A Clinical Perspective on Cell Markers in Acute Lymphocytic Leukemia1

Two hundred consecutive new patients, with acute lymphocytic leukemia (ALL) have been studied with a battery of five cell marker assays to determine if a classification system with prognostic significance can be developed; 182 have been classified among four groups as follows: 33 T-cell, 3 B-cell, 126 common, and 20 undifferentiated ALLs. Patients with Tcell disease are likely to have unfavorable clinical prognostic features and a poor response to therapy. Rare patients with Bcell disease are closely related clinically to non-Hodgkin's lymphoma. Those with common ALL infrequently have unfa vorable clinical features and have a superior outcome to that of T-cell patients. Children with undifferentiated markers seem to respond less well to treatment than do those with common ALL, yet may not be identifiable as poor risk by clinical features. What remains to be resolved with further observation is whether these marker patterns are more reliable indicators of prognosis than the usual clinical determinants predisposing to treatment failure (high white blood cell count, mediastinal mass, and central nervous system disease). At the present time, it appears that in the absence of poor-risk clinical prognostic features, patients with common ALL are more likely to have lasting remissions than those with erythrocyte-rosette-positive T-cell disease or those with ALL that is undifferentiated by markers. Introduction The recognition of prognostic factors in ALL2 has led to a search for biological correlates of clinical behavior and for ways to better identify important subgroups of this disease. Lymphoblast cell markers are of interest to different investiga tors for different reasons. For the leukemia therapist, the vital questions are these. How important to treatment planning is the marker phenotype? Which markers are important? What, if anything, do cell markers add to our expectations of treatment response based upon clinical prognostic factors? These ques tions have to be addressed within the setting of evolving treatment programs. Therapy, of course, is the most important prognostic indicator of all; it is conceivable that certain "im portant" variables might have little impact if a better therapy were available. The ongoing study at St. Jude was stimulated by early results linking E-rosette-positive ALL to poor outcome (7) and by the work of Greaves ef al. (4), suggesting that certain lymphoid differentiation antigens may contribute to an improved classification of ALL. Our investigation seeks to de termine the relationships among blast cell membrane markers, 1 Presented at the Conference on Cell Markers in Acute Leukemia, March 4 and 5, 1980, Bethesda, Md. Supported by Cancer Center Support (CORE) Grant CA-21765 from the National Cancer Institute, by Leukemia Program Project Grant CA-20180 and by American Lebanese Syrian Associated Charities. 2 The abbreviations used are: ALL, acute lymphocytic leukemia; E-rosette, erythrocyte-forming rosette; Slg, surface ¡mmunoglobulin; AML. acute myelocytic leukemia; CNS, central nervous system. clinical prognostic features, and response to therapy in child hood ALL. Materials and Methods A battery of surface marker tests has been used to study bone marrow blast cells. The E-rosette test demonstrates lymphoblasts with T-cell properties. By our criteria, to be considered E-rosette positive, more than 2% of the patient's blast cells must form heat-stable rosettes with sheep erythrocytes. Slg is sought by direct immunofluorescence using a polyvalent anti-human immunoglobulin serum. All samples with Slg-positive lymphoblasts have contained more than 40% fluorescent cells, whereas Slg-negative cases have consistently had fewer than 10%. With these 2 studies, at most 15% of newly diagnosed patients can be classified as to cell type, and the remainder have been labeled "null cell" or "non-T, non-B" ALL. We have used 3 heterologous antisera to membrane antigens to characterize these patients further. All 3 antisera were prepared in rabbits. An anti-T serum was produced against human fetal thymocytes. After appropriate absorptions, it bound to 60% of blood lymphocytes and reacted with Slg-negative cells in blood and tonsil cell preparations. A second anti-T reagent was produced using for immunization the blast cells from a patient who had reacted strongly to the first antiserum. The anti-la (B) serum is commercially available and was prepared by the method of Billing ef al. (2). The antigenic complex detected is rare on T-lymphoblasts but widely distributed otherwise. It is present in the majority of non-T, non-B ALLs, but also in those that are Slg positive and in about two-thirds of AML cases. Our anti-ALL serum was prepared against the blast cells of an Erosette-negative, Slg-negative, anti-T-negative, anti-la-positive patient. After absorptions, it failed to react with blood lymphocytes, normal bone marrow, or AML cells. It is reactive with the majority of non-T, non-B cases and detects the common ALL antigen originally described by Greaves (4, 9). Each of the 3 antisera is used in an indirect immunofluorescence assay. The presence of greater than 45% flu orescent cells in the marrow sample is considered a positive result for purposes of assignment of membrane phenotype. Our battery of 5 surface marker studies has been applied to the investigation of new leukemia patients. This report concerns the clinical features at diagnosis and early follow-up for 200 consecutive children diagnosed to have ALL at St. Jude during a 27-month period from November 1977 to February 1980. These patients have been prospectively classified according to the scheme in Table 1. T-cell disease is that which is E-rosette positive and/or anti-T positive. Cells from Erosette-positive patients seem always to bear T-cell antigens, but the converse is not true, and some ALLs with T-cell properties have not been E-rosette forming. B-cell ALL is that which is Slg positive. Com mon ALL includes those patients whose cells have reactivity with both the anti-ALL and la sera. Rarely (less than 10% of cases), the ALL antigen is detected but with weak or absent la expression. For the purposes of this clinical report, these individuals are included with the common ALLs. Finally, there is an undifferentiated group in which blast cells do not express any markers in detectable quantities or in which only la expression is observed. Results and Discussion Of the 200 patients, 182 have been classified according to this scheme (Table 2). The 18 unclassified patients were those 4794 CANCER RESEARCH VOL. 41 on April 15, 2017. © 1981 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from Clinical Study of Cell Markers in ALL from whom insufficient blast cells were available to perform all the tests, although all but one were tested for E-rosettes and were found to be negative. These patients are not considered for further analysis, although clinical features and early followup data indicate that they are most likely common ALLs. Of those classified, there were 33 T-cell, 3-B cell, 126 common, and 20 undifferentiated ALLs. B-cell acute leukemia is ob viously rare. The 3 patients in this series had distinct lymphomatous disease and lymphoblasts with L3 morphology (5). Two failed treatment early, and one remains in complete remission at 2 years; they will not be considered for further discussion. The remainder of this presentation will focus upon compari sons among the 179 patients classified into the major groups of T, common, and undifferentiated. There is no significant difference among these 3 groups according to age or sex. However, blacks comprise 25% of the undifferentiated group, yet less than 10% of the whole series. In our experience, the clinical features at diagnosis most reliably associated with treatment failure are high WBC, mediastinal mass, or CMS involvement. The relationship of these features to cell marker patterns was determined and major trends are quite apparent. With respect to WBC (Table 3), there is a disproportionate number of very high counts among T-cell patients and of low counts among common ALL patients. Nineteen of 33 T-cell patients had WBC over 50,000/cu mm, while 113 of 126 common ALLs had counts below 50,000. Rarely does one see a low WBC in a T-cell patient or a high one in a common ALL patient. The undifferentiated group tends to fall into an inter mediate range, although both extremes of high and low have been seen. Turning to other clinical features, 17 of the 18 with mediastinal masses are T-cell patients, as are 5 of 8 with Table 1 ALL surface marker classification T-cell B-cell Common ALL Undifferentiated E-rosette positive and/or T positive • Slg positive • ALL positive • la positive • All negative, ' or la positive only Table 2 Surface markers and clinical features T-cellB-cellCommonUndifferentiatedUnclassifiedTotalNo. of patients3331262018200Boys213631213112Girls120638588Blacks4085118 pretreatment CNS involvement (Table 4). Thus, high-risk clini cal features are strikingly prominent in patients with T-cell disease, and infrequent in common and undifferentiated ALL. During the accumulation of this series, there were 4 patients originally diagnosed to have ALL, in whom subsequent cytochemical staining characteristics indicated a diagnosis of AML or acute monocytic leukemia. In all 4 cases, the cell marker pattern was undifferentiated. These patients are not included in the follow-up data presented below. Table 5 illustrates early outcome results for the 111 ALL patients classified up to June 1979. The period of follow-up ranges from 8 to 26 months. These patients are being treated according to St. Jude Total Therapy Study IX, a program which is evaluating variations in the intensity of remission induction and consolidation; present analysis of this protocol, which includes a further 160 patients entered prior to this cell marker study, shows no difference among its 3 arms. The most recent 64 patients in the marker series are excluded from further consideration because they have been treated according to a different protocol and follow-up is very short. Of the 25 T-cell patients analyzed, 4 failed to respond to initial induction therapy, and 8 others have experienced marrow relapse. Following initial treatment, failure in T-cell ALL survival is often short, and 10 of these 12 have already died. Of the 74 common ALL patients, there were 2 early deaths from sepsis, but only one failure to respond to induction if an adequate trial could be given. Hematological relapse has occurred in 13 patients. Three patients with common ALL have unfortunately died during remission, 2 from cytomegalovirus pneumonia and one from Reye's syndrome. In the undifferentiated group, there Table 3 Surface markers and clinical features No. of patientsWBC(x103)Total>10050-10010-50<10T-cell3316395Common126944370Undifferen tiated204097Total1792976182 Table 4 Surface markers and clinical features T-cell Common UndifferentiatedTotalNo. of patients33 126 20179Mediastinal mass17 1 018CNS5 2 18Change of diagnosis0