SMN that may be partly due phamide, which was associated with a 5.4% risk of to the younger age of ALL patients at transplant, but our developing a SMN.10 In the MRC studies (I and II) for analysis failed to corroborate this finding. treatment of myeloma with single-agent melphalan or Treatment of SMNs has met with limited success. Con- cyclophosphamide, melphalan-treated patients had a higher trol was possible only in those cancers that are recognised incidence of SMN.8,9 In accordance with these obser- as being relatively curable (in our series, basal cell carci- vations, it may be assumed that melphalan and TBI con- noma, medulloblastoma and breast cancer). Surgery is one ditioning will be more carcinogenic and lead to a signifi- of the important treatment modalities for most solid can- cantly increased risk of SMN. This was not observed in our cers, and patients with chronic GVHD face the problem of series. With only 12 patients developing second malig- impaired healing of surgical wounds, making treatment nancies the statistical power to identify differences is obvi- even more difficult. ously low but to detect a difference of 2–4% between two This study has excluded patients with haematological regimens with a power of 95% almost 6000 patients will malignancies post transplant since it has been well estab- have to be followed up for 10 or more years which may lished that the incidence of MDS and AML is increased in be practically difficult. Although cyclophosphamide/TBI patients undergoing both allogeneic and autologous trans- was in use prior to melphalan/TBI in our unit, it is plantation. We have observed a similar trend, with a 2% important to note that in the melphalan/TBI group more risk of developing MDS in patients treated for myeloma.25 patients received higher doses of TBI as compared to the The combined risk of haematological malignancies and cyclophosphamide/TBI group. SMNs post allogeneic transplantation is thus likely to be Those who received 9.5 Gy had an incidence similar to high. those who received 10.5 Gy. As most of our patients In conclusion, allogeneic bone marrow transplant has received single fraction TBI this finding was surprising curative potential and has saved many lives in the last 25 because the experiences from Hiroshima and Chernobyl years. Its use is increasing, and even though it is associated

Bone Marrow Transplantation Second malignant neoplasms after allogeneic stem cell transplantation S Kulkarni et al 370 with some serious long-term sequelae, the potential advan- be more potent leukemogen than cyclophosphamide. Ann tages far outweigh these risks. Presently, there seems to be Intern Med 1996; 105: 360–367. no alternative to radiation and high-dose alkylating agents 9 Cuzick J, Erskine S, Edelman D, Galton DA. Comparison of for transplant conditioning. Although we did not find an incidence of the myelodysplastic syndrome and acute myeloid influence of GVHD on the risk of SMNs it is possible that leukaemia following melphalan and cyclophosphamide treatment for myelomatosis. Br J Cancer 1987; 55: 523–529. immune dysfunction may contribute and selective immuno- 10 Chou RH, Wong GB, Kramer JH et al. Leukaemia following suppression with more effective treatment of chronic chemotherapy for breast cancer. Cancer Res 1990; 50: GVHD may help reduce the risk. Melphalan is likely to 2741–2746. have better anti-leukaemic activity, lower incidence of early 11 Powles R, Milan S, Horton C et al. A prospective haemato- transplant-related complications and is increasingly used oncology patient database: structure and description. Bone for transplant conditioning therapy. Hence, it was reassur- Marrow Transplant 1996; 17 (Suppl. 1): S140. ing to know that it does not appear to be associated with 12 Helenglass G, Powles RL, McElwain TJ et al. Melphalan and an increased incidence of second malignant neoplasms. total body irradiation (TBI) versus cyclophosphamide and TBI Although choice of conditioning regimen does not seem to as conditioning for allogeneic matched sibling bone marrow play a role in development of SMNs long-term follow-up transplants for acute myeloblastic leukaemia in first remission. Bone Marrow Transplant 1988; 3: 21–31. is required. Diligent monitoring involving long-term fol- 13 Mehta J, Powles R, Treleaven J et al. Long term follow-up low-up should thus be undertaken in all long-term trans- of patients undergoing allogeneic bone marrow transplantation plant survivors so that SMNs are detected as early as for acute myeloid leukaemia in first complete remission after possible. cyclophosphamide-total body irradiation and cyclosporine. Bone Marrow Transplant 1996; 18: 741–746. 14 Hjiyannakis P, Mehta J, Milan S et al. Melphalan, single-fraction total-body irradiation and allogeneic bone marrow trans- Acknowledgements plantation for acute leukaemia: review of transplant related mortality. Leuk Lymphoma 1997; 25: 565–572. et al This work was supported by Bud Flannagan Leukaemia Fund and 15 Powles R, Singhal S, Treleaven J . Identification of David Adams Leukaemia Fund. patients who may benefit from prophylactic immunotherapy after bone marrow transplantation for acute myeloid leukaemia on the basis of lymphocyte recovery early after transplantation. Blood 1998; 91: 3481–3486. References 16 Parkin DM, Whelan SL, Ferlay J et al. (eds). Cancer Inci- dence in Five Continents, vol. 7. International Agency for Research on Cancer: Lyon, France (IARC scientific publi- 1 Lowsky R, Lipton J, Fyles G et al. Secondary malignancies cation No. 143). after bone marrow transplantation in adults. J Clin Oncol 17 Kaplan EL, Meier PJ. Non parametric estimation from incom- 1994; 12: 2187–2192. plete observations. Am Stat Assoc 1958; 53: 457–481. 2 Kolb HJ, Guenther W, Duell T et al. Cancer after bone mar- 18 Cox DR. Regression models and life tables. J Roy Stat Soc row transplantation. IBMTR and EBMT/EULEP Study Group 1972; 34B: 187–220. on Late Effects. Bone Marrow Transplant 1992; 10 (Suppl. 19 Swift PS, Cowan MJ, Wara WM et al. Toxicities of total body 1): 135–138. irradiation for pediatric bone marrow transplantation. Int J 3 Bhatia S, Ramsay NK, Steinbuch M et al. Malignant neo- Radiat Oncol Biol Phys 1996; 34: 843–851. plasms following bone marrow transplantation. Blood 1996; 20 Kolb HJ, Poetscher C. Late effects after allogeneic bone mar- 87: 3633–3639. row transplantation. Curr Opin Hematol 1997; 4: 401–407. 4 Witherspoon RP, Deeg HJ, Storb R et al. Secondary malig- 21 Yoshimoto Y. Cancer risk among children of atomic bomb nancies after marrow transplantation for leukemia or aplastic survivors. A review of RERF epidemiologic studies. Radiation anemia. Transplantation 1994; 57: 1413–1418. Effects Research Foundation. JAMA 1990; 264: 596–600. 5 Traweek ST, Slovak ML, Nademanee AP et al. Myelodyspla- 22 Shimizu Y, Schull WJ, Kato H. Cancer risk among atomic sia and acute myeloid leukemia occurring after autologous bomb survivors. The RERF Life Span Study. Radiation bone marrow transplantation for lymphoma. Leuk Lymphoma Effects Research Foundation. JAMA 1990; 264: 601–604. 1996; 20: 365–372. 23 Pierce DA, Shimizu Y, Preston DL et al. Studies of the mor- 6 Curtis RE, Rowlings PA, Deeg HJ et al. Solid cancers after tality of atomic bomb survivors. Report 12, Part I. Cancer: bone marrow transplantation. New Engl J Med 1997; 336: 1950–1990. Radiat Res 1996; 146: 1–27. 897–904. 24 Levine EG, Bloomfield CD. Leukemias and myelodysplastic 7 Powles RL, Smith CL, Tiley C et al. The anti-leukaemic syndromes secondary to drug, radiation, and environmental action of melphalan and total body irradiation in bone marrow exposure. Semin Oncol 1992; 19: 47–84. transplantation. Haematol Rev 1990; 4: 23–26. 25 Saso R, Kulkarni S, Powles R et al. Secondary MDS/AML in 8 Greene MH, Harris EL, Gershenson DM et al. Melphalan may patients treated for myeloma. Blood 1998; 92 (Suppl. 1): 455a.

Bone Marrow Transplantation