| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regular Article |
From the Institute of Pathology,*
the Department of
Surgery,
Universitäts-Frauenklinik,
University of
Basel, Basel, Switzerland; the Cancer Genetics
Branch,
National Human Genome Research
Institute, National Institutes of Health, Bethesda, Maryland;
Kreiskrankenhaus,¶
Lörrach, Germany; and
Frauenklinik,||
Rheinfelden, Germany
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The translation of basic research findings to clinical applications is now becoming dramatically more challenging, with the introduction of high-throughput genomics and proteomics technologies.9 For example, in a single cDNA microarray experiment, one is able to determine the expression status of 50,000 human genes. These technologies often require fresh tissues, which makes it difficult to directly apply them in clinical studies. Formalin-fixed archival tissues provide a means to validate such genomic and proteomic screening in large sets of histologically well-characterized tumors with clinical endpoints. However, testing of even a small fraction of the human gene and protein targets is beyond the scope of traditional molecular pathology technologies. Not only are these techniques slow and tedious, but the availability of tissue is often rate-limiting. For example, one can only cut at most 300 sections from a typical archival tissue block. In the case of smaller tumors, or previously used precious research materials, the number of sections is often much smaller.
Our recently developed tissue microarray (TMA) technology has the potential to significantly accelerate studies seeking for associations between molecular changes and clinical endpoints.10 In this technology, 0.6 mm tissue cylinders are punched from hundreds of different primary tumor blocks and subsequently brought into a recipient tissue microarray block. Sections from such array blocks can then be used for simultaneous in situ analysis of hundreds or thousands of primary tumors on DNA, RNA, and protein level. The high speed of arraying, the lack of a significant damage to donor blocks, and the regular arrangement of arrayed specimens greatly facilitating automated analysis are the most significant advantages of the TMA technology over previous concepts of analyzing multiple different tissues in one paraffin block.11 To test the utility of our "tissue chip" approach for finding associations between molecular changes and clinical endpoints, we used breast cancer as a model system. The TMA analysis of the well-established prognostic markers ER and PR as well as of p53, another suggested prognostic parameter, suggests that tissue chips provide a means for rapid screening of the prognostic significance of molecular markers and may help to translate genomic and proteomics information to clinical applications.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Samples from 611 breast carcinomas had previously been included in a breast cancer TMA. The carcinomas of 553 patients of which follow-up data (tumor-specific survival and treatment information) could retrospectively be evaluated were included in this study. These patients had a median age of 61 (range, 33 to 97) years. They were treated for primary breast cancer at the University Hospital in Basel (Switzerland), Women’s Hospital Rheinfelden (Germany), and the Kreiskrankenhaus Lörrach (Germany) between 1985 and 1994. The mean follow-up time was 65.8 months (range, 1 to 151). A systemic therapy had been performed in 273 patients including 172 with hormonal therapy alone, 52 with cytotoxic therapy alone, and 49 having both hormonal and cytotoxic treatment. Formalin-fixed, paraffin-embedded tumor material was available from the Institute of Pathology, University of Basel. The pathological stage, tumor diameter, and nodal status were obtained from the primary pathology reports. All slides from all tumors were reviewed by one pathologist (J.T.) to define the histological grade according to Elston and Ellis12 and the histological tumor type. The series included 405 ductal, 77 lobular, 16 medullary, 14 mucinous, 11 cribriform, 11 tubular, 7 papillary, 4 apocrine, 3 clear cell, 1 metaplastic, 1 atypical medullar, 1 large cell, 1 small cell, and 1 neuroendocrine cancer. Among 553 tumors, 27.8% were grade 1, 42.9% were grade 2, and 29.3% were grade 3. The local tumor stage was pT1 in 39.5%, pT2 in 46.3%, pT3 in 4.9%, and pT4 in 9.3%. The stage could not be unequivocally determined from the pathology reports in 6 tumors. Axillary lymph nodes had been examined in 519 patients (the nodal stage (pN) was: 52.4% pN0, 39.3% pN1, and 8.3% pN2). Stage, grade, and nodal status were strongly associated with tumor-specific survival of our patients (p < 0.0001 each).
Tissue Microarray Construction
Tumor samples were arrayed as previously described.10
Briefly, H & E-stained sections were made from each block to
define representative tumor regions. Tissue cylinders with a diameter
of 0.6 mm were then punched from selected areas of each "donor"
block using a custom-made precision instrument (Beecher Instruments,
Silver Spring, MD) and brought into a recipient paraffin block. The TMA
blocks were constructed in four copies each containing one sample from
a different region of all tumors. One sample was taken from the center
(Figure 1)
and three samples were taken
from different peripheral areas of the tumors. After the TMA
construction, large sections were cut from the "donor" blocks of
532 tumors having sufficient material available.
|
Three conventional "large" sections from all tumors and three
sections from each of the four different replica TMA blocks were used
for immunostaining. Standard indirect immunoperoxidase procedures
(ABC-Elite, Vector Laboratories, Burlingame, CA) in combination with
monoclonal antibodies were used for detection of p53 (DO-7, prediluted
DAKO, Glostrup, Denmark), estrogen receptor (ER ID5, 1:1000, DAKO), and
progesterone receptor (NCL-PGR, 1A6, 1:600, NOVOCASTRA Laboratories
Ltd, Newcastle-upon-Tyne, UK). A microwave pretreatment was performed
for p53 (30 minutes at 90°C) retrieval. Diaminobenzidine was used as
a chromogen. Tumors with known positivity were used as positive
controls. The primary antibody was omitted for negative controls. The
same scoring criteria were applied in TMA and in large sections. All
slides were manually read by one pathologist (J.T.). Tumors were
considered positive for ER and PR if an unequivocal nuclear positivity
was seen in at least 10% of tumor cells. To define a tumor as p53
positive, moderate staining intensity was requested in 
20% of
tumor cell nuclei. These cutoff values were arbitrarily selected before
the beginning of the study based on previous
suggestions.13,14
An only faint p53 staining was scored
negative because such a staining can often be seen in non-neoplastic
cells, for example, in the basal cell layer of squamous epithelium or
urothelium. Examples of positive and negative tumors are given in
Figure 2
.
|
Contingency table analysis and 
2
tests
were used to study the relationship between immunohistochemical results
on large section and on TMAs. Survival curves were plotted according to
Kaplan-Meier. A log-rank test was applied to examine the relationship
between ER, PR, or p53 positivity and tumor-specific survival. Patients
were censored at the latest date when they were seen alive or at
the date of their non-tumor-related death.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
. Loss of ER expression was strongly
associated with poor prognosis, regardless of whether the statistical
analysis was done on data obtained from individual TMAs, combinations
of multiple TMAs, or on conventional large sections from each tumor
(Figure 4)
.
|
|
, so that
4 different TMAs provided essentially the same frequency as observed on
large sections. Loss of PR expression was also strongly associated with
poor prognosis both when evaluated from individual TMAs, combinations
of multiple TMAs, and on large sections from each tumor (Figure 5)
.
|
. If the results from multiple arrays were
combined, there was a slight increase in the overall positivity rate,
up to 24.1%. However, the frequency of p53 positivity on large
sections (42.8%) was almost twice as high as on TMAs. Unexpectedly,
the results obtained on individual TMAs or by combining results from
multiple TMAs were much more strongly linked to poor prognosis than the
results obtained on large sections (Figure 6)
. Further analysis showed that tumors
that were positive both on TMAs and large sections had a poor
prognosis. However, the prognosis of these 111 tumors that were only
positive on large sections but not on arrays was equally good as in
tumors that were p53 negative on large sections (data not shown). A
review of the discrepant cases showed that differences between TMA and
large section data were not due to tissue heterogeneity in most cases.
The vast majority of discrepant cases had a faint-to-moderate p53
staining in 15 to 30% of tumor cells in large sections. These tumors
were often considered negative on array sections because the criteria
of positivity (at least 20% of cells with unequivocal positivity) were
usually not reached in the small arrayed samples. The same tissues were
often considered positive on large sections, where the criteria for
positivity were frequently met at least in small tumor areas.
Reanalysis of the prognostic significance of the large section p53
staining revealed that a large percentage of positive cells (Figure 6G)
and a strong staining intensity (Figure 6H)
were strongly linked to
poor prognosis. A moderate staining intensity or fraction of
positive cells between 20% and 50% was not indicative of poor
prognosis, although such tumors were considered p53 positive in our
previously selected definition.
|
. The fraction of tumors with
heterogeneous findings was 8.8% for ER, 28.9% for PR, and 11.3% for
p53. These heterogeneous tumors tended to have an intermediate clinical
outcome as compared with a homogeneously positive or negative finding
(Figure 7, B–D)
.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This study was specifically designed to address the most obvious limitation of the TMA technique: the sampling of large, potentially heterogeneous tumors. The substantial heterogeneity of tumors is often evident both at the morphological and genetic level.1,15-22 This heterogeneneity is thought to represent the genetic instability of tumors and forms the basis for the current concepts of progression and clonal evolution of cancer. We observed that the frequency of ER positivity in the entire material was virtually the same when measured from a single TMA section (each tumor represented by a 0.6 mm diameter tissue spot), as compared to entire sections of breast cancers. For PR, the concordance was slightly lower (88%) and three samples from each tumor were required to achieve the same level of positivity as large section analyses. However, it must be kept in mind that large sections often represent a small fraction (ie, 0.004 x 10 x 10 mm) of large tumors (ie, 30 x 30 x 30 mm). The question to what extent TMA data can reproduce large section data are therefore much less important than whether clinicopathological associations can be reproduced or newly detected on TMAs. In this study, all TMA analyses provided highly significant association with prognosis for ER, PR, and p53. Possible heterogeneity or fixation differences between central or peripheral tumor regions did not affect these results even though the fraction of positivity was generally lower in samples from tumor center and periphery 1 than in periphery 2 and 3. It appears that the high number of tumors that can be included in a TMA study compensates for some false negative results which may be equally frequent in all subgroups of one arrayed tumor set. Therefore, in the case of large study materials, a single sample from each tumor may often be sufficient to derive information on clinical associations. This is also supported by our previous observations, where TMA analyses made it possible to reproduce numerous clinicopathological associations that were previously reported in the literature using conventional techniques based on large tissue specimens.10,23,24
In this study, 83 tumors with a borderline p53 staining (15% to 30% positive cells) on large sections were called negative on TMAs but considered positive in the large section analysis. The significant associations with clinical outcome detected on all four replica TMAs but not in the large section analysis prompted a reanalysis of large sections revealing that the prognosis of breast cancer patients was dependent on the fraction of p53-positive cells and their staining intensity. The obvious difficulties in the quantitation of p53 staining may be a reason for the observed discrepancies in the literature. While the majority of previous studies had reported an association between nuclear p53 accumulation and poor prognosis in breast cancer,13,14 several other studies had not confirmed this observation.25-27 Our p53 data also showed that data obtained on TMAs can be superior to large section data. Several technical issues apparently compensate for some loss of information due to the small tissue size. The staining of a single TMA slide provides a much greater degree of consistency and standardization than the immunostaining of hundreds of individual slides. Furthermore, quantitation of immunostainings is markedly easier on arrayed samples than on large sections. For example, it is possible to directly compare staining intensities of the different specimens on the same TMA slides, thereby improving the subjective interpretation of staining intensities. Most of all, the interpretation is, by default, limited to a small predefined area in arrayed samples. This facilitates a reproducible application of the selected scoring criteria because the entire tissue is always used for interpretation and the subjective selection of one tumor area for decision making is avoided. In the future, the TMA technology may help to optimize and standardize the interpretation of immunostainings, which is currently subjective and poorly reproducible and often leads to major discrepancies in studies investigating clinical associations for novel biomarkers. An exchange of stained or unstained TMA slides between laboratories reporting controversial data would help to rapidly unmask technical or interpretational reasons for conflicting study results.
Our data suggest that taking multiple punches from each tumor not only increases the number of interpretable tumors but also allows the distinction of three subgroups (positive, negative, and heterogeneous). The similar prognosis of tumors with a heterogeneous finding for PR and p53 as observed for homogeneously positive tumors suggests that false negative staining in one or several arrayed samples due to regional fixation problems may the reason for heterogeneous findings in some of these tumors. The similarly poor prognosis of ER heterogeneous tumors as found for ER negative tumors could theoretically be explained by an insufficient response of heterogeneous tumors to hormonal therapy. In this study, the use of four different samples per tumor did also result in a marked increase of interpretable tumors if data from multiple replica arrays were combined. However, analysis failure in up to 30% of arrayed samples was caused by technical problems related to the early generation of tissue arrays. Improved array making will reduce the need of multiple samples per tumor to increase the number of cases available for evaluation. The current success rate in our laboratory for arraying breast cancers is now greater than 90%.
In summary, our data suggest that TMAs can be successfully used to establish associations between molecular changes and clinical endpoints. Array-based tissue analysis is a rapid, cost-effective, and tissue-saving method for high-throughput clinicopathological studies. Molecular markers that appear most promising based on TMAs constructed from retrospectively collected sets of tumors will then need to be prospectively validated using molecular analyses of larger tissue specimens available from prospective clinical trials.
|
Footnotes |
|---|
Joachim Torhorst and Christoph Bucher contributed equally to this paper.
Accepted for publication September 17, 2001.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  S.-M. Sheen-Chen, C.-C. Huang, R.-P. Tang, F.-F. Chou, and H.-L. Eng Prognostic Value of Signal Transducers and Activators of Transcription 3 in Breast Cancer Cancer Epidemiol. Biomarkers Prev., September 1, 2008; 17(9): 2286 - 2290. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Kyndi, F. B. Sorensen, H. Knudsen, M. Overgaard, H. M. Nielsen, and J. Overgaard Estrogen Receptor, Progesterone Receptor, HER-2, and Response to Postmastectomy Radiotherapy in High-Risk Breast Cancer: The Danish Breast Cancer Cooperative Group J. Clin. Oncol., March 20, 2008; 26(9): 1419 - 1426. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. B. Moeder, J. M. Giltnane, M. Harigopal, A. Molinaro, A. Robinson, K. Gelmon, D. Huntsman, R. L. Camp, and D. L. Rimm Quantitative Justification of the Change From 10% to 30% for Human Epidermal Growth Factor Receptor 2 Scoring in the American Society of Clinical Oncology/College of American Pathologists Guidelines: Tumor Heterogeneity in Breast Cancer and Its Implications for Tissue Microarray Based Assessment of Outcome J. Clin. Oncol., December 1, 2007; 25(34): 5418 - 5425. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Melck, H. Masoudi, O. L. Griffith, A. Rajput, G. Wilkins, S. Bugis, S. J. M. Jones, and S. M. Wiseman Cell Cycle Regulators Show Diagnostic and Prognostic Utility for Differentiated Thyroid Cancer. Ann. Surg. Oncol., December 1, 2007; 14(12): 3403 - 3411. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. J. Ullenhag, A. Mukherjee, N. F.S. Watson, A. H. Al-Attar, J. H. Scholefield, and L. G. Durrant Overexpression of FLIPL Is an Independent Marker of Poor Prognosis in Colorectal Cancer Patients Clin. Cancer Res., September 1, 2007; 13(17): 5070 - 5075. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Hu, X. He, K. A. Baggerly, K. R. Coombes, B. T.J. Hennessy, and G. B. Mills Non-parametric quantification of protein lysate arrays Bioinformatics, August 1, 2007; 23(15): 1986 - 1994. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. J. Duncan, P. Rolland, S. Deen, I. V. Scott, D. T.Y. Liu, I. Spendlove, and L. G. Durrant Loss of IFN{gamma} Receptor Is an Independent Prognostic Factor in Ovarian Cancer Clin. Cancer Res., July 15, 2007; 13(14): 4139 - 4145. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. S. Lee, S.-B. Cho, H. E. Lee, M. A Kim, J. H. Kim, D. J. Park, J. H. Kim, H.-K. Yang, B. L. Lee, and W. H. Kim Protein Expression Profiling and Molecular Classification of Gastric Cancer by the Tissue Array Method Clin. Cancer Res., July 15, 2007; 13(14): 4154 - 4163. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A Lugli, I Zlobec, K Baker, P Minoo, L Tornillo, L Terracciano, and J R Jass Prognostic significance of mucins in colorectal cancer with different DNA mismatch-repair status J. Clin. Pathol., May 1, 2007; 60(5): 534 - 539. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Rolland, Z. Madjd, L. Durrant, I. O Ellis, R. Layfield, and I. Spendlove The ubiquitin-binding protein p62 is expressed in breast cancers showing features of aggressive disease Endocr. Relat. Cancer, March 1, 2007; 14(1): 73 - 80. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. S. Lam, A. J. Pantuck, A. S. Belldegrun, and R. A. Figlin Protein Expression Profiles in Renal Cell Carcinoma: Staging, Prognosis, and Patient Selection for Clinical Trials Clin. Cancer Res., January 15, 2007; 13(2): 703s - 708s. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Zimpfer, S. Schonberg, A. Lugli, C. Agostinelli, S. A Pileri, P. Went, and S. Dirnhofer Construction and validation of a bone marrow tissue microarray J. Clin. Pathol., January 1, 2007; 60(1): 57 - 61. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Bertucci, D. Birnbaum, and A. Goncalves Proteomics of Breast Cancer: Principles and Potential Clinical Applications Mol. Cell. Proteomics, October 1, 2006; 5(10): 1772 - 1786. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Su, M. J. Shrubsole, R. M. Ness, Q. Cai, N. Kataoka, K. Washington, and W. Zheng Immunohistochemical Expressions of Ki-67, Cyclin D1, {beta}-Catenin, Cyclooxygenase-2, and Epidermal Growth Factor Receptor in Human Colorectal Adenoma: A Validation Study of Tissue Microarrays. Cancer Epidemiol. Biomarkers Prev., September 1, 2006; 15(9): 1719 - 1726. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Dolled-Filhart, A. McCabe, J. Giltnane, M. Cregger, R. L. Camp, and D. L. Rimm Quantitative In situ Analysis of {beta}-Catenin Expression in Breast Cancer Shows Decreased Expression Is Associated with Poor Outcome. Cancer Res., May 15, 2006; 66(10): 5487 - 5494. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. M. Callagy, P. D. Pharoah, S. E. Pinder, F. D. Hsu, T. O. Nielsen, J. Ragaz, I. O. Ellis, D. Huntsman, and C. Caldas Bcl-2 Is a Prognostic Marker in Breast Cancer Independently of the Nottingham Prognostic Index Clin. Cancer Res., April 15, 2006; 12(8): 2468 - 2475. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. P. Linke, T. M. Bremer, C. D. Herold, G. Sauter, and C. Diamond A Multimarker Model to Predict Outcome in Tamoxifen-Treated Breast Cancer Patients Clin. Cancer Res., February 15, 2006; 12(4): 1175 - 1183. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A Jubb and K Hillan Expression of HIF-1{alpha} in human tumours J. Clin. Pathol., December 1, 2005; 58(12): 1344 - 1344. [Full Text] [PDF]
|
|
|
|
|
|
B. Willipinski-Stapelfeldt, S. Riethdorf, V. Assmann, U. Woelfle, T. Rau, G. Sauter, J. Heukeshoven, and K. Pantel Changes in Cytoskeletal Protein Composition Indicative of an Epithelial-Mesenchymal Transition in Human Micrometastatic and Primary Breast Carcinoma Cells Clin. Cancer Res., November 15, 2005; 11(22): 8006 - 8014. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. R. Nambiar, S. R. Boutin, R. Raja, and D. W. Rosenberg Global Gene Expression Profiling: A Complement to Conventional Histopathologic Analysis of Neoplasia Vet. Pathol., November 1, 2005; 42(6): 735 - 752. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Harigopal, A. J. Berger, R. L. Camp, D. L. Rimm, and H. M. Kluger Automated Quantitative Analysis of E-Cadherin Expression in Lymph Node Metastases Is Predictive of Survival in Invasive Ductal Breast Cancer Clin. Cancer Res., June 1, 2005; 11(11): 4083 - 4089. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. J. Chang, B. C. Yoo, S.-B. Lim, S.-Y. Jeong, W. H. Kim, and J.-G. Park Metabotropic Glutamate Receptor 4 Expression in Colorectal Carcinoma and Its Prognostic Significance Clin. Cancer Res., May 1, 2005; 11(9): 3288 - 3295. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Handerson, R. Camp, M. Harigopal, D. Rimm, and J. Pawelek {beta}1,6-Branched Oligosaccharides Are Increased in Lymph Node Metastases and Predict Poor Outcome in Breast Carcinoma Clin. Cancer Res., April 15, 2005; 11(8): 2969 - 2973. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. M. Sheehan, V. S. Calvert, E. W. Kay, Y. Lu, D. Fishman, V. Espina, J. Aquino, R. Speer, R. Araujo, G. B. Mills, et al. Use of Reverse Phase Protein Microarrays and Reference Standard Development for Molecular Network Analysis of Metastatic Ovarian Carcinoma Mol. Cell. Proteomics, April 1, 2005; 4(4): 346 - 355. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Alvaro, M. Lejeune, M. T. Salvado, R. Bosch, J. F. Garcia, J. Jaen, A. H. Banham, G. Roncador, C. Montalban, and M. A. Piris Outcome in Hodgkin's Lymphoma Can Be Predicted from the Presence of Accompanying Cytotoxic and Regulatory T Cells Clin. Cancer Res., February 15, 2005; 11(4): 1467 - 1473. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Jacquemier, C. Ginestier, J. Rougemont, V.-J. Bardou, E. Charafe-Jauffret, J. Geneix, J. Adelaide, A. Koki, G. Houvenaeghel, J. Hassoun, et al. Protein Expression Profiling Identifies Subclasses of Breast Cancer and Predicts Prognosis Cancer Res., February 1, 2005; 65(3): 767 - 779. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Al-Kuraya, P. Schraml, J. Torhorst, C. Tapia, B. Zaharieva, H. Novotny, H. Spichtin, R. Maurer, M. Mirlacher, O. Kochli, et al. Prognostic Relevance of Gene Amplifications and Coamplifications in Breast Cancer Cancer Res., December 1, 2004; 64(23): 8534 - 8540. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. J. Berger, R. L. Camp, K. A. DiVito, H. M. Kluger, R. Halaban, and D. L. Rimm Automated Quantitative Analysis of HDM2 Expression in Malignant Melanoma Shows Association with Early-Stage Disease and Improved Outcome Cancer Res., December 1, 2004; 64(23): 8767 - 8772. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Th. Went, S. Dirnhofer, M. Bundi, M. Mirlacher, P. Schraml, S. Mangialaio, S. Dimitrijevic, J. Kononen, A. Lugli, R. Simon, et al. Prevalence of KIT Expression in Human Tumors J. Clin. Oncol., November 15, 2004; 22(22): 4514 - 4522. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. L. Camp, M. Dolled-Filhart, and D. L. Rimm X-Tile: A New Bio-Informatics Tool for Biomarker Assessment and Outcome-Based Cut-Point Optimization Clin. Cancer Res., November 1, 2004; 10(21): 7252 - 7259. [Abstract] [Full Text] [PDF]
|
|
