Document JJDxd33Vo4xgj0mXr5pDBDO6O

-10 -u- "YVp9L.a 4lbIstcalLkruJLoansectt0n llrorathed d o n oitht Llkrp7 et theHealth Sduunn UnhadtJdltm*luMedlcalCratv Morphologic and Cytochemical Observations on the Overt Leukemic Phase of Therapy-related Leukemia JAMES W. VARDIMAN, M.D., ANA COELHO, M.D.. HARVEY M. GOLOMB, M.D., AND JANET ROWLEY, M.D. comparison of morphologic and cytochemical features of the leukemic cells of patients with overt acute nonlymphocytic leukemia following therapy for primary malignancies (2O-ANLL) 10 -!me of patients with A N L L de novo revealed several differnces. N O Auer rods were seen in cells of patients with 2"- ANLL, but they were seen in 54 (42%) of 129 cases of A N L L de novo. Only one (4.7%) of 2 1 patients with 2"-ANLL was classified as M-4 by FAB criteria, compared to 46 (36%) of 129 patients with A N L L de novo classified as M-4. Only one (10%) of 10 patients with 2"-ANLL had more than 10% peroxidase-positive leukemic cells, whereas 19 (95%) of 20 of a control group of patients with A N L L de novo of similar FAB subtypes had more than 10% positive cells. Two of 11 (18%) 2O-ANLL patients had 10% or more naphthol-AS-D chloroacetate esterase-positive cells, in contrast to nine (47%)of 19 patients with A N L L de novo. Nonspecific esterase activity IPS ?resent in 10% or more of cells in only one (10%) of 10 patients with 2"-ANLL studied, as compared to seven ( 3 5 6 ) of 20 of the control group. Previous reports indicate that pa- tients with 2O-ANLL have cytogenetic abnormalities and therapeutic responses that differ from those in ANLL de novo; our findings indicate that there are morphologic and cytochemical differences as well. (Key words: 2O-ANLL; Therapy-related IeukemiK Cytochemistry) A m J Clin Pathol 1983; 79: 525530 THERAPY-RELATED or secondary acute nonlymphxytic leukemia (2O-ANLL) and dysmyelopoietic syndromes (DMPS) are being recognized with increasing frequency as leukemic processes with characteristic morphologic, clinical, and cytogenetic features.1.3.8.10.12.18.20.23.25.26.29-31 Frequently, in patients with 2"-ANLL, a preleukemic phase with hyperplasia and dysplasia in the myeloid, megakaryocytic, and erythroid series evolves into overt leukemia that is refractory to treatment. Most patients with leukemia following cytotoxic therapy for malignant or nonmalignant disease hax2 nonrandom chromosome abnormalities. usually consisting of loss of part or all of chrornosomc number 5 or number 7 Or.both.1.12.18."-26 In our morphologic and cytochemical analysis of 26 Received July IS, 1982; received revised manuscript and accepted publication November 8, 1982. Presented at the meeting of the American Society of Clinical Pathologists. March 1982, New Orleans, Louisiana. SWpoRed in part bya grant from the Geometric Data COrpontion. .+Ldress reprint requests to Dr. Vardiman: Department of Pathol: OS'. University of,Chicago Hospitals and Clinics, 950 E. 59th Street. 414. Chicago. Illinois 60637. Departments of Pathology and Medicine, The University of Chicago Hospitals and Clinics, Chicago, Illinois patients who developed 2"-ANLL or DMPS following treatment for malignant disorders, we noted considerable differences from a group of 129patients with ANLL de novo. The differencesseen during the overt leukemic phase form the basis of this report. Materials and Methods At the University of Chicago Hospitals and Clinics, 150 cases of ANLL have been studied from 1970 through July 1981. Of these, 129 had ANLL de novo, and the other 2 1 cases were considered to represent overt 3"-ANLL, Le., leukemias that occurred following chemotherapy or radiotherapy or both, for malignant diseases. Five additional patients who developed a dysmyelopoietic syndrome following therapy for a malignant disease, but who did not develop overt acute leukemia, are not included in this series. All available peripheral blood smears, bone marrow aspirates, and bone marrow biopsy specimens for the entire group of ANLL patients were reviewed, and 500 cell count differentials were performed on the bone marrow aspirates. Each case of acute leukemia was clas- sified according to the FAB riter ria;'-^.^^ only those cases in which the numbers of blasts in the marrow exceeded 30% of the nucleated marrow cells are included in this report. Since 1975, cytochemical reactions, including 'peroxidase, naphthol AS-D chloroacetate esterase, a-naphthy1 butyrate, and/or a-naphthyl acetate with or without NaF, have been applied to 83 of the 129 cases of ANLL de novo. Standard cytochemical procedures were sed."*'^*'^ In the group of 21 patients with 2"-ANLL, 13 patients had cytochemical studies recent enough to be analyzable for the purposes of this report, and they were compared with a control group of 20 patients with ANLL dc novo who were in similar F A B subgroups and who had cytochemical procedures performed within the =me period. At least 100 bone marrow cells were analyzed for each cytochemical reaction. The number of immature cells ,. 0002-9 173/83/0500/0525$0 I . IO 0 American Society of Clinical Pathologhts 525 NOTICE: This materialmay be protected b) cmvrirht law (Titte 17. United States Code).. 526 - VARDIMAN ET..A!.. -AJ.C.P. May 1981 demonstrating positive reactions was counted, and the intensity of the reaction was quantitated on a scale of I + to 4+ positivity. A cell was scored as 1+ for peroxidase if a definite, single peroxidase-positive granule was present in the cytoplasm; 2+ if more than one granule was present, but less than one-fourth of the cytoplasm was occupied by positive granuJes; 3+ if onefourth to one-half of the cytoplasm. showed reaction product; and 4+ if the reaction was intense and positive granules were present throughout the cytopla5m. Cells that were considered to be immature leukemic cells were tallied; the degree of chromatin condensation was used as the primary criterion of immaturity. Nevertheless, the tally probably reflects not only blasts but also some promyelocytes, since these two cell types cannot always be distinguished in the cytochemical procedures. Reactivity in neutrophils was tallied separately, but in some cases, neutrophils were so scarce that a meaningful interpretation was not possible. A similar scoring procedure was used for the naphthol AS-D chloroacetate esterase (NASDCA) and acid phosphatase (AcP) procedures. Although both a-naphthyl butyrate and a-naphthyl acetate esterases with and without NaF reactions were available for most patients, more 2"ANLL patients had been studied with the butyrate substrate; we chose, therefore, that reaction to assess the nonspecific esterase (NSE). (In our laboratory, a-naphthy1 butyrate and a-naphthyl acetate give similar results.) Primitive cells demonstrating NSE activity were scored according to the intensity of the reaction that was noted diffusely in the cells. The number of immature cells showing NSE activity, as well as the total number of positive cells, was scored. No case was considered to represent acute myelomonocytic leukemia (M-4) unless there were more than 20%monocytoid cells in the blood and/or bone marrow, and unless this finding was confirmed by a similar number of nonspecific esterase-positive cells. The periodic-acid-Schiff reaction was scored in both the immature leukemic cells and the erythroid precursors. Transmission electron microscopy (TEM) was used for the study of cases in which the nonlymphocytic nature of the leukemic process was in doubt because of poor cytochemical reactivity of the neoplastic cells. U1trastructural cytochemical studies were not performed routinely. Results Clinical The clinical, morphologic, and cytogenetic findings in all except four of the 2 1 patients with 2"-ANLL have been detailed el~ewhere.'~**~B*ri'e~fly, 14 had been treated previously for Hodgkin's disease (HD), three for small cleaved follicular center cell lymphoma (PDLI, two for multiple myeloma (MM), and two for carcinoma. Sixteen patients had received both radiotherapy and chemotherapy; three, radiotherapy alone; and two, chemotherapy alone. The mean age at the onset of leukemia was 43 years. A preleukemic period was documented in 13 (62%)of the patients. Examination of the peripheral blood and bone marrow smears and biopsy specimens of patients during the preleukemic period demonstrated many features in common with a myelodysplastic syndrome, very similar to the features of refractory anemia with excess blasts (RAEB), or RAEB in transformation, as described in 'the FAB classificat i ~ n . ~ .T~h.e' n~umbers of immature cells and the degree of dysplasia increased as each patient evolved toward an overt leukemic phase, defined as more than 30% blasts in the bone marrow. In the 129 patients with ANLL de novo, the mean age at the time of diagnosis was 48 years. A preleukemic phase was documented in very few patients, if any. Classification of Lertkania Of the 129 patients with ANLL de novo, 28 (22%) were classified as M-l by the FAB criteria, 34 (26%)as M-2, and 10 (8%)as M-3. Acute myelomonocytic leukemia (M-4) was diagnosed in 46 (36%)ofthese patients, whereas M-5 was found in five (4%). Erythroleukemia (M-6) was present in five (4%) patients (Table 1). One case was unclassifiable because of extensive marrc.;..:n' ecrosis. The mean blast count in these patients was 54%. with a range of 30-90%. The mean marrow cellularity was 88%, with a range of 30-100%. as evaluated from marrow biopsy specimens. Classification of the cases of 2"-ANLL according to the FAB criteria was difficult in approximately 70% of the cases. These cases showed atypical features, such 3s clear evidence of trilineage involvement (70%), man! micromegakaryocytes and/or evidence of megaksVQcytic proliferation (70%), and moderate to markeLi).increased marrow reticulin (47%). Additionally, the aPpearance of very primitive blasts in a background Of dysplastic myeloid and erythroid elements often gave the impression that the leukemia might later be characterized by primitive, poorly differentiated cells if il were allowed to run its course. The mean marrow ceilularity was 87%, as judged from bone marrow biops!'. with a range of 50-100%.The mean blast count at the time of overt leukemia was 53%, with a range of jo95%, for those patients in whom countable aspirata were available. It frequently was difficult to decide \'hen the disease had evolved from a myelodysplastic St3IC Vd. 19 * YInto c *8u; sed : ?ow fi %titatic ~+ N L tta(-1s9M%%M class1 hazurc ever, clud 'eigh .stud ing I or h brot relia beer one ofA the Of M- I I: 10 r 10% oft had rem im~ by. in Pos diff in 1 THERAPY-RELATED LEUKEMIA 527 into overt leukemia. Although the number of blasts was used as the basis for this decision, the presence of mar- row fibrosis in some cases did not permit accurate quantit2:ion of the blasts. Of the 2 1 patients who developed ~h i L secondary to therapy, four ( 19%)were classified 35 M-1; 11 (52%)were M-2, one (5%) was M-3; and four (19%) were M-6. Only one (5%) patient was classified 3s M-4, and no cases of acute monocytic leukemia (Mj) were detected (Table 1). Auer rods were present in 54 (42%)of the 219 patients with ANLL de novo but were not seen in any of our patients with 2O-ANLL. In the single case of 2O-ANLL classified as M-3, more than 50% of the cells had heavy az.xophilicgranulation, and some had the bi-lobed, renifom nuclei typical of promyelocytic leukemia; however, Auer rods were not detected on the marrow aspirate smears. Cytochemical Findings Only the findings in 13 2O-ANLL patients are included in this report, since the procedures on the other eight patients either had been performed before this study began and were considered unreliable for the scoring method used because the reaction product had faded, or had been performed on touch preparations from fibrotic marrows, and there were too few cells to score reliably. Of the 13 patients with 2"-ANLL, four had been classified as M- 1 , seven as M-2. one as M-4, and one as M-6. Twenty patients with similar classifications of ANLL de novo were studied by cytochemistry during the same period, and these served as a control group. Of :he 20 patients with ANLL de novo, three were M- 1, nine were M-2, six were M-4, and two were M-6. In the group of 2O-ANLL patients. only one (10%) of IO patients studied for peroxidase activity had more than 10% peroxidase-positive immature leukemic cells (64% of the cells were positive), and seven of the 10 patients had only 3-970 peroxidase-positive cells (Table 2). The remaining two patients had only 2% peroxidase-positive immature cells, and the diagnosis of M- 1 was confirmed by ?EM studies. In contrast, 19 (95%)of the 20 patients in the control group had more than 10% peroxidasePositive immature leukemic cells (Fig. 1). No significant differencebetween the two groups was found, however, in the percentages of peroxidase-positive or peroxidasenegative neutrophils. The intensity of the peroxidase reaction also differed between the 2O-ANLL patients and the control group. Only one (10%)of 10 of the patients with 2"-ANLL had any immature cells scored as 3+. Although eight of the 10 had some cells that were scored as high as 2+, more t h m 50%of the peroxidase-positive cells in five patients W e only l+. In contrast, 14 of the 20 patients with Table 1. Distribution of Patients in the Present Study According to the FAB Classification ANLL de novo 2O-ANLL Type Number of Patients z Number of Patienu 8 M- I M-2 M-3 M-4 M-5 M-6 Unclassified 28 22 34 26 10 8 46 36 54 54 I 4 11 1 -1 4 19 52 5 -5 19 Total Auer rods I29 54 42 21 0 0 ANLL de novo had peroxidase-positive immature cells that were scored as 3+ or higher, and in 11 cases, more than 50% of the immature cells were scored as 2+ or higher. The naphthol AS-D chloroacetate esterase reaction was positive in more than 10% of the immature leukemic cells in two ( 18%)of 1 1 2O-ANLL patients (Table 2). In one of these patients classified as M-2, 26% of the cells were positive with the NASDCA reaction, and 64% were positive with the peroxidase reaction. In the second patient, there were 28% NASDCA-positive cells, but only 9% peroxidase-positive cells. In seven of the remaining 2"-ANLL patients, some positive cells were present, but the NASDCA reactions were very weak, and all positive cells in each ofthese patients were scored as only l + . In the control group, nine (47%)of 19 patients studied with NASDCA had more than 10%positive immature cells, and 16 (84%)of these had at least some cells that were scored higher than 2+. A differencein the number of peroxidase-positivecells and NASDCA-positive cells was noted in both groups of patients. One patient with 2O-ANLL and two patients in the control group had more positive cells with NASDCA than with peroxidase; in general, however, in both groups of patients, far fewer cells were positive with the NASDCA than were positive with the peroxidase Table 2. Results of Cytochemical Studies 2O-ANLL Control Group with ANLL de novo > IO% Peroxidase-positive cells > IO8 Chloroacetate-esterasepositive cells >io% Nonspecific esterase-positive cells > 10% Acid-phosphatase-positive cells 1/10 (10%) 2/1 I (18%) 1/13 (8%) 8 / 8 ( 100%) 19/20(95%) 9/19 (47%) 6/20 (3046) 20/20 (100%) 528 VARDIMAN 7 AL. -AJ.C.P. May 1983 FIG. 1 (IeJ). The cytochemical differences that were typical between 2"-ANLL and ANLL d e novo are illustrated in this photomicrograph. Peroxidase-positive blasts are readily identified in the case of ANLL d e novo (Fig. IA, /ar Icfi),whereas the blasts from a patient with 2O-ANLL (Fig. IE. IeJ, cenfer) show no reactivity. (A neutrophil serves as a positive control.) Peroxidase (X1,ooO). FIG. 2 (righf).In this study, it was generally noted that there was a greater sensitivity of the peroxidase reaction than the NASDCA reaction in leukemic cells. This figure compares the peroxidase reaction (Fig. L4.righf. cenfer)with the NASDCA reaction (Fig. 2 E. far righf)from the same patient with M-I ANLL de novo. Most of the immature cells show some peroxidase activity. but only a more mature granulocytic cell shows NASDCA activity. Peroxidase and NASDCA (X1,OOO). reaction. Fifty-four per cent of the patients in both groups had percentages of peroxidase-positivecells more than twice the percentage of NASDCA-positive cells (Fig. 2). Only one 2"-ANLL patient had more than 10%positive cells with NSE (Table 2); 46% of the total nucleated marrow cells were found to be positive, and the morphologic classification was M-4. In the control group, all six of the patients classified as M-4 had more than 20% NSE-positive cells, and none of the remaining patients had more than 10%positive cells. Diffusely distributed acid phosphatase activity was present in all eight patients with 2"-ANLL who were studied for that reaction, and in aI1 20 patients in the control group. In five (62%) of the cases of 2"-ANLL, more than 50% AcP-positive cells were tallied, whereas 16 (80%)of the patients in the control group had more than 50% AcP-positive cells. In general, the intensity of the reaction also was stronger in the control group. In 60% of patients with 2"-ANLL, 2-8% of cells interpreted as nonerythroid blasts were shown to have PAS positivity in the form of fine granules; similar activity was present in 7 1% of the control group. However, PAS activity in from 2-8% of erythroid cells was seen in 50% of the patients with therapy-related leukemia, but only 29% of the patients with ANLL de nuvo showed similar reactivity. Discussion The myelodysplastic and leukemic process following chemotherapy and/or radiotherapy has been character- ized as a disorder involving the hematopoietic stem cells. with ensuing abnormalities in the myeloid, monocytic, erythroid, and megakaryocytic series.'0*'2~20*R23ec~e3n0t evidence indicates that, in some cases, lymphoid cells may participate in the leukemic process as well.24More than half of the patients who develop 2"-ANLL have a preleukemic phase of variable duration that is characterized by cytopenia, macrocytic anemia, and severe dysplasia in the cells of the peripheral blood and bone marrow. Morphologically,these changes resemble R --.EB. or RAEB in transformati~n.'"~*C'h~romosome abnormalities, most often consisting of hypoploidy, with abnormalities frequently involving chromosome number 5 or number 7 or both, are present almost in~ariably.~~.D'~u.r*i~ng the preleukemic phase, there is often a gradual increase in the number of immature cells until an overt leukemia emerges. Comparison of the morphologic and cytochernic31 features of the overt leukemic phase of 2"-ANLL n i t h those of ANLL de novo revealed several differenxs in the patients studied at our institution. Patients with 2"ANLL often were more difficult to classify by FAB cnteria, and 70% showed evidence of considerable t d i n eage involvement. There were fewer patients with therapy-related disease, however, who showed significan? participation of monocytes in the leukemic process. and Auer rods appeared to be uncommon. In addition, the numbers of leukemic cells that demonstrated cytochemical reactivity for myeloid differentiation enzymes. .Wh as myeloperoxidase and naphthol AS-D chloroacs[ate esterase, were fewer in 2"-ANLL than in ANLL de wl'** In several recent reports, the number of patients with enc I in of PZ ot in 01 1; vat. 19 NO. 5 THERAPY-RELATED LEUKEMIA 529 2O-ANLL who had either M-4 or M-5 leukemia ap- may have some bearing on the morphologic expression peared to be fewer than expected,12.20*22,2a3l*th2o7ugh of the process recently has been highlighted by Rowley Griinwald and Rosner state that M-4 is the most com- and associates2' in a report that relates the morphologic m n type of 2"-ANLL following therapy for Hodgkins type of ANLL to the karyotypic changes in 503 patients disease.I6 In a detailed morphologic report by Foucar with ANLL de novo. In that study, patients who were and associates,I2two of 15 patients with 2"-ANLL were lacking a chromosome 5 or 7 or both-the changes most described as having a minor monocytic component in frequently observed in 2"-ANLL-were noted to have their leukemia. McKenna and co-workers2' reported on the M-I, M-2, or M-6 type of leukemia, whereas patients 1 3 2O-ANLL patients who were studied ultrastructur- with M-4 or M-5 rarely showed this karyotypic pattern. ally. In only two cases promonocytes predominated; None of our patients with treatment-related leukemia abnormalities in the monocytic series were noted in ad- had Auer rods, although they were found in 42% of our ditional patients, however, providing evidence for a pan- patients with ANLL de novo, an incidence similar to myelocytic disease process. In a report by Pedersen-Bjer- that reported in the literature for other series of patients g m d and colleague^?^ two of 19 patients with overt 2"- with ANLL de no~o?.'~-~I't*i's~difficult to ascertain ANLL were classified as M-4, according to the FAB how many cases of 2"-ANLL reported in the literature criteria, and none were classified as M-5. Mitelman and were observed to have Auer rods; however, we found Jh. associates2' found that in acute nonlymphocytic leu- 160 cases that were described adequately, and Auer rods LL kemia related to occupational exposure to toxic agents, were mentioned in only 17 (1 1%). 1 monocytic varieties of ANLL were found in only 17.4% Of our larger group of 83 patients with ANLL de novo on he of exposed patients, compared to 42.4% of nonexposed studied by cytochemistry, 7% have shown fewer than ell patients. Likewise, the incidence of karyotypic abnor- 10%peroxidase-positive leukemic cells, and some have malities, including missing chromosomes 5 and 7, was required additional studies, such as electron microscopy, hisher in the exposed than in the nonexposed group in for confirmation of the nonlymphoid origin of their leu- us, that study and thus somewhat similar to the incidence kemia; these cases have been in either the M-1 or M-5 ic, in patients with 2O-ANLL. An analysis of 60 cases of category. The finding in our patients with 2"-ANLL that :nt overt 2"-ANLL from the Fourth International Work- the numbers of cells that were reactive in the peroxidase :US shop on Chromosomes in Leukemia showed 10 (16%) and naphthol AS-D chloroacetate esterase reactions )re cases of M-4, whereas 179, or 21%, of 860 ANLL de were smaller than in the control group indicates that ve novo cases were classified as M-4,I3although this differ- ( I ) the leukemic cells are of more primitive granulocytic ence is not statistically significant. origin, ( 2 )that they are of diverse origin with light mi- However, a review of 147 reported cases of 2'-ANLL, croscopic features indistinguishable from those of my- in which sufficient data are available for determination eloblasts, or ( 3 )that their enzyme content is abnormal. of the leukemic subtype, showed that 28% of the cases The finding, in 2"-ANLL, of decreased numbers of were classified as myelomonocytic or monocytic types. cells positive for the cytochemical markers usually pres- Most of these cases were not classified according to the ent in myeloid leukemias is not unexpected, since the FAB criteria, however, and very few were studied cy- secondary bone marrow neoplasia represents a pan- tochemically. One notable exception is the study of myelosis, with participation of cells of the erythroid and Kapadia and colleaguest8 in which nine (45%) of 20 megakaryocytic as well as granulocytic series. Evolution cases were classified as M-4. Nevertheless, in that report, of such a process may well result in a leukemia of very only "occasional"leukemic cells were described as pos- primitive cells derived from precursor cells of diverse itive with the nonspecific estefase reaction in patients origin, as well as from stem cells. Sultan2' has described Classified as M-4. four patients with treatment-related leukemia who had 11 would appear, therefore, that careful analysis of a significant numbers of megakaryoblasts in their marrow larger group of patients is needed before we can deter- and peripheral blood, which could be identified only by __. mine whether our observation of a decreased incidence application of the ultrastructural platelet peroxidase ofmonocyticvarieties of leukemia in 2"-ANLL, as com- technic described by Breton-Gorius.6 Additionally, Pared with ANLL de novo, is valid and. whether this McKenna2' described two patients with immature cells observation bears any relationship to cytogenetic findings or is of any significant clinical value. The incidence of myelomonocytic leukemia and monocytic leukemia that proved in ultrastructural studies to be basophils, but which could not be identified as such by light microscopy. Thus, it appears that the immature cells in - In our patients with ANLL de novo is similar or slightly patients with overt 2"-ANLL may be very primitive and @eater than that reported by other ~ ~ o u P s , ~ind. i~- ~ , o' f~diverse origin. Our cytochemical studies support this cating that we do not underdiagnose this type of leu- view. Ultrastructural cytochemistry, which was not used kemia. routinely in our cases, may lead to a better definition That the cytogenetic changes observed in leukemia of the cell types involved. I_ __ -- JJU VARDIMAN ET AL. -AJ.C.P. May 1983 In this study, the peroxidase reaction appeared to be more sensitive than the naphthol A S D chloroacetate esterase reaction in detecting leukemic cells of granulocytic origin. In the control group of 20 patients with ARLL de novoStudied at the same time asour 2"-ANLL patients, only one (5%) patient had fewer than 10%peroxidase-positive immature cells, whereas 53% had fewer than 10%NASDCA-positive cells. The poor cytochemical reactivity of the blasts in our patients with 2"-ANLL is similar to that described in chronic myelogenous leukemia (CML), in which, even in a blast. crisis involfing morphologically granulocytic "cells,-only a small percentage of the blasts may show ..peroxidase activity.9 There are other similaritiesbetween the blast phase ofCML and that of 2"-ANLL, including a preceding myeloproliferative phase, rarity of Auer rods;?he presence of nonrandom chromosome changes, and resistance to therapy. These two diseases probably reflect stem cell abnormalities with similar terminal manifestations. Acknowledgmenis.The authors wish to thank E. Salciunas and C. Groves for their technical assistance, P. Lindberg for typing the manuscript, and E. Land for editing the text. References 1. Andenon RL, Bagby GC, Richert-Boe K, et al: Therapy-related preleukemic syndrome. Cancer 1981;47: 1867-1871 2. Anner RM, Lennon JM, Drewinko B: Prognostic significance of morphologic and cytochemical markers in adult acute leukemia. Am J Clin Pathol 1978; 69:494499 3. Auclerc G, Jacquillat C,Auclerc MF, et al:Post-therapeutic acute leukemia. Cancer 1979; 44:2017-2025 4. Bennett JM, Catovsky D, Daniel M, et al: Proposals for the clas- sification of the acute leukaemias: French-American-British (FAB) co-operative group. Br J Haematol 1976; 33:451358 5. Bennett JM, Catovsky D, Daniel MT, et al: Proposals for the classification of the myelodysplastic syndromes. Br J Haematol 198%5 ~ 1 8 9 - 1 9 9 6. BretonGorius J, Reyes F, Duhamel G, et al: Megakaryoblastic acute leukemia: Identification by the ultrastructural demonstration of platelet peroxidase. Blood 1978; 51:45-60 7. Burns CP, Armitage JO, Frey AL, et al: Analysis of the presenting features of adult acute leukemia: The French-American-British classification. Cancer 198 1; 479460-2469 8. Cadman EC, Capizzi RL, Bertino JR: Acute nonlymphocytic leukemia: A delayed complication of Hodgkins disease therapy: Analysis of 109 cases. Cancer 1977; 40:1280-1296 9. Catovsky D, OBrien M, Cherchi M, et al: Ultrastructural, cytochemical and surface marker analysis of cells during blast crisis of chronic granulocytic leukaemia. Boll 1st Sieroter Milan 1978; 57:344-354 10. Dohy H, Genot JY, Imbert M, et al. Myelodysplasia and leukaemia related to chemotherapy and/or radiotherapy-A haematological study of 13 cases. Clin Lab Haemat 1980 2:111I19 1I . Enck RE, Bauman AW, Bennett JM: Adult acute leukemia: The Rochester (NY)experience. Arch Intern Med 1976; 136:12561261 12. Foucar K. McKenna RW, Bloomheld CD. et al: Therapy-related leukemia: A panmyelosk. Cancer 1979; 43: 1285-1'96 13. Fourth International Workshop on Chromosomes in Leukemia, Cancer Genetics and Cytogenetics: In press 14. Glick AD, Paniker K, Flexner JM, et al: Acute leukemia ofadslts: Ultrastructural. cytochemical and histologic observations in 100 cases. Am J Clin Pathol 1980; 73:459-470 IS. Gralnick HR, Galton DAG, Catovsky D. et al: Classification of acute leukemias. Ann Intern Med 1977; 87:740-753 16. Griinwald HW, Rosner E Acute myeloid leukemia following treatment of Hodgkin's disease: A review. Cancer 1982; 50676-683 17. Janckila Al,Li C. Lam Y et al: The cytochemistry of tartrate resistant acid-phosphatase: Technical considerations. Am J a i n Pathol 1978; 70:45-55 18. Kapadia SB. Krause J R Ellis LD, et ai: Induced acute nonlym. phocytic leukemia following long-term chemotherapy: A Study of 20 cases. Cancer 1980; 45:1315-1321 19. Kaplow LS: Simplified myeloperoxidase stain using benzidine dihydrochloride. Blood 1965; 26:215-219 20. McKenna RW, Parkin JL. Foucar K, et al: Ultrastructural char. acteristics of therapy-related acute nonlymphocytic leukemia: Evidence for a panmyelosis. Cancer I98 I ; 48:725-737 21. Mertelsmann R, Thaler HT, To L, et al: Morphological classification, response to therapy, and survival in 263 adult patients with acute nonlymphoblastic leukemia. Blood 1980; 56:773- 78 I 12. Mitelman F, Brandt L, Nilsson PG: Relation among occupational exposure to potential mutagenic/carcinogenic agents, clinical findings and bone marrow chromosomes in acute nonlymphocytic leukemia. Blood 1978; 52:1229- I237 23. Pedersen-Bjergaard J, Philip P. Monensen BT, et al: Acu~rnonlymphocytic leukemia, preleukemia, and acute myeloproliferative syndrome secondary to treatment of other malignant diseases: Clinical and cytogenetic characteristics and results of in virro culture ofbone marrow and HLA typing. Blood 1981; 57:7 12-723 24. Prentic AG, Smith AG, Brandstock W. Mixed lymphoblasticmyelomonoblastic leukemia in treated Hodgkin's disease. Blood 1980; 56~129-133 25. Rowley JD. Golomb HM, Vardiman J: Nonrandom chromosomal abnormalities in acute nonlymphocytic leukemia in patients treated for Hodgkin's disease and non-Hodgkin Iymphomas. Blood 1977; 50:759-770 26. Rowley JD, Golomb HM, Vardiman JW: Nonrandom ckromcsome abnormalities in acute leukemia and dysmyelopoietic syndromes in patients with previously treated malignant disease. Blood 1981; 58:759-767 27. Rowley JD, Alimena G, Garson OM, et al: A collaborative stud! of the relationship of the morphological type of acute nonlymphocytic leukemia with patient age and karyotype. Blood 198': 59: 1013-1022 28. Sultan C, Deregnaucourt J, Lo YW, et al: Distribution of250 cases of acute myeloid leukaemia (AML) according to the classification and response to therapy. Br J Haematol 1981: 47:545-55 I 29. Sultan C. Sigaux F, Imbert M, et al: Acute myelodyspla'i2. myelofibrosis: A report of eight cases. Br J Haematsi :ssl: 49~11-I6 30. Vardiman JW, Golomb HM, Rowley JD, et ai: Acute nonb'mphocytic leukemia in malignant lymphoma: A morpholoPc study. Cancer 1978; 42:229-242 3I. Weiden PL, Lerner KG, Gerdes A, et al: Pancytopenia and leukemia in Hodgkins disease: Report of 3 cases. Blood 19!': 42571-577 32. Yam LT, Li CY, Crosby WH: Cytochemical identification O.F monocytes and granulocytes. Am J Clin Pathol 1971; 55:251- 290 marrow biopsy; Pathol tigatiol