Neuroblastoma detection using serum proteomic profiling:
a novel mining technique for cancer?
John A. Sandovalb, Lacey E. Dobroleckid, Jeffrey Huangc, Jay L. Grosfelda,
Robert J. Hickeyd, Linda H. Malkasd,*
aSection of Pediatric Surgery, Indiana University School of Medicine and Riley Children’s Hospital, Indianapolis,
IN 46202, USA
bDepartment of Surgery, Indiana University School of Medicine and Riley Children’s Hospital, Indianapolis, IN 46202, USA
cDepartment of Bioinformatics, Indiana University of Informatics, Indiana University School of Medicine and Riley Children’s
Hospital, Indianapolis, IN 46202, USA
dDivision of Hematology/Oncology, Indiana University School of Medicine and Riley Children’s Hospital, Indianapolis,
IN 46202, USA
Background: Serum proteins in neuroblastoma (NB), such as neuron-specific enolase and lactate
dehydrogenase, are used as nonspecific markers of disease severity. In this study, we have generated
serum protein profiles that correlate with NB by applying proteomic technologies to uncover, at the
protein level, serum polypeptide expression patterns in patients with NB.
Methods: Surface-enhanced laser desorption/ionization–time-of-flight mass spectrometry was used to
generate protein expression spectra in human NB (group I, n = 18) and healthy children (group II, n = 17)
sera. Groups I and II mass spectral data were compared after baseline subtraction. Peaks with high
signal-to-noise ratios were selected and grouped into bins with various intervals along mass-to-charge
axis. Two-sample t test and 3-fold cross validation were used to identify differential biomarkers between
groups I and II.
Results: Significant differentially expressed proteins were identified between groups I and II (P V.05).
The discriminatory features (proteomic patterns) of cancer from normal sera were successfully identified
using the classification algorithm. The average classification performance after 3-fold cross validation
Conclusion: Surface-enhanced laser desorption/ionization–time-of-flight mass spectrometry is suitable
for preliminary assessment of NB and could potentially provide a noninvasive diagnosis of NB. We
propose that surface-enhanced laser desorption/ionization provides a novel method for NB diagnosis
because direct observations of spectral differences between normal and NB sera are possible.
D 2006 Elsevier Inc. All rights reserved.
0022-3468/$ – see front matter D 2006 Elsevier Inc. All rights reserved.
Presented at the 57th Annual Meeting of the Section on Surgery of the American Academy of Pediatrics, Washington, DC, October 7-9, 2005.
*Corresponding author. Division of Hematology/Oncology, Indiana University Cancer Research Institute, Indianapolis, IN 46202, USA. Tel.: +1 317
278 8260; fax: +1 317 274 8046.
E-mail address: email@example.com (L.H. Malkas).
Journal of Pediatric Surgery (2006) 41, 639–646
Early detection of neuroblastoma (NB) has the promise
of influencing the survival of children with advanced-stage
disease. Mass screening programs were initiated in the
1970s as an early detection method aimed at detecting
and treating NB, which would otherwise present at an
advanced stage in older children. However, the overall
impact of these programs has shown limited success in
reducing mortality [1-3]. To effectively impact the death
rate associated with NB, it is necessary to continue
searching for methods of early diagnosis for this child-
A prewarning system for NB ideally includes the
identification of molecular markers that qualitatively and
unequivocally differentiate cancer from normal cells. As
such, a wealth of biologic information is available
concerning molecular alterations that occur in NB. Conven-
tional genetic prognostic markers include MYCN copy
number, TRKA, TRKB, telomerase, and CD44 [4-8]. A host
of other differentially expressed genes in advanced NB are
related to patient prognosis and have been reviewed by
others [9,10]. Moreover, the development of DNA array
technologies has had an enormous impact in the identifica-
tion of numerous genes and in understanding the biochem-
ical role of gene products that are expressed in NB.
However, the biologic significance of almost all differen-
tially expressed genes remains to be demonstrated.
A complementary array technology to DNA in cancer
investigation is protein profiling. Because proteins derived
from gene sequences are ultimately responsible for cellular
biologic function, characterizing the protein complement,
or proteome, of a cell, tissue, or fluid has been proposed
to be a significant component in cancer research .
Our laboratory has an active effort in evaluating proteomic
profiles in NB using 2D-PAGE and mass spectrometry
(MS). Escobar et al  profiled nuclear proteins among
NB cell lines and Sandoval et al  showed novel proteins
excreted from NB. Moreover, because serum is an
important archive of proteins that are shed or secreted in
disease, it is becoming clear that serum proteins are a
valuable source of information that may signal differences
between cancer and disease-free individuals. As a result,
studies have corroborated the existence of serum protein
patterns that can differentiate several human cancers from
One attractive feature of clinical proteomics is the ability
to analyze easily obtained biofluids such as serum and urine.
Using current improvements in protein analyzing technol-
ogies, an innovative approach to biofluid proteome inves-
tigation is known as surface-enhanced laser desorption/
ionization–time-of-flight MS (SELDI–TOF MS). The tech-
nique of SELDI–TOF MS combines the principle of protein
chip retention chromatography with MS [19-21]. The result
is a profile of a population of proteins in a sample according
to the molecular weight and the net electrical charge of
the individual proteins. With regards to cancer, SELDI–TOF
MS approaches have successfully found new biomarkers
and achieved high sensitivity and specificity for the
diagnosis of the following malignancies: bladder, prostate,
ovary, breast, liver, neck, lung, and pancreas [22-31].
The purpose of this study was to determine whether
SELDI–TOF MS–based proteomic approaches could be
used to identify a serum protein signature in patients with
NB compared with healthy children. Our hope is to generate
an NB serum profile using novel serum proteomic
technologies that may potentially contribute to a quick and
1. Materials and methods
Subjects were selected based on criteria approved by the
Indiana University School of Medicine Institutional Review
Board. A total of 35 sera specimens were prospectively
collected. Group I (n = 18) consisted of children either
diagnosed with NB or completing medical treatment
(chemotherapy or radiation) for NB. Group II (n = 17)
constituted control samples collected from healthy children.
This group consisted of children who underwent an out-
patient surgical procedure (thyroglossal duct cyst removal,
hemangioma excision or chipexy, inguinal hernia repair,
hydrocele repair, and cutaneous lesion excision (lipomas,
dermoids, scar revisions, etc]) at the JW Riley Hospital for
Children. Any evidence of prior malignancy (other than
NB) or active/chronic infection classified as exclusion
criteria for both groups in this study. Three milliliters of
blood were collected in red top tubes from both groups and
immediately centrifuged at 2500 rpm for 15 minutes. Serum
(1 mL) was separated into 100-lL aliquots and stored at
?808C until surface-enhanced laser desorption/ionization
1.2. Surface-enhanced laser desorption/
ionization–time-of-flight mass spectrometry
serum proteomic analysis
On the day of the analysis, the serum was placed on ice
and thawed completely. The samples were centrifuged
at 13,200 rpm for 10 minutes at 48C. Any samples that
were hemolyzed or greatly lipemic were not used for
analysis. All of the samples were processed in duplicate on
immobilized metal affinity (IMAC30) chips bound with
copper. After testing several other chip types, these were
found to produce the most robust spectra. The IMAC30
chips were placed in a ProteinChip bioprocessor and 50 lL
of 0.1mol/L CuSO4 was incubated on the chips for
5 minutes with vigorous shaking. Then, the chips were
rinsed with deionized water and neutralized for 5 minutes
with 50 lL of 0.1 mol/L sodium acetate (pH 4.0). After
another rinse with deionized water, the chips were
equilibrated by washing twice with 150 lL of binding
buffer (1 ? PBS + 0.3 mol/L NaCl, pH 7.4). Next, the
J.A. Sandoval et al. 640
chips were loaded with 90 lL of binding buffer to which
10 lL of serum was added. The serum was incubated
with the chips for 1 hour with vigorous shaking. After
an hour, the serum was removed and the chips were
washed 3 times with 150 lL of binding buffer. A final
rinse with deionized water was performed, the bioprocessor
was removed, and the chips were allowed to air dry.
Once the chips were dry, 0.7 lL of a 50% solution of
a-cyano-5-hydroxycinnamic acid diluted in 50% (vol/vol)
acetonitrile and 0.5% trifluoroacetic acid (vol/vol) was
applied to each spot. The a-cyano-5-hydroxycinnamic acid
solution was allowed to dry completely before it was added
for a second time.
1.3. Data processing and analysis
Raw SELDI mass spectral data, obtained from Groups I
and II, were corrected by baseline subtraction and filtering,
which are used to facilitate removal of background noise
from the spectrum and low-intensity signals occurring
frequently, respectively. Candidate peaks were selected
based on the local maximum at which signal-to-noise
(S/N) ratios are significant. Depending on the mass
sensitivity of SELDI (eg, 0.2%), peaks across different
samples within this range are aligned by merging them into
a bin, and the highest intensity within the bin was used as
representation [32,33]. A 2-samplet test was performed, and
P values were computed to determine the significance of
discriminating power for each biomarker based with respect
to a constant value of .05 (P V .05). Three-fold cross-
validation (data set was divided into 3 subsets, and the
experiment was repeated 3 times) using the nearest neighbor
classifier was performed to evaluate the prediction power of
our proposed method.
We analyzed the simultaneous relative abundance of
many protein/peptide sequences in serum using high-
throughput SELDI–TOF MS from a total of 35 patient
samples between NB (n = 18) and healthy children (n = 17).
Neoroblastoma patient demographics are shown in Table 1.
In our NB patient population, the median age was 24.6 F
17.9 months (range, 2-70 months) with the male-to-female
ratio of 1:1.5. Advanced-stage NB (stage III and IV)
comprised 77.7% of cases and 4 patients (22.2%) were
low-risk NB (I, IIB, and IVs). Most tumors originated from
the adrenal gland (77.7%), whereas other primary tumor
sites included the pelvis (11.1%) and thoracic cavity
(11.1%). Forty-four percent of sera analyzed was before
chemotherapy (prechemotherapy samples), 44.4% had
completed chemotherapy treatment (postchemotherapy
samples), and 11.1% of NB samples evaluated had no
chemotherapy as part of the treatment algorithm. The
non-diseased control population was collected from healthy
children with a median age of 33.7 F 43.3 months
(range, 1-168 months) and male-to-female ratio 1:1.4 (data
The SELDI spectral output from NB and healthy
samples were standardized and evaluated following our
algorithm to select mass-to-charge (m/z) values that
correspond to local peaks in the spectra. We evaluated a
large molecular weight range (50-10,000 d) for potential
serum biomarkers. Fig 1A shows the difference of averaged
intensities between NB and control samples from 3 divided
molecular weight ranges: 50 to 2500 d (top panel), 2500 to
5000 d (middle panel), and 5000 to 10000 d (bottom
panel). In each panel, we observe 2 views of the averaged
mass spectra data; a mass chromatogram view, in which a
plot of the ion current intensity over of a range of m/z
values is shown as peaks; or a gel-like display, a gray scale
representation of the mass spectra data simulating electro-
phoretic separation. Comparing the NB and nondisease
spectra, we discern pattern peak differences as the
molecular weight range increases from 50 to 10,000 d. In
addition, Fig. 1B shows a representative mass spectral (top)
and gel (bottom) views of 5 differentially proteins (arrows)
identified in NB and control SELDI data analyzed between
the molecular mass ranges of 2750 to 3000 d. The
overexpression of 4 proteins in the NB sera relative to
the control samples and 1 protein underexpressed in the NB
samples, as compared with controls, are evident in the mass
spectral and gel views.
The mass spectral view in Fig. 2 shows a mass
chromatogram (top) and gel (bottom) view of a molecular
weight protein peak at 4634 d between 4 experimental NB
samples and 4 controls. One clearly observes the expression
of this protein in the 4 control samples and the relative
absence of this protein in NB. With regard to this under-
expressed protein in NB, the t value of this protein area
(m/z range, 4625-4670) is a negative value of ?3.3814 and
Neuroblastoma (NB) patient demographic information
Age (mo) Sex StagePrimary
Pre or post
Neuroblastoma detection using serum proteomic profiling641
molecular weight ranges of 50 to 2500 d (top), 2500 to 5000 d (middle), and 5000 to 10,000 d (bottom). In each panel, the intensity of the
proteins is represented as a mass chromatogram view or a gray-scale image (gel view) image of SELDI MS data. (B) A SELDI MS peak trace
shows the relative intensity vs m/z of detected proteins between 2750 to 3000 d from NB and control samples. Both the mass spectral view
and the gel-like representations show 4 protein peaks that were up-regulated in NB (arrows over blue NB peaks) and 1 other peak that was
underexpressed as compared with control (arrow over red control peak).
(A) Average SELDI–TOF MS patterns of NB and control subject sera. Three panels of mass spectra data are shown from the
J.A. Sandoval et al.642
P value of .00093, which is statistically significant in
discriminating 2 sample groups.
Three-fold crossvalidation was performed to determine
the predictive accuracy of the proteomic pattern generated
by our biomarker detection algorithm, the average classifi-
cation accuracy with 87.26% of cases in the test sample
being correctly assigned to the correct corresponding groups
based on protein profile, whereas the specificity and
sensitivity are 80.05% and 94.44%, respectively, based on
the method of nearest neighboring classifier. Both sensitiv-
ity and specificity are relatively high, which shows our
proposed method is robust in differentiating both NB and
normal sample groups.
Substantial gaps exist in the early detection of
advanced-stage NB. Efforts to identify high risk NB
by current screening techniques have been unsuccessful
in altering the natural course of disease. Thus, a major
challenge is finding reliable and cost-effective methods
that distinguish NB subpopulations requiring aggressive
treatment from low-risk patients. Recent innovations in
proteomic-based technologies have had a significant
impact on the approaches to identify protein biomarkers
in cancer. Because of the clinical relevance of serum
proteins to cancer, serum proteomic analysis is a prom-
ising approach in the identification of panels of proteins
that may be indicative for malignancy. The primary
objective of this study was to investigate whether serum
protein profiling using SELDI–TOF MS was a feasible
approach to differentiate sera of patients with NB from
We identified differentially expressed proteins between
NB (group I, n = 18) and nondiseased children (group II,
n = 17) (P V .05). The use of the proposed statistical
analysis methods to identify an informative set of potential
protein markers provided a rigorous evaluation to eliminate
random and systemic errors associated with biomarker
discovery methodologies. Although we are not the first to
describe SELDI serum profiling in NB, our results are
important in that we are the first to report a representative
NB serum profile over a large molecular weight range
(50-10,000 d) .
Because early NB mass screening trials have not
demonstrated the ability to decrease mortality and adverse
related health effects, there has been reconsideration of the
usefulness of such screening programs for NB [35,36].
Moreover, because research for NB screening has not been
generally advocated for, investigation into early detection
methods has received considerable support . One critical
issue regarding NB screening is related to catecholamine use
as a single marker for early diagnosis. An important
consideration in single marker measurement for disease is
whether the quantification of a single biomarker is
sufficiently sensitive and specific for screening in the
general public . In this regard, proteomic serum
profiling as a cancer detection strategy is a departure from
the traditional assay modality of measuring a specific tumor
analyte. By identifying multiple serum proteins that can
serve as a signature for NB, a panel of independent
NB-related proteins is less susceptible to the influence of
genetic and environmental factors than the level of a single
marker protein. Because NB heterogeneity and genetic
differences between individuals tend to confound what
might be clear disease associations, multiple markers
affected by the disease may provide a better indication of
biomarker in NB. The shaded area in the mass spectral view and the outlined area in the gel view correspond to peak 4634; the panels show
the up-regulation of protein 4634 kd in control samples and down-regulation of this protein in NB. The t-value assigned to this protein is
?3.3814 correlates with the lower abundance of this protein in NB.
Differential expression of 4634-Da peak in NB and controls. Representative spectra (left) and gel views (right) of a selected serum
Neuroblastoma detection using serum proteomic profiling 643
a malignant condition. Our data are significant in that we
show multiple distinct SELDI sera proteins between
children with NB vs healthy controls that may be uniquely
expressed in the circulation.
Nevertheless, despite progress and support of serum
profiling methods, a major obstacle facing discovery-based
serum proteomics is whether SELDI-based technology can
accurately and reliably diagnose disease states. Major
controversies are related to data reproducibility and validity
generated from serum proteomics studies [39-41]. Another
area of concern relates to the inability of SELDI to identify
the marker protein behind the mass peak. Because data
from the National Cancer Institute/Early Detection
Research Network have shown support for the reproduc-
ibility emerging from SELDI sources, a task for the future
of SELDI serum profiling is addressing current short-
comings through standardized protocols and advancing
SELDI sources with high-performance MS peptide sequenc-
ing analyzers [42-44]. With regard to our data, we plan
reproducibility/validation studies on larger NB cohorts,
with separate experiments between different NB stages, and
the implementation of traditional proteomic methods
consistent with affinity chromatography, 2D-PAGE, and
liquid chromatography linked tandem MS to identify
individual marker proteins.
In summary, we show promising initial results in
identifying a serum proteomic signature for NB. Using a
ProteinChip array system and bioinformatics approaches to
profile and compare a large number of serum proteins in
NB and healthy children, we defined a panel of biomarkers
that could differentiate the presence of NB from non-
disease. Because this information will be used to create a
panel of biomarkers needed to unambiguously identify
patients with NB, a foreseeable progression will be to
identify the serum proteins that compose the NB protein
panel through traditional proteomic techniques. Currently,
we are in the process of evaluating a larger NB patient
sera cohort with the inclusion of other solid childhood
malignancies such as Wilms’ tumor, rhabdomyosarcoma,
and osteosarcoma to validate the biomarker patterns
discerned from these analyses. Because recent strides in
serum proteomic technologies show substantial promise in
the ability to detect a variety of malignancies, the future
of early high-risk NB detection may include a nonin-
vasive blood test based on pattern recognition serum
This work was supported in part by research award
CA57350 and CA83199 from the National Institutes of
Health to LHM. The authors thank both the Vera Bradley
Foundation and the ANNA foundation (Indianapolis, Ind)
for their continuous support of our research.
 Woods WG, Gao RN, Shuster JJ, et al. Screening of infants and
mortality due to neuroblastoma. N Engl J Med 2002;346(14):1041-6.
 Schilling FH, Spix C, Berthold F, et al. Neuroblastoma screening at
one year of age. N Engl J Med 2002;346(14):1047-53.
 Yamamoto K, Ohta S, Ito E, et al. Marginal decrease in mortality and
marked increase in incidence as a result of neuroblastoma screening at
6 months of age: cohort study in seven prefectures in Japan. J Clin
 Brodeur GM, Seeger RC, Schwab M, et al. Amplification of N-myc in
untreated human neuroblastomas correlates with advanced disease
stage. Science 1984;224(4653):1121-4.
 Nakagawara A, Arima-Nakagawara M, Scavarda NJ, et al. Associ-
ation between high levels of expression of the TRK gene and
favorable outcome in human neuroblastoma. N Engl J Med 1993;
 Nakagawara A, Azar CG, Scavarda NJ, et al. Expression and function
of TRK-B and BDNF in human neuroblastomas. Mol Cell Biol 1994;
 Hiyama E, Hiyama K, Yokoyama T, et al. Correlating telomerase
activity levels with human neuroblastoma outcomes. Nat Med 1995;
 Favrot MC, Combaret V, Lasset C. CD44—a new prognostic marker
for neuroblastoma. N Engl J Med 1993;329(26):1965.
 Ohira M, Oba S, Nakamura Y, et al. A review of DNA microarray
analysis of human neuroblastomas. Cancer Lett 2005;228:1-5.
 Takita J, Ishii M, Tsutsumi S, et al. Gene expression profiling and
identification of novel prognostic marker genes in neuroblastoma.
Genes Chromosomes Cancer 2004;40(2):120-32.
 MacNeil JS. Better biomarkers for the diagnostics labyrinth. Genome
 Escobar MA, Hoelz DJ, Sandoval JA, et al. Profiling of nuclear
extract proteins from human neuroblastoma cell lines: the search for
fingerprints. J Pediatr Surg 2005;40(2):349-58.
 Sandoval JA, Hoelz DJ, Woodruff HA, et al. Novel peptides
secreted from human neuroblastoma: useful clinical tools? J Pediatr
 Mobley JA, Lam YW, Lau KM, et al. Monitoring the serological
proteome: the latest modality in prostate cancer detection. J Urol
 Feldman AL, Espina V, Petricoin III EF, et al. Use of proteomic
patterns to screen for gastrointestinal malignancies. Surgery
 Poon TC, Yip TT, Chan AT, et al. Comprehensive proteomic profiling
identifies serum proteomic signatures for detection of hepatocellular
carcinoma and its subtypes. Clin Chem 2003;49(5):752-60.
 Li J, Zhang Z, Rosenzweig J, et al. Proteomics and bioinformatics
approaches for identification of serum biomarkers to detect breast
cancer. Clin Chem 2002;48(8):1296-304.
 Petricoin EF, Ardekani AM, Hitt BA, et al. Use of proteomic patterns
in serum to identify ovarian cancer. Lancet 2002;359(9306):572-7.
 Hutchens TW, Yip TT. New desorption strategies for the mass
spectrometric analysis of macromolecules. Rapid Commun Mass
 Weinberger SR, Dalmasso FA, Fung ET. Current achievements using
ProteinChip Array Technology. Curr Opin Chem Biol 2001;6:86-91.
 Weisner A. Detection of tumor markers with ProteinChip technology.
Curr Pharm Biotechnol 2004;5:45-67.
 Vlahou A, Schellhammer PF, Mendrinos S, et al. Development of a
novel proteomic approach for the detection of transitional cell
carcinoma of the bladder in urine. Am J Pathol 2001;158:1491-501.
 Adam BL, Qu Y, Davis JW, et al. Serum protein fingerprinting
coupled with a pattern-matching algorithm distinguishes prostate
cancer from benign prostate hyperplasia and healthy men. Cancer Res
J.A. Sandoval et al.644
 Wright Jr G, Cazares LH, Leung SM, et al. Proteinchip (R) surface
enhanced laser desorption/ionization (SELDI) mass spectrometry: a
novel protein biochip technology for detection of prostate cancer
biomarkers in complex protein mixtures. Prostate Cancer Prostatic Dis
 Petricoin EF, Ardekani AM, Hitt BA, et al. Use of proteomic patterns
in serum to identify ovarian cancer. Lancet 2002;359:572-7.
 Vlahou A, Schorge JO, Gregory BW, et al. Diagnosis of ovarian
cancer using decision tree classification of mass spectral data.
J Biomed Biotechnol 2003;2003:308-14.
 Vlahou A, Laronga C, Wilson L, et al. A novel approach
toward development of a rapid blood test for breast cancer. Clin
Breast Cancer 2003;4:203-9.
 Poon TC, Yip TT, Chan AT, et al. Comprehensive proteomic
profiling identifies serum proteomic signatures for detection
of hepatocellular carcinoma and its subtypes. Clin Chem 2003;
 Wadsworth JT, Somers KD, Stack Jr BC, et al. Identification of
patients with head and neck cancer using serum protein profiles. Arch
Otolaryngol Head Neck Surg 2004;130:98-104.
 Zhukov TA, Johanson RA, Cantor AB, et al. Discovery of
distinct protein profiles specific for lung tumors and pre-
malignant lung lesions by SELDI mass spectrometry. Lung
 Koopmann J, Zhang Z, White N, et al. Serum diagnosis of pancreatic
adenocarcinoma using surface-enhanced laser desorption and ioniza-
tion mass spectrometry. Clin Cancer Res 2004;10:860-8.
 Baggerly KA, Morris JS, Wang J, et al. A comprehensive approach to
the analysis of matrix-assisted laser desorption/ionization–time of
flight proteomics spectra from serum samples. Proteomics
 Coombes KR, Fritsche Jr HA, Clarke C, et al. Quality control and
peak finding for proteomics data collected from nipple aspirate fluid
by surface-enhanced laser desorption and ionization. Clin Chem
 Combaret V, Bergeron C, Brejon S, et al. Protein chip array profiling
analysis of sera from neuroblastoma patients. Cancer Lett 2005;228:
 Woods WG, Tuchman M, Robison LL, et al. A population-based
study of the usefulness of screening for neuroblastoma. Lancet
 Tsubono Y, Hisamichi S. A halt to neuroblastoma screening in Japan.
N Engl J Med 2004;350(19):2010-1.
 Philip T. Early detection of neuroblastoma in infants: research? Yes.
Routine screening? No. Med Pediatr Oncol 1999;33(4):355-6.
 Bates SE, Longo DL. Tumor markers: value and limitations in
the management of cancer patients. Cancer Treat Rev 1985;12(3):
 Diamandis EP. Analysis of serum proteomic patterns for early cancer
diagnosis: drawing attention to potential problems. J Natl Cancer Inst
 Baggerly KA, Morris JS, Edmonson SR, et al. Signal in noise:
evaluating reported reproducibility of serum proteomic tests for
ovarian cancer. J Natl Cancer Inst 2005;97(4):307-9.
 Liotta LA, Lowenthal M, Mehta A, et al. Importance of communi-
cation between producers and consumers of publicly available
experimental data. J Natl Cancer Inst 2005;97(4):310-4.
 Semmes OJ, Feng Z, Adam BL, et al. Evaluation of serum protein pro-
filing by surface-enhanced laser desorption/ionization time-of-flight
mass spectrometry for the detection of prostate cancer: I. Assessment
of platform reproducibility. Clin Chem 2005;51(1):102-12.
 Tang N, Tornatore P, Weinberger SR. Current developments in SELDI
affinity technology. Mass Spectrom Rev 2004;23(1):34-44.
 Kwapiszewska G, Meyer M, Bogumil R, et al. Identification of
proteins in laser-microdissected small cell numbers by SELDI-TOF
and tandem MS. BMC Biotechnol 2004;4(1):30.
Pat Donahoe, MD (Boston, MA): A lovely presentation and
a very powerful tool for the future. You have identified
the presence of 5 proteins. Can you generate sequence
data from which to identify them?
John A. Sandoval, MD (response): The protein profiles
shown here correspond to a subset of proteins within the
serum that collectively constitute a pattern that differ-
entiates neuroblastoma from healthy children. That being
said, many of the small proteins and peptides that form
these patterns may be fragments of larger proteins. One
of the limitations of the mass spectrometry SELDI data is
that the identification of the proteins or protein fragments
that produce these spectra is difficult, as even accurate
mass measurements cannot unambiguously identify a
protein. Nevertheless, the serum protein patterns them-
selves are considered to be a molecular fingerprint for
disease, even in the absence of identification of the
Andrew Davidoff, MD (Memphis, TN): But do you think
that any of those peaks are LDH or ferritin? This might
serve as a sort of positive control of your technique.
John A. Sandoval, MD (response): We evaluated a low
molecular weight range in this study from 10 to 50 d.
Lactate dehydrogenase has a molecular weight of approx-
imately 140,000 d, and ferritin has a molecular weight of
approximately 450,000 d. Both of these intact proteins are
obviously out of range to be detected by our analysis. But
serum preparation involves the activation of proteolytic
cascades for blood clotting, which allows small proteins
and peptides fragments to be released from parent
proteins. So, theoretically, we may not have captured the
complete polypeptides of LDH or ferritin but isolated
protein fragments from these parent molecules.
Agostino Pierro, MD (London, UK): I enjoyed your paper
very much. In the third conclusion, you said that you
could use this novel technique to see the response to
treatment. One of the factors that we commonly see in
neuroblastoma is that sometimes after chemotherapy for
stage III or IV disease, we find only calcified tissue with
no viable tumor. Have you looked at the pattern of
protein expression before and after treatment in the same
patients to see if there is a difference?
John A. Sandoval, MD (response): We actually had two
patients that we were able to follow prior to and after
chemotherapy. But we did not compare the mass spectra
between these samples. One of the forthcoming goals we
will accomplish is comparing proteomic changes from
the sera of chemotherapy from patients with advanced
stage neuroblastoma. The data generated from these
Neuroblastoma detection using serum proteomic profiling645
studies may be useful in monitoring the emergence of Download full-text
David Hackam, MD (Pittsburgh, PA): I very much enjoyed
your presentation. I have a question regarding the gold
standard, to which this assay would be compared. At the
beginning of your presentation, you discussed various
biomarkers related to the tumor itself that are important
prognosticators. I wonder whether you may determine
whether known tumor markers that are typically associ-
ated with either a good or bad prognosis may be detected
in the serum, and whether these may be measured using
the MALDI-TOFF assay? Have you measured, for
instance, the release of tumor-derived n-myc in the
serum of patients with tumors versus control patients,
and have you correlated this with outcome?
John A. Sandoval, MD (response): In this study, our aim
was to determine whether molecular profiling of the
serum proteome could discriminate between neuroblas-
toma and healthy children. We did not attempt to
correlate genetic markers with our protein patterns. As
alluded to earlier, we are assimilating a larger sample
size to distinguish between early- and late-stage neuro-
blastoma and predict disease progression. In the valida-
tion stages of this biomarker-based work, we will
plan correlative studies with gold standard markers
J.A. Sandoval et al.646