Cytogenetic profile of patients with acute myeloid leukemia and central nervous system disease

Abstract

BACKGROUND:

Acute myeloid leukemia (AML) infrequently involves the central nervous system (CNS). This study was undertaken in patients with AML to determine whether cytogenetic findings predict CNS involvement.

METHODS:

The medical records of 1354 patients with AML who were treated at The University of Texas MD Anderson Cancer Center between January 2000 and December 2008 were reviewed. Forty patients (3%) had CNS involvement at time of presentation or disease recurrence, of whom 37 had conventional cytogenetics performed on bone marrow aspirate material. Demographics, treatment, and status at last follow-up were collected.

RESULTS:

Eleven patients (30%) had a diploid karyotype, and 14 patients (38%) had complex cytogenetics. Only 5 of the 40 patients had CNS disease at diagnosis, and the remaining patients had CNS disease at relapse. Patients who developed CNS disease were younger (P = .019), had a higher white blood cell (WBC) count at diagnosis (P = .001), had higher lactate dehydrogenase level (LDH) levels (P < .0001), and had higher percentages peripheral blast cells (P = .024) at diagnosis compared with the rest of the population. In addition, patients with CNS disease had higher rates of chromosome 16 inversion (P < .001), chromosome 11 abnormality (P = .005), and trisomy 8 (P = .02) and had a tendency toward complex cytogenetics (P = .2) compared with the control group (patients who had AML with no CNS involvement).

CONCLUSIONS:

Patients with AML and CNS disease often had higher LDH levels and WBC counts at diagnosis, and they often presented with chromosome 16 inversion and chromosome 11 abnormalities. The current study indicated that the overall survival of patients with AML who had CNS involvement is poor. Cancer 2012;. © 2011 American Cancer Society.

Conventional cytogenetic analysis of acute leukemia is a mainstay of the diagnostic workup. Cytogenetic results have been integrated, in part, into the current World Health Organization (WHO) classification of both acute myeloid leukemia (AML) and acute lymphoblastic leukemia (ALL).1 Three important multicenter clinical trials by Cancer and Leukemia Group B (CALGB), the United Kingdom Medical Research Council (MRC), and the Southwest Oncology Group (SWOG)2-5 have demonstrated the importance of an initial cytogenetic assessment on the outcome of patients with AML. Cytogenetic abnormalities commonly are being divided into 3 prognostic groups: favorable, intermediate, and unfavorable. Treatment regimens for patients with AML are tailored, in part, on the basis of risk factors, including cytogenetic data.

The risk of central nervous system (CNS) involvement in patients with acute leukemia is far greater in patients with ALL than it is in patients with AML. Consequently, cytogenetic data have been correlated with CNS disease in patients with ALL,6-8 but there is a paucity of literature on patients with AML. Conversely, clinical presentation and treatment of CNS disease associated with AML have been addressed in the literature.9, 10

In this study, our objective was to identify patients with AML who had CNS involvement and to review clinicopathologic findings with a particular emphasis on the results of conventional cytogenetic analysis. We also compared this group with a large group of patients who presented with AML without CNS involvement.

MATERIALS AND METHODS

The Institutional Review Board of The University of Texas MD Anderson Cancer Center approved this study. We reviewed the medical records of 1354 patients with AML who were referred to The University of Texas MD Anderson Cancer Center between January 1, 2000, and December 31, 2008. Cytogenetic data were recorded at diagnosis for all 1354 patients. We recorded the basic laboratory evaluation, response to therapy, and outcome of the whole group. The following parameters were specifically collected from the medical record: cytogenetic profile, age, sex, white blood cell (WBC) count at presentation, therapy delivered, and response to therapy. From this group of 1354 patients, we identified 40 patients with AML who also had CNS involvement at the time of either presentation or relapse. These 40 patients are the focus of this study. For those patients who had CNS relapse, we identified the exact site of involvement, the method of diagnosis, and the therapy delivered. Cytogenetic abnormalities were subdivided into 3 groups—diploid, low risk, and complex—according to the WHO classification, and we noted the specific presence of −5 or −7, t(15;17), +8, inversion of chromosome 16 (inv[16]), t(8;21), trisomy 8, and chromosome 11 abnormalities.

CNS disease was confirmed by either the demonstration of blast cells within the cerebrospinal fluid according to criteria set for European Organization for Research and Treatment of Cancer (EORTC) Study 5888111 or by magnetic resonance imaging (MRI) results strongly suggestive of leptomeningeal spread. In 11 of 25 patients (44%) who had CNS disease demonstrated by blast cells in the cerebrospinal fluid, circulating blasts were present in the peripheral blood at the time of lumbar puncture. Nine of those patients also had MRI results strongly suggestive of leptomeningeal disease, which helped to exclude contamination of the CNS from peripheral blood. Conventional cytogenetic analysis of the bone marrow was performed as described previously.12 The cytogenetic results were reported according to the International System for Human Cytogenetic Nomenclature.13 Cytogenetic analysis was not performed consistently on patients at relapse. Cytogenetic abnormalities were grouped according to published WHO criteria into complex, intermediate/diploid, and low risk. First, we identified the presence of complex cytogenetics (defined as ≥3 abnormalities), and then we subgrouped them according to the presence of known, specific cytogenetic abnormalities, such as −5 or −7, inv(16), t(15;17), chromosome 11 abnormality, t(8;21), and trisomy 8.

For comparison, patients who had AML with CNS involvement were compared with 1314 adult patients who had AML without CNS disease who were treated at our institution from 2000 through the end of 2008. A chi-square test and regression logistic analysis were used to analyze statistically significant differences in categorical variables (cytogenetic abnormalities). Statistical analysis was performed using SPSS software (version 19.0; SPSS Inc., Chicago, Ill).

RESULTS

The patient characteristics are summarized in Table 1. Of the 40 patients studied, 5 patients (12.5%) had CNS disease at the time of diagnosis, and 35 patients (87.5%) developed CNS disease at a later stage (either isolated CNS relapse or CNS relapse with bone marrow relapse).

Table 1. Patient Characteristics

Characteristic	AML Patients Without CNS Disease, n=1314	AML Patients With CNS Disease, n=40	Pa
Abbreviations: AML, acute myeloid leukemia; CNS, central nervous system; CR, complete response; LDH, lactate dehydrogenase; NS, nonsignificant; WBC, white blood cells. a P values were determined with the chi-square test.
Sex, no.
Men	752	21	NS
Women	562	19
Mean age, y	60	48	.019
Mean WBC, ×10^3/μL	18.81	36.54	.001
Mean platelets, ×10^3/μL	73.09	77.98	.16
Mean hemoglobin, g/dL	8.6	8.6	.45
Mean peripheral blood blasts, %	23.62	39.16	.024
Mean bone marrow blasts, %	44.39	55.47	.47
Mean LDH, IU/L	1322	2423	<.0001
Patients who achieved a CR after induction, %	51	72	.054
Patients with primary refractory disease, %	26	20	.7

CNS presentation (Table 2) consisted of headaches in 18 patients, cranial nerve palsies in 21 patients, and motor deficit in 15 patients. Of 40 patients, 25 (62.5%) had CNS disease diagnosed by lumbar puncture, 36 (90%) had CNS disease diagnosed by MRI, and 21 had disease diagnosed by both CSF examination and MRI. MRI findings consisted of diffuse leptomeningeal enhancement in 12 patients, nonspecific enhancement in 17 patients, mass in 6 patients, and hydrocephalus in 1 patient.

Table 2. Central Nervous System Disease Characteristics

Characteristic	No. of Patients (%)
Abbreviations: CNS, central nervous system.
Concomitant bone marrow disease at the time of CNS disease diagnosis	26 (65)
CNS disease at time of initial leukemia diagnosis	5 (12.5)
CNS disease diagnosed by lumbar puncture	25 (62.5)
CNS disease diagnosed by magnetic resonance imaging	36 (90)

Patients who developed CNS disease presented with a higher WBC count at diagnosis (P = .001), a higher lactate dehydrogenase (LDH) level (P < .0001), a higher percentage of peripheral blood blasts (P = .024), and younger age (P = .019) compared with the 1314 patients who did not develop CNS disease. There was no significant difference between the 2 groups with regard to hemoglobin level, platelet count, percentage of marrow blasts, or sex.

Thirty-seven of the 40 patients (92.5%) had bone marrow cytogenetics performed. Their cytogenetic profiles were compared with the profiles from 1314 adults who had AML without CNS disease who were treated at our institution: Those results are summarized in Table 3. Patients with CNS disease were more likely to have inv(16) (16% compared with 4%; P < .001), chromosome 11 abnormality (12.5% compared with 3.7%; P = .005), and trisomy 8 (12% compared with 7%; P = .02) and had a tendency toward complex cytogenetics (38% compared with 26%; P = .2).

Table 3. Comparison of Cytogenetic Profiles Between Patients With and Without Central Nervous System Disease

	No. of Patients (%)
Characteristic	Patients With CNS Disease in the Current Study, n=37	AML Patients Without CNS Disease, n=1314	Pa
Abbreviations: AML, acute myeloid leukemia; CNS, central nervous system; cyto: cytogenetics; Inv, inversion. a P values were determined with the chi-square test.
Cytogenetics
Normal	11 (30)	466 (36)	.331
Complex: ≥3 abnormalities	14 (38)	348 (26)	.231
Simple	15 (37)	500 (38)	.994
−5 or −7	3 (8)	301 (23)	.021
Simple cyto	0 (0)	201 (15)	.007
Complex cyto	3 (8)	100 (8)	.979
t(15;17)	2 (5)	66 (5)	.995
Simple cyto	2 (5)	55 (4)	.801
Complex cyto	0 (0)	11(1)	.561
+8	5 (12)	94 (7)	.201
Simple cyto	2 (5)	66 (5)	.995
Complex cyto	3 (7)	28 (2)	.025
Inv(16)	6 (16)	47 (4)	<.001
Simple cyto	6 (16)	40 (3)	<.001
Complex cyto	0 (0)	7 (1)	.643
t(8;21)	1 (3)	36 (3.5)	.7902
Simple cyto	0 (0)	35 (2.6)	.296
Complex cyto	1 (3)	1 (0.9)	<.001
11Q	5 (12.5)	49 (3.7)	.005
Simple cyto	3 (7)	30 (2.3)	.035
Complex cyto	2 (5.5)	19 (1.4)	.035

The median time from achieving a complete response (CR) to developing CNS disease was 14 months (range, 2-72 months). Responses to initial treatment were as follows: 29 of 40 patients who developed CNS disease achieved a CR (73%) compared with 776 of 1166 patients (67%) who did not develop CNS disease (P = .054), and 11 of 40 patients (28%) who developed CNS disease were refractory to treatment compared with 368 of 1166 patients (32%) who did not develop CNS disease (P = .7).

It is worth noting that 34 patients had treatment failure shortly after achieving a CR, and these patients were not included in the CR group. We also omitted from the response analysis 136 patients who were diagnosed at The University of Texas MD Anderson Cancer Center but who did not receive therapy at our institution. We also examined responses to initial therapy according to cytogenetics: One hundred twenty-four of 271 patients (46%) with complex cytogenetics and without CNS disease achieved a CR compared with 10 of 14 patients with CNS disease (71% 3 of 10 patients had inv[16], 1 of 10 patients had t[15;17], and 4 of 10 had trisomy 8). In the CNS disease group, complete remission after initial therapy was achieved in 6 of 6 patients with inv(16), in 1 of 2 patients with t(15;17), in 1 of 1 patient with t(8;21), and in 4 of 5 patients with trisomy 8.

A better correlation with response was observed when both the specific cytogenetics and the level of complexity were taken into consideration. For instance, among the patients with −5 or −7 abnormality, 102 of 256 patients (40%) were refractory to therapy, including 67 patients (66%) who had complex cytogenetics. In the group with CNS disease, 3 patients had −5 or − 7 and complex cytogenetics, and those 3 patients did not respond to therapy. Conversely, patients with inv(16), trisomy 8, t(15;17), and t(8;21) tended to be associated with simple cytogenetic arrangements (44 of 51 patients [86%], 70 of 99 patients [71%], 57 of 68 patients [84%], and 36 of 37 patients [97%], respectively). If they were associated with a complex arrangement, then it did not negatively affect their response to therapy, and the majority of refractory patients with inv(16), trisomy 8, t(15;17), and t(8;21) presented with simple cytogenetics arrangements (≤2 abnormalities).

Table 4 summarizes the treatment regimens used for CNS disease and their outcome. Twenty-eight of 40 patients (70%) received intrathecal chemotherapy, and 23 patients (57.5%) received systemic chemotherapy. Thirty-five patients (87%) received either whole-brain/base of skull or craniospinal radiation. Most patients received a combination of 1 of the other treatments. Only 5 of 40 patients with CNS disease survived for ≥18 months. Three of those patients had isolated CNS relapse, 1 patient had CNS disease at the time of the initial AML diagnosis, and 1 patient had disease recurrence in bone marrow and the CNS. All 5 patients received intrathecal chemotherapy, 4 underwent stem cell transplantation, and 4 received radiation to the brain or skull base.

Table 4. Details Treatment Received by 40 Patients With Central Nervous System Disease

Treatment for CNS Disease	No. of Patients (%)	No. of Patients Who Cleared CSF and/or MRI Findings of Leptomeningeal Disease
Abbreviations: CSF, cerebrospinal fluid; CNS, central nervous system; MRI, magnetic resonance imaging.
Intrathecal chemotherapy	28 (70)	14
Without CNS radiation	7 (17)	2
With CNS radiation	21 (83)	12
Brain or base of skull radiation	25 (62)	12
Craniospinal radiation	10 (25)	5
Systemic chemotherapy (any)	23 (57)	11
Systemic cytarabine	17 (42)	9

The median overall survival for the entire group of 40 patients with CNS disease was 3.17 months (range, 0.06-108 months). No particular cytogenetic abnormality predicted overall survival. Five patients survived for ≥18 months. Three of those patients had an isolated CNS relapse, 1 patient had CNS disease at the time of initial AML diagnosis, and 1 patient had a relapse in bone marrow and the CNS. All 5 of those patients received intrathecal chemotherapy, 4 underwent stem cell transplantation, and 4 received radiation to the brain or skull base.

DISCUSSION

Conventional cytogenetic analysis is an important pretreatment prognostic factor for patients with AML.2, 14, 15 However, the literature on cytogenetic abnormalities in patients with AML and CNS involvement is scarce, most likely because CNS involvement is uncommon in patients with AML. The frequency of CNS disease in the patients with AML in our current study supports the literature on this point. Over the course of 8 years at our institution, during which 1354 patients with AML were treated at our hospital, only 40 patients (3%) had CNS disease. Thirty-seven of those patients had conventional cytogenetic analysis performed on bone marrow aspirate material. We also had access to data from a large number of patients who had AML without CNS disease, most of whom also had cytogenetic data, thereby allowing us to correlate cytogenetic findings with the risk of CNS disease as well as with overall survival.

We observed that patients who had AML with CNS disease had a higher frequency of inv(16), chromosome 11 abnormalities, trisomy 8, and complex cytogenetics compared with patients who had AML without CNS disease. The association of these cytogenetic abnormalities observed in our study also has been observed and reported by several other investigators. Holmes et al,16 in a retrospective analysis of patients with acute myelomonocytic leukemia who had inv(16) (p13q22) and who attended The University of Texas MD Anderson Cancer Center between 1976 and 1983, reported an incidence of CNS disease of 35% (9 of 26 patients). The median time to develop CNS disease from first CR was 19 months. All 9 patients had CNS disease at relapse, which is in accordance with our finding. Conversely, other investigators reported a lower incidence of ≤17%, which was attributed to the benefit of using high-dose cytosine arabinoside as a prophylactic CNS therapy.17-19 In 1 study by the Children's Cancer Group, it was demonstrated that patients with chromosome 11 abnormalities were more likely to develop CNS disease.20 In our study, 12.5% of patients with CNS disease had chromosome 11 abnormalities compared with 3.7% of patients without CNS disease (P = .005). In another study, Chang and colleagues evaluated 101 patients who had AML and extramedullary infiltrates.21 Patients with chromosome 11 abnormalities, specifically, 11q23, had a higher rate of extramedullary disease (69.2% with CNS disease vs 26.5% without CNS disease), and this difference was statistically significant. When the site of extramedullary disease was analyzed, however, 11q23 abnormalities were associated significantly with lymphadenopathy and gum hypertrophy, but not with CNS disease.

Our study demonstrated that complex cytogenetics tends to be more common in patients who developed CNS disease (38% vs 26%; P = .2). Multiple studies have reported a worse outcome for patients with AML who have with complex cytogenetics.2, 3, 22, 23

One interesting finding of our study was the statistically significant increased risk of CNS disease in patients who had high WBC counts (P = .001) and LDH levels (P < .0001) at diagnosis. These patients also tended to be younger and had higher peripheral blast cell levels at diagnosis. A high WBC count has been associated with adverse survival in patients with ALL,8, 24, 25 but few studies have examined the initial characteristics of patients with AML. Our finding is in accordance with several published studies,23, 26 suggesting that a high WBC count at diagnosis (>100 × 10⁹/L) is associated with a worse disease-free survival.

In summary, patients with AML and CNS disease often have a high WBC count, a high LDH level, inv(16), and chromosome 11 abnormality features at the time of diagnosis. Our study can serve as a descriptive study of what may predispose patients with AML to develop CNS disease. Confirmation of our results by other groups is important. At this time and in view of the low incidence of CNS disease (3%), it would not be reasonable to make any therapeutic changes to patients who present with these risk factors, but such patients may warrant close evaluation and follow-up for CNS disease.

FUNDING SOURCES

No specific funding was disclosed.

CONFLICT OF INTEREST DISCLOSURES

The authors made no disclosures.