Leukemia & Lymphoma

ISSN: 1042-8194 (Print) 1029-2403 (Online) Journal homepage: https://www.tandfonline.com/loi/ilal20

Elderly acute lymphoblastic leukemia: a Mayo

Clinic study of 124 patients

Kevin C. Miller, Aref Al-Kali, Mithun V. Shah, William J. Hogan, Michelle A.

Elliott, Kebede H. Begna, Naseema Gangat, Mrinal M. Patnaik, David S.
Viswanatha, Rong He, Patricia T. Greipp, Lisa Z. Sproat, James M. Foran,
Mark R. Litzow & Hassan B. Alkhateeb

To cite this article: Kevin C. Miller, Aref Al-Kali, Mithun V. Shah, William J. Hogan, Michelle A.
Elliott, Kebede H. Begna, Naseema Gangat, Mrinal M. Patnaik, David S. Viswanatha, Rong He,
Patricia T. Greipp, Lisa Z. Sproat, James M. Foran, Mark R. Litzow & Hassan B. Alkhateeb (2018):
Elderly acute lymphoblastic leukemia: a Mayo Clinic study of 124 patients, Leukemia & Lymphoma,
DOI: 10.1080/10428194.2018.1509318

To link to this article: https://doi.org/10.1080/10428194.2018.1509318

Published online: 02 Oct 2018.

Submit your article to this journal

Article views: 144

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=ilal20
LEUKEMIA & LYMPHOMA
https://doi.org/10.1080/10428194.2018.1509318

ORIGINAL ARTICLE: CLINICAL

Elderly acute lymphoblastic leukemia: a Mayo Clinic study of 124 patients

Kevin C. Millera , Aref Al-Kalib , Mithun V. Shahb, William J. Hoganb , Michelle A. Elliottb,
Kebede H. Begnab, Naseema Gangatb, Mrinal M. Patnaikb , David S. Viswanathac, Rong Hec,
Patricia T. Greippd, Lisa Z. Sproate, James M. Foranf, Mark R. Litzowb and Hassan B. Alkhateebb
a
Mayo Clinic School of Medicine, Mayo Clinic, Rochester, MN, USA; bDivision of Hematology, Mayo Clinic, Rochester, MN, USA;
c
Division of Hematopathology, Mayo Clinic, Rochester, MN, USA; dDepartment of Laboratory Medicine and Pathology, Mayo Clinic,
Rochester, MN, USA; eDivision of Hematology and Medical Oncology, Mayo Clinic, Phoenix, AZ, USA; fDivision of Hematology/
Oncology, Mayo Clinic, Jacksonville, FL, USA

ABSTRACT ARTICLE HISTORY

Poor outcomes in elderly acute lymphoblastic leukemia (ALL) are well recognized, but the con- Received 28 April 2018
tributors are ill-defined. We characterized 124 patients
60 years old at our institution. The Revised 18 July 2018
majority (n ¼ 102, 82%) were treated with intensive chemotherapy. Of these, 8/102 (8%) died Accepted 24 July 2018
within the first 100 days; 92/102 (90%) achieved complete remission (CR/CRi). Only 31/124 (25%)
KEYWORDS
patients underwent allogeneic hematopoietic stem cell transplantation. The median overall sur- Elderly; acute lymphoblastic
vival (OS) for the entire cohort was 19.8 months. In a multivariate analysis, ECOG performance leukemia (ALL); survival;
status
2, high white blood cell count, and high lactate dehydrogenase (at time of diagnosis) BCR-ABL1; transplant
negatively influenced OS (p<.01). In a subgroup analysis of the intensive treatment group,
BCR-ABL1þ patients had markedly better OS (hazard ratio 0.3, 95% CI 0.1–0.7; p<.01). In sum-
mary, despite few early deaths and a high CR/CRi rate, elderly ALL continues to have a poor
prognosis, underscoring the need for more effective therapies.

Introduction more, many elderly ALL patients are excluded from

clinical trials or never referred [9,16–18].
Acute lymphoblastic leukemia (ALL) is an aggressive
While this paradigm is intuitive, most insights about
hematologic malignancy associated with children [1].
elderly ALL are undergirded by studies which are lim-
Nevertheless, 4/10 cases of ALL in the United States
ited by low patient numbers or do not reflect current
occur in adults (>20 years old), and patients

60 years old make up roughly one third of adult ALL treatment strategies in the United States [2–7,9,11].
cases [2]. Unfortunately, outcomes in this older popu- This is problematic because the rise of modern inten-
lation are poor [3–11], and progress in the past sive chemotherapy regimens (e.g. hyper-CVAD and
30 years has been minimal [12,13]. A recent study UKALL XII/E2993), tyrosine kinase inhibitors (TKIs) and
using Surveillance, Epidemiology and End Results reduced-intensity alloHSCT, as well as improved sup-
Program-Medicare data demonstrated a 1 and 3 year portive care suggest that the landscape of elderly ALL
overall survival (OS) of 43% and 27%, respectively, for has changed in the 21st century [8,19–22]. For
elderly patients with ALL [14]. example, a study from MD Anderson [20] demon-
Poor survival in elderly ALL is generally thought to strated improved OS in older patients treated with
be the product of compounding adverse risk factors. hyper-CVAD compared to historical controls (5 year OS
Commonly, older patients present with comorbidities 20% versus 9%, p¼.046). The negative prognostic
and/or poor performance status, leaving little room influence of BCR-ABL1-positivity has also been chal-
between risking treatment-related mortality and limit- lenged, likely reflecting the impact of TKI-based com-
ing intensive treatment altogether. The latter, com- bination therapy [22]. Furthermore, the answers to
bined with unfavorable disease biology and low rates many important questions in elderly ALL remain
of allogeneic hematopoietic stem cell transplantation murky, such as the prognostic impact of comorbidities
(alloHSCT) sow the seeds for relapse [9,15]. What is [23,24] and clinical outcomes after alloHSCT [25].

CONTACT Hassan B. Alkhateeb [email protected] Division of Hematology, Department of Medicine, Mayo Clinic, 200 1st St SW,
Rochester, MN 55905, USA
Supplemental data for this article can be accessed here.
ß 2018 Informa UK Limited, trading as Taylor & Francis Group
2 K. C. MILLER ET AL.

There are projected to be almost 1000 patients last follow-up. Survival curves were compared using

60 years old who will be diagnosed with ALL in the the log-rank test. A Cox proportional hazard model
United States in 2018 [2]. Unquestionably, the devel- was used for both univariate and multivariate analyses.
opment of more effective therapeutic strategies for Using a forward selection method, covariates with
this subset of patients continues to be an urgent p<.10 in univariate analysis were included in multivari-
need, but the factors driving poor outcomes remain ate models. Finally, to analyze alloHSCT outcomes, the
ill-defined. cumulative incidence of relapse and non-relapse mor-
tality (NRM) were analyzed as competing events. Non-
relapse mortality was defined as death from any cause
Materials and methods
without evidence of leukemia after alloHSCT.
Study design Moreover, predictors of relapse and NRM were
assessed using Fine-Gray proportional hazard models.
After institutional review board approval, K.C.M. and
For alloHSCT, OS and DFS were assessed from the
H.B.A. conducted a systematic chart review of all
date of transplant. p<.05 was the cutoff for statistical
patients diagnosed with acute lymphoblastic leuke-
significance. JMP 10.0 (SAS Institute Inc., Cary, NC) and
mia/lymphoma (ALL) according to the 2008 World
EZR 1.36 were used for the statistical analysis [27].
Health Organization (WHO) criteria at the three Mayo
Clinic Cancer Center sites (Rochester, Florida and
Arizona). We identified 124 patients who were Results

60 years old at the time of diagnosis from 1 January
Patient characteristics
2000 to 31 December 2016. For patients diagnosed
prior to 2008, the 2008 WHO criteria were retroactively We identified 124 patients with de novo elderly ALL.
applied. Burkitt and Burkitt-like leukemia/lymphoma The median age at time of diagnosis was 67 (range,
were excluded, as well as chronic myeloid leukemia in 60–86); 74 (60%) were males. Eighty-one (65%)
blast crisis. patients were deceased at the time of last follow-up:
The Eastern Cooperative Oncology Group 61 (75%) died with either persistent or relapsed ALL,
Performance Status (ECOG PS) at diagnosis was 14/81 (17%) while in CR/CRi, and 6/81 (7%) due to an
abstracted from the medical record by K.C.M. In sev- unknown cause. Among the 43 (35%) patients who
eral cases (n ¼ 6, 5%), ECOG PS was missing from the were alive at the time of last follow-up, the median
medical record, and was retroactively assigned by time of clinical follow-up was 30 months (interquartile
H.B.A. based on the patients’ disease presentation, range [IQR], 16–78).
after being blinded to their clinical outcomes. The Baseline characteristics are detailed in Table 1. At
Charlson Comorbidity Index (CCI) scores at the time of time of diagnosis, 108 patients (87%) had B-ALL, of
diagnosis were calculated by K.C.M. from information which 40 (37%) were Philadelphia chromosome-posi-
in the medical records [26]. The primary diagnosis of tive (BCR-ABL1þ). Only 16 (13%) patients had T-ALL.
ALL was not included in the CCI score (i.e. a patient Overall, 24 (19%) patients had an ECOG PS
2, and 26
with no comorbidities would have a score of 0). (21%) had a CCI
3. Cytogenetic analysis demon-
strated that 31/95 (33%) patients had a monosomal
karyotype, 5/95 (5%) had high hyperdiploidy (51–65
Statistical analysis
chromosomes), 5/95 (5%) had low hypodiploidy/near
The Mann-Whitney U test was used to compare con- triploidy (30–39, or 60–78 chromosomes, respectively),
tinuous variables between groups, and the Fisher’s and 1/95 (1%) had a mix of hyper-diploid and hypo-
Exact Test was used to compare categorical variables. diploid clones [28,29]. In total, 28/95 (29%) patients
OS was analyzed using the Kaplan-Meier method, had complex cytogenetics (
5 structural abnormal-
calculated from the date of diagnosis until either the ities). Additionally, fluorescence in situ hybridization
time of death or time of last follow-up, as of 25 (FISH) studies revealed 17/48 (35%) patients harbored
October 2017. Disease-free survival (DFS) was calcu- a CDKN2A deletion, and 6/95 (6%) had a KMT2A (MLL)
lated in patients who achieved complete remission rearrangement.
(CR; defined as 5% leukemic blasts in the bone mar- Interestingly, 20 (16%) patients had a history of
row) or complete remission with insufficient hemato- exposure to leukemogenic chemotherapy. We defined
logical recovery (CRi; defined as CR without full this as therapy-related ALL, a rarely described phe-
recovery of platelets and/or neutrophils), from the nomenon [30–32]. Thirteen (65%) of these patients
time of achievement of CR/CRi until relapse, death, or had an antecedent treated hematologic malignancy.
 ELDERLY ALL: MAYO STUDY OF 124 PATIENTS 3

Table 1. Patient characteristics. was hyper-CVAD (alternating eight courses of hyper-

Elderly ALL patients, n 124 fractionated cyclophosphamide, vincristine, doxorubi-
Rochester, n (%) 71 (57)
Arizona, n (%) 37 (30) cin, and dexamethasone with high doses of
Florida, n (%) 16 (13) methotrexate and cytarabine, followed by mainten-
Male, n (%) 74 (60)
Median age at time of diagnosis, n (range): 67 (60–86)
ance with 6-mercaptopurine, vincristine, methotrexate
60–69, n (%) 83 (67) and prednisone) [22], which was used in 62/102 (61%)
70–79, n (%) 34 (27)

80, n (%) 7 (6)
patients. Rituximab was added to hyper-CVAD in the
ECOG Performance Status (PS), n (%): 21/62 (34%) patients with
20% of blasts expressing
0–1 100 (81) CD20; a median of eight doses of rituximab were

2 24 (19)
Charlson Comorbidity Index (CCI), n (%): administered (range, 1–8). Moreover, TKIs were added
0 42 (34) to hyper-CVAD in 17/62 (27%) patients. On the other
1 25 (20)
2 31 (25) hand, asparaginase-based regimens were used in

3 26 (21) 24/102 (24%) patients. The remaining 16/102 (13%)
Serum Creatinine (mg/dL), median (range) 1.1 (0.4–4.0)
Serum Albumin (g/dL), median (range) 3.6 (2.0–4.7) were treated with other intensive regimens. No
Total Bilirubin (mg/dL), median (range) 0.5 (0.2–3.6) patients were enrolled in clinical trials up-front.
B symptoms, n (%) 47/109 (43)
Lymphadenopathy, n (%) 24/101 (24) Further, 16/87 (18%) patients had central nervous sys-
Pleural effusion, n (%) 9/103 (9) tem (CNS) leukemia identified by lumbar puncture at
Splenomegaly, n (%) 17/101 (17)
CNS disease (i.e. positive lumbar puncture), n (%) 16/87 (18) diagnosis; all 16 (100%) received intrathecal chemo-
Lactate dehydrogenase (IU/L), median (range) 362 (124–8840) therapy. Of the other eighty-six patients with no
WBC (109/L), median (range) 6.3 (0.5–338.7)
High WBC (
30  109/L), n (%) 21/115 (18)
known evidence of CNS leukemia, 80 (93%) were
Hemoglobin (g/dL), median (range) 10 (2.2–18.5) treated with prophylactic intrathecal chemotherapy.
Platelets (109/L) , median (range) 48 (4–750)
Absolute neutrophil count (109/L), median (range) 1.4 (0.02–32.0)
Just 22/102 (22%) patients in the intensive treatment
Peripheral blasts, median % (range) 29 (0–95) group had an up-front dose reduction of at least one
Marrow blasts, median % (range) 87 (0–100) chemotherapy agent. Unsurprisingly, these 22 patients
Marrow cellularity, median % (range) 90 (10–100)
Cell lineage type, n (%): were significantly older than the full-dose treatment
B-ALL 108 (87) group (median age 69 versus 65, p<.01). However, all
T-ALL 16 (13)
Cytogenetics/FISH, n (%): 102 (100%) intensively treated patients received
Philadelphia chromosome-positive (BCR-ABL1þ) 40/108 (37) anthracyclines. Of the 24 patients who received aspar-
Monosomal karyotype 31/95 (33)
High hyperdiploidy 5/95 (5) aginase, 7 (29%) experienced dose-limiting toxicities:
Low hypodiploidy/near tripoidy 5/95 (5) coagulopathy (n ¼ 3), liver injury (n ¼ 3) and delayed
Mix of hyper-diploid and hypo-diploid clones 1/95 (1)
Complex cytogenetics (
5 abnormalities) 28/95 (29) hypersensitivity skin rash (n ¼ 1).
CDKN2A deletion 17/48 (35) CR/CRi was achieved in 92/102 (90%) of the
KMT2A (MLL) rearrangement 6/95 (6)
t(1;19) (TCF3-PBX1) 3/95 (3) patients treated intensively (CR in 47, CRi in 27, and
IKZF1 deletion 2/95 (2) CR with unavailable peripheral blood count informa-
ETV6 deletion 2/95 (2)
t(5;14) (IL3-IGH) 1/95 (1)
tion in 18). Four (4%) patients required second-line
chemotherapy to achieve CR/CRi. All 29 (100%) BCR-
ABL1þ patients achieved CR/CRi, as opposed to 62/72
Another 6 (30%) patients had solid tumors treated
(86%) of BCR-ABL1-negative patients (p¼.06). TKIs were
with leukemogenic chemotherapy. Finally, one patient
incorporated for all 29 (100%) intensively treated BCR-
had autoimmune disease treated with cyclophospha-
ABL1þ patients. The achievement of CR/CRi differed
mide (detailed in Supplementary Table S1).
significantly by age group: 73/77 (95%) patients ages
Considering both solid and hematologic malignancies,
60–69, 18/22 (82%) patients ages 70–79 and 1/3 (33%)
37 (30%) patients had a prior cancer diagnosis.
patient
80 years old achieved CR/CRi (p<.01). The
median time to CR/CRi was 30 days (IQR, 24–53).
Of the 10/102 (10%) patients who did not achieve CR/
Intensive treatment
CRi, 7 (7%) had primary refractory disease, and 3 (3%)
The majority of patients (n ¼ 102/124, 82%) were died early, prior to bone marrow assessment.
treated with intensive remission induction chemother- Among patients who received intensive treatment,
apy, while 22/124 (18%) received low-intensity or pal- 5 (5%) died within 60 days, and a total of 8 (8%) died
liative treatment up-front. In the intensive treatment within 100 days; these deaths were related to subdural
group, the median time to initiate therapy was 4 days hematoma (n ¼ 3), infection (n ¼ 2), treatment discon-
(range, 0–35). The most common induction regimen tinuation for psychiatric reasons (n ¼ 1), and unknown
4 K. C. MILLER ET AL.

Figure 1. Overall survival in elderly ALL patients by treatment group, classified as intensive (n ¼ 102) or low-intensity/pallia-
tive (n ¼ 22).

causes (n ¼ 2). The median OS in the intensive median OS was 5.7 months (IQR, 1.7–12.2) (Figure 1).
treatment group was 23.2 months (IQR, 10.9–87.3) Ten (45%) patients died within 100 days of diagnosis.
(Figure 1). OS did not significantly differ between
patients who achieved CR versus CRi (data
Relapse
not shown).
In total, relapse occurred in 42 (42%) of the 99
patients who achieved CR/CRi within a median time of
Low-intensity/palliative treatment
10.9 months (range, 1.2–72.8). The median DFS in the
For several reasons, including advanced age, comor- intensive treatment group (n ¼ 92) was 18.8 months
bidities, poor performance status and personal prefer- (IQR, 8.2–50.3), while the low-intensity/palliative treat-
ence, 22/124 (18%) patients opted for low-intensity or ment group (n ¼ 7) had a median DFS of 7.7 months
palliative treatment up-front. This included less-inten- (IQR, 6.8–8.6) (p<.01). Six (14%) patients had an extra-
sive chemotherapy (e.g. vincristine and steroids, ± TKIs; medullary relapse, including four with CNS involve-
n ¼ 9), TKIs (±concomitant steroids; n ¼ 6), or direct ment, and two with leukemia cutis. After relapse, 31
enrollment into hospice care (±steroids; n ¼ 7). The (74%) patients were treated with salvage chemother-
low-intensity/palliative treatment group was signifi- apy with a wide range of regimens. On the other
cantly older than the intensive treatment group hand, 11 (26%) patients enrolled into hospice immedi-
(median age 73 versus 65, p<.01). Moreover, a signifi- ately. The median OS from the time of relapse was
cantly higher proportion had an ECOG PS
2 (n ¼ 12, 8.8 months (IQR, 4–20) for patients who received sal-
55%) and CCI
3 (n ¼ 12, 55%) (p<.01). Eleven (50%) vage treatment versus 1.3 months (IQR, 1–4) for those
patients in the low-intensity/palliative treatment group went directly to hospice (p<.01). When considering all
were BCR-ABL1þ; 9/11 (82%) received TKIs, while the 42 patients, the median OS from the time of relapse
other two went to hospice care. was 5.6 months (IQR, 2.1–11.7). At 1 year post-relapse,
Seven (32%) patients achieved CR/CRi (3 CR, 3 CRi, 1 the probability of OS was 22.8% (95% CI, 11.0–37.1%).
CR with unknown peripheral blood counts), within a
median time of 54 days (range, 23–128); 4/7 (57%)
Hematopoietic stem cell transplantation
were BCR-ABL1þ. The seven patients who achieved CR/
CRi had a median OS of 12.2 months (IQR, 12.1–13.8), Altogether, 32 (26%) patients underwent hematopoi-
compared to just 2.3 months (IQR, 0.9–7.2) for the 15 etic stem cell transplantation (HSCT): alloHSCT in 31,
patients who did not achieve CR/CRi (p<.01). For the and autologous HSCT in 2 (one patient had an autolo-
entire low-intensity/palliative treatment group, the gous transplant followed by alloHSCT after relapse).
 ELDERLY ALL: MAYO STUDY OF 124 PATIENTS 5

Figure 2. The cumulative incidence of relapse and non-relapse mortality in 31 elderly ALL patients who underwent allogeneic
hematopoietic stem cell transplantation (alloHSCT).

From here, only the 31 (25%) patients who underwent Three (9%) patients died within 100 days of
alloHSCT will be considered. The median age at alloHSCT, from early relapse (n ¼ 2; both were in CR1)
alloHSCT was 64 (range, 60-71). The majority of these and acute graft versus host disease (GVHD) of the
patients (n ¼ 25, 81%) had B-ALL, of which 9/25 (36%) gastrointestinal tract (n ¼ 1). By Day þ100, grade II-IV
were BCR-ABL1þ. Only 6/31 (19%) had T-ALL. Twenty- GVHD occurred in 16/31 (51.6%) and grade III–IV
seven (87%) patients were in their first complete GVHD occurred in 4/31 (12.9%). In total, 11/31 (35.5%)
remission (CR1), while 4 (13%) were in CR2. The patients developed chronic GVHD: mild (n ¼ 3), moder-
median number of induction regimens prior to ate (n ¼ 4) and severe (n ¼ 4).
alloHSCT was 1 (range, 1–3), and the median time to The 1 year and 3 year cumulative incidence of
alloHSCT from date of diagnosis was 5.1 months relapse was 29% and 43%, respectively; the 1 year and
(range, 2.6–28.1). The median follow-up from Day 0 3 year cumulative incidence of NRM was 29% and
was 20 months (IQR, 18–35) in surviving patients. 32%, respectively (Figure 2). In a Fine-Gray propor-
The median hematopoietic cell transplantation- tional hazard model for competing events (relapse or
NRM), age at time of alloHSCT was a significant pre-
comorbidity index (HCT-CI) score was 1 (range, 0–4),
dictor for NRM (Hazard Ratio [HR] 1.30 per year, 95%
and 13 (43%) patients had an HCT-CI score
3. The
CI 1.04–1.62, p ¼ .02) (Supplementary Table S2). The
median Karnofsky Performance Status (KPS) was 90%
median OS from the time of alloHSCT was 15.8 months
(range, 70–100). Thirteen (43%) patients had a
(IQR, 7.6-not reached [NR]), with a 1 year and 3 year
matched-related donor (median age 59, range 50–72),
OS of 55% and 38%, respectively. The 1 year and
while 17 (57%) had a matched-unrelated donor
3 year DFS post-alloHSCT were 42% and 26%,
(median age 26, range 22–42). Twenty-seven (93%)
respectively.
were 10/10 HLA matched, while 2 (7%) were 9/10
matched (the other 2 patients had unknown matching
status). All patients received T-cell replete peripheral Predictors of survival
blood stem cell grafts. Female to male transplant The entire elderly ALL cohort had a median OS of 19.8
occurred in 6/29 (21%), and 12/27 (48%) were cyto- months (IQR, 9.4–56.6). Median OS was significantly
megalovirus mismatched. Most patients (n ¼ 27, 90%) affected by age at time of diagnosis: 23.2 months for
received fludarabine/melphalan-based reduced-inten- patients ages 60–69 (IQR, 11.7–87.3), 12.1 months for
sity conditioning regimens; 3 (10%) patients received patients ages 70–79 (IQR, 5.6–34.1) and 9.6 months
myeloablative conditioning regimens, as defined by for patients
80 years old (IQR, 1.3–17.4)(p¼.02)
Center for International Blood and Marrow Transplant (Figure 3). There was no difference in OS for therapy-
Research (CIBMTR) criteria [33]. related ALL compared to the rest of the cohort.
6 K. C. MILLER ET AL.

Figure 3. The impact of age at time of diagnosis on overall survival for the entire cohort of elderly ALL patients, split into ages
60–69 (n ¼ 83), 70–79 (n ¼ 34), and
80 (n ¼ 7).

Table 2. Univariate and multivariate analysis for predictors of overall survival in elderly ALL.
Univariate analysis Multivariate analysisb
Variable No. (%) Hazard ratio (95% CI) p Value Hazard ratio (95% CI) p Value
Female 50/124 (40) 0.73 (0.46–1.14) .17
Age at diagnosis – 1.06a (1.02–1.10) .0034 1.02a (0.97–1.08) .41
Therapy-related ALL 20/124 (16) 0.87 (0.42–1.61) .67
Prior solid or hematologic cancer 37/124 (30) 0.74 (0.43–1.20) .24
ECOG Performance Status (PS)
2 at diagnosis 24/124 (19) 3.18 (1.87–5.20) <.0001 3.41 (1.65–6.56) .0012
Charlson Comorbidity Index (CCI)
3 at diagnosis 26/124 (21) 2.50 (1.49–4.07) .0008 1.90 (0.99–3.58) .0538
WBC
30  109/L at diagnosis 21/115 (18) 1.65 (0.92–2.81) .0911 2.47 (1.05–5.44) .0386
Lactate dehydrogenase (IU/L) at diagnosis – 1.0004a (1.0002–1.0005) .0004 1.0004a (1.0002–1.0006) .0003
Serum albumin <3 g/dL at diagnosis 13/67 (19) 1.55 (0.77–2.87) .21
CNS leukemia involvement at diagnosis 16/87 (18) 1.32 (0.63–2.53) .44
T-cell phenotype 16/124 (13) 1.13 (0.56–2.05) .72
Philadelphia chromosome-positive (BCR-ABL1þ) 40/123 (33) 0.74 (0.45–1.17) .20
CDKN2A deletion 17/48 (35) 1.79 (0.80–3.91) .15
Monosomal karyotype 31/95 (33) 0.71 (0.29–1.45) .36
Complex cytogenetics (
5 abnormalities) 28/95 (29) 1.26 (0.72–2.13) .41
a
Hazard ratio per unit change in the regressor.
b

Aforward selection method using p < .10 was used to select variables for the multivariate model.
Bold denotes p<.05.

In a univariate Cox proportional hazard model for all subgroup analysis of the 102 patients who received
124 patients, age, ECOG PS
2, CCI
3 and lactate intensive treatment (Supplementary Table S3). In a
dehydrogenase (LDH) at time of diagnosis negatively univariate analysis, we found that BCR-ABL1þ patients
influenced OS (p<.05) (Table 2). There was also a trend had significantly better OS. This maintained signifi-
towards poor OS in patients with a white blood cell cance in a multivariate model (HR 0.29, 95% CI,
count (WBC)
30  109/L at time of diagnosis (p¼.09). 0.10–0.72, p<.01) which included gender, ECOG PS
When these variables were included in a multivariate
2, WBC
30  109/L and LDH. Additionally, ECOG PS
model, ECOG PS
2, WBC
30  109/L and LDH nega-
2, WBC
30  109/L and LDH continued to nega-
tively influenced OS (p<.05). Additionally, there was a tively influence OS in this model.
trend toward poor OS in patients with a CCI
3 (HR To further delineate the effect of the Philadelphia
1.90, 95% CI 0.99–3.58, p¼.05). chromosome, we compared the presenting variables of
Since the low-intensity/palliative treatment group BCR-ABL1þ (n ¼ 29) to BCR-ABL1-negative B-ALL (n ¼ 57)
was older, and had a higher proportion of patients in the intensive treatment group (Supplementary
with advanced ECOG PS and high CCI, we did a Table S4). The characteristics of the two groups were
 ELDERLY ALL: MAYO STUDY OF 124 PATIENTS 7

Figure 4. Overall survival for Philadelphia chromosome-positive (BCR-ABL1þ) B-ALL (n ¼ 29) compared to Philadelphia chromo-
some-negative (BCR-ABL1-negative) B-ALL (n ¼ 57), only considering patients who received intensive treatment.

comparable, except BCR-ABL1-negative patients had a findings suggest the improved efficacy of contempor-
lower ANC (p<.01) and a lower hemoglobin (p ¼ 0.04) ary induction regimens. Interestingly, we classified 20
at diagnosis. Seven (25%) of the BCR-ABL1þ patients (16%) cases as therapy-related ALL. While these
had an up-front chemotherapy dose reduction. There patients did not exhibit different outcomes in the pre-
was no significant difference in alloHSCT utilization sent study, the proportion (16%) was much higher
between the two groups (35% versus 30%, p ¼ .81). than in younger adults (5–7%) [30–32].
Nevertheless, there was a trend towards OS advantage The Philadelphia chromosome (BCR-ABL1þ) was
in BCR-ABL1þ compared to BCR-ABL1-negative B-ALL present in 37% of B-ALL patients, which is within
(median OS 38.8 versus 20.9 months, p¼.05) (Figure 4). range of several recent reports, but lower than some
Interrogating this further, BCR-ABL1þ patients had a groups have shown in the past (i.e. approaching 50%)
higher rate of CR/CRi (100% versus 86%, p¼.047) and a [11,29,34,38,39] Interestingly, BCR-ABL1-positivity was a
lower of rate of relapse (21.4% versus 45.8%, p¼ .048) favorable prognostic factor when considering the 102
compared to BCR-ABL1-negative B-ALL patients. patients treated intensively, corroborating findings
from Byun et al. [23]. This magnifies the question of
how to approach BCR-ABL1þ elderly ALL. Several
Discussion
recent clinical trials have investigated using TKIs with
Herein, we present the detailed natural history of 124 less-intensive chemotherapy (i.e. vincristine and/or ste-
patients with elderly ALL at our multi-site institution. roids) to treat BCR-ABL1þ adult ALL with promising
To the best of our knowledge, this is the second larg- results [40–44]. However, we classified the few
est elderly ALL cohort that has been described in the patients who received these regimens in the present
United States. Not surprisingly, our study illustrates study as low-intensity/palliative given the lack of long-
that elderly patients have poor survival which worsens term follow-up data. To date, there are no studies
with increasing age. Moreover, the proportion of long- comparing these approaches to intensive chemother-
term survivors was comparable to the two largest pro- apy (e.g. hyper-CVAD þ TKI) specifically in the elderly,
spective cohorts of elderly ALL, i.e. UKALL XII/E2993 but this would be salient.
[8] and the MD Anderson experience with Hyper- Furthermore, we report the outcomes of 31
CVAD [20]. patients (25% of our cohort) who underwent alloHSCT,
A major concern when treating elderly ALL is the mostly with RIC regimens while in CR1. We found the
risk of early mortality. This was relatively low in our cumulative incidence of relapse and NRM at 3 years to
intensive treatment group (5% by Day 60, and 8% by be 43% and 32%, respectively, which is within range
Day 100) [8,11,23,24,34–37]. In addition, a high propor- of a recent CIBMTR study [21]. Thus, while it appears
tion (90%) of patients achieved CR/CRi. Both these that RIC alloHSCT can lead to long-term survival in
8 K. C. MILLER ET AL.

a subset of elderly ALL patients, it must be chosen emerged. Hopefully these combination therapies and
with considerable discretion given the age-related risk other novel approaches will help to turn the tide for
of NRM [45,46]. Also relapse after alloHSCT was com- this devastating malignancy.
mon, calling attention to the need for more effective
treatment strategies both pre- and post-transplant.
Acknowledgements
The role of alloHSCT in BCR-ABL1þ ALL remains a vex-
ing question [44], but was impossible to assess in We would like to acknowledge the Mayo Clinic Acute
Leukemia and Myeloid Disease Group, the patients and
our cohort.
their families.
Our study also illustrates the dismal prognosis of
relapsed elderly ALL: at 1 year post-relapse, more than
3=4 of patients were deceased. Hopefully, the recently Prior presentation
approved biologic agents blinatumomab and inotuzu- Parts of this manuscript were presented in abstract form at
mab ozogamicin will make a significant impact in this the 22nd Congress of the European Hematology Association,
setting [47,48]. These drugs are rapidly moving into Madrid, Spain, 2017; at the 59th Annual Meeting of the
front-line therapy as well. For example, Kantarjian American Hematology Association, Atlanta, Georgia, USA,
2017; and at the Center for International Blood & Marrow
et al. [49] recently published the results of a phase II
Transplant Research/American Society for Blood and Marrow
trial in elderly BCR-ABL1-negative B-ALL which com- Transplantation Tandem Meetings, Salt Lake City, Utah,
bined inotuzumab ozogamicin with lower-intensity USA, 2018.
‘mini-hyper-CVD’. The CR/CRi rate was 98% with few
early deaths in this study, and the 3-year OS was quite Potential conflict of interest: Disclosure forms provided
impressive at 56% (95% CI, 39–79%). by the authors are available with the full text of this article
Our analysis was limited by being retrospective in online at https://doi.org/10428194.2018.1509318.
nature. As such, treatment decisions were heteroge-
neous and likely influenced by numerous unquantifi- ORCID
able factors. Also, because this study was conducted
Kevin C. Miller http://orcid.org/0000-0002-5534-5570
at a multi-site tertiary academic center, there was Aref Al-Kali http://orcid.org/0000-0002-0824-3715
undoubtedly a referral bias. Indeed, it has been specu- William J. Hogan http://orcid.org/0000-0002-5841-4105
lated that many patients with elderly ALL may never Mrinal M. Patnaik http://orcid.org/0000-0001-6998-662X
be referred for treatment; true outcomes in the com- Mark R. Litzow http://orcid.org/0000-0002-9816-6302
munity may be considerably more sobering [9]. Finally,
we did not have next-generation genetic sequencing
(NGS), Philadelphia-like, or minimal residual disease References
data available for the patients in our study. Since the
[1] Terwilliger T, Abdul-Hay M. Acute lymphoblastic leu-
majority of patients in our study died from relapsed kemia: a comprehensive review and 2017 update.
disease, this data would be of particular interest to Blood Cancer J. 2017;7:e577.
investigate whether elderly ALL has distinctive bio- [2] Howlader NNA, Krapcho M, Miller D, et al. SEER
logical characteristics that could be leveraged Cancer Statistics Review, 1975–2014. Bethesda, MD:
therapeutically. National Cancer Institute. Available from: https://seer.
cancer.gov/csr/1975_2014/, based on November 2016
In conclusion, elderly ALL is a heterogeneous and
SEER data submission, posted to the SEER web site,
difficult to treat. Our study highlights the important April 2017.
role that advanced age, poor performance status and [3] Pagano L, Mele L, Casorelli I, et al. Acute lympho-
comorbidities play in manifesting adverse outcomes. blastic leukemia in the elderly. A twelve-year retro-
However, it is likely that these simple metrics merely spective, single center study. Haematologica.
2000;85:1327–1329.
scratch the surface of patients’ readiness for intensive
[4] Thomas X, Olteanu N, Charrin C, et al. Acute lympho-
treatment. There is mounting evidence in numerous blastic leukemia in the elderly: the Edouard Herriot
other hematologic malignancies that comprehensive Hospital experience. Am J Hematol. 2001;67:73–83.
frailty assessments predict outcomes, and it is hard to [5] Annino L, Goekbuget N, Delannoy A. Acute lympho-
imagine this would be different in elderly ALL [50]. blastic leukemia in the elderly. Hematol J.
2002;3:219–223.
Finally, even though we illustrated a dismal picture for
[6] Brandwein JM, Gupta V, Wells RA, et al. Treatment of
elderly ALL, it appears that hope is on the horizon elderly patients with acute lymphoblastic leukemia-
now that clinical trials using biologic agents in com- evidence for a benefit of imatinib in BCR-ABL positive
bination with lower-intensity chemotherapy have patients. Leukemia Res. 2005;29:1381–1386.
 ELDERLY ALL: MAYO STUDY OF 124 PATIENTS 9

[7] Shin DY, Kim I, Kim KH, et al. Acute lymphoblastic leu- with acute lymphoblastic leukemia in the era of
kemia in elderly patients: a single institution’s experi- potent tyrosine kinase inhibitors. Haematologica.
ence. Korean J Intern Med. 2011;26:328–339. 2017;102:e187–e190.
[8] Sive JI, Buck G, Fielding A, et al. Outcomes in older [23] Saillard C, Etienne A, Charbonnier A, et al. Evaluation
adults with acute lymphoblastic leukaemia (ALL): of comorbidity indexes in the outcome of elderly
results from the international MRC UKALL XII/ patients treated for acute lymphoblastic leukemia.
ECOG2993 trial. Br J Haematol. 2012;157:463–471. Leuk Lymph. 2014;55:2211–2212.
[9] Gokbuget N. How I treat older patients with ALL. [24] Goekbuget N, Beck J, Brueggemann M, et al.
Blood. 2013;122:1366–1375. Moderate intensive chemotherapy including CNS-
[10] Marks DI. The challenges of managing older patients prophylaxis with liposomal cytarabine is feasible and
with acute lymphoblastic leukemia. Am Soc Clin effective in older patients with Ph-negative Acute
Oncol Educ Book Am Soc Clin Oncol Meet. 2015; Lymphoblastic Leukemia (ALL): results of a prospect-
35:e343–e351. ive trial from the german Multicenter Study Group for
[11] Kozlowski P, Lennmyr E, Ahlberg L, et al. Age but not Adult ALL (GMALL). Blood. 2012;120:1493–1493.
Philadelphia positivity impairs outcome in older/eld- [25] Rosko AE, Wang H-L, de Lima M, et al. Reduced inten-
erly patients with Acute Lymphoblastic Leukemia in sity conditioned allograft yields favorable survival for
the Swedish population. Eur J Haematol. older adults with B-cell acute lymphoblastic leukemia.
2017;99:141–149. Am J Hematol. 2017;92:42–49.
[12] Pulte D, Gondos A, Brenner H. Improvement in sur- [26] Charlson ME, Pompei P, Ales KL, et al. A new method
vival in younger patients with acute lymphoblastic of classifying prognostic comorbidity in longitudinal
leukemia from the 1980s to the early 21st century. studies: development and validation. J Chronic Dis.
Blood. 2009;113:1408–1411. 1987;40:373–383.
[13] Guru Murthy GS, Venkitachalam R, Mehta P. Trends in [27] Kanda Y. Investigation of the freely available easy-to-
survival outcomes of B-lineage acute lymphoblastic use software ’EZR’ for medical statistics. Bone Marrow
leukemia in elderly patients: analysis of Surveillance,
Transplant. 2013;48:452–458.
Epidemiology, and End Results database. Leuk
[28] Kenderian SS, Al-Kali A, Gangat N, et al. Monosomal
Lymph. 2015;56:2296–2300.
karyotype in Philadelphia chromosome-negative acute
[14] Li S, Molony JT, Chia VM, et al. Mortality among eld-
lymphoblastic leukemia. Blood Cancer J. 2013;3:e122.
erly patients newly diagnosed with acute lympho-
[29] Moorman AV, Chilton L, Wilkinson J, et al. A popula-
blastic leukemia (ALL) using 100% Medicare ALL data.
tion-based cytogenetic study of adults with acute
J Clin Oncol. 2016;34(15):7022–7022
lymphoblastic leukemia. Blood. 2010;115:206–214.
[15] Yilmaz M, Kantarjian H, Jabbour E. Treatment of acute
[30] Aldoss I, Dagis A, Palmer J, et al. Therapy-related ALL:
lymphoblastic leukemia in older adults: now and the
cytogenetic features and hematopoietic cell trans-
future. Clin Adv Hematol Oncol. 2017;15:266–274.
plantation outcome. Bone Marrow Transplant. 2015;
[16] Talarico L, Chen G, Pazdur R. Enrollment of elderly
50:746–748.
patients in clinical trials for cancer drug registration: a
[31] Abdulwahab A, Sykes J, Kamel-Reid S, et al. Therapy-
7-year experience by the US food and drug adminis-
related acute lymphoblastic leukemia is more fre-
tration. J Clin Oncol. 2004;22:4626–4631.
[17] Scher KS, Hurria A. Under-representation of older quent than previously recognized and has a poor
adults in cancer registration trials: known problem, lit- prognosis. Cancer. 2012;118:3962–3967.
tle progress. J Clin Oncol. 2012;30:2036–2038. [32] Swaika A, Frank RD, Yang D, et al. Second primary
[18] Bindu Kanapuru M, Harpreet Singh M, Myers A, et al. acute lymphoblastic leukemia in adults: a SEER ana-
Enrollment of older adults in clinical trials evaluating lysis of incidence and outcomes. Cancer Med.
patients with hematologic malignancies – the Food 2018;7:499–507.
and Drug Administration (FDA) Experience [abstract]. [33] Bacigalupo A, Ballen K, Rizzo D, et al. Defining the
Atlanta (GA): American Society of Hematology; 2017. intensity of conditioning regimens: working defini-
[19] Muffly L, Pasquini MC, Martens M, et al. Increasing tions. Biol Blood Marrow Transplant.: J Am Soc Blood
use of allogeneic hematopoietic cell transplantation Marrow Transplant. 2009;15:1628–1633.
in patients aged 70 years and older in the United [34] Roberts KG, Gu Z, Payne-Turner D, et al. High fre-
States. Blood. 2017;130:1156–1164. quency and poor outcome of Philadelphia chromo-
[20] O’Brien S, Thomas DA, Ravandi F, et al. Results of the some-like acute lymphoblastic leukemia in adults. J
hyperfractionated cyclophosphamide, vincristine, Clin Oncol. 2017;35:394–401.
doxorubicin, and dexamethasone regimen in elderly [35] Poch Martell M, Atenafu EG, Minden MD, et al.
patients with acute lymphocytic leukemia. Cancer. Treatment of elderly patients with acute lympho-
2008;113:2097–2101. blastic leukaemia using a paediatric-based protocol.
[21] Rowe JM, Buck G, Burnett AK, et al. Induction therapy Br J Haematol. 2013;163:458–464.
for adults with acute lymphoblastic leukemia: results [36] Ribera JM, Garcia O, Oriol A, et al. Feasibility and
of more than 1500 patients from the international results of subtype-oriented protocols in older adults
ALL trial: MRC UKALL XII/ECOG E2993. Blood. and fit elderly patients with acute lymphoblastic leu-
2005;106:3760–3767. kemia: results of three prospective parallel trials from
[22] Byun JM, Koh Y, Shin D-Y, et al. BCR-ABL translocation the PETHEMA group. Leukemia Research. 2016;41:
as a favorable prognostic factor in elderly patients 12–20.
10 K. C. MILLER ET AL.

[37] Sancho J-M, Ribera J-M, Xicoy B, et al. Results of the residual disease receive a hematopoietic stem cell
PETHEMA ALL-96 trial in elderly patients with transplant in first remission?. Best Practice Res Clin
Philadelphia chromosome-negative acute lympho- Haematol. 2016;29:345–350.
blastic leukemia. Eur J Haematol. 2007;78:102–110. [45] Mohty M, Labopin M, Volin L, et al. Reduced-intensity
[38] Larson RA. Management of acute lymphoblastic leu- versus conventional myeloablative conditioning allo-
kemia in older patients. Semin Hematol. 2006;43: geneic stem cell transplantation for patients with
126–133. acute lymphoblastic leukemia: a retrospective study
[39] Burmeister T, Schwartz S, Bartram CR, et al. Patients’ from the European Group for Blood and Marrow
age and BCR-ABL frequency in adult B-precursor ALL: Transplantation. Blood. 2010;116:4439–4443.
a retrospective analysis from the GMALL study group. [46] Marks DI, Wang T, Perez WS, et al. The outcome of
Blood. 2008;112:918–919. full-intensity and reduced-intensity conditioning
[40] Vignetti M, Fazi P, Cimino G, et al. Imatinib plus ste- matched sibling or unrelated donor transplantation in
roids induces complete remissions and prolonged sur- adults with Philadelphia chromosome-negative acute
vival in elderly Philadelphia chromosome-positive lymphoblastic leukemia in first and second complete
patients with acute lymphoblastic leukemia without remission. Blood. 2010;116:366–374.
additional chemotherapy: results of the Gruppo [47] Kantarjian HM, Stein AS, Bargou RC, et al.
Italiano Malattie Ematologiche dell’Adulto (GIMEMA) Blinatumomab treatment of older adults with
LAL0201-B protocol. Blood. 2007;109:3676–3678. relapsed/refractory B-precursor acute lymphoblastic
[41] Foa R, Vitale A, Vignetti M, et al. Dasatinib as first-line leukemia: Results from 2 phase 2 studies. Cancer.
treatment for adult patients with Philadelphia 2016;122:2178–2185.
chromosome-positive acute lymphoblastic leukemia. [48] Kantarjian HM, DeAngelo DJ, Stelljes M, et al.
Blood. 2011;118:6521–6528. Inotuzumab ozogamicin versus standard therapy for
[42] Chalandon Y, Thomas X, Hayette S, et al. Randomized acute lymphoblastic leukemia. N Engl J Med.
study of reduced-intensity chemotherapy combined 2016;375:740–753.
with imatinib in adults with Ph-positive acute [49] Kantarjian HM, Ravandi F, Short NJ, et al. Inotuzumab
lymphoblastic leukemia. Blood. 2015;125:3711–3719. ozogamicin in combination with low-intensity chemo-
[43] Rousselot P, Coude MM, Gokbuget N, et al. Dasatinib therapy for older patients with Philadelphia chromo-
and low-intensity chemotherapy in elderly patients some-negative acute lymphoblastic leukaemia: a
with Philadelphia chromosome-positive ALL. Blood. single-arm, phase 2 study. Lancet Oncol. 2018;19:
2016;128:774–782. 240–248.
[44] Litzow MR. Should anyone with Philadelphia chromo- [50] Abel GA, Klepin HD. Frailty and the management of
some-positive ALL who is negative for minimal hematologic malignancies. Blood. 2018;131:515–524.