Pediatr Blood Cancer 2010;55:1096–1102

Prognostic Significance of Early Lymphocyte Recovery in Pediatric Osteosarcoma

Colin Moore, MD,1 Don Eslin, MD,2 Alejandro Levy, MD,2 Jessica Roberson, MD,2 Vincent Giusti, MD,2 andRobert Sutphin, MD2,*

Background. Evidence suggests early lymphocyte recovery afterchemotherapy predicts superior outcome for patients with cancer,a phenomenon not previously investigated in osteosarcoma. Thisstudy determined the prognostic significance of early lymphocyterecovery for pediatric patients with osteosarcoma. Procedures. Wereviewed data of 19 consecutive patients treated for osteosarcoma atour institution from1997 to 2007. After initial chemotherapy, patientswere separated into two groups: early versus late lymphocyte recov-ery, using a threshold absolute lymphocyte count of >800 cells/�lon day 14 (ALC-14). Results. The 5-year overall survival (OS) forour cohort was 73.7% [±10.1 standard error (SE)]. Thirteen patients(68%) had an ALC-14>800 cells/�l, with 12/13 alive and 5-year OSof 92.3% (±7.4 SE). In contrast, six patients (32%) had an ALC-14<800 cells/�L, with 1/6 alive and 5-year OS of 33.3% (±19.2

SE). The difference is statistically significant (P =0.0013, log-ranktest). Two patients presented with multifocal disease at diagnosis,had late lymphocyte recovery and died. One patient presented withmetastatic disease, had early lymphocyte recovery and is alive. Sixpatients developed relapsed disease with a 5-year OS of 33.3%(±19.2 SE). The majority (5/6) of patients with relapsed diseasedied while on active therapy. The only survivor in this group hadan ALC-14>800 cells/�l and recently completed relapse therapy.Conclusions. These data demonstrate that early lymphocyte recoveryrepresents a significant prognostic indicator for osteosarcoma. Earlyidentification and risk stratification therapy based on the ALC-14threshold may improve outcomes and our knowledge of this disease.Pediatr Blood Cancer. 2010;55:1096–1102. © 2010 Wiley-Liss, Inc.

Key words: absolute lymphocyte count; osteosarcoma; prognostic factor

INTRODUCTION

Pediatric malignancy is the second most common cause of child-hood mortality [1]. Osteosarcoma represents the most commonbone tumor in pediatric populations, with an annual incidence inthe United States of approximately four per million children [2].Osteosarcoma has an approximate 5-year overall survival (OS) of70% [3–6]. OS in patients without metastatic disease at time of pre-sentation is currently reported as 75–80% [4,7]. In patients withmetastases at time of diagnosis, the OS has been reported as low as10–30% [3,7].

Traditional therapeutic protocols for osteosarcoma involve neo-adjuvant chemotherapy and tumor resection. The most activechemotherapeutic agents against osteosarcoma currently includeadriamycin, cisplatin, high-dose methotrexate, and ifosfamide [8,9].Current standard of care therapy includes pre-surgical chemother-apy, followed by tumor resection with limb salvage and/oramputation, followed by post-surgical chemotherapy based on his-tological response to neo-adjuvant therapy.

Previous studies have shown the value of prognostic indica-tors for survival in patients with osteosarcoma, which include: age,location of primary tumor, histological type, tumor size, metas-tasis at presentation, histological response, alkaline phosphataselevel, Her-2/neu expression, and P-glycoprotein level [7,10–12].However, with the exception of histologic response, none of theseindicators have been consistently significant across studies. Themost agreed-upon prognostic indicator is the degree of histolog-ical response to chemotherapy, which is evaluated by extent oftumor necrosis at time of surgical excision, has remained the onlyprognostic indicator in use for osteosarcoma. Surgical resection isplanned in most studies after the first two months of treatment,making use of this risk factor late in treatment. Protocols have usedthis prognostic indicator to direct post-surgical therapy, althoughrisk stratification based on this factor has shown minimal improve-ment on the overall and event free survival (EFS) for these patients[7,13–15]. Thus, osteosarcoma remains one of the few pediatricmalignancies without a significant indicator to stratify patients earlyin therapy.

Evidence suggests early lymphocyte recovery after initialchemotherapy predicts superior outcome for patients with multi-ple forms of malignancies including acute myelogenous leukemia,acute lymphoblastic leukemia, Ewing sarcoma, breast cancer, mul-tiple myeloma, and non-Hodgkin lymphoma [16–21]. However,early lymphocyte recovery has yet to be investigated as a prognosticindicator in patients with osteosarcoma.

The objective of this study was to determine the prognostic sig-nificance of early lymphocyte recovery for pediatric patients newlydiagnosed and treated for osteosarcoma. Furthermore, our aimsincluded the evaluation of a prognostic factor which could be usedto evaluate the effectiveness of therapy earlier in treatment. Currenttherapy effectiveness is evaluated by the extent of tumor necrosisafter several courses of treatment, therefore no true early indicatorsare available.

METHODS

Patient Selection

Patients with a diagnosis of osteosarcoma, with or withoutmetastatic disease or multifocal bone disease at presentation, wereselected for evaluation. We reviewed the medical records of consec-utive patients diagnosed and treated at our institution from 1997to 2007. All diagnoses were confirmed histologically and onlypediatric patients under the age of 21 treated at our facility were

1Department of Pediatric Medical Education, Arnold Palmer MedicalCenter, Orlando Health, Orlando, Florida; 2Pediatric Hematol-ogy/Oncology, Center for Children’s Cancer and Blood Disorders atArnold Palmer Hospital, M. D. Anderson Cancer Center Orlando,Orlando, Florida

Conflict of interest: Nothing to declare.

*Correspondence to: Robert Sutphin, Pediatric Hematology/Oncology,M. D. Anderson Cancer Center Orlando, Orlando, FL 32806.E-mail: robert.sutphin@orlandohealth.com

Received 19 November 2009; Accepted 4 May 2010

© 2010 Wiley-Liss, Inc.DOI 10.1002/pbc.22673Published online 23 August 2010 in Wiley Online Library(wileyonlinelibrary.com).

Lymphocyte Recovery in Osteosarcoma 1097

included. The patients evaluated had been treated on one of severalstandard treatments depending on the era of diagnosis (POG-9351,POG-9754, AOST-0331) in which primary chemotherapeutic agentsconsisted of either: (1) methotrexate, adriamycin, and cisplatin(MAP), or (2) methotrexate, adriamycin, cisplatin, with ifosfamideand etoposide (MAPIE). Additional agents used in some patientprotocols included zinecard and muramyl tripeptide (MTP-PE). Allbut four patients received G-CSF therapy to aid in bone marrowrecovery between courses of therapy. Many patients included in thestudy also had resection of their primary tumor at an outside facilitywhere tumor necrosis evaluation was performed. This informationwas included with the tumor necrosis data obtained from surgeriesat our own institution.

Data Collection

The study included 22 consecutive patients with the diagnosisof osteosarcoma (median age 14 years: range 5–20 years). All datawere obtained from reviewing existing patient records. Hematologicdata were obtained on CBCs performed between days 13 and 15after initial chemotherapy. Three charts had incomplete CBC datafor analysis on day 14 and were excluded from this analysis. Patientswere separated into two groups: early versus late lymphocyte recov-ery. A threshold of absolute lymphocyte count of >800 cells/�l onday 14 (±1 day) after the first course of chemotherapy (ALC-14)was used to identify patients with early lymphocyte recovery. Clin-ical data obtained from review of medical records included: age,gender, chemotherapy utilized, presence of metastasis or multifo-cal disease at diagnosis, ALC, absolute neutrophil count (ANC),platelet count on day 14, percentage of tumor necrosis, relapse, andsurvival data (Table I).

Data Analysis

For analysis, OS was defined as the time period from diagno-sis and start of treatment at our center until the most recent followup or death. EFS was also analyzed, with an event defined as pro-gressive disease, relapse or death. The data were analyzed using theKaplan–Meier methods and log-rank tests to calculate cumulativeOS and EFS. A Cox proportional hazards regression model wasused in a univariate fashion to compare potential risk factors to sur-vival and hazard ratios, which were estimated with 95% confidenceintervals. Risk factors evaluated included: age, gender, ALC-14,chemotherapy regimen used, tumor necrosis, and use of G-CSF.Tumor necrosis was defined as good histological response at time ofresection (>90% necrosis), or poor response (<90%) as in previousstudies [6,13,22,23].

RESULTS

A total of 19 osteosarcoma patients were used in the final dataanalysis. Overall, the patients had a median age of 14.0 and medianfollow up of 6.3 years (Table I). The 5-year OS for our cohortwas 73.7% [±10.1 standard error (SE)] and the 5-year EFS was63.2% (±11.1 SE). The patient population had a median ALC-14of 1,092 cells/�l and a mean of 1,420 cells/�l. Our data provedsignificant for OS using ALC as a continuous variable from athreshold range of ALC-14 of 700–1,700 (Table II). An ALC-14of >800 cells/�l proved to be the most powerful prognostic value

for both OS and EFS and was chosen as the threshold value for thisstudy.

After the threshold was identified, the patients were divided intotwo groups, those with early lymphocyte recovery (ALC-14 > 800)and those with late lymphocyte recovery (ALC-14 < 800). Thirteenpatients (68%) had an ALC-14 > 800 cells/�L, with 12/13 alive atlast follow up. The 5-year OS and EFS for this group of patients were92.3% (±7.4 SE) and 84.6% (±10.0 SE), respectively (Fig. 1). Incomparison, six patients (32%) had an ALC-14 < 800 cells/�L, with1/6 alive at last follow-up. The 5-year OS and EFS for this group was33.3% (±19.2 SE) and 16.7% (±15.2 SE), respectively. The differ-ence between ALC-14 > 800 versus ALC-14 < 800 was statisticallysignificant for OS and EFS (P = 0.0013 and 0.0143, respectively)using the log-rank test (Table III). Seventeen (89%) patients hadtumor necrosis data available for analysis. The 5-year OS and EFSfor those with good histological response (>90% necrosis, n = 12)was 91.7% (±8.0 SE) and 75.0% (±12.5 SE), respectively (Fig. 2).Patients with poor histological response (<90% necrosis, n = 5) hada 5-year OS and EFS which was identical at 60.0% (±21.9 SE). Thedifference was not statistically significant for either OS or EFS inthis group (P = 0.23 and 0.55). Upon further univariate analysis, noother risk factor analyzed showed a significant difference on eitherOS or EFS in the study group (Table III).

By evaluating lymphocyte recovery and tumor response together,all (9/9) patients with both early lymphocyte recovery and goodtumor response are currently alive and 3/4 patients with earlylymphocyte recovery but poor tumor response are alive. When eval-uating the patients with late lymphocyte recovery, but good tumorresponse, 1/3 of these patients are alive, while the only patient whohad both late lymphocyte recovery and poor tumor response didnot survive. Two patients had multifocal disease at diagnosis didnot have surgery for local control and therefore did not have tumornecrosis data. Due to a limited number of events, multivariate sta-tistical analysis was unable to be performed. Comparative analysisusing a Cox proportional hazards regression model of potential riskfactors studied was utilized in a univariate fashion. Of all factorsanalyzed, only ALC-14 was significantly predictive with a relativerisk (RR) of 19.8 (P = 0.027) for OS and an RR of 5.9 (P = 0.062)for EFS (Table IV).

Only one patient presented with metastatic disease at diagno-sis (patient 7, lung) and two patients had multifocal bone disease(patients 5, 19). Of these three patients with very high risk factorsat diagnosis, the patient with lung metastases had early lymphocyterecovery and is alive and well, while the two patients with multifocaldisease had late recovery and both died of disease. Six patients sub-sequently developed relapsed disease with a 5-year OS of 33.3%(±19.2 SE). The majority (5/6) of patients with relapsed diseasedied while on active therapy. All deaths in this group had an ALC-14 < 800 cells/�l. The only survivor had early lymphocyte recovery(patient 18). He recently completed relapse therapy and is 3.7 yearssince diagnosis.

DISCUSSION

The role of lymphocytes in antitumor therapy has been anincreasing topic of discussion in current research. Ray-Coquardet al. [24] recently evaluated the prognostic value of lymphopeniaat diagnosis in patients with non-Hodgkin lymphoma, metastaticbreast cancer, and advanced soft tissue sarcoma. They found thatlymphopenia, defined as an ALC of <1,000 cells/�l at time of diag-

Pediatr Blood Cancer DOI 10.1002/pbc

1098 Moore et al.

TABLEI.DemographicandClinicalDataObtained

Prim

ary

Tum

orO

vera

llE

FSPa

tient

Age

Gen

der

tum

orT

reat

men

tM

etas

tatic

/mul

tifoc

alG

-CSF

WB

C14

AL

C14

Plt1

4A

NC

14ne

cros

issu

rviv

al(y

ears

)(y

ears

)Fi

rste

vent

Stat

us

110

.5Fe

mal

eFe

mur

MA

PN

oN

o0.

873

662

010

06.

41.

9R

elap

seD

ead

215

Mal

eFe

mur

MA

PN

oN

o1.

31,

235

7316

910

012

.212

.2A

live

312

.3M

ale

Fem

urM

AP

No

Yes

0.6

576

6412

100

11.4

11.4

Aliv

e4

6Fe

mal

eFe

mur

MA

PN

oN

o1.

41,

372

470

606.

36.

3A

live

517

.8M

ale

Mul

tifoc

alM

API

ER

ib,s

kull,

spin

eN

o3.

559

545

91,

820

a0.

60.

6D

ead

617

.8Fe

mal

eFe

mur

MA

PIE

No

Yes

2.1

798

7781

955

2.6

2.1

Rel

apse

Dea

d7

5.8

Fem

ale

Tib

iaM

AP

Lun

gY

es1.

599

049

420

100

8.8

8.8

Aliv

e8

17M

ale

Tib

iaM

AP

No

Yes

7.7

3,85

021

12,

618

508.

68.

6A

live

98.

1M

ale

Hum

erus

MA

PN

oY

es2.

31,

035

241

644

301.

20.

6R

elap

seD

ead

1013

.5Fe

mal

eT

ibia

MA

PIE

No

Yes

21.8

2,39

826

02,

398

100

8.2

8.2

Aliv

e11

8.3

Mal

eFi

bula

MA

PIE

No

Yes

1.6

912

6333

610

06.

06.

0A

live

1212

.3M

ale

Tib

iaM

AP

No

Yes

1.3

1,09

265

117

100

5.8

5.8

Aliv

e13

10.9

Fem

ale

Fem

urM

API

EN

oY

es3.

91,

853

112

1,05

310

07.

97.

9A

live

1416

.8M

ale

Tib

iaM

API

EN

oY

es40

.22,

814

219

33,7

6810

07.

87.

8A

live

1514

Fem

ale

Tib

iaM

API

EN

oY

es54

.92,

196

336

4,95

920

7.7

7.7

Aliv

e16

14.1

Fem

ale

Hum

erus

MA

PIE

No

Yes

4.3

688

129

2,92

499

2.8

1.4

Rel

apse

Dea

d17

14.9

Fem

ale

Fibu

laM

AP

No

Yes

3.2

1,71

848

582

997

4.1

4.1

Aliv

e18

16.5

Mal

eFe

mur

MA

PN

oY

es2.

41,

464

259

288

953.

71.

5R

elap

seA

live

1915

.1M

ale

Mul

tifoc

alM

AP

Tib

ia,i

liac

cres

tY

es1.

566

092

315

a1.

91.

4R

elap

seD

ead

MA

P,m

etho

trex

ate,

adri

amyc

in,

and

cisp

latin

;M

API

E,

met

hotr

exat

e,ad

riam

ycin

,ci

spla

tin,

ifos

fam

ide,

and

etop

osid

e;A

LC

and

AN

Car

elis

ted

asce

lls/�

l;W

BC

and

Plat

elet

are

liste

das

Kce

lls/�

l.a N

osu

rger

ype

rfor

med

.

Pediatr Blood Cancer DOI 10.1002/pbc

Lymphocyte Recovery in Osteosarcoma 1099

TABLE II. Evaluation of ALC-14 5-Year Overall Survival at Various Threshold Values

600 800 1,000 1,200 1,400

No. of patients above ALC-14 threshold 17 13 11 9 7Five-year OS 76.5% 92.3% 90.9% 100% 100%No. of patients below ALC-14 threshold 2 6 8 10 12Five-year OS 50.0% 33.3% 50.0% 50.0% 58.3%P-value 0.406 0.001 0.021 0.006 0.029

TABLE III. Univariate Analysis of Factors Associated With Survival in Osteosarcoma

Factor No. of patients (deaths) Five years overall survival (SE) Pa Five years event free survival (SE) P

Age (year)>14 10 (4) 60.0 (15.5) 0.40 50.0 (15.8) 0.23<14 9 (2) 88.9 (10.5) — 77.8 (13.8) —Gender

Male 9 (3) 70 (14.5) 0.90 60.0 (15.5) 0.62Female 10 (3) 77.8 (13.8) — 66.7 (15.7) —

ALC-14≥800 13 (1) 92.3 (7.4) 0.001 84.6 (10.0) 0.004<800 6 (5) 33.3 (19.2) — 16.7 (15.2) —

Tumor necrosis>90% 12 (2) 91.7 (8.0) 0.23 75.0 (12.5) 0.55<90% 5 (2) 60.0 (21.9) — 60.0 (21.9) —

G-CSFYes 15 (4) 73.3 (11.4) 0.39 66.0 (12.2) 0.52No 4 (2) 70.5 (21.7) — 50.0 (25.0) —

ChemotherapyMAP 11 (3) 81.8 (11.6) 0.68 63.6 (14.5) 0.97MAPIE 8 (3) 62.5 (17.1) — 62.5 (17.1) —

Platelets>75 12 (5) 58.3 (14.2) 0.21 85.7 (13.2) 0.12<75 7 (1) 100 — 50.0 (14.4) —

ANC≥500 9 (4) 60.0 (15.5) 0.42 60.0 (15.5) 0.70<500 10 (2) 88.9 (10.5) — 66.7 (15.7) —

aP-value from the log-rank test.

nosis, was highly predictive of survival in all three tumor typesstudied. Similar studies have evaluated not only the pre-therapy roleof the lymphocyte, but also its ongoing role throughout therapy.

Previous investigations into the role of lymphocyte recovery inpediatric cancers have shown similar prognostic indications. In eval-uating 24 high risk patients with Ewing Sarcoma, De Angulo et al.

[17] concluded that early lymphocyte recovery (>500 cells/�l) byday 15 was shown to be an independent factor in predicting improvedOS. These findings were confirmed by DuBois et al. [21] in theirevaluation of 41 patients with high risk Ewing sarcoma. In a sepa-rate study, De Angulo et al. [16] also showed significance for earlylymphocyte recovery in ALL and AML. These studies evaluated the

Fig. 1. Overall survival (A) and event free survival (B) for patients with early lymphocyte recovery (ALC-14 > 800, n = 13), versus late lymphocyterecovery (ALC-14 < 800, n = 6).

Pediatr Blood Cancer DOI 10.1002/pbc

1100 Moore et al.

Fig. 2. Overall survival (A) and event free survival (B) for patients with good histological response (tumor necrosis >90%, n = 11), versus poorresponse (<90%, n = 6).

level of lymphocyte recovery at approximately two weeks after ini-tiation of therapy, a concept that was utilized in this study. Previousstudies, however, have been unclear as to the underlying mecha-nism behind this phenomenon. It has been postulated that increasinginherent leukocyte antitumor activity in patients with malignancies,in addition to adjuvant chemotherapy, may play a large role in thisphenomenon [25–28].

Many studies have evaluated prognostic indicators in pediatricosteosarcoma. Some of the more traditional factors identified havebeen the presence or absence of metastatic disease at time of diagno-sis and histological response to chemotherapy by evaluating tumornecrosis at time of resection [6,7,10,11]. In the present study thepatient with metastatic disease at time of diagnosis had early lym-phocyte recovery and survived at 5 years. This survivor howeveronly had metastatic disease to the lungs, which has been shown tohave a superior outcome versus bony metastasis [3,6] Whether hostimmune functions play a role in the site of metastatic disease isunclear and would require larger studies for evaluation.

Despite a small sample size, our cohort is a fair representationof patients with both low and high risk osteosarcoma. Our analysisof other potential prognostic indicators showed no significant vari-able other than ALC-14. One way in which our cohort differed fromlarger studies was that tumor necrosis did not reach significance asan indicator of survival. This is likely due to the small sample size.There were only five patients with poor histologic response thatwere evaluable, making a comparison with those with good histo-logic response difficult. Previous studies have had a wide range ofboth OS and EFS for good responders (OS 77.8–94.7%; EFS 67.6–84.4%) versus poor responders (OS 51.6–56%; EFS 29.2–63%)[6,13,29,30]. Though our data did not reach significance, the OSand EFS of both our good (91.7% and 75%, respectively) and poor(60% for both) responders’ falls within the previously published

ranges. Because the trend of both the OS and EFS for histologicresponse in our study is similar to previous studies, we believe thisstrengthens the fact that the ALC-14 is a strong prognostic indicatorand deserves further study.

Improving the long-term survival for patients at higher risk, suchas those with late lymphocyte recovery, poor histologic response orrelapse, will require a novel therapeutic approach. New therapies inosteosarcoma that have been evaluated are various and include usingpeptides, cytokines, dendritic cells, tumor-infiltrating lymphocytes,L-muramyl tripeptide (L-MTP-PE), and 153Sm-EDTMP directedradiotherapy [4,31–33]. Recent studies have shown an increasingrole of natural killer (NK) cell and antitumor therapy. Wang et al.[28] evaluated the role of IL-21 in antitumor activity in melanomaand fibrosarcoma and concluded the effectiveness of the therapy wasbased on induced NK cell function. Similarly, Liu et al. [34] con-cluded that their investigational therapy of IFN-�-endostatin-basedgene radiotherapy showed an antitumor effect that was likely relatedto IFN-� stimulated cytotoxic T-lymphocyte and NK cell activa-tion. In addition, a recent study has evaluated the use of immunemodulation in patients with osteosarcoma with addition of adjuvantchemotherapy using MTP-PE, an agent that activates monocytesand macrophages with tumoricidal activity. Meyers et al. showed theaddition of MTP-PE to chemotherapy resulted in enhanced OS. Theaddition of MTP-PE to standard chemotherapy resulted in improve-ment in 6-year OS from 70% to 78% in their cohort [4]. Thesetherapies point out the benefit of incorporating the immune systeminto treatments and may explain why early lymphocyte recovery issignificant for improved survival.

Despite the limitations in statistical analysis, a trend can be seenwith improved OS in patients with both early lymphocyte recov-ery and good tumor response (9/9 alive) compared to those withearly lymphocyte recovery but poor tumor response (3/4 alive), late

TABLE IV. Comparative Analysis of ALC-14 Versus Tumor Necrosis Associated With Survival in Osteosarcoma

Overall survival Event free survivalFactor RR Pa CI RR P CI

ALC-14>800 1.0 — — 1.0 — —<800 19.8 0.027 1.42–275 5.5 0.062 0.93–33.2Tumor necrosis>90% 1.0 — — 1.0 — —<90% 8.1 0.10 0.665–99.4 1.5 0.639 0.26–9.26

aP-value from Cox’s proportional hazards regression model.

Pediatr Blood Cancer DOI 10.1002/pbc

Lymphocyte Recovery in Osteosarcoma 1101

lymphocyte recovery, but good tumor response (1/3 alive) or latelymphocyte recovery and poor tumor response (0/1 alive). By eval-uating patients using both lymphocyte recovery and tumor necrosiswe may be able to further stratify patients combining a known prog-nostic indicator and our novel indicator. Future large prospectivestudies should utilize all previously identified potential prognos-tic factors including tumor location, histological type, tumor size,histological response, and early lymphocyte recovery in order toestablish a risk stratification system for osteosarcoma.

Risk stratification after initiation of therapy is not a novel conceptin cancer therapies. The use of minimal residual disease is currentlybeing evaluated in patients being treated for acute lymphoblasticleukemia after its identification as a significant prognostic indicatorof OS and EFS in these patients [35–38]. Risk stratification prior toinitiation of therapy in other pediatric sarcoma patients has alreadybeen shown to provide many benefits to patients including decreasedtherapy for lower risk patients, without compromising survival [39].Future studies using this early risk stratification after initiation oftherapy in patients with osteosarcoma could potentially improvepatient outcome and quality of life. We realize the limitations ofsuch a small cohort. We were unable to show significance with tumornecrosis, the most widely accepted prognostic indicator in use today,but our trends follow other larger studies and therefore we believethe strong significance of the ALC-14 will prove to be extremelyuseful in developing a risk stratification system for osteosarcoma.Future studies with larger cohorts should be designed to assess themost prognostic level of the ALC-14 in order to provide the first stepin early risk stratification for these patients. As with other pediatriccancers, risk stratification then may lead to improved treatments forspecific subsets of patients with osteosarcoma.

REFERENCES

1. Singh GK, Yu SM. US childhood mortality, 1950 through 1993:Trends and socioeconomic differentials. Am J Public Health1996;86:505–512.

2. Mirabello L, Troisi RJ, Savage SA. International osteosarcoma inci-dence patterns in children and adolescents, middle ages and elderlypersons. Int J Cancer 2009;125:229–234.

3. Meyers PA, Heller G, Healey JH, et al. Osteogenic sarcoma withclinically detectable metastasis at initial presentation. J Clin Oncol1993;11:449–453.

4. Meyers PA, Schwartz CL, Krailo MD, et al. Osteosarcoma: Theaddition of muramyl tripeptide to chemotherapy improves overallsurvival—A report from the Children’s Oncology Group. J ClinOncol 2008;26:633–638.

5. Damron TA, Ward WG, Stewart A. Osteosarcoma, chondrosar-coma, and Ewing’s sarcoma: National Cancer Data Base Report.Clin Orthop Relat Res 2007;459:40–47.

6. Bielack SS, Kempf-Bielack B, Delling G, et al. Prognostic factorsin high-grade osteosarcoma of the extremities or trunk: An analysisof 1,702 patients treated on neoadjuvant cooperative osteosarcomastudy group protocols. J Clin Oncol 2002;20:776–790.

7. Bacci G, Ferrari S, Bertoni F, et al. Long-term outcome forpatients with nonmetastatic osteosarcoma of the extremity treatedat the istituto ortopedico rizzoli according to the istituto ortopedicorizzoli/osteosarcoma-2 protocol: An updated report. J Clin Oncol2000;18:4016–4027.

8. Bielack S, Carrle D, Casali PG. Osteosarcoma: ESMO clinical rec-ommendations for diagnosis, treatment and follow-up. Ann Oncol2009;20:137–139.

9. Chou AJ, Geller DS, Gorlick R. Therapy for osteosarcoma: Wheredo we go from here? Paediatr Drugs 2008;10:315–327.

10. Bramer JA, van Linge JH, Grimer RJ, et al. Prognostic factors inlocalized extremity osteosarcoma: A systematic review. Eur J SurgOncol 2009;35:1030–1036.

11. Pakos EE, Nearchou AD, Grimer RJ, et al. Prognostic factors andoutcomes for osteosarcoma: An international collaboration. Eur JCancer 2009;45:2367–2375.

12. Zhou H, Randall RL, Brothman AR, et al. Her-2/neu expres-sion in osteosarcoma increases risk of lung metastasis and canbe associated with gene amplification. J Pediatr Hematol Oncol2003;25:27–32.

13. Meyers PA, Heller G, Healey J, et al. Chemotherapy for non-metastatic osteogenic sarcoma: The Memorial Sloan-Ketteringexperience. J Clin Oncol 1992;10:5–15.

14. Provisor AJ, Ettinger LJ, Nachman JB, et al. Treatment of non-metastatic osteosarcoma of the extremity with preoperative andpostoperative chemotherapy: A report from the Children’s CancerGroup. J Clin Oncol 1997;15:76–84.

15. Bacci G, Forni C, Ferrari S, et al. Neoadjuvant chemotherapy forosteosarcoma of the extremity: Intensification of preoperative treat-ment does not increase the rate of good histologic response to theprimary tumor or improve the final outcome. J Pediatr HematolOncol 2003;25:845–853.

16. De Angulo G, Yuen C, Palla SL, et al. Absolute lymphocyte countis a novel prognostic indicator in ALL and AML: Implicationsfor risk stratification and future studies. Cancer 2008;112:407–415.

17. De Angulo G, Hernandez M, Morales-Arias J, et al. Early lympho-cyte recovery as a prognostic indicator for high-risk Ewing sarcoma.J Pediatr Hematol Oncol 2007;29:48–52.

18. Porrata LF, Gertz MA, Inwards DJ, et al. Early lymphocyte recoverypredicts superior survival after autologous hematopoietic stem celltransplantation in multiple myeloma or non-Hodgkin lymphoma.Blood 2001;98:579–585.

19. Porrata LF, Ingle JN, Litzow MR, et al. Prolonged survivalassociated with early lymphocyte recovery after autologoushematopoietic stem cell transplantation for patients with metastaticbreast cancer. Bone Marrow Transplant 2001;28:865–871.

20. Porrata LF, Litzow MR, Tefferi A, et al. Early lymphocyte recov-ery is a predictive factor for prolonged survival after autologoushematopoietic stem cell transplantation for acute myelogenousleukemia. Leukemia 2002;16:1311–1318.

21. DuBois SG, Elterman K, Grier HE. Early lymphocyte recovery inEwing sarcoma. J Pediatr Hematol Oncol 2007;29:351–352.

22. Bacci G, Ferrari S, Bertoni F, et al. Histologic responseof high-grade nonmetastatic osteosarcoma of the extremity tochemotherapy. Clin Orthop Relat Res 2001;386:186–196.

23. Whelan JWS, Uscinska B, et al. Localised extremity osteosarcoma:Mature survival data from two European Osteosarcoma Intergrouprandomized clinical trials. Proc Am Soc Clin 2000;19:1281a.

24. Ray-Coquard I, Cropet C, Van Glabbeke M, et al. Lymphopeniaas a prognostic factor for overall survival in advanced carcinomas,sarcomas, and lymphomas. Cancer Res 2009;69:5383–5391.

25. Castriconi R, Dondero A, Cilli M, et al. Human NK cell infu-sions prolong survival of metastatic human neuroblastoma-bearingNOD/scid mice. Cancer Immunol Immunother 2007;56:1733–1742.

26. Brentjens RJ, Santos E, Nikhamin Y, et al. Genetically targeted Tcells eradicate systemic acute lymphoblastic leukemia xenografts.Clin Cancer Res 2007;13:5426–5435.

27. Suck G. Novel approaches using natural killer cells in cancer ther-apy. Semin Cancer Biol 2006;16:412–418.

28. Wang G, Tschoi M, Spolski R, et al. In vivo antitumor activ-ity of interleukin 21 mediated by natural killer cells. Cancer Res2003;63:9016–9022.

Pediatr Blood Cancer DOI 10.1002/pbc

1102 Moore et al.

29. Fuchs N, Bielack SS, Epler D, et al. Long-term results of theco-operative German-Austrian-Swiss osteosarcoma study group’sprotocol COSS-86 of intensive multidrug chemotherapy andsurgery for osteosarcoma of the limbs. Ann Oncol 1998;9:893–899.

30. Lee JW, Kim H, Kang HJ, et al. Clinical characteristics and treat-ment results of pediatric osteosarcoma: The role of high dosechemotherapy with autologous stem cell transplantation. CancerRes Treat 2008;40:172–177.

31. Mori K, Redini F, Gouin F, et al. Osteosarcoma: Current sta-tus of immunotherapy and future trends (Review). Oncol Rep2006;15:693–700.

32. Anderson P, Kopp L, Anderson N, et al. Novel bone cancer drugs:Investigational agents and control paradigms for primary bone sar-comas (Ewing’s sarcoma and osteosarcoma). Expert Opin InvestigDrugs 2008;17:1703–1715.

33. Loeb DM, Garrett-Mayer E, Hobbs RF, et al. Dose-finding studyof 153Sm-EDTMP in patients with poor-prognosis osteosarcoma.Cancer 2009;115:2514–2522.

34. Liu LL, Smith MJ, Sun BS, et al. Combined IFN-gamma-endostatingene therapy and radiotherapy attenuates primary breast tumor

growth and lung metastases via enhanced CTL and NK cell acti-vation and attenuated tumor angiogenesis in a murine model. AnnSurg Oncol 2009;16:1403–1411.

35. Moghrabi A, Levy DE, Asselin B, et al. Results of the Dana-FarberCancer Institute ALL Consortium Protocol 95-01 for children withacute lymphoblastic leukemia. Blood 2007;109:896–904.

36. Borowitz MJ, Devidas M, Hunger SP, et al. Clinical significanceof minimal residual disease in childhood acute lymphoblasticleukemia and its relationship to other prognostic factors: A Chil-dren’s Oncology Group study. Blood 2008;111:5477–5485.

37. Flohr T, Schrauder A, Cazzaniga G, et al. Minimal residual disease-directed risk stratification using real-time quantitative PCR analysisof immunoglobulin and T-cell receptor gene rearrangements in theinternational multicenter trial AIEOP-BFM ALL 2000 for child-hood acute lymphoblastic leukemia. Leukemia 2008;22:771–782.

38. Vrooman LM, Silverman LB. Childhood acute lymphoblas-tic leukemia: Update on prognostic factors. Curr Opin Pediatr2009;21:1–8.

39. Dantonello TM, Int-Veen C, Harms D, et al. Cooperative trial CWS-91 for localized soft tissue sarcoma in children, adolescents, andyoung adults. J Clin Oncol 2009;27:1446–1455.

Pediatr Blood Cancer DOI 10.1002/pbc