Protean clinical manifestations in children with leukemias containing MLL-AF10 fusion.

Translocations involving the MLL gene on chromosome 11q23 occur in 5-10% of human leukemias, and involve fusion with more than 30 different partner genes. The MLL-AF10 fusion produced by the t(10;11)(p12;q23) or ins(10;11)(p12;q23q13) occurs in a small percentage of acute leukemias, most commonly acute myelogenous leukemia (AML) of the M5 FAB subtype. We report two cases of AML (M5a and M0) and one case of acute lymphoblastic leukemia containing MLL-AF10 fusion. Each case had varied clinical characteristics, despite expressing similar MLL-AF10 fusion transcripts. Including the three cases described in this report, we identified a total of 38 cases of leukemia with MLL-AF10 fusion. Approximately one-third of these are not M5 AML. Taken together, these findings emphasize that while the sentinel molecular event may be identical in a disease, the clinical presentation and outcome can vary widely.

Oncogenesis in utero: fetal death due to acute myelogenous leukaemia with an MLL translocation.

The incidence of translocations involving the 11q23 gene MLL is markedly increased in leukaemias that occur in infants <1 year of age. Epidemiological and molecular data have demonstrated that at least some of these translocations occur in utero. In this report we describe a case of fetal death at 36 weeks of gestation. At autopsy the fetus was found to have widely disseminated acute myelogenous leukaemia (AML), FAB subtype M5. Molecular cytogenetic studies of nuclei recovered from paraffin-embedded tissue sections demonstrated that the leukaemic cells contained an MLL translocation. This is the first detailed report, to our knowledge, of fetal death due to acute leukaemia, and directly demonstrates oncogenesis in utero.

CD19 selection improves the sensitivity of B cell lymphoma detection.

Reinfusion of residual tumor cells into B cell non-Hodgkin's lymphoma (B-NHL) patients during autologous transplantation may be an important cause of disease relapse. Determining the extent to which B-NHL cells are present in autologous progenitor cell products and if the presence of residual B-NHL cells is predictive of relapse will require extremely sensitive methods of detecting rare B-NHL cells. We attempted to improve the sensitivity of polymerase chain reaction (PCR)-based detection of rare B-NHL cells by preselecting CD19+ cells using an immunomagnetic column. To measure detection sensitivity, we prepared samples containing different levels of B-NHL cell contamination by mixing B-NHL cell lines containing the chromosomal translocation t(14;18) bcl-2/JH) with control leukapheresis samples. DNA extracted from each CD19-selected sample and from each matched nonselected sample was added to a PCR to amplify the bcl-2/JH breakdown junction. CD19 preselection improved the sensitivity of detection of t(14;18)-positive B-NHL cells 115-fold, so that B-NHL cells at a concentration of 1 tumor cell per 1 x 10(6) hematopoietic cells were detected in every specimen evaluated. t(14;18)-positive cells were not detected in any of 13 control leukapheresis specimens. We conclude that a combination of CD19 preselection and PCR amplification may improve the sensitivity of detection of rare lymphoma cells by two orders of magnitude without a significant decrease in specificity.

Part II. Telomerase expression in cerebrospinal fluid specimens as an adjunct to cytologic diagnosis.

The diagnosis of meningeal carcinomatosis hinges on the cytologic examination of cerebrospinal fluid (CSF), which has a known low sensitivity for the identification of malignant cells. Often only 'suspicious' or 'atypical' diagnoses can be rendered, and specimens are commonly unsatisfactory for evaluation due to poor morphologic preservation. Telomerase is widely expressed in most brain metastases, medulloblastomas, lymphomas, oligodendrogliomas, and is expressed focally in glioblastomas. Little is known about the level of telomerase expression in these tumors, except for brain metastases, where a four-fold variation in telomerase levels exists. In our laboratory, as few as ten carcinoma cells can be detected by a sensitive polymerase chain reaction-based assay, the telomeric repeat amplification protocol (TRAP), for telomerase, but it was unclear whether varying levels of telomerase expressed by different types of metastases would influence detection. Using the TRAP protocol, we studied 281 CSF samples from a wide variety of patients with neurologic and non-neurologic conditions for telomerase expression. An adjusted specificity of 90% and a sensitivity of 64% were achieved for detection of malignant cells in CSF by telomerase expression. The TRAP assay for telomerase detection may serve as an adjunct to the traditional examination of CSF. Neither previously documented four-fold variation in the levels of telomerase expression in brain metastases, high CSF protein levels nor high white blood cell counts precluded detection of malignant cells in CSF.

The malignant lymphomas of Kenya: morphology, immunophenotype, and frequency of Epstein-Barr virus in 73 cases.

Recent information is limited regarding pathological features of the malignant lymphomas of Africa, other than Burkitt's lymphoma. In this study, we apply modern techniques and nomenclature to classify 73 lymphomas from a central histopathology laboratory serving 40 mission hospitals in Kenya. We were particularly interested in the frequency of recently recognized lymphomas and the incidence of Epstein-Barr virus in various lymphoma subtypes. Malignant lymphomas accounted for 12% of all surgical pathology specimens processed in the laboratory over the 21-month period included in the study. Patient age ranged from 4 to 97 years (median, 35 years). The male-to-female ratio was 2.5:1. Sixty lymphomas (82%) were non-Hodgkin's, and 13 (18%) were Hodgkin's disease. Of the non-Hodgkin's lymphomas (NHLs), 52 (87%) were B-lineage, including 21 (35% of NHLs) Burkitt's lymphomas (only one from the jaw), 11 (18%) diffuse large B cell lymphomas; nine (15%) small lymphocytic lymphomas, six (10%) Burkitt's-like lymphomas, two (3%) follicular lymphomas (two of two expressed bcl-2 protein; one of two showed bcl-2 major breakpoint region rearrangement), two (3%) mantle cell lymphomas, and one extranodal marginal zone lymphoma. Of the eight T cell lymphomas, six were precursor T-cell type, and the remaining two were peripheral T cell lymphomas, unspecified. The median age of the 13 patients (18% of lymphomas) with Hodgkin's disease was 23 years (range, 9 to 97 years). Six were nodular sclerosis, four were mixed cellularity, one case each was lymphocyte depletion, lymphocyte predominance, and unclassified Hodgkin's disease. Hodgkin's cells in 6 of the 12 nonlymphocyte predominance cases were positive for CD20, and in three of the six for CD45 as well. Epstein-Barr virus was identified using in situ hybridization for EBER 1 in the malignant cells of 22 of 39 informative lymphomas, including each of 17 Burkitt's lymphomas, and three of seven diffuse large B cell lymphomas. Of note, none of five Burkitt's-like lymphomas expressed EBER 1. One of two informative cases of peripheral T cell lymphoma, and four of nine cases of Hodgkin's disease were EBER 1 positive. In summary, T cell lymphomas and recently recognized B-lineage non-Hodgkin's lymphoma subtypes do not appear to be particularly common in East Africa.

Immunocytochemical detection of breast cancer cells in marrow and peripheral blood of patients undergoing high dose chemotherapy with autologous stem cell support.

Detection of small numbers of breast cancer cells is important in staging the disease and can be helpful in assessing the efficacy of purging regimens prior to autologous stem cell infusion. Immunohistochemical methods are potentially useful and broadly applicable for this purpose since they are simple to perform, sensitive, and may be quite specific. We have used a combination of four monoclonal antibodies [260F9, 520C9, 317G5 (Baxter Corp); BrE-3 (Dr. R. Ceriani)] against tumor cell surface glycoproteins in a sensitive immunocytochemical assay to identify breast tumor cells in bone marrow and peripheral blood. Immunostained cytospin preparations were fixed prior to staining to preserve cytological details of immunopositive cells. After immunostaining, slides were counterstained with hematoxylin to confirm the identify of labeled cells. In cytocentrifuge experiments in which small numbers of CAMA human breast tumor cells were added to bone marrow mononuclear cells, a linear relationship between the number of tumor cells added and the number of tumor cells detected was obtained over a broad range of tumor cell concentrations. The probability of detecting tumor cells was dependent on the number of cytocentrifuge slides examined. When ten slides (5 million cells) were examined, the probability of detecting tumor at a concentration of 4 tumor cells per million bone marrow mononuclear cells was 98%. In clinical specimens, tumor cells were detected in marrow aspirates from 73 of 240 (30%) patients undergoing autologous transplantation, including 70 (37%) of 190 patients with clinical stage IV disease, 0 of 7 patients with clinical stage III disease, and 3 of 43 (7%) patients with clinical stage II disease. Seventy-three of 657 peripheral blood specimens from 26 of 155 patients (17%) contained breast cancer cells with counts ranging from 1 to 97 tumor cells per million leukocytes. Tumor cells were most frequently found in the blood of patients with stage IV disease [21 of 107 (20%)] but were also found in a substantial number [5 of 44 (11%)] of patients with stage II disease. Positive selection of CD34-positive hematopoietic progenitor cells as well as negative purging methods such as incubation with 4-hydroxyperoxy-cyclophosphamide (4-HC) were evaluated with respect to tumor cell depletion. Selection of CD34-positive progenitor cells from bone marrow or peripheral blood resulted in log reduction of 1 to > 4 tumor cells reinfused at autologous transplantation. A lesser log reduction (up to 1) was demonstrated following 4-HC purging. We conclude that properly performed and controlled immunocytochemical staining of bone marrow and peripheral blood cytospins is a sensitive and simple way to detect and quantitate breast cancer cells in hematopoietic specimens harvested for autotransplantation and that CD34-positive progenitor cell selection results in significant reduction in the number of breast cancer cells reinfused with marrow or peripheral blood stem cells.

Bone marrow involvement by lobular carcinoma of the breast cannot be identified reliably by routine histological examination alone.

The aims of this study were twofold: (1) to evaluate the ability of pathologists to recognize infiltration of bone marrow core biopsy specimens by breast carcinoma, particularly lobular carcinoma, using routine hematoxylin-eosin (HE) sections; and (2) if indicated, to determine the reasons for difficulties in diagnosis. Thirty-six bone cores obtained before bone marrow harvest were involved by breast carcinoma and were confirmed by pancytokeratin immunostains. Thirty of the 36 were ductal carcinomas and six were lobular carcinomas. Fourteen negative bone core biopsy specimens (from patients with breast cancer or lymphoma) were included as controls. These 50 bone cores were reviewed by three surgical pathologists. Lobular carcinoma was correctly identified in only 39% of positive specimens as compared with 88% for ductal carcinoma. After instruction, sensitivity for the detection of lobular carcinoma improved to 61% but at the expense of an unacceptably high rate of false-positive diagnoses (18%). None of the three pathologists was able to achieve both high sensitivity and high specificity in recognizing lobular carcinoma in the bone marrow. Lobular carcinoma was difficult to detect because of tumor cell size similar to hematopoietic cells, infiltration as single cells, presence of bland cytological features, and paucity of tissue reaction to the tumor. Although the number of cases of bone marrow involved by lobular carcinoma is small, these findings suggest that pancytokeratin stains should be performed routinely in the evaluation of bone core biopsy specimens from patients with lobular carcinoma, and probably from patients with ductal carcinoma whose HE-stained bone core biopsy specimens are considered negative for tumor.

Central nervous system angiocentric, angiodestructive T-cell lymphoma (lymphomatoid granulomatosis).

Most neurosurgeons and neurologists still consider lymphomatoid granulomatosis (LG) as a type of central nervous system (CNS) vasculitis. However, new insights, primarily from the hematologic literature, have cast doubt on the benign character of this disorder. In this article we (1) report a case of an 18-year-old woman with diffuse CNS disease and no mass lesion who developed multiple small cortical infarcts and dementia secondary to multifocal angiocentric, angiodestructive lymphoma; we (2) review other cases of LG with predominant CNS involvement; we (3) summarize the current understanding of LG, which is now considered to be a premalignant or overt angiocentric, angiodestructive T-cell lymphoma rather than a non-neoplastic vasculitis: the importance to neurologic surgeons and neurologists is that while pulmonary involvement in LG is generally the most prominent finding, patients may present with early or dramatic CNS disease; and we (4) note that although dementia is uncommon in young adults, this report adds yet another rare condition to the long differential list of dementia in this age group.

Absence of HTLV-I DNA sequences in cutaneous T-cell lymphoma/mycosis fungoides.

Recently, the presence of deleted forms of the HTLV-I provirus was described in 6 out of 6 cases of CTCL. We investigated whether the presence of these viral genomes could be verified in 20 patients with CTCL using the polymerase chain reaction (PCR) and Southern blot analysis. Only one of the 20 cases showed a band corresponding to the pX region of HTLV-I. These data indicate that in the majority of CTCL cases, sequences closely related to HTLV-I are not present, or their copy number is below the limit of detection employed in this study.

Impact of chromosomal translocations on prognosis in childhood acute lymphoblastic leukemia.

The presence of a chromosomal translocation in the leukemic cells at diagnosis of acute lymphoblastic leukemia (ALL) in children is associated with a high risk for treatment failure. We have reexamined the relationship between translocations and prognosis in 146 children with ALL who received risk-based therapy such that high-risk patients were treated with intensive drug schedules. In univariate analysis, multiple factors were associated with a relatively poor event-free survival (EFS) including age less than 2 years or greater than 10 years (combined group), WBC count greater than 10 x 10(9)/L, French-American-British (FAB) morphologic classification L2, absence of common ALL antigen (CALLA, CD10) expression, absence of hyperdiploidy with a chromosome number of 50 to 60, and presence of the specific translocations t(4; 11)(q21;q23) or t(9;22)(q34;q11) (combined group). However, there was no disadvantage with respect to EFS in patients with translocations compared with those who lacked translocations (73% at 4 years in both groups). Furthermore, when patients with specific cytogenetic abnormalities for which the prognostic significance has been well established (hyperdiploid 50 to 60, t(4;11), and t(9;22] were removed from the analysis, the remaining group with other translocations had a better EFS than the remaining group lacking translocations, although this was not statistically significant (81% v 65% at 4 years, P = .24). In a multivariate analysis, a model including WBC count and FAB classification was the strongest predictor of EFS. The presence or absence of translocations was not an independent predictor of EFS and did not contribute to the ability of any model to predict EFS. In conclusion, when effective intensive therapy is used to treat childhood ALL with high-risk clinical features, categorization of patients on the basis of chromosomal translocations without attention to the specific abnormality is not useful as a prognostic factor.

Morphology in Ki-1(CD30)-positive non-Hodgkin's lymphoma is correlated with clinical features and the presence of a unique chromosomal abnormality, t(2;5)(p23;q35).

Ten patients with strongly Ki-1(CD30)-positive non-Hodgkin's lymphoma (NHL) were identified at our institution during the past 5 years. Based on morphology, the lymphomas of five of these patients were classified as anaplastic large-cell lymphoma (ALCL); the lymphomas of four patients lacked the morphologic features of ALCL (non-ALCL); and the lymphoma of one patient was unclassifiable. Significant clinical and cytogenetic differences were observed between patients with ALCL and those with non-ALCL. The patients with ALCL tended to be young at the time of diagnosis. They presented with peripheral lymphadenopathy, and two of the five patients had skin involvement. An identical reciprocal translocation involving chromosomes 2 and 5 [t(2;5)(p23;q35)] was observed in lymph nodes from each of the two ALCL patients whose chromosomes were studied. Four of the five patients with ALCL are alive and in complete remission 10-27 months after receiving systemic chemotherapy. In contrast, the patients with non-ALCL were heterogeneous with respect to clinical findings. All of the non-ALCLs were histologically aggressive; however, their morphology varied. The t(2;5) was absent in the lymphoma specimens from each of three non-ALCL patients studied. Three of the four patients died within 17 months after receiving systemic chemotherapy. Thus, differences in morphology are correlated with differences in the clinical findings, karyotype, and outcome in Ki-1-positive NHL.

The t(2;5)(p23;q35): a recurring chromosomal abnormality in Ki-1-positive anaplastic large cell lymphoma.

In lymphoid neoplasms, nonrandom cytogenetic abnormalities correlate with clinical, morphologic and immunophenotypic features. A subtype of non-Hodgkin's lymphoma, which expresses the Ki-1 antigen (CD30) and has distinct morphologic and clinical features, has recently been described. We now report the association of a reciprocal translocation involving the short arm of chromosome 2 (band p23) and the long arm of chromosome 5 (band q35), t(2;5)(p23;q35), with Ki-1 positive anaplastic large cell lymphoma. Rearrangement of the genes that are located at the breakpoints on chromosomes 2 and 5 may be a critical step in the pathogenesis of this lymphoma.

An (8;14)(q24;q11) translocation involving the T-cell receptor alpha-chain gene and the MYC oncogene 3' region in a B-cell lymphoma.

We describe a t(8;14)(q24;q11) involving the T-cell receptor alpha-chain gene (TCRA) and the 3' region of the MYC protooncogene in a B-cell lymphoma. The B-cell origin of this tumor was determined by its histological architecture, by immunophenotypic analysis, and by Southern analysis of immunoglobulin gene rearrangements. An identical fragment encompassing the translocation breakpoint junction was detected through Southern analysis using both a TCRAJ and a MYC probe. The other alleles at the TCRAJ and MYC loci were in the germline configuration. Restriction enzyme and nucleotide sequencing analyses revealed that the breakpoint junction on chromosome 8 lies approximately 700 base pairs (bp) downstream of the 3' end of the third MYC exon; on chromosome 14, the break is located 12.6 kilobases (kb) downstream of the 3' end of the C delta fourth exon. A heptamer-like consensus sequence on chromosome 14 adjacent to the translocation breakpoint implies the involvement of recombinase activity. However, no consensus sequences were found on chromosome 8 within 140 bp in either direction from the breakpoint. It is possible that this translocation involving MYC occurred during an attempt at an inappropriate rearrangement of the TCRA locus in a cell of B-cell lineage.

The histopathologic diagnosis and subclassification of Hodgkin's disease.

Although Hodgkin's disease is considered a distinct clinical entity, it exhibits a wide range of histologic and cytologic features. It is important to recognize histologic subtypes and their variants for diagnostic reasons, as well as for their clinical significance. An appreciation of the histologic diversity of Hodgkin's disease may also be critical for the interpretation of data regarding the origin of the RS cell.

Clinicopathologic manifestations and breakpoints of the t(3;5) in patients with acute nonlymphocytic leukemia.

Acquired chromosomal rearrangements in acute nonlymphocytic leukemia (ANLL) have been linked to specific clinicopathologic features that suggest new disease subtypes. In this collaborative study, we report five patients with ANLL and a t(3;5) in their leukemic cells. At diagnosis, four of the patients had a t(3;5) as their sole karyotypic anomaly; the remaining patient had additional structural and numerical abnormalities. Careful cytogenetic analysis indicated that the breakpoints of this rearrangement are 3q25.1 and 5q34, in contrast to the various breakpoints reported in earlier studies (3q21----3q25 and 5q31----5q35). The karyotypic, morphologic, and clinical characteristics of this group, as well as those of 14 previously reported patients with the t(3;5), were compared to identify any features that might warrant consideration of a specific syndrome. The available information indicates a worldwide distribution and a nearly equal male:female ratio for patients with this translocation. The median age of the group, 37 years, was younger than that of all patients with ANLL, 49 years. A preceding myelodysplastic syndrome was observed in three patients. The limited numbers of observations on leukocyte count, hemoglobin level, and platelet count precluded meaningful comparison with data for ANLL patients in general. Although each FAB morphologic subtype, except M3, occurred in patients with a t(3;5), the frequency of M6 was much greater than expected. Bone marrows from each of the five patients we report showed increased numbers of megakaryocytes; trilineage dysplasia was observed in the marrow of each of the four patients for whom it could be assessed. Taken together, these findings suggest that the t(3;5) may affect cells capable of differentiation into multiple lineages.

Der(5)t(5;7)(q11.2;p11.2): a new recurring abnormality in malignant myeloid disorders.

Complete or partial monosomy for the long arm of chromosomes 5 and/or 7 is frequently observed in malignant cells from patients with a therapy-related myelodysplastic syndrome (t-MDS) or therapy-related acute nonlymphocytic leukemia (t-ANLL). Partial monosomy is usually the result of a chromosomal deletion; however, unbalanced translocations have also been observed. We have identified one such translocation in three patients who had either t-ANLL or a primary MDS. The genetic consequences of this translocation [-5,-7,+der(5)t(5;7)(q11.2;p11.2)] are partial monosomy for the long arm of chromosome 5 and complete monosomy for the long arm of chromosome 7. Thus, this rearrangement may represent a new, recurring abnormality that is associated with malignant myeloid disorders.

Relapse after interferon alfa-2b therapy for hairy-cell leukemia: analysis of prognostic variables.

Sixty-nine patients with hairy-cell leukemia (HCL) were treated with interferon alfa-2b (IFN) in a single-institution study. The dose used was 2 x 10(6) U/m2 self-administered subcutaneously three times weekly, for a planned treatment duration of 12 to 18 months. Of the 68 evaluable patients, the major response rate was 75%, with 13% complete responses (CRs) and 62% partial responses (PRs). An additional eleven patients (16%) had minor responses (MRs). Duration of response was denoted as failure-free survival (FFS), defined as the time from the end of IFN therapy to a need for further antileukemic therapy. Of the 60 responding patients followed after discontinuation of IFN, 27 have relapsed, requiring further therapy. The median actuarial FFS for these 60 patients is 25.4 months. All but five patients are alive, and the actuarial overall survival for the 69 patients is 91% +/- 4% at 4 years from the start of IFN. The best indicators of relapse were the neutrophil alkaline phosphatase (NAP) score and degree of residual bone marrow hairy cells (%HCL) at the completion of therapy. Patients with NAP less than 30 (n = 21) had the best prognosis (median FFS, 30.4 months), while those with NAP greater than or equal to 30 and %HCL less than or equal to 30 (n = 21) or %HCL greater than 30 (n = 16) had intermediate and poor prognoses, respectively (median FFS, 23.5 and 12.4 months) (P = .0005). Fourteen of the relapsing patients are evaluable for response to a second course of IFN, with seven PRs and four MRs. Stratified randomized trials are indicated to determine the role of maintenance therapy for responding patients.

Evaluation of Leu-M5 (CD11c) in hairy cell leukemia by the alkaline phosphatase anti-alkaline phosphatase technique.

The concomitant presence of B antigens and of the antigen recognized by the monoclonal antibody Leu-M5 (CD11c) on neoplastic lymphoid cells has been reported to be largely restricted to hairy cell leukemia (HCL). The authors studied Leu-M5 reactivity of neoplastic cells from 59 patients whose specimens were referred with a stated diagnosis of HCL by using the alkaline phosphatase anti-alkaline phosphatase technique on peripheral blood (PB) and bone marrow (BM) specimens. Tartrate-resistant acid phosphatase (AcP-T) activity was also studied. In 49 patients, HCL had been confirmed previously by BM biopsy, and specimens were evaluated for disease status during or after therapy with interferon (IFN) or 2'-deoxycoformycin. The remaining ten patients were newly referred for confirmation of the diagnosis of HCL before therapy. In all 55 patients in whom the BM biopsy demonstrated HCL, virtually every leukemic cell was Leu-M5 reactive, and the reaction proved, in some cases, to be helpful in the detection of small numbers of hairy cells in PB or BM preparations. AcP-T reactivity was demonstrated in the neoplastic cells of 52 of these 55 patients, including all but 3 of those receiving IFN, and was helpful in confirming persistent leukemia when interpretation of BM biopsy sections was difficult because the numbers of hairy cells were small. However, in four of the ten newly referred patients, BM biopsy showed features of splenic lymphoma with villous lymphocytes, rather than HCL. The neoplastic cells of these four patients were of B-cell origin and in three were Leu-M5 reactive. The authors' study indicates that Leu-M5 is present in nearly all hairy cells, but its presence in conjunction with other B-cell markers is not specific for HCL.

Hairy-cell leukemia. Morphologic, cytochemical, and immunologic features.

Hairy-cell leukemia is a unique lymphoproliferative disorder that has fascinated clinicians and researchers for nearly 3 decades. It can, in most cases, be diagnosed correctly on the basis of a bone marrow biopsy specimen, and the diagnosis is supported by the demonstration of tartrate-resistant acid phosphatase activity in the leukemic cells. Nevertheless, as outlined, rare "variant" cases may pose diagnostic problems, and other lymphoproliferative disorders may have features that closely mimic those of HCL. In virtually all cases of HCL, the hairy cells are of B-cell lineage, representing a relatively mature B cell that is closely related to plasma cells. Treatment of HCL with splenectomy, IFN-alpha, or 2' deoxycorformycin has led to excellent responses in most cases. Much of the progress in our understanding of the nature of this rare leukemia has been made possible through the cooperation of community physicians with referral institutions, or groups of institutions, that are interested in studying HCL. Despite the remarkable advances in the treatment of this disorder over the last 5 years, and the ready availability of IFN-alpha, much remains to be learned regarding the biology of hairy cells, and further advances in therapy still must be made. The needed information can be obtained only through continued referral of patients to centers with expertise and an interest in HCL.

Specific chromosomal abnormalities in acute nonlymphocytic leukemia correlate with drug susceptibility in vivo.

Specific chromosomal abnormalities are independent predictors of response to therapy in acute nonlymphocytic leukemia (ANLL) de novo. In a series of 149 patients with ANLL, we sought to determine whether the t(8;21), t(15;17), t(9;11) or other abnormalities of the long arm of chromosome 11, inv(16) or t(16;16), inv(3) or t(3;3), trisomy 8, and abnormalities of chromosome 5 (-5/5q-) or of chromosome 7 (-7/7q-) identify differences in susceptibility to chemotherapy drugs in vivo. The immediate outcome of the first cycle of remission induction chemotherapy was analyzed for patients in each cytogenetic subgroup as an index of the drug susceptibility of the leukemia cells in vivo. Patients with t(8;21), inv(16), t(16;16), or 11q abnormalities had high rates of complete remission after initial therapy (60-100%), whereas patients with -7/7q- or -5/5q- had low initial response rates (0-36%), suggestive of drug resistance in vivo. In general, cytogenetic groups with high initial complete remission rates ("drug sensitive") also had long disease-free survivals; those groups with low initial remission rates ("drug resistant") had short remission durations even if these patients eventually achieved complete remission with further therapy. Patients with acute promyelocytic leukemia (APL), all of whom had the t(15;17), were the exception; despite low initial remission rates, they had long disease-free survivals, possibly due to a more rapid cytotoxic effect of chemotherapy on the clonogenic APL cells than on the more numerous malignant promyelocytes. We conclude that the prognostic importance of specific chromosomal abnormalities in ANLL resides in part in differing susceptibilities to chemotherapy.