Open Access Highly Accessed Research article

Risk estimation of distant metastasis in node-negative, estrogen receptor-positive breast cancer patients using an RT-PCR based prognostic expression signature

Andrew Tutt12*, Alice Wang3, Charles Rowland3, Cheryl Gillett4, Kit Lau3, Karen Chew5, Hongyue Dai6, Shirley Kwok3, Kenneth Ryder2, Henry Shu3, Robert Springall4, Paul Cane7, Blair McCallie3, Lauren Kam-Morgan8, Steve Anderson8, Horst Buerger9, Joe Gray10, James Bennington11, Laura Esserman12, Trevor Hastie13, Samuel Broder3, John Sninsky3, Burkhard Brandt9 and Fred Waldman115

Author Affiliations

1 Breakthrough Breast Cancer Research Unit, King's College, London, UK

2 Guy's Hospital, London, UK

3 Celera, LLC, Alameda, CA, USA

4 Guy's and St Thomas' Hospital Breast Research Tissue & Data Bank, London, UK

5 Comprehensive Cancer Center, University of California San Francisco, San Francisco, CA, USA

6 Rosetta Inpharmatics, A wholly owned subsidiary of Merck & Co. Inc., Seattle, WA, USA

7 Cellular Pathology Department, St Thomas' Hospital, London, UK

8 Laboratory Corporation of America, Triangle Park, NC, USA

9 Institute of Tumor Biology, University Medical Center, University of Hamburg, Hamburg, Germany

10 Life Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, CA, USA

11 Department of Pathology, California Pacific Medical Center, San Francisco, CA, USA

12 Carol Franc Buck Breast Cancer Center, University of California San Francisco, San Francisco, CA, USA

13 Departments of Statistics, and Health Research & Policy, Stanford University, Stanford, CA, USA

For all author emails, please log on.

BMC Cancer 2008, 8:339  doi:10.1186/1471-2407-8-339

Published: 21 November 2008

Additional files

Additional file 1:

Concordance for ER status by immunhistochemistry and RT-PCR. We evaluated the concordance of ER status determined by RT-PCR and IHC for the Guy's untreated patients. Of the 287 patients, 97.6% and 96.9% were ER-positive by IHC and RT-PCR assay, respectively. The cutoff point for ER positivity of RT-PCR assay was prespecified as 1.0 based on prior studies [ref. [1]]. The concordance between RT-PCR and IHC results was high. The Kappa coefficient and 95% confidence limits were 0.80 (0.57, 1.00).

Format: PDF Size: 45KB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data

Additional file 2:

Selection of Candidate Genes profiled in the training set. We selected 197 candidate genes from the published literature and microarray-based gene expression profiling experiments. The candidate genes included the 70-gene panel described in van't Veer et al, 104 genes further analyzed from Rosetta dataset (Dai et al), the Paik et al reported 16-gene panel (excluded the housekeeping genes from 21-gene panel), and 24 ER related genes. The 197 candidate genes and their accession number are listed.

Format: PDF Size: 77KB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data

Additional file 3:

Normalization gene selection. The expression level of 6 housekeeping genes (HSK) was determined on 150 breast cancer tissue samples. Gene stability was evaluated using the geNorm program [Ref. [2]] which relies on the principle that the expression ratio of two ideal reference genes would be identical in all samples, regardless of experimental conditions or cell type. The program calculates the gene stability measure (M) which is the average pair-wise variation for a gene compared with all other tested control gene. Genes with lower M values are more stable. The results show that PPIG, SLU7, and NUP214 were the most stable housekeeping genes in this sample set.

Format: PDF Size: 17KB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data

Additional file 4:

mRNA Enhancement. The RNAs used to profile the genes in the training set were enriched whereas the RNAs in the validation and tamoxifen treatment sets were used directly. To ensure that the profiling methods produced similar results, we compared the enriched and unenriched expression profile of 14 genes in 50 training set samples. Good correlation (R2 = 0.9931) of gene expression levels was observed between the metastasis scores generated with the enriched vs. unenriched samples. Perfect agreement was obtained on risk category calls.

Format: PDF Size: 16KB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data

Additional file 5:

Genes selected by supervised principal component procedure. Table of genes selected by supervised principal component procedure.

Format: PDF Size: 31KB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data

Additional file 6:

Coefficients of the 14 genes in the principle component. Table and graphs of coefficients of the 14 genes in the principle component

Format: PDF Size: 46KB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data

Additional file 7:

Algorithm for computing Metastasis Score (MS). Algorithm for computing Metastasis Score (MS). The relative changes in gene expression are calculated by ΔΔCt method.

Format: PDF Size: 75KB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data

Additional file 8:

Selection bias assessment. Selection bias assessment in the training set and validation set.

Format: PDF Size: 74KB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data

Additional file 9:

Clinical and pathological characteristics of patients from tamoxifen-treated set. Table of clinical and pathological characteristics of patients from tamoxifen-treated set.

Format: PDF Size: 57KB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data

Additional file 10:

Correlation between the Ki-67 LI and the MS and the Ki-67 mRNA level. (a) Correlation of Ki-67 LI with MS, (b) Correlation of Ki-67 LI with Ki-67 mRNA level. R2 = 0.30 and 0.18, were observed for Ki-67 LI versus MS, and Ki-67 LI versus Ki-67 mRNA level, respectively.

Format: PDF Size: 22KB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data

Additional file 11:

Ingenuity pathway analysis of prognostic 14-gene signature. The primary network identified below indicates the majority of the 14 genes of the signature are associated with known important tumor suppressors (TP53 and CDKN2A) and cell cycle control (CDC2). Nodes are displayed using various shapes that represent the functional class of the gene product (diamond-enzymes, ovals-transcription factors, triangles-kinase, circles-others). A solid line indicates a direct interaction while a dashed line indicates an indirect interaction. A line without an arrowhead indicates binding. Genes from 14-gene signature are highlighted in grey. Note that thirteen out of 14 genes of the signature are in the network.

Format: PDF Size: 139KB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data