Correspondence 2118 that the mechanism of 11q23 translocation in this patient is similar to that seen in t-AMLs associated with topo II poisons. Although this entity is presumably rare, the true frequency with which therapy- related lymphoblastic lymphoma occurs following aggressive, multi- agent chemotherapy can only be ascertained by careful follow-up of cancer survivors.

Acknowledgements

This work was supported in part by grants from the NIH (CA 15606 and 73773). PDA is a scholar of the Leukemia Society of America. S Thandla1,2 Department of 1Pediatrics, 6Cancer Genetics, M Alashari3,4 Roswell Park Cancer Institute; and the DM Green1,5 Departments of 3Pathology and 5Pediatric PD Aplan1,2,5,6 Hematology-Oncology, Children’s Hospital of Buffalo, Buffalo, New York; and Departments of 4Pathology and 6Pediatrics, School of Medicine and Biomedical Sciences, University at Buffalo, State University of New York, Buffalo, New York, USA

Figure 2 The MLL rearrangement was not constitutional. BamHI References digested DNA extracted from the patient’s peripheral blood lympho- cytes (PBL) prior to the leukemic phase was hybridized to the 0.7B 1 Pui CH, Ribiero RC, Hancock MI, Rivera GK, Evans WE, Raimondi MLL probe. The MOLT 4 cell line is used as a germline control. Size SC, Head DR, Behm FG, Mahmoud MH, Sandlund JT, Crist WM. standards are as indicated. A single germline fragment is seen in both Acute myeloid leukemia in children treated with epipodophyllotox- the control and the patient PBL sample indicating that the MLL ins for acute lymphoblastic leukemia. New Engl J Med 1991; 325: rearrangement seen in the tumor cells is not due to a restriction frag- 1682–1687. ment length polymorphism. 2 Secker-Walker LM, Moorman AV, Bain BJ, Mehta AB. Secondary acute leukemia and myelodysplastic syndrome with 11q23 abnor- malities. Leukemia 1998; 12: 840–844. treatment with an alkylating agent and an epipodophyllotoxin. While 3 Felix CA. Secondary leukemias induced by topoisomerase-targeted it is possible that therapy-related lymphoblastic lymphoma has a drugs (review). Biochim Biophys Acta 1998; 1400: 233–255. longer latency period than therapy-related leukemia, wide variations 4 Wong FL, Boice JD Jr, Abramson DH, Tarone RE, Kleinman RA, in latency periods for secondary malignancies with the 11q23 translo- Stovall M, Goldman MB, Seddon JM, Tarbell N, Fraumeni JF Jr, cation following similar treatment regimens have been reported.2 Li FP. Cancer incidence after retinoblastoma. Radiation dose and What is the cause for the second malignancy in this patient? There sarcoma risk. JAMA 1997; 278: 1262–1267. was no history of a hereditary predisposition to malignancies in the 5 Swinnen LJ, Mullen JM. Treatment of organ transplant related lym- family. To our knowledge there is no association between primary phomas. Hematol Oncol Clin North Am 1997; 11: 963–973. hepatocellular carcinoma and lymphoblastic lymphoma and the 6 Eguiguren JM, Ribiero RC, Pui CH, Hancock MI, Pratt CB, Head possibility that these were two unrelated de novo malignancies cannot DR, Crist WM. Secondary non-Hodgkin’s lymphoma after treatment be eliminated. However, the known association between 11q23 trans- for childhood cancer. Leukemia 1991; 10: 908–911. locations and therapy-related acute leukemias make it likely that the 7 Smith MA, McCaffrey RP, Karp JE. The secondary leukemias: chal- lymphoblastic lymphoma in this patient was a direct result of an MLL lenges and research directions JNatlCancerInst1996; 88: 407–418. translocation induced by treatment with DNA damaging agents. The 8 Pedersen-Bjergaard J. Acute lymphoid leukemia with t(4;11)(q21;q23) mechanisms(s) of illegitimate recombination leading to MLL following chemotherapy with cytostatic agents targeting at DNA-topo- rearrangements is not well understood nor is the role of topo-II in this isomerase II (editorial). Leukemia Res 1992; 16:733–735. process. These topics have recently been reviewed.3 9 Rubnitz JE, Camitta BM, Raimondi SC, Carroll AJ, Borowitz MJ, We conclude that therapy-related lymphoblastic lymphoma can be Shuster JJ, Link MP, Pullen DJ, Downing JR, Behm FG, Mahmoud associated with 11q23 translocations involving the MLL gene. While H, Pui C-H. Childhood acute lymphoblastic leukemia with the the case described here had a slightly longer latency period compared MLL-ENL fusion and t(11;19)(q23;p13.3): favorable prognosis in T with the t-AMLs associated with the topo II poisons, it seems likely cell cases. Blood 1998; 92: 312a.

Extensive bone marrow necrosis associated with multiple myeloma

TO THE EDITOR extensive BMN is characterized by the development of fever, bone pain, a leukoerythroblastic peripheral blood film, variable degrees of Bone marrow necrosis (BMN), defined morphologically by the pancytopenia and elevations of lactate dehydrogenase and alkaline destruction of hematopoietic tissue, including the stroma, with preser- phosphatase. BMN has been reported in patients with different malig- vation of the bone, is a rare syndrome of unknown etiology. Clinically, nancies, sickle cell disease, disseminated intravascular coagulation (DIC), antiphospholipid syndrome (APS), as well as infections.1 There are case reports linking certain medications including fludarabine, interferon alfa to the development of BMN as well.2,3 Here we report Correspondence: J Michaeli, Memorial Sloan-Kettering Cancer Center, two cases of multiple myeloma (MM) with extensive BMN. 1275 York Avenue, New York, NY 10021, USA; Fax: 212 717 3119 In case 1, a 45-year-old man was diagnosed with Durie–Salmon Received 25 May 1999; accepted 5 July 1999 stage III MM in June 1995. A serum immunoelectrophoresis and Correspondence 2119 nephelometric immunoglobulin quantitation showed 31 g/l of IgD. Uri- nary excretion of ␭ light chain was 3.0 g/24 h. The patient was treated with VAD (vincristine, doxorubicin and dexamethasone) and oral pulse dexamethasone therapy. He had a good response to the treatment and his serum IgD decreased to 0.052 g/l. He did well until December 1996, when he was admitted to an outside hospital with worsening fatigue and sudden onset of lower back pain without associated neurol- ogic symptoms. Two days following the completion of the 4-day course of dexamethasone, the patient had a sudden onset of severe thoracic and lumbar back pain. His physical examination was significant for mild spinal and paraspinal tenderness throughout the thoracic and lum- bar regions. His pertinent laboratory values are detailed in Table 1. The patient was admitted to our hospital and received high-dose dexame- thasone and pain control. MRI of the spine was normal. On the second hospitalization day, he reported continued, though somewhat improved pain throughout the spine, as well as similar intensity pain in the hips bilaterally. Laboratory values were significant for progressive pancytop- enia, a markedly increased LDH of 17920 U/l (reference range 90–250 U/l), an elevated alkaline phosphatase of 326 U/l (reference range, 45– 115 U/l), and phosphate of 6.2 mg/dl (reference range 2.6–4.5 mg/dl) Figure 1 Focus of viable poorly differentiated plasma cells (arrow) (Table 1). A BM aspirate was attempted but the marrow was inaspirable. and necrotic plasma cells around osseous trabeculae. A BM biopsy showed atypical plasma cells and extensive necrosis; nor- mal hematopoietic elements were virtually absent. Dexamethasone and oxymorphone were continued, filgrastim and erythropoietin injections were begun. The pain resolved subsequently. This was accompanied by a marked decrease in the serum LDH and phosphate levels at dis- charge (Table 1). Case 2 is a 45-year-old man with Durie–Salmon stage III light chain MM diagnosed in November 1994. He was treated with pulse oral dexamethasone and monthly pamidronate, as well as palliative radi- ation therapy to his thoracolumbar spine (T6 to L4) at a total dose of 3000 cGy. His pain resolved and his performance status improved sig- nificantly. He then received four cycles of VAD followed by three cycles of high-dose cyclophosphamide (3 g/m2) for consolidation and mobilization of peripheral stem cells in preparation for autologous stem cell transplant. He remained in complete remission until June 1997, when he developed severe cervical pain and was found to have new compression fracture at T5 with marked myelomatous involvement of C2 associated with a prevertebral soft tissue mass. He was placed on a Philadelphia brace and radiation therapy was initiated with a total of 3000 cGy. Despite the initiation of dexamethasone therapy, his 24-h urine protein increased from 0.1 to 7.4 g. His laboratory values were notable for the markedly increased LDH level and hyperphosphatemia Figure 2 Higher power view of the same area showing viable (Table 1). The patient was admitted for pain control and started with plasma cells (arrow) within a necrotic plasma cell infiltrate. dexamethasone pulse and radiation therapy to the C1–C5 spine with a total of 2200 cGy. On the sixth hospitalization day, his hemoglobin level decreased to 8.7 g/dl. A BM biopsy showed a small focus of viable can cause BMN. Rather BMN in these patients occurred at the time of and immature plasma cells with extensive necrosis; normal hematopo- relapse with MM. Their clinical presentations share many of the well- ietic elements were virtually absent (Figures 1 and 2). The patient had described features typical of BMN: bone pain, increased alkaline phos- a good response to the pulse dexamethasone therapy. He then received phatase, and marked elevated LDH level associated with characteristic high-dose melphalan (140 mg/m2) followed by autologous stem cell BM findings on biopsy. Although the first patient developed severe pan- transplant. After a prolonged hospitalization course, the patient was cytopenia at the time of presentation, the second patient developed discharged in a stable condition. His laboratory values on the discharge only anemia and mild thrombocytopenia. Interestingly, hyperphosphat- day are shown in Table 1. emia was present at presentation in both patients. This apparently was To our knowledge, this appears to be the first description of the not reported previously with BMN. The hyperphosphatemia was transi- association of BMN with MM in the English medical literature. In the ent and might represent the high cell turnover in BMN. While extensive two patients described, there was no evidence for the presence of other BMN associated with hemotologic malignancies is almost invariably a potential triggers of BMN, such as infection, DIC or exposure to chemo- fatal complication, the prompt initiation of therapy in our patients therapeutic agents like fludarabine and interferon-␣ that presumably resulted in complete restoration of hematopoiesis, relief of bone pain

Table 1 Laboratory values before, at the presentation of BMN and after treatment

Lab values Case 1 Case 2

Pre-Tx At BMN Post-Tx Pre-Tx At BMN Post-Tx

WBC (× 109/l) 4.2 1.1 3.7 8.5 6.4 4.2 Hemoglobin (g/dl) 8.5 7.7 7.7 12.2 8.7 10.0 Platelet (× 109/l) 60 24 26 139 145 166 LDH (U/l) 212 17920 570 188 4980 272

PO4 (mg/dl) 4.6 6.2 4.1 4.2 6.4 3.6 Alka phos (U/l) 115 326 266 95 183 108 Correspondence 2120 and normalization of LDH, and phosphate levels. Thus, our observation References clearly indicates that BMN is potentially reversible and stresses the importance of recognizing this clinical syndrome and the need for 1 Norgard MJ, Carpenter JT, Conrad ME. Bone marrow necrosis and immediate therapeutic intervention. degeneration. Arch Intern Med 1979; 139: 905–911. The pathogenesis of BMN is poorly understood. A common notion 2 Aboulafia DM, Demirer T. Fatal bone marrow necrosis following is that the various processes that trigger BMN eventually lead to bone fludarabine administration in a patient with indolent lymphoma. marrow and endothelial cell injury, which in turn leads to microvas- Leuk Lymphoma 1995; 19: 181–184. cular occlusion. Interestingly, in an animal model, extensive BMN can 3 Kumakura S, Ishikura H, Kobayashi S. Bone marrow necrosis and be induced in a reproducible manner by the injection of foreign pro- the Lambert–Eaton syndrome associated with interferon alfa treat- tein into a sensitized animal.4 It has also been postulated that tumor ment. New Engl J Med 1998; 338: 199–200. necrosis factor ␣ (TNF-␣) may play a role in the pathogenesis of 4 Yoshida M, Hirata M, Inada K. Hemorrhage and necrosis in mouse BMN.3,5 This may be mediated through endothelial cell injury, which bone marrow induced by endotoxin. Jpn J Exp Med 1973; 43: would cause vascular obstruction in the bone marrow and subsequent 393–402. hypoxia resulting in BMN. In support of this hypothesis, elevated TNF- 5 Knupp C, Pekala PH, Cornelius P. Extensive bone marrow necrosis ␣ levels have been detected in the BM of patients with BMN.3 Both in patients with cancer and tumor necrosis factor activity in plasma. the MM cells and/or stromal cell secrete multiple cytokines including Am J Hematol 1988; 29: 215–221. interleukin 6 (IL-6) and TNF-␣.6,7 It is plausible, therefore, that the 6 Treon SP, Anderson KC. Interleukin-6 in multiple myeloma and high tumor burden in these two patients with advanced, relapsing related plasma cell dyscrasias. Curr Opin Hematol 1998; 5: 42–48. MM, was associated with high expression of TNF-␣ in the BM 7 Michaeli J, Choy CG, Zhang X. The biologic features of multiple microenvironment, which subsequently led to the endothelial injury, myeloma. Cancer Invest 1997; 15: 76–84. resulting in the development of BMN. AX Zhu1,2 1Myeloma and 2Hematology Service, R Niesvizky1,2 Division of Hematologic Oncology, E Hedrick2 Department of Medicine and F Fata13Department of Pathology, DA Filippa3 Memorial Sloan-Kettering Cancer Center, J Michaeli1,2 New York, NY, USA