We would like to thank Dr. Alejandro Best-Rocha; Dr. Fiona Craig, director of the University of Pittsburgh Medical Center Flow Cytometry Laboratory; Dr. Urvashi Surti, director of the Magee-Womens Hospital Cytogenetics Laboratory; and Ms. Lori Schmitt, pathology manager of the Children's Hospital of Pittsburgh at the University of Pittsburgh Medical Center, for all their help in this study.
Although the literature reports a low incidence of Burkitt lymphoma (BL) as a post-transplant lymphoproliferative disorder (PTLD), this entity appears to be different from other monomorphic PTLDs (M-PTLDs), both in its aggressive clinical presentation and its distinct pathologic profile.
Patients with BL, diagnosed in the post-transplant setting, (patients aged ≤18 years) were retrieved from the pathology archives at Children's Hospital of Pittsburgh of the University of Pittsburgh Medical Center from 1982 to 2010. Clinical outcomes were obtained along with pathologic review.
Twelve patients with pediatric BL in the post-transplant setting (9 boys, 3 girls) were retrieved. The patients displayed a monomorphic population of small to intermediate-sized, noncleaved, lymphoid elements with a “starry-sky” pattern. The immunophenotype for patients available to the study was CD20+ (n = 9/10), CD10+ (n = 8/8), bcl-6+ (n = 11/11), with a near 100% Ki-67/MIB-1 proliferation index (n = 7/7), and negative for TdT (n = 7/7). Most pretransplant Epstein-Barr virus titers were negative (n = 8/10), with post-transplant titers positive in all tested patients (n = 11), and with positive Epstein-Barr–encoded RNA in situ hybridization in most cases (n = 9/11). The median time from transplantation to diagnosis was 52 months (range, 6-107 months). Nine patients were currently alive after immediate antineoplastic chemotherapy, with median disease-free time of 93 months from diagnosis (range, 2-199 months).
Post-transplant lymphoproliferative disorder (PTLD) is a widely encompassing group of disorders ranging from atypical lymphoid or plasmacytoid proliferations to malignant lymphomas that are the direct result of immunosuppression in a recipient of a solid-organ, bone-marrow, or stem-cell allograft. Pediatric transplant recipients are at increased risk of developing a PTLD because many of them are Epstein-Barr virus (EBV) seronegative at the time of transplantation, which makes them vulnerable to developing a primary EBV infection while on potent immunosuppressive medications.1-3 The histological features of PTLDs can range from early EBV-associated infectious mononucleosis-like proliferations to malignant lymphoma.4 The majority of monomorphic PTLDs (M-PTLD) are composed of transformed B-cells that fit the criteria of a non-Hodgkin lymphoma (NHL), predominately of the diffuse large B-cell type (DLBCL), and only rarely do patients present with a Burkitt lymphoma (BL) phenotype. However, newer data are emerging to suggest that many PTLDs are distinct from their NHL/B-cell lymphoma counterparts and share a molecular profile with memory or activated B cells of postgerminal center origin, with the exception of those PTLDs retaining a germinal center profile, such as BL.5
BL is an aggressive, mature B-cell lymphoma, composed of a monomorphic population of small to intermediate-sized, noncleaved, lymphoid cells with high mitotic and proliferation rates and a characteristic “starry-sky” appearance imparted by the scattered tingible body macrophages engulfing apoptotic debris of the tumor. Ancillary studies with immunophenotyping and cytogenetic analysis confirm the diagnosis, separating it from a diffuse large B-cell lymphoma (DLBCL). Although the c-MYC translocation is nonspecific and can be seen in up to 10% of DLBCL, the expression profile of nearly 100% Ki-67/MIB-1 proliferation index, positive CD10, and bcl-6 and negative bcl-2, with the morphologic “starry-sky” pattern is diagnostic for BL. Three main clinical variants of BL have a various propensity to express EBV.4 The highest expression is seen in the endemic type of equatorial Africa (>95% of cases), whereas the North American sporadic type (20%-30%) and immunodeficiency-associated HIV type (25%-40%) have a lower association with EBV expression.6-8 However, previous patient series of BL occurring in the post-transplant setting have a higher incidence of EBV expression (>70%), greater than that reported for immunodeficiency-associated BL, and more akin to the endemic type and other PTLDs.7-17
The aims of this study were to characterize the patients with BL-PTLD at our institution, a large pediatric transplant center, and to delineate what specific clinical, pathologic, and/or genetic aspects of this disease are discrete.
MATERIALS AND METHODS
Pediatric NHL lymphoma patients who fit the morphologic criteria of BL and who were diagnosed in the post-transplant setting (patients aged 18 years or younger) were retrieved from the surgical pathology archives of the Children's Hospital of the University of Pittsburgh Medical Center Department of Pathology (Pittsburgh, Pennsylvania) from 1982 to 2010. Clinical data were retrieved from the patients' medical records and/or from the Transplant Surgery Division records. This study was approved by the University of Pittsburgh Institutional Review Board (IRB number, PRO09100103) as an expedited study with a waiver of informed consent.
All available histologic and immunohistological sections from the formalin-fixed, paraffin-embedded, tissue blocks were independently reviewed by 3 pathologists (J.P., M.R.M., and R.J.) to confirm the diagnosis of Burkitt lymphoma, based on morphologic criteria set forth according to the 2008 World Health Organization classification of hematopoietic and lymphoid tumors.4
Immunohistochemistry and In Situ Hybridization
Patients with previous immunohistochemistry and in situ hybridization with Epstein Barr-encoded RNA (EBER-1 or EBER-2) were reviewed to confirm their original diagnoses. Briefly, methods for these studies are outlined below. Immunohistochemistry was performed on 4- to 5-μm sections prepared from formalin-fixed, paraffin-embedded, tissue blocks. Immunoperoxidase stains were performed on those cases that had missing or incomplete slides by using a panel of antibodies that included CD20, CD10, bcl-6, and Ki-67 with appropriate dilutions as recommended by the manufacturers. A semiquantitative evaluation was applied as follows: (−), no staining; (+), 1%-10%; (++), 11%-50%; (+++), 51%-75%; (++++), >75%. In situ hybridization for EBV-encoded small ribonucleic acid (EBER-2) was performed using a 30-mer digoxigenin-labeled oligonucleotide probe (EBER kit, Ventana Medical Systems, Tucson, Ariz) as previously described.18, 19 Before February 2009, EBER-1 (biotinylated probe, custom made; Phoenix Biotechnologies, San Antonio, Texas) was used for clinical diagnosis. The procedure was carried out using the Fisher Microprobe (Thermo Fisher Scientific, Waltham, Massachusetts) system, which uses a capillary gap to allow for use of small reagent volumes. Briefly, tissues were deparaffinized and hydrated to water. Endogenous peroxidases were quenched with a methanol-hydrogen peroxide solution. Enzymatic digestion was performed using a commercially prepared pepsin solution at 110°C. The probe was denatured at 110°C, with hybridization at 47°C. After hybridization, nonspecific binding was reduced by using a series of stringent saline sodium citrate washes. The avidin-biotin complex was bound to the biotin portion of the probe (Vector Vectastain Elite ABC kit; Vector Laboratories, Burlingame, California) and observed using 3,3′-Diaminobenzidine (DAB) as the chromogen (ScyTek Laboratories, Logan, Utah). A semiquantitative evaluation and staining was applied as follows: (−), no staining; (+), 1%-10%; (++), 11%-50%; (+++), 51%-75%; (++++), >75%.
Classic cytogenetic karyotype and/or fluorescent in situ hybridization (FISH) studies were performed as part of the clinical workup on a subset of the patients.
The samples (either bone marrow aspirate or mass lesion) were processed by standard cytogenetic techniques. Briefly, cultures were set up in complete Marrow Max medium (Invitrogen by Life Technologies, Carlsbad, California). The unstimulated cultures were incubated for 24 and 48 hours, and mitogen-stimulated cultures were incubated for 9 hours in a 37°C humidified environment with 5% carbon dioxide until harvest. Before harvest, the cultures were treated with ethidium bromide for 90 minutes and then with colcemid (25 μL) for 30 minutes. After treatment, cells were exposed to hypotonic solution (0.075 mol/L potassium chloride), fixed with methanol/acetic acid (3:1), spread onto a slide, and G-banded (Giemsa stain from Sigma; Sigma-Aldrich, St. Louis, Missouri), and 20 metaphases were analyzed. Karyograms were described according to the International System for Human Cytogenetics Nomenclature (ISCN 2009).
Fluorescent in situ hybridization analysis
FISH studies were performed in 6 cases. Briefly, FISH studies were performed using Vysis DNA probes (Abbott Molecular, Des Plaines, Illinois) with primarily the MYC break-apart dual color rearrangement probe used to identify the presence of the c-MYC gene rearrangement. In positive cases, the partner translocation was determined with MYC/IGH@ dual-color fusion probes (1RIG1Y). Rarely, if MYC/IGH@ (8q24/14q32) was found to be negative, variant MYC translocation with IgL (22q11) or IgK (2p11) was performed to determine by FISH analysis whether there was a clonal IGL or IGK gene rearrangement. Briefly, FISH studies were performed on formalin-fixed, paraffin-embedded, tissue sections, following manufacture's protocol. After overnight hybridization and subsequent washing, the slides were analyzed by using a fluorescence microscope (Olympus, Center Valley, Pennsylvania). A total of 200 interphase cells were recorded with cutoff values of >3%. A 1%-3% range of false-positive signals for break-apart probes has been reported.20 Selected images were captured using an Iris image analysis workstation (MetaSystems, Boston, Massachusetts).
Diagnostic flow cytometry reports, performed at the time of original clinical diagnosis, were reviewed. Diagnostic flow cytometry was attempted on fresh tissue for 6 cases and obtained results on 5 cases with sufficient number of cells. Two of these cases had flow cytometry performed only on the concurrent involved bone marrow, with histologic confirmation. Samples were processed by using standard flow cytometric cytogenetic techniques with cell-suspension immunophenotypic studies performed for 2 gated regions based on cell size and complexity. Percentage viability was also determined (70%-98%). Before 2002, 3-color fluorochrome tubes (and thereafter 4-color tubes) were used with tagged antibodies to detect the following combinations of antigens :
CD45-FITC/CD14-PE/IgGa-FITC/IgG2a-PE or CD45-FITC/CD14-PE/CD33-PC5 Or CD45-PCP Cy5. 5/CD14-FITC/CD34-APC/CD19-PE/CD3-FITC; CD20- PerCP/CD10-FITC/CD22-PE; CD10-PE/CD19-PerCP or PC5/CD5-FITC; CD38-FITC/CD22-PE/CD20-PerCP; CD10-APC; CD43-FITC/CD5-PE/CD19-PerCP Cy 5.5; IgG2a-PE/IgG2a-FITC/HLA-DR-PE or HLA-DR-PE/CD15-FITC or CD15-FITC/CD33-PE/CD117-PerCP Cy5.5/HLA-DR-APC; CD19-PC5/CD13+33-PE/CD34-FITC; CD7-FITC/CD13+33-PE/CD19-PerCP/CD56-APC/IgG2a-PE/CD13+33-PE/CD34-FITC; FMC-7-FITC/CD23-PE-A/CD19-PerCP Cy5.5/CD5-APC; CD10-FITC/CD13+33-PE/CD19-PerCP/CD34-APC; CD2-FITC/CD8-PE/CD4-PerCP or CD2-FITC/CD8-PE/CD3-PerCP Cy5.5/CD4-APC; CD3-PerCP/CD7-FITC/CD16+56+-PE; CD117-PE/CD56-PC5; KAPPA-FITC/LAMBDA-PE/CD20-PerCP or KAPPA-FITC/LAMBDA-PE/CD19-PerCP/CD5-APC or KAPPA-FITC/LAMBDA-PE/CD20-PerCP/CD10-APC; BCL2-FITC/CD10-PE/CD20-PerCP Cy5.5; CD16-FITC/CD13- PE/CD45-PerCP/CD11b-APC; CD36-FITC/CD64-PE/CD45-PerCP Cy5.5/CD34-APC; CD16&57-FITC/CD7-PE-A/CD3-PerCP Cy5.5/CD56-APC; (cytoplasmic) IgG1-PerCP/IgG1-FITC/cCD3-PerCP/cTdT-FITC or (cytoplasmic) cTdT-FITC/cCD3-PerCP/cCD34-PE/MPO-FITC or (cytoplasmic) cTdT-FITC/MPO-PE-A/Ccd3-PerCP Cy 5.5/CD34-APC.
Clinical Findings, Including EBV Status
Twelve patients with pediatric BL in the post-transplant setting were retrieved during a 28-year period (Table 1 lists all 12 patients as BL-1 through BL-12). There were 10 patients with available racial information, all of whom were Caucasian. Age at transplant ranged from 9 weeks to 16 years (median age, 2.9 years). Allograft organs included liver (n = 5), heart (n = 5), small bowel (n = 1), and kidney (n = 1). Median time from transplant to BL-PTLD diagnosis was 52 months (range, 6-107 months). Most frequent signs and symptoms at diagnosis included rapidly enlarging extranodal or nodal masses and abdominal pain. Two patients developed BL-PTLD within their liver allograft (Table 1); other frequent sites included the head and neck (n = 4) and abdomen (n = 4), including the small bowel, with 2 sites in the native kidney, 1 of which also included dissemination in the small bowel, adrenal and subcutaneous tissue. Three patients (BL-6, 7, and 8) developed allograft failure that required a second transplant after diagnosis. Retransplantation occurred 26 months (BL-6), 20 months (BL-7), and 66 months (BL-8) after the BL-PTLD diagnosis; none had allograft involvement.
Table 1. Clinical Findings for Transplant Patients With Burkitt Lymphoma Presenting as a Post-Transplant Lymphoproliferative Disorder
Age at TX
Abbreviations and comments are listed as they appear in the table columns rather than alphabetically . TX, transplant; BL, Burkitt lymphoma; PTLD, post-transplant lymphoproliferative disorder; LN, lymph node; GI, gastrointestinal tract; TX-PTLD, months from transplant to BL-PTLD; PTLD Outcome, months from initial BL-PTLD to recurrence or last follow-up; EBV Donor, donor Epstein-Barr virus serology for the allograft organ; POS, positive; NA, no data available; E, equivocal; LRD, living related donor; NEG, negative; EBV Pre-Tx, recipient Epstein-Barr virus serology before transplantation; EBV Post-Tx, recipient Epstein-Barr virus serology after transplantation; EBV VL, quantitative Epstein-Barr virus viral load 1 month before PTLD, divided into clinically relevant quartiles at Children's Hospital of Pittsburgh: I <150 copies/mL, II 150-15,999 copies/mL, III 16,000-799,999 copies/mL, IV >800,000 copies/mL. Group III and IV are historically associated with increased risk of either having or developing EBV-PTLD. Rx, therapy for BL-PTLD; CI, cessation immunosuppression; CT, lymphoma-specific chemotherapeutic regimen. Chemotherapy Regimens: V, vincristine; M, methotrexate; ARA-C, cytarabine; P, prednisone, D, doxorubicin; CY, cyclophosphamide; R, rituximab; L, leucovorin; E, etoposide.
Subcutaneous back lesions, kidney, adrenal, small bowel, adenopathy
R, V, P, D, CY
Alive; in therapy
Five of the patients had a prior diagnosis of PTLD. Three patients (BL-3, 5, and 7) had a previous polymorphous PTLD (P-PTLD), diagnosed 4 months to 84 months after transplant and 17-72 months preceding BL-PTLD diagnosis. There were 2 patients (BL-2 and 9) with early EBV mononucleosis-type PTLD, diagnosed 5-40 months after transplant and 23-24 months before BL-PTLD presentation. Donor EBV status was known in 8 patients (positive, n = 5; negative, n = 2; and equivocal, n = 1). There were no records for the other 4 patients (Table 1). Pretransplantation-recipient EBV serology was negative in 8 of the 10 patients who had clinical data. Subsequent post-transplant EBV serology was positive in all 11 patients who had clinical data available (Table 1). Whole-blood EBV viral loads estimated by quantitative PCR testing21 are reported by our institution into clinically relevant quartiles, with quartile groups III (16,000-799,999 copies/mL) and IV (>800,000 copies/mL) historically associated with increased risk of either having or developing an EBV-PTLD. Of the 8 patients who had quantitative EBV viral-load testing 1 month before their diagnosis, 6 of them had high levels, according to our institution's clinical quartiles III and IV (Table 1). In addition, all of the heart transplant patients (BL-1, 3, 4, 9, and 12) were chronically and/or persistently high viral-load carriers, with quartile III-IV viral load levels. Before their BL-PTLD diagnoses, 1 of the heart allograft patients (BL-3) had a P-PTLD and another (BL-9) had an early EBV-associated PTLD. BL-3 was a high viral load carrier before both his initial and recurrent BL-PTLD. The nonheart allograft patients had more variable viral-load levels. BL-5 had an early P-PTLD 3 months after transplantation and 31 months before BL-PTLD. During this diagnosis, he had a high viral load (quartile III) but then had subsequent variable viral-load levels (quartiles II to III) in the months before his BL-PTLD, including a low viral load (6000-11,000 copies/mL) 1 month before his BL-PTLD. BL-10 and the other liver allograft patients had low viral loads before diagnosis, with BL-11 having an isolated elevation at the time of BL-PTLD diagnosis (Table 1).
Ancillary Testing, Including EBER and Clonality Status
Two patients (BL-6 and 7) had incomplete pathologic data for reporting an immunophenotypic profile; however, BL-7 had strong EBER expression (Fig. 1), had molecular confirmation with a positive t(8;14) by karyotype, and had evidence of a c-MYC rearrangement by the break-apart FISH probe (Table 2). BL-6 was the oldest patient in the series and had limited data, but these data did include the histopathologic diagnosis of BL in post-transplant setting, which was reviewed and confirmed by all 3 pathologists. Unfortunately, tissue was not available to perform additional studies. Overall, the patients displayed an immunophenotype positive for CD20 (n = 9/10) (Fig. 1), CD10 (n = 8/8), and bcl-6 (n = 11/11), and negative for TdT (n = 7/7) and bcl-2 (n = 3/3). Interestingly, 1 patient (BL-7) had negative CD20 expression by paraffin-section immunohistochemistry 1 month after initial diagnosis. At the original diagnosis, IgM and both cytoplasmic kappa and lambda light chains were expressed. All cases stained for Ki-67/MIB-1 showed a strong, diffuse, staining pattern indicating a very high (>95%) proliferative index (n = 7/7) (Fig. 1). In situ EBER expression was present in 9 of 11 patients tested (Fig. 1). Of the 2 patients with negative EBER expression, 1 had a repeated test that showed persistent absence of staining with EBER-2 probe. These 2 patients (BL-5 and 10) also had low quantitative EBV viral load at diagnosis.
Table 2. Pathologic Findings for Transplant Patients With Burkitt Lymphoma Presenting as a Post-Transplant Lymphoproliferative Disorder
Cytogenetics (Karyotype and FISH)
Immunophenotype compiled from paraffin immunohistochemistry and flow cytometry. c is cytoplasmic vs s, which is surface light-chain expression. (-) means negative. (+) means 1-10%. (++) means 11-50%. (+++) means 51-75%. (++++) means >75%.
Abbreviations: BM@DX, concurrent bone marrow status at diagnosis; CNS@DX, concurrent CSF status at diagnosis; EBER, Epstein Barr encoded RNA; EBV, Epstein Barr virus; FISH, fluorescent in situ hybridization; NA, no data available; NEG, negative; POS, positive; PTLD, post-transplant lymphoproliferative disorder; P-PTLD, polymorphous-PTLD.
From concurrent bone marrow (BM) at tissue diagnosis.
Flow cytometry demonstrated a CD20+/CD10+/TdT-immunophenotype in all 5 patients analyzed (BL-1, 3, 5, 9, and 10), with 4 patients demonstrating cytoplasmic kappa restriction; 1 patient (BL-5) had no surface immunoglobulin expression.
Cytogenetic testing was performed on 8 patients, either with classical karyotype and/or with FISH studies (Table 2). Six patients had a t(8;14) and 1 had a t(8;22). There was 1 negative c-MYC translocation patient (BL-5) for which FISH was repeated and confirmed negative with commercial probes. This patient, as mentioned above, also had variable EBV viral load at diagnosis (quartile II-III in the months before diagnosis) and negative EBER staining of the tissue; however, the classic BL morphology and immunophenotype, including a >95% Ki-67 proliferation index, positive staining for CD20, CD10, and bcl-6, with negative TdT staining, confirmed the diagnosis, which was reviewed by all pathologists. This patient was alive without disease after receiving chemotherapy.
Three male patients died (BL-3, 6, and 10). One died shortly after original disease presentation (BL-10), another died shortly after recurrence in his kidneys (BL-3), and the third, with limited records, died 11 years after diagnosis (BL-6). None of these patients had clinical evidence of unresponsive and/or progressive disease at time of death. BL-3 had a heart allograft and developed recurrence in his kidneys and bone marrow 5 years after initial presentation (bone marrow was negative at initial presentation). He was one of the heart transplant patients with a chronically high viral load who developed P-PTLD 4 months after transplantation and 72 months before his initial BL-PTLD diagnosis. He underwent lymphoma-specific chemotherapy both at initial presentation (records of exact regimen were unobtainable) and at recurrence (Table 1). He died 1 month after recurrence from complications of chemotherapy with sepsis and multiorgan failure. At autopsy, there was no evidence of residual BL-PTLD. BL-10, who had a liver allograft, had involvement of both the allograft and bone marrow at the time of diagnosis. He died 3 months later from complications of graft failure and cardiac arrest. An autopsy was not performed, and no information on residual disease was available, but he was noted to be clinically responsive to treatment. There are limited records for BL-6, who developed BL-PTLD in his native small bowel. His records indicate that he died 11 years after diagnosis from complications related to immunosuppressive therapy but unrelated to diagnosis of BL-PTLD; no autopsy was performed. BL-6 was the oldest patient in our series, and records do not indicate the specific type of medical therapy he received beyond the surgical removal of BL-PTLD.
PTLD is more frequently seen in the pediatric transplant population given the EBV-naive state of most children before transplantation and subsequent primary infection while in a post-transplant, immunosuppressed state. As a PTLD, BL type is a rare form of M-PTLD with distinct clinical and pathologic features. It also has features distinguishing it from its lymphoma counterpart in the nontransplant setting but should be treated as a high-grade lymphoma with immediate lymphoma-specific chemotherapy and cessation of immunosuppressant medications.
A few patients had interesting findings that are described here. One patient had IgM staining indicating early B-cell origin, but 1 month after stopping immunosuppressive therapy, the resected specimen had no CD20 staining by paraffin immunohistochemistry (BL-7). Unfortunately, a CD20 immunohistochemical stain was not preformed on the original tissue. The clinical records do not state whether the patient received rituximab (anti-CD20 monoclonal) therapy, which would likely explain the CD20 loss, as has been reported by others.22 Interestingly, others have reported loss of CD20 expression secondary to B-cell downregulation, as a result of EBV infection in rare cases of PTLD. This patient was positive for EBER in situ expression at diagnosis, and if no rituximab was administered, then this may be the likely explanation.23 Another interesting finding was that of a small bowel allograft recipient (BL-5), who had a negative for c-MYC rearrangement and absence of EBER in situ expression, despite the classic BL morphologic and immunohistochemical findings, including a proliferation rate >95% in the tissue. Others have shown that classic Burkitt morphological features with a very high proliferation fraction should still be considered as a BL, even when there is a negative MYC translocation.24 This group has found an alternative mechanism of c-MYC upregulation in these rare negative cases not identified by the commercial FISH probes.24 Moreover, studies have shown that those cases morphologically diagnosed BL, with a very high proliferation index and negative c-MYC, who were treated with less aggressive chemotherapy for a DLBCL, tended to have worse outcomes.25 This underscores the need to clinically treat these patients more aggressively. At our institution, patients with BL in the post-transplant setting are immediately treated with a lymphoma-specific multidrug chemotherapy regimen along with a cessation of immunosuppression during treatment, which is in contrast to the standard initial immunoreductive therapy for P-PTLDs and some M-PTLDs. BL-5 was treated as a BL initially with antineoplastic chemotherapy and is alive without disease, which supports the multiparameter approach for diagnosis, even with negative FISH studies.
In our patient series, we show a very high expression of EBV, with positive EBER in 82% of patients, which is slightly higher than the observed frequency of 72% of EBV(+) pediatric and adult BL-PTLD cases in the literature.5, 9-17 This expression of EBV is much higher than that reported for both sporadic and immunodeficiency-associated BL, primarily HIV-associated BL.6-8 The interplay of active EBV infection in the immunosuppressed state with the development of BL-PTLD is noteworthy. Existing data show that certain chromosome 8 breakpoint locations (in which the c-MYC regulatory region remains intact) are more common in endemic equatorial African EBV(+) BL when compared with sporadic EBV(+) BL and EBV(−) BL.26 There appears to be a distinct mechanism in endemic BL in which EBV may induce an effect outside of the regulatory regions of the c-MYC gene, which could help explain some of the clinical differences in sporadic versus endemic BL. Of interest in our patients is the similarly high prevalence of EBV in BL-PTLD and endemic BL, compared with sporadic and immunodeficiency-associated BL. A question to be further investigated is whether there is an interplay between EBV and alternative chromosome 8 breakpoints in both endemic BL and BL-PTLD or whether this is unique to the endemic, non-PTLD, BL.26
While BL-PTLD is a rare entity with little in-depth investigation in the literature,9, 10, 12-14, 27 we suggest it is a discrete form of PTLD, with a high EBV expression, but it should be treated as a high-grade lymphoma. At our institution, BL-PTLD historically represents a small, but significant, proportion of PTLD cases. BL-PTLD represents 15% of our PTLD patients for pediatric heart, lung, and heart-lung transplants from 1982 to 2009, with a 1.1% overall incidence among our pediatric transplant heart-lung recipients; 14% of our pediatric renal PTLD patients from 1989 to 1995, with a 1.6% overall incidence among our pediatric transplant kidney recipients; and 5.6% of all liver PTLD patients from 1989 to 1996, with a 0.71% overall incidence among our pediatric liver-transplant recipients, during this time period.28 BL in the post-transplant setting is a more aggressive type of PTLD. BL-PTLD does not respond to a trial of decreased immunosuppression like P-PTLD and some M-PTLDs, but BL-PTLD does require cessation of conventional immunosuppression during treatment with multiagent lymphoma-specific chemotherapy.29, 30 In those patients with complete medical data, chemotherapy was the immediate therapy of choice, in addition to cessation of conventional immunosuppression during this treatment (Table 1). Although BL does have a high cure rate, 25% of our transplant patients died (n = 3), all of whom were boys. Two died of infection, and the third died from complications of liver failure. However, none were thought to have died of inadequate treatment or progressive-residual disease—although only 1 patient (BL-3) had definitive confirmation by postmortem examination. The autopsy findings of this patient did not reveal evidence of residual lymphoma despite having a disease relapse 1 month before his death. Interestingly, this relapse occurred 5 years after the original diagnosis of PTLD-BL, which is not typical for sporadic BL because this tumor has the highest risk of relapse within the first year.4 The reason for the longer time to relapse in this patient is uncertain, but it may be a reflection of the different pathogenesis unique to BL as a PTLD. Also of note, 2 of the boys who died had bone marrow involvement at diagnosis; the status of the bone marrow of the other patient (BL-6) was not known. In our patient series, there is 1 other living patient with bone marrow involvement (BL-11); however, his follow-up is too recent to fully comment on survival. None of the patients had central nervous system involvement. Our findings suggest that bone marrow involvement (stage 4 disease) in the post-transplant setting remains a poor prognostic factor, similar to its lymphoma counterpart, despite receipt of lymphoma-specific chemotherapy.
In this pediatric population, the time from transplant to the initial development of BL-PTLD was longer than other types of pediatric PTLD, likely because polymorphic disease predominates in children and tends to occur early after transplantation.31 In our patient series, the median time to BL-PTLD presentation was 52 months after transplant (range, 6-107 months), much longer when compared with a median time of 5.25 months (range, 1.5-73 months) in pediatric PTLD without any cases of BL, from a group of 691 children who received transplants at The Hospital for Sick Children in Toronto, Canada.32 In our liver allograft recipients, the median time to presentation was 63 months after transplant (range, 6-101 months), compared with a median of 5.25 months to PTLD in the Canadian pediatric liver patients.32 In our heart allograft recipients, the median time to presentation was 56 months after transplant (range, 29-107 months), compared with 9.5 months to PTLD in the Canadian heart patients.32 Our reported longer interval for BL-PTLD after transplant is similar to other published pediatric case reports for BL-PTLD, ranging from 24-72 months post-transplant.9-11 One hypothesis for this longer lag time in BL-PTLD is that there is a more complex disease pathophysiology needed to induce the genetic dysregulation of the c-MYC oncogene that accumulates over time in contrast to the immediate cytotoxic T-lymphocyte immunosuppression and subsequent EBV B-cell proliferation of memory B-cells in other PTLDs. Five patients did have some type of PTLD proliferation (early EBV associated and P-PTLD) before the BL-PTLD presentation (Table 2), including 2 of the heart transplant patients (BL-3 and 9). All of our heart transplant patients appeared to be chronically high EBV viral-load carriers, whereas the liver and small bowel transplant patients had more variable viral-load data. The 2 liver patients (BL-2 and 7) with early PTLD did not have complete viral-load quantitative levels for further comment. The small bowel patient with early PTLD (BL-5) had isolated high viral-load levels before both his early PTLD and subsequent BL-PTLD; however, he was not a chronically high viral-load carrier. We consider the differing results between cardiac-transplant and liver/small-bowel transplant patients to be more indicative of different risks or implications of high EBV load between cardiac and liver recipients, which may reflect differing levels of immunosuppression in the respective organ systems as well as the more narrow therapeutic window between infection and susceptibility for rejection that is seen in cardiac recipients compared with liver recipients. A previous study at our institution did show that both early PTLD and high chronic EBV viral loads were risk factors for late-onset PTLD, including BL in our heart transplant population and for those BL patients who had chronically high viral loads before the onset of BL.33 Although the current study was not specifically designed to assess risk factors for development of BL, these data suggest that prior PTLD and chronic high EBV-load status are risk factors for the development of BL in some pediatric solid-organ transplant populations.
BL-PTLD also appears distinct from other PTLDs because despite having a germinal center (GC) phenotype (bcl-6+, CD10+), it maintains a high expression of EBV. In the literature on adult patients, the majority of PTLD patients who have a GC-immunophenotype are typically EBV-negative.5, 15 Furthermore, EBV-associated PTLDs are thought to have a pattern of type III latency in which EBNA-2 and LMP1 (EBV-transforming antigens) are highly expressed, which gives rise to many of the virally associated effects, such as inducing B-cell differentiation and downregulating expression of bcl-6 and CD10; whereas EBV-positive BL has a latency pattern that lacks expression of these 2 antigens (LMP1 and EBNA-2), having only EBNA-1 expression.15, 34 Interestingly, EBV-negative PTLD patients with a GC-immunophenotype appear to share more similar characteristics to BL-PTLD patients, including a later onset after transplant with a more aggressive clinical presentation.15 Also, new molecular data with gene clustering expression chips suggest that many of the M-PTLDs are more closely related to early and P-PTLDs than to their NHL counterparts, with the exception of BL and a minority of DLBCL, which retain a germinal center phenotype.5, 35 New molecular evidence with cDNA microarrays on 21 PTLDs (n = 12 M-PTLDs) and 39 B-NHLs showed that gene expression clustering analysis could not distinguish between P-PTLD and M-PTLD with a post-GC phenotype, suggesting for the first time a clonal relation between these morphologically distinct entities derived from a similar activated/memory B-cell linage with maturation beyond the GC stage.5 This study and others suggest that BL-PTLD is distinct from other M-PTLDs, given its GC phenotype, increased clinical aggressiveness, longer post-transplant period to onset, and need for high-grade treatment. However, the high expression of EBV in BL-PLTLD cannot entirely separate it from other M-PTLDs. Further molecular expression analysis will need to be performed specifically in larger groups of BL-PTLD patients, endemic, sporadic, and immunodeficiency-associated BL, compared with other M-PTLDs to make definite conclusions as to the possible different origin of these clonal proliferations.
In conclusion, we have reported on 12 BL patients with post-transplant lymphoproliferative disorder in our pediatric transplant population. In our institution, BL-PTLD represents a small, but significant, fraction of PTLD cases. Limitations in this study include its retrospective nature and limited pathologic and clinical data for a subset of the patients. Given the small number of patients during an almost 3-decade period, we are limited in the scope of our hypothesis on the relation of BL as a PTLD versus a NHL occurring in the post-transplant setting. There is some molecular evidence in favor of the notion that BL in the post-transplant setting is not a distinct form of PTLD but rather is more akin to malignant lymphoma.5 However, the overall clinicopathological evidence, including the high incidence of EBV expression, suggest that BL-PTLD is an aggressive and distinct PTLD. Furthermore, although not the primary aim of the study, our data also suggest that prior PTLD and chronically high EBV-load status may be risk factors for the development of BL. However, regardless of the origin or specific classification type, we and others have indicated that there is a strong association between immediate lymphoma-specific chemotherapy and favorable outcome. Further work to understand the genetic expression profile of BL-PTLDs and the interplay of EBV will offer a new perspective to the biology of this unique PTLD, as currently only a few of these patients have been previously included in published molecular studies, with only a minority of pediatric transplant cases.