Enhanced Risk Stratification for Children and Young Adults with B-Cell Acute Lymphoblastic Leukemia: A Children's Oncology Group Report.

Current strategies to treat pediatric acute lymphoblastic leukemia rely on risk stratification algorithms using categorical data. We investigated whether using continuous variables assigned different weights would improve risk stratification. We developed and validated a multivariable Cox model for relapse-free survival (RFS) using information from 21199 patients. We constructed risk groups by identifying cutoffs of the COG Prognostic Index (PICOG) that maximized discrimination of the predictive model. Patients with higher PICOG have higher predicted relapse risk. The PICOG reliably discriminates patients with low vs. high relapse risk. For those with moderate relapse risk using current COG risk classification, the PICOG identifies subgroups with varying 5-year RFS. Among current COG standard-risk average patients, PICOG identifies low and intermediate risk groups with 96% and 90% RFS, respectively. Similarly, amongst current COG high-risk patients, PICOG identifies four groups ranging from 96% to 66% RFS, providing additional discrimination for future treatment stratification. When coupled with traditional algorithms, the novel PICOG can more accurately risk stratify patients, identifying groups with better outcomes who may benefit from less intensive therapy, and those who have high relapse risk needing innovative approaches for cure.

Distribution of common acute lymphoblastic leukemia antigen in nonhematopoietic tissues.

The common acute lymphoblastic leukemia antigen (CALLA), as defined by J-5 murine monoclonal antibodies, was detected on renal tubular and glomerular cells from fetal and adult donors by an indirect immunoperoxidase technique. CALLA could also be detected on epithelial cells of the fetal small intestine and on myoepithelial cells of adult breast but not on myoepithelial cells of the salivary gland. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis of immunoprecipitated 125I-labeled membrane antigens from dissociated renal cells demonstrated that the antigen migrated as a 90,000 mol wt antigen rather than the 98,000-100,000 mol wt antigen noted on CALLA-positive tissue culture cell lines. The data suggest that the determinant defined by the J-5 monoclonal antibody is neither a lymphoid cell-specific differentiation antigen nor a leukemia-specific antigen.

Acute bilineal leukemia: a rare disease with poor outcome.

Most cases of acute leukemia can be assigned to the myeloid, B or T lineage. In a few cases, definitive assignment cannot be achieved because blasts express antigens of more than one lineage. A subset of these, referred to as acute bilineal leukemias (aBLLs), is characterized by the presence of more than one population of blasts, each comprising a single lineage. We identified 19 cases of aBLL, including 10 mixed T and myeloid (T-My) and nine mixed B and myeloid (B-My); no mixed B and T cases were identified. Cytogenetic data were available for 16 patients. Three of seven patients with B-My had a t(9;22)(q34q11.2), two had 11q23 translocations and one had del(9). Two of nine patients with T-My had 2p13 translocations; five had other unrelated abnormalities. Of 16 patients with outcome data, only six achieved complete remission and only two remain free of disease 2.5 and 4.5 years after chemotherapy or stem cell transplantation. aBLL is a rare disease that combines B or T and myeloid blasts. Cytogenetic abnormalities of t(9;22) and 11q23 are common in, and may be restricted to, B-My cases, while T-My cases have frequent but generally non-recurring abnormalities. Both types of aBLL are associated with poor outcome.

2006 Bethesda International Consensus recommendations on the flow cytometric immunophenotypic analysis of hematolymphoid neoplasia: medical indications.

The clinical indications for diagnostic flow cytometry studies are an evolving consensus, as the knowledge of antigenic definition of hematolymphoid malignancies and the prognostic significance of antigen expression evolves. Additionally the standard of care is not routinely communicated to practicing clinicians and diagnostic services, especially as may relate to new technologies. Accordingly there is often uncertainty on the part of clinicians, payers of medical services, diagnostic physicians and scientists as to the appropriate use of diagnostic flow cytometry. In an attempt to communicate contemporary diagnostic utility of immunophenotypic flow cytometry in the diagnosis and follow-up of patients with hematolymphoid malignancies, the Clinical Cytometry Society organized a two day meeting of international experts in this area to reach a consensus as to this diagnostic tool. This report summarizes the appropriate use of diagnostic flow cytometry as determined by unanimous approval of these experienced practitioners.

Quantitative analysis of bone marrow CD34 cells in aplastic anemia and hypoplastic myelodysplastic syndromes.

Aplastic anemia (AA) and hypoplastic myelodysplastic syndromes (hMDS) are often difficult to distinguish. However, an accurate diagnosis is important because the prognosis and treatment of these diseases may differ. CD34+ hematopoietic progenitors are central to the pathogenesis of both disorders; they are the targets of the autoimmune attack in AA and neoplastic transformation in MDS. The aim of this study was to assess whether bone marrow CD34+ cell numbers could be used in differentiating between AA and hMDS. The percentage of bone marrow CD34+ cells was normal or increased (mean -3.5+0.5%, range 1-7%) in 15 of 35 patients studied, and low (mean -0.13 +/- 0.02%, range 0.02-0.36%) in 20 of 35 patients. All patients with a normal or increased percentage of CD34+ cells were ultimately diagnosed with hMDS based on the detection of clonal cytogenetic abnormalities or progression to refractory anemia with excess blasts/acute myeloid leukemia. All patients with low marrow CD34+ cell numbers met standard clinical criteria for AA and have not demonstrated neoplastic transformation with follow-up. Quantification of marrow CD34+ cells may serve as an important tool for distinguishing between AA and hMDS.

High concordance from independent studies by the Children's Cancer Group (CCG) and Pediatric Oncology Group (POG) associating favorable prognosis with combined trisomies 4, 10, and 17 in children with NCI Standard-Risk B-precursor Acute Lymphoblastic Leukemia: a Children's Oncology Group (COG) initiative.

Chromosome aberrations have a major role in pediatric acute lymphoblastic leukemia (ALL) risk assignment. The Children's Cancer Group (CCG) and the Pediatric Oncology Group (POG) independently assessed the significance of trisomy for chromosomes 4, 10, and 17 in National Cancer Institute (NCI) Standard- and High-Risk ALL. Data from 1582 (CCG) and 3902 (POG) patients were analyzed. Eight-year event-free survivals (EFS) of 91% (CCG) and 89% (POG) (P < 0.001) were achieved in patients assigned to NCI Standard Risk whose leukemic cells had simultaneous trisomies 4, 10, and 17. Both groups showed the degree of favorable prognostic importance increased with the actual number of favorable trisomies. POG analyses also demonstrated hyperdiploidy (> or =53 chromosomes) was less of an independently significant prognostic factor in the absence of these key trisomies. This finding supported conclusions from previous CCG and POG studies that specific trisomies are more important than chromosome number in predicting outcome in pediatric B-precursor ALL. In NCI Higher Risk patients, the number of favorable trisomies was not prognostically significant, but showed the same trend. Moreover, specific trisomies 4, 10, and 17 remain associated with favorable prognosis in Standard-Risk B-precursor ALL, even in the context of very different treatment approaches between the groups.

Monoclonal antibodies against human pancreatic adenocarcinoma: distribution of DU-PAN-2 antigen on glandular epithelia and adenocarcinomas.

This report describes the distribution of antigen DU-PAN-2, defined by a monoclonal antibody raised against human pancreatic carcinoma cells, on a variety of tumors and nonneoplastic tissues. With the use of an immunoperoxidase technique, the antigen was detected on 16 of 16 pancreatic carcinomas, on 5 of 5 gallbladder or bile duct carcinomas, on the great majority of stomach adenocarcinomas, and, less commonly, on adenocarcinomas of other primary sites. Substantial intratumor heterogeneity of antigen expression was noted. DU-PAN-2 antigen was present on many types of normal glandular epithelial cells but often was more weakly expressed than on the corresponding tumors. The immunomorphology of staining, coupled with biochemical information known about the antigen, supports the notion that the DU-PAN-2 antigen is a mucin-like substance. Its relative restriction of expression on different types of glandular epithelium suggests that DU-PAN-2 antibody might be a useful reagent for helping to determine the site of origin of adenocarcinomas.

Phase I study of an anti-breast cancer immunotoxin by continuous infusion: report of a targeted toxic effect not predicted by animal studies.

260F9 Monoclonal antibody-recombinant ricin A chain, an immunotoxin reactive with approximately 50% of breast carcinomas, was given by continuous iv infusion at a dose of 50 micrograms/kg per day or 100 micrograms/kg per day. Five patients with refractory breast cancer received treatment for from 6 to 8 days. Severe toxic effects, including marked fluid overload and debilitating sensorimotor neuropathies, occurred in most patients. Immunoperoxidase studies suggested that 260F9 monoclonal antibody targeting of the Schwann cells may have induced demyelination and subsequent neuropathy. This is the first report of a targeted toxic effect due to an immunoconjugate.

A novel ribonucleoprotein complex defined by monoclonal antibodies to NIH 3T3 cells transfected with human pancreatic adenocarcinoma DNA.

Three monoclonal antibodies elicited to NIH 3T3 cells transfected with DNA from a human pancreatic adenocarcinoma cell line recognized a novel ribonucleoprotein complex. Minimally, this ribonucleoprotein complex contained a Mr 240,000 protein (by sodium dodecyl sulfate-polyacrylamide gel electrophoresis) and two RNA species with apparent sizes of 1.5 and 3.0 kilobases (by formaldehyde agarose gel electrophoresis). In addition to a cytoplasmic and nuclear subcellular localization, the RNA antigen was secreted from human tumor cell lines and NIH 3T3 cells transfected with pancreatic tumor DNA (inhibitable by monensin) and was apparently not a viral or Mycoplasma contaminant. The ribonucleoprotein antigen was detected in some normal tissues by immunoperoxidase but was not found in or secreted from in vitro cultured normal human fibroblasts, nontransfected or spontaneously transformed NIH 3T3 cells, or normal peripheral blood leukocytes.

Tissue distribution and cellular location of the antigens recognized by human monoclonal antibodies 16.88 and 28A32.

Results from a previous study (M. V. Haspel et al., Cancer Res., 45: 3951-3961, 1985) indicated that it was possible to isolate a high proportion of human monoclonal antibodies reactive with cell surface, tumor-associated antigens when the hybridomas were obtained from fusions utilizing peripheral blood lymphocytes from patients immunized with autologous tumor cells. The assignment of membrane reactivity was made from immunoperoxidase studies which used air-dried, nonpermeabilized Cytospin preparations of colon tumor cells. Tumor specificity was assessed by immunohistological assays by using frozen sections of normal and malignant human tissues. We now describe a series of studies using two of these antibodies, 16.88 and 28A32, in which further information was obtained concerning the tumor specificity and cellular location of the target antigens reactive with these monoclonal antibodies. Data were acquired from a variety of experimental techniques which included quantitative and qualitative immunofluorescence on live and permeabilized cells, RBC-rosetting assays, immunoperoxidase studies on Cytospin and frozen tissue sections, and immunoblot procedures. These studies show that the 16.88 and 28A32 human monoclonal antibodies bind to antigens which (a) are located in the cell cytoplasm and are not expressed in detectable levels on the cell surface, and (b) are found in many normal and malignant cell types.

Glucocorticoid receptors in immunological subtypes of childhood acute lymphocytic leukemia cells: a Pediatric Oncology Group Study.

Glucocorticoid receptors were quantitated by a whole cell method in cells from 593 children with acute leukemia at the time of diagnosis. Leukemia cells were also immunologically typed and divided into early pre-B- (not reactive with antibodies to T-lymphocyte antigens, surface immunoglobulin-negative, cytoplasmic immunoglobulin-negative), pre-B- (not reactive with antibodies to T-lymphocyte antigens, surface immunoglobulin-negative, cytoplasmic immunoglobulin-positive), B- (not reactive with antibodies to T-lymphocyte antigens, surface immunoglobulin-positive), and T- (reactive with antibodies to T-lymphocyte antigens) subtypes. There was a median of 9.7 X 10(3) sites per cell in the 359 with early pre-B-acute lymphocytic leukemia, a median of 8.1 X 10(3) sites per cell from 103 patients with pre-B-cell leukemia, and a median of 4.0 X 10(3) sites per cell from 116 patients with T-cell leukemia. The distributions per cell were significantly different among these 3 groups (P less than 0.0001). The 15 patients with B-cell disease had a median of 3.2 X 10(3) sites per cell. At the time of analysis, remission induction data are available for most of these patients. Within the early pre-B- group 291 patients with a median receptor number of 9.9 X 10(3) achieved remission, while 13 with a median receptor number of 4.8 X 10(3) did not. These distributions were significantly different (P = 0.034). Within the pre-B- and T-cell groups the distributions of receptor numbers for responders and non-responders were not significantly different. We conclude that each immunological subtype has characteristic receptor distribution. High receptor number within the null group is associated with the ability of the patient to achieve remission; however, the range of values within each patient group is too broad to use this assay as a predictor of response for any individual patient.

Antigens of human pancreatic adenocarcinoma cells defined by murine monoclonal antibodies.

We have elicited and characterized the serological specificity of five murine monoclonal antibodies (DU-PAN-1, 2, 3, 4, and 5) to a human pancreatic tumor cell line, HPAF. The antibodies are not detecting HLA-associated antigens since all of the monoclonals failed to react with human lymphoid and myeloid cell lines and uncultured cells. All of the monoclonals except DU-PAN-5 reacted with four of five pancreatic tumor cell lines and two of two uncultured pancreatic tumors. An immunoperoxidase technique was used to determine the presence of the antigens detected by the monoclonal antibodies in frozen sections of tumor and adult and fetal normal tissues. DU-PAN-1 antigen was detected on pancreatic tumors, and a transitional cell carcinoma of the bladder, but was detected on no other adult or fetal normal tissues including pancreas. One of the antigens (DU-PAN-2) defined by the monoclonals was present on pancreatic ductal epithelial cells and showed a restricted distribution on tumor cells from some other carcinoma patients and on cells from certain fetal tissues. DU-PAN-3 antigen was present on adult and fetal pancreatic cells and certain tumor cells but could not be detected on cells of other fetal or adult normal tissues. DU-PAN-4 and DU-PAN-5 antigens have a more widespread distribution on normal or tumor cell types.

Triacylglycerol turnover in Tetrahymena pyriformis. Relation to phospholipid synthesis.

The metabolic function of triaclyglycerol in Tetrahymena pyriformis was investigated by prelabeling endogenous lipid with a 14C-labeled short chain fatty acid, and then following the disappearance of radioactivity from triacylglycerol and its appearance in other products. In 90 min, up to 85% of the label in triacylglycerol turns over, and although some radioactivity appears in CO2 and glycogen, most of the label appears in phospholipid. Starvation of the cells, as well as resuspension in enriched medium or provision of acetate all block triacylglycerol breakdown, while supplementation of the medium with pyruvate does not. Prelabeling lipid with [3H] glycerol shows that some of the transfer of material from triacylglycerol to phospholipid involves transfer of the glycerol backbone, although transfer of triacylglycerol fatty acids directly to phospholipid probably also occurs. In addition, the catabolism of triacylglycerol occurs by a "last-in-first-out" mechanism, indicating some form of compartmentation of triacylglycerol in this cell. The results demonstrate an important metabolic interrelationship between triacylglycerol catabolism and phospholipid synthesis and raise the question, in this cell at least, of the validity of considering triacylglycerol only as a fuel storage form.

Gastrointestinal involvement and multiple lymphomatous polyposis in mantle-zone lymphoma.

The clinical, histologic, and immunologic features of three cases of lymphoma presenting with gastrointestinal symptoms and involving the gastrointestinal tract were studied. Two of the cases had involvement of long segments of the bowel with polypoid lesions, a rare presentation of gastrointestinal lymphoma referred to as multiple lymphomatous polyposis (MLP). All cases were classified as mantle-zone lymphoma (MZL), a follicular variant of intermediate lymphocytic lymphoma (ILL) characterized by the proliferation of small atypical lymphoid cells as wide mantles surrounding benign-appearing germinal centers. The patients were typical of other patients with MZL in that they were male, middle age or older, and their clinical courses were not aggressive despite the presence of disease at an advanced stage. Our review of the literature suggests that there is an inordinate number of cases of MZL with gastrointestinal involvement. We also note that most reports of cases of MLP have described histologic lesions remarkably similar to what we have observed. We conclude that MZL may have predilection for involvement of the gastrointestinal tract, that this involvement is often in the manner of MLP, and that most cases of MLP probably have MZL or ILL.

Secondary acute myeloid leukemia in children with acute lymphoblastic leukemia treated with etoposide.

PURPOSE: To describe the occurrence of secondary acute myeloid leukemia (AML) in children with acute lymphoblastic leukemia (ALL) treated with etoposide (VP-16). PATIENTS AND METHODS: Two hundred five consecutive children with early B-lineage ALL were treated according to the Dallas/Fort Worth (DFW) protocol between January 1986 and July 1, 1991. Therapy included a four-drug induction followed by consolidation and continuation phases of nightly oral mercaptopurine (6-MP) and repetitive courses of divided-dose oral methotrexate (dMTX) and asparaginase (L-asp). Three doses of VP-16 and cytarabine (Ara-C) were given during consolidation and later, during continuation, two doses were given 3 to 4 days apart, every 9 weeks. Intrathecal (IT) chemotherapy was given throughout the treatment period. RESULTS: Two hundred three of the 205 patients entered remission. Only eight of these 203 children have had a bone marrow relapse (ALL). However, 10 other children have developed secondary AML 23 to 68 months following the diagnosis of ALL. Overall event-free survival (EFS) at 4 years is 79.3% +/- 5.1%, with a risk of secondary AML at 4 years of 5.9% +/- 3.2%. CONCLUSION: This experience provides strong evidence for a link between epipodophyllotoxin therapy and secondary AML since none of these children received alkylating agent therapy or irradiation. This serious complication raises concern as to the appropriate use of epipodophyllotoxins in the treatment of childhood ALL.

Clinicopathologic aspects of E rosette negative T cell acute lymphocytic leukemia: a Pediatric Oncology Group study.

The Pediatric Oncology Group has studied 1,367 patients with non-B cell acute lymphocytic leukemia (ALL) of whom 186 (14%) had blasts that reacted with previously well-characterized heteroantisera, recognizing a T-lymphocyte specific surface membrane antigen (PT+). In 87 of these T cell cases, the leukemic cells failed to form at least 20% of sheep erythrocytic rosettes at 4 degrees C. Comparison of clinicopathologic features among PT-, E-PT+, and E+ groups of patients revealed significant differences among them. E-PT+ patients were older than PT- patients, had higher white blood cell counts (WBCs) and were more likely to have a mediastinal mass, and thus contained a higher proportion of poor-risk patients. However, the E-PT+ patients were also significantly different from the more traditionally-defined E+ patients in that they had lower WBCs and hemoglobin levels, and less frequent lymphadenopathy or mediastinal mass. In many respects, then, E-PT+ patients were intermediate in character between PT- and E+ patients. Our findings support the notion that further subclassification of ALL using antibodies recognizing lineage-specific surface determinants will permit recognition of groups of patients with distinct clinicopathologic features that may differ in prognosis or response to therapy. Such classification also focuses future biological studies on the pathogenesis of leukemias to immunologically well-defined subgroups of lymphoid neoplasms.

Consolidation therapy with antimetabolite-based therapy in standard-risk acute lymphocytic leukemia of childhood: a Pediatric Oncology Group Study.

PURPOSE: To develop antimetabolite-based consolidation regimens that minimize acute and long-term toxicities and improve the survival rate of children with standard-risk B-lineage acute lymphocytic leukemia (ALL). PATIENTS AND METHODS: Seven hundred twenty-seven eligible patients with standard-risk early pre-B ALL were registered onto the study. Seven hundred sixteen patients attained a complete remission (CR) after induction therapy. Of these, 114 patients were randomized to a different regimen and were the subject of a separate report. Six hundred two patients were randomized to receive one the following regimens: intermediate-dose methotrexate (IDMTX) with leucovorin rescue on weeks 7, 10, 13, 16, 19, and 22 (regimen A); regimen A plus asparaginase (ASP) administered intramuscularly (i.m.) weekly for 24 weeks (regimen B); or regimen A plus a 24-hour infusion of cytarabine (AraC) with each IDMTX (regimen C). After consolidation, patients were placed on maintenance therapy through week 156. Regimens A and C were opened in February 1986, and regimen B in May 1987. Comparisons are based on concurrently randomized patients (May 1987 to January 1991 between regimens A and B, and February 1986 to January 1991 between regimens A and C). RESULTS: The 5-year continuous CR (CCR) rates were not significantly different: A versus B, 78.1% (3.9 +/- SE) versus 83.3% +/- 3.5% and A versus C, 79.4% +/- 3.2% versus 83.5% +/- 2.9%; P by one-sided log-rank tests were .27 and .34, respectively. Significant treatment differences were not found with regard to sex, rate of testicular and CNS relapse, or CNS complications. During consolidation, regimen C had significantly more bacterial infections (P = .0032) and days spent in the hospital (P < .001) compared with regimen A. CONCLUSION: We were unable to show a statistical advantage of adding either ASP or AraC to IDMTX in terms of improvement in event-free survival (EFS).

Intensive oral methotrexate protects against lymphoid marrow relapse in childhood B-precursor acute lymphoblastic leukemia.

PURPOSE: To describe the use of combination chemotherapy, including divided-dose oral methotrexate (dMTX), for children with B-precursor acute lymphoblastic leukemia (ALL). dMTX produced prolonged MTX exposure on an outpatient basis. PATIENTS AND METHODS: Two hundred forty-three patients were treated from January 1986 to May 1992. dMTX was given weekly during consolidation and biweekly for the first 16 months of continuation therapy with mercaptopurine (6-MP) and asparaginase (L-ASP). Initially, etoposide (VP-16) and cytarabine (Ara-C) pulses were included. Treatment continued for 30 months with single-dose weekly MTX replacing dMTX during continuation, part 2. Unexpected acute neurotoxicity was eliminated by the addition of leucovorin. VP-16 and Ara-C were omitted in the face of acute myelogenous leukemia (AML). RESULTS: Two hundred thirty-nine patients entered remission: 16 had a lymphoid marrow relapse, two each with testicular or CNS relapse; 19 a CNS relapse; 16 secondary AML; three other second malignancies; two withdrew for transplant; three died in remission; 16 withdrew because of noncompliance, and nine withdrew with toxicity. Event-free survival (EFS) at 4 years was 73 +/- 4%; 81 +/- 4% for 150 patients with better risk features and 60 +/- 7% for 93 with high-risk features. Lymphoid marrow relapse-free survival in the standard- and high-risk patients was 94 +/- 3% and 86% +/- 6%, respectively. The most common adverse event was secondary AML in the standard-risk group and isolated CNS relapse in the high-risk group. CONCLUSION: This therapy produced an overall EFS similar to other published regimens, but the pattern of failures is very different, with few patients having a lymphoid marrow relapse. These data suggest that highly effective therapy for children with ALL can be delivered on an outpatient basis using a regimen featuring repetitive dMTX.

Prognostic significance of CD34 expression in childhood B-precursor acute lymphocytic leukemia: a Pediatric Oncology Group study.

We studied the blasts from 795 children greater than 1 year of age with newly diagnosed, untreated B-precursor acute lymphoblastic leukemia (ALL) for expression of the hematopoietic stem cell-associated antigen CD34. All cases were confirmed as B-lineage lymphoblastic leukemia by virtue of expression of CD19 and/or CD22, lack of T-cell antigens, and lack of surface-membrane immunoglobulin (Ig). The CD34 antigen was present in at least 10% of blast cells in 587 (73.8%) of the patients. There was no significant difference in presenting clinical characteristics between CD34+ and CD34- patients save for an increased incidence of CNS involvement at diagnosis in the latter. Patients with CD34+ leukemia were more likely to have blasts expressing CD22, CD9, and CD13 antigens but were less likely to coexpress CD20. Patients with pre-B (cytoplasmic mu) ALL were significantly more likely to lack CD34 on their blasts, while children with hyperdiploid ALL were more likely to be CD34+. Although remission induction rates were not significantly different between patients with CD34+ and CD34-ALL (P = .23), event-free survival was shorter for patients with CD34- leukemia (P = .0014). Even though CD34 expression was associated with certain other known prognostically favorable variables including hyperdiploidy and lack of cytoplasmic Ig, it had an independent favorable effect on treatment outcome, even after adjusting for competing prognostic factors.

Acute lymphoid leukemia in adolescents: clinical and biologic features predict a poor prognosis--a Pediatric Oncology Group Study.

Analysis of remission induction rates for 1,768 children (1.5 to 11 years) and 425 adolescents (greater than or equal to 11 years) with acute lymphoid leukemia (ALL), and of event-free survival times for 570 children and 147 adolescents, disclosed that adolescents fared significantly worse by both measures of treatment outcome (P = .0001). Adolescents with either T cell or non-T cell ALL entered remission significantly less often than did children (P = less than .02 and P = less than .001, respectively). Within each of the major immunophenotypes of ALL, adolescents had shorter duration of continuous complete remission: early pre-B (non-B, non pre-B, non-T) (P = .001), pre-B (P = .05), and T (P = .027). We compared the clinical characteristics of adolescents and children, and lymphoblast characteristics present at diagnosis to account for the inferior prognosis of adolescent patients. Adolescents had a higher incidence of T cell ALL (P = .0001) and thus a higher incidence of all T cell-associated characteristics. Adolescents with non-T, non-B ALL were more likely to be male (P = .044), and to have higher leukocyte counts (P = .002) and lower levels of IgG (P = .0003), IgA (P = .0001), and IgM (P = .002). Their leukemic cells had lower PAS scores (P = .0001), a higher incidence rate of L2 morphology by French-American-British (FAB) criteria (P = .001), common ALL antigen negativity (P = .0001), and hypodiploid or pseudodiploid karyotypes (P = .004). These findings clearly indicate an increased incidence of prognostically unfavorable clinical and biologic features in adolescents with ALL, providing a biologic explanation for their poor prognosis.