Skip to main content
NCBI home page
The Journal of Molecular Diagnostics : JMD logoLink to The Journal of Molecular Diagnostics : JMD
. 2015 Jul;17(4):374–381. doi: 10.1016/j.jmoldx.2015.02.002

Comparing Platforms for Messenger RNA Expression Profiling of Archival Formalin-Fixed, Paraffin-Embedded Tissues

Svitlana Tyekucheva ∗,†,∗, Neil E Martin ‡, Edward C Stack §,, Wei Wei ‖, Vinod Vathipadiekal ‖, Levi Waldron ∗∗, Michelangelo Fiorentino ††, Rosina T Lis §,, Meir J Stampfer ‡‡,§§,¶¶, Massimo Loda §,¶,‖‖,∗∗∗,†††, Giovanni Parmigiani ∗,†, Lorelei A Mucci ‡‡,§§, Michael Birrer ‖
PMCID: PMC4483460  PMID: 25937617

Abstract

Archival formalin-fixed, paraffin-embedded (FFPE) tissue specimens represent a readily available but largely untapped resource for gene expression profiling–based biomarker discovery. Several technologies have been proposed to cope with the bias from RNA cross-linking and degradation associated with archival specimens to generate data comparable with RNA from fresh-frozen materials. Direct comparison studies of these RNA expression platforms remain rare. We compared two commercially available platforms for RNA expression profiling of archival FFPE specimens from clinical studies of prostate and ovarian cancer: the Affymetrix Human Gene 1.0ST Array following whole-transcriptome amplification using the NuGen WT-Ovation FFPE System V2, and the NanoString nCounter without amplification. For each assay, we profiled 7 prostate and 11 ovarian cancer specimens, with a block age of 4 to 21 years. Both platforms produced gene expression profiles with high sensitivity and reproducibility through technical repeats from FFPE materials. Sensitivity and reproducibility remained high across block age within each cohort. A strong concordance was shown for the transcript expression values for genes detected by both platforms. We showed the biological validity of specific gene signatures generated by both platforms for both cohorts. Our study supports the feasibility of gene expression profiling and large-scale signature validation on archival prostate and ovarian tumor specimens using commercial platforms. These approaches have the potential to aid precision medicine with biomarker discovery and validation.

With the advent of precision medicine, there is a growing interest in developing prognostic and predictive molecular signatures to help guide the care of patients with cancer.1 Among approaches taken to realize the goal of identifying the ideal treatment for the patient at the correct time are mutation analysis,2 immunohistochemical staining,3 and mRNA expression analysis.4–6

Such studies require access to tissue specimens with detailed clinical annotation and often long-term follow-up evaluation. These clinical samples, including biopsy and surgical specimens, typically are formalin-fixed and embedded in paraffin to allow for morphologic assessment by a pathologist. Although this process preserves morphologic features of the tumors, it also makes RNA expression profiling more challenging relative to fresh-frozen tissue. RNA in formalin-fixed, paraffin-embedded (FFPE) samples is subject to degradation, fragmentation, and cross-linking, which typically limits library preparation and gene expression assessment.7 Importantly, traditional quality control measurements for RNA, such as RNA integrity number, are not necessarily predictive of the success of corresponding gene expression assays.8

The large numbers of well-annotated FFPE tumor tissue samples currently archived remain a vast and underused resource in the genomic study of cancer. Notably, most large clinical and epidemiologic cohorts only collect FFPE samples. Given this wealth of archival material from patients with known outcomes and the continued FFPE processing of new clinical specimens, there is a need to develop and test reliable methods for profiling mRNA expression in FFPE materials.

Several platforms have been developed in recent years to assess mRNA expression from FFPE tissue including whole-transcriptome amplification9,10 and direct assessment using multiplexed color-coded probes.11 As an initial step to developing prognostic and predictive mRNA signatures from archival tumor specimens, we performed head-to-head comparisons of gene expression profiles from prostate and ovarian cancer FFPE specimens from large-scale epidemiologic studies and clinical trials representative of a wide variety of fixation times, block ages, and block storage conditions using two platforms: the NuGen WT-Ovation FFPE System V2 + Affymetrix GeneChip Human Gene 1.0 ST Array [NuGen (NuGen, Inc., San Carlos, CA) + Affymetrix (Affymetrix, Santa Clara, CA)] and the NanoString nCounter Cancer panel [NanoString (NanoString Technologies, Seattle, WA)].

Materials and Methods

Prostate Tissue Samples

Archival FFPE radical prostatectomy specimens were collected from treating institutions for men diagnosed with prostate cancer who had enrolled in the prospective Physicians' Health Study.12,13 Participants provided informed consent to collect biospecimens, and the study was approved by the Institutional Review Boards at the Harvard School of Public Health and Partners Health Care. Pathology was reviewed centrally to confirm the cancer diagnosis and to provide consistent histopathologic review including Gleason scores. Areas of high-density tumor were identified, and two to three 0.6-mm punches were taken from both tumor and adjacent nontumor prostatic tissue for RNA extraction. The cores were deparaffinized using 800 μL Citrisolv (Fisher Scientific, Pittsburgh, PA) at 60°C for 20 minutes followed by 1.2 mL Citrisolv:absolute alcohol (2:1) at room temperature for 10 minutes. Cores subsequently were washed with absolute alcohol, dried at 55°C, and incubated overnight at 45°C in 300 μL lysis buffer [10 mmol/L NaCl, 500 mmol/L Tris (pH 7.6), 20 mmol/L EDTA, 1% sodium dodecyl sulfate] containing 1 mg/mL proteinase K (Ambion, Austin, TX). RNA was extracted using the RecoverAll Total Nucleic Acid Isolation kit (Ambion). After tissue digestion, following the manufacturer's protocol, samples were incubated in 10× DNase (Ambion) and column-purified to elute RNA. The concentration was determined using the Nanodrop 1000 (Fisher Scientific) and RiboGreen RNA Assay Kit (Ambion). Samples and technical replicates were randomized and assigned study identifications to blind laboratory personnel.

Ovarian Tissue Samples

Ovarian cancer specimens were obtained from the Gynecologic Oncology Group tissue bank (Columbus, OH). All specimens were archival FFPE tissue samples obtained from women with advanced-stage epithelial ovarian cancer in clinical trial GOG218, a randomized phase III trial testing the impact of the addition of bevacizumab to standard chemotherapy for the up-front treatment of advanced-stage epithelial ovarian cancer.14 The specimens were obtained at the primary debulking surgery.

The pathology was confirmed by the central Gynecologic Oncology Group review to ensure the correct histology and percentage of tumor. Ten-micron sections were made on positively charged slides. RNA was extracted from paraffin scrolls using the RNeasy FFPE kit (Qiagen, Valencia, CA) with modifications. In brief, 1 mL of xylene was added to a 25-mm paraffin scroll, vortexed vigorously for 10 seconds, and centrifuged at full speed for 2 minutes. Supernatant then was removed and 1 mL of ethanol (100%) was added to the pellet, mixed by vortexing, and centrifuged at full speed for 2 minutes. Supernatant was removed and the tube was incubated at room temperature for 10 minutes with the lid opened. RNA subsequently was extracted using reagents provided in the RNeasy FFPE extraction kit according to the manufacturer's instructions. Finally, purified RNA was eluted from the RNeasy MinElute spin column (Qiagen) using 30 mL RNase-free water.

Experimental Design

We profiled seven paired tumor and adjacent normal prostate tissue samples from three patients with a Gleason score of 8, one patient with a Gleason score of 7, and three patients with a Gleason score of 6 disease on the NuGen + Affymetrix platform, and a subset of five pairs (two with a Gleason score of 6 and three with a Gleason score of 8) on NanoString. Block ages for prostate cancer specimens ranged from 11 to 21 years. We profiled two to three technical replicates for each sample with varying RNA input amounts. For ovarian cancer we selected five serous carcinoma and six clear cell carcinoma samples, and block ages for this cohort ranged from 4 to 7 years. All of these samples were profiled on both NuGen + Affymetrix and NanoString, with either two or three technical replicates. Complete information on all samples and technical replicates that were analyzed are presented in Supplemental Table S1.

Messenger RNA Expression Profiling

Array

For array-based mRNA profiling, we first performed whole-transcriptome amplification using the WT-Ovation FFPE System V2 (NuGen). This approach initiates amplification at the 3′ end as well as randomly throughout the transcriptome, improving the performance in severely degraded FFPE samples. After isothermal amplification, 50 or 100 ng of total RNA was amplified to 4 to 7 μg of biotinylated cDNA complementary to the original mRNA. The amplification step has been optimized for RNA extracted from FFPE specimens15,16 and has shown comparable differential expression profiles to corresponding fresh-frozen tissues.17 After amplification, we hybridized 3.75 μg of amplified cDNA to the GeneChip Human Gene 1.0 ST Array (Affymetrix). The 1.0 ST Array profiles expression of >28,000 genes with an average of 26 probes per gene.

NanoString

We used the NanoString nCounter platform18 to capture and count 230 cancer-related human genes using a prebuilt kit supplied by the manufacturer. Following the manufacturer's instructions, we aliquoted 100 to 200 ng total RNA in 5 μL to initiate analysis.

Data Preprocessing

We normalized the Affymetrix data across samples and batches using Robust Multichip Average method.19,20 For the NanoString platform the data were processed according to the manufacturer's recommendations. Briefly, background subtracted counts were multiplied by scaling factors proportional to the sum of counts for spiked-in positive control probes to account for individual assay efficiency variation, and to the geometric average of the housekeeping gene probes (CLTC, GAPDH, GUSB, HPRT1, PGK1, and TUBB) to account for variability in the mRNA content. Background signal was calculated as a median value of the negative hybridization control probes. Normalized counts were log-transformed for downstream analysis.

Raw and preprocessed data were submitted to the GEO database (http://www.ncbi.nlm.nih.gov/geo; accession number GSE54809).

Data preprocessing and statistical analysis were performed using statistical software packages R version 3.1.1 (http://www.r-project.org) and Bioconductor version 2.14 (http://www.bioconductor.org).21 Power analysis was performed using G*Power version 3.1.3 (http://www.gpower.hhu.de).22

Results

We sought to evaluate several potential variables related to the performance of RNA expression assays. We assessed the sensitivity and reproducibility of the gene expression measurements across technical replicates of RNA extracted from the same tissue with respect to the following: i) RNA input levels, ii) FFPE block age, and iii) tumor grade and histology. We also used tumor grade and histology to assess the biological validity of the expression profiles.

Sensitivity and Reproducibility Analysis

For each platform and each cohort we calculated the percentages of probes detectable above the background. In each platform, background probes comprise sequences not found in the human genome. We observed no significant association between the proportion of detectable probes and the block age, as quantified by the Pearson correlation coefficients (Table 1). The wider range in the percentage of probes present for the NuGen + Affymetrix data on the prostate cohort likely can be explained by differences in the instrument calibrations between the two batches in which the samples were assayed. One batch had higher intensities, which were corrected analytically by Robust Multichip Average method normalization (Supplemental Figure S1). With sample sizes of 24 (Ovarian and prostate NanoString studies, and Ovarian NuGen + Affymetrix) and 30 (Prostate NuGen + Affymetrix), we had a power of 0.86 and 0.93, respectively, to detect whether 30% or more of the variance in the percentage of the probes present was explained by the block age. Therefore, failure to find a significant correlation between the percentage of the probes present and block age was suggestive of the fact that block age indeed does not explain much of the variation in the percentage of the probes within each study.

Table 1.

Ranges of the Percentages of the Probes with the Signal above the Background, Correlation of the Percentage of Present Probes with Block Age, and Ranges of Correlations between Technical Replicates for Each Cohort and Gene Expression Platform

Platform Percent present Correlation with block age Correlations between technical replicates
Prostate samples
 NuGen + Affymetrix 0.17–0.58 0.02 0.91–0.98
 NanoString 0.40–0.79 −0.11 0.88–0.97
Ovarian samples
 NuGen + Affymetrix 0.40–0.68 0.06 0.94–0.98
 NanoString 0.83–0.93 −0.04 0.95–0.99
Open in a new tab

In the prostate samples with sufficient replicates at each RNA quantity we did not observe a statistically significant difference between the percentage of probes present at 50 versus 100 ng of input RNA for the NuGen + Affymetrix platform (analysis of variance power of 0.85 to detect whether 40% of variance in percentage of the present probes is explained by RNA quantity). A lower percentage of probes in the samples with smaller RNA input quantity was present in the NanoString data (Tukey honest significant difference test for 200 ng versus 100 ng, P = 0.04). For this analysis we had an analysis of variance power of 0.85 to detect whether 45% of variance in the percentage of probes present above the background is explained by RNA quantity. For ovarian samples only two replicates were of lower quantity, therefore the formal testing was not performed.

The Pearson correlation coefficients between all pairs of technical replicates within each platform ranged from 0.88 to 0.99 (Table 1 and Figure 1, A–D). The consistency of gene expression levels, as measured by the magnitude of the correlation, between technical replicates did not change with block age. We had a power of 0.85 to detect the following percentages of the concordance explained by the block age: 32% and 35% for prostate NuGen + Affymetrix and NanoString, respectively, and 41% for ovarian samples on both platforms. The consistency of gene expression levels depended on the input RNA amounts for the NanoString platform (Supplemental Figure S2). This can be explained by higher input amounts corresponding to a higher percentage of detectable probes. We had a power of 0.85 to detect 45% of the concordance explained by RNA input amounts. We also found lower correlations between replicates among the ovarian samples that presented with more extensive tissue necrosis. The presence of necrosis was not measured in the prostate samples although it usually is quite rare. Because correlation coefficients could be driven by outlying values, we also assessed concordance for each pair of the replicates by considering the fraction of genes that are either below or above certain expression intensities for both members of the pair (Figure 1, E–H). Lower concordance was observed at lower, less reliably measured, expression intensities, as expected. Only a small fraction (<2%) of genes with medium to high expression intensities in one replicate had low expression in the other replicate. As expected, the gene-wise correlations (Figure 1, I–L) between the replicate pairs tended to be lower for genes with lower average intensities than for genes with higher average intensities, suggesting that gene expression quantification from FFPE tissues still was reliable when genes were stratified by expression level.

Figure 1.

Figure 1

Open in a new tab

A–D: Correlations between technical replicates as a function of block age for all platforms and cohorts. Pairs of technical replicates from the same individual are shown by the same symbol. For prostate specimens normal tissue pairs are shown in gray, tumor is shown in black. E–H: Percentage of the genes that are either above or below the intensity threshold in each sample in a technical replicate pair. Each pair of replicates is plotted with an individual curve. J–L: Scatterplots of correlations between gene expression values for each gene represented on a platform across technical replicates. I and K: Presented as smoothed scatters in which color depth represents the density of the points in the plot for better visibility.

Concordance between NuGen + Affymetrix and NanoString

By using NetAffx annotations (Affymetrix), we mapped all 236 genes surveyed on the NanoString Cancer panel assay to a total of 256 transcript clusters on the Affymetrix GeneChip Human Gene 1.0 ST array. For these genes we calculated correlations between the samples assayed on both platforms. For this analysis, we averaged transcript expression values for technical replicates within each platform. Most of the genes had high positive correlations, with lower correlations predominantly found for the genes with lower expression levels (Figure 2, A and B). Of the 256 correlations, 77 (30%) in prostate and 33 (13%) in ovarian were less than 0.3. Among these genes with lower correlation, 18 mapping pairs, corresponding to 15 unique gene symbols, were common in both diseases: BCR, CASP10, CEBPA, CSF3, CYP1A1, FLT3, GATA1, HRAS, LMO2, MLH1, MLL, MPL, TFE3, WEE1, and WNT10B. We performed a permutation analysis and obtained distributions of the correlations between pairs of genes measured on two platforms under the null scenario of spurious correlations between the unrelated measurements, which can be high for relatively small sample sizes. Null scenario correlations were computed for 100 permutations of the sample labels for one of the platforms. From this analysis we show that for both prostate and ovarian cohorts the observed correlations between the platforms were higher than expected by chance (Figure 2, C and D).

Figure 2.

Figure 2

Open in a new tab

A and B: Bubble chart showing the relationship between the correlations of the gene expression values on the NuGen + Affymetrix and NanoString platforms and the average expression levels of those genes. Each circle represents a gene and the size is proportional to the absolute value of the correlations. Red represents positive correlations and blue represents negative correlations. C and D: Density plots of the observed correlations between the gene expression values on the NuGen + Affymetrix and NanoString platforms (solid lines) and the correlations under the null hypothesis of no association between the expression values measured on the NuGen + Affymetrix and NanoString platforms obtained by permuting sample labels (dashed lines).

Biological Validation

To provide evidence that both platforms provide biologically useful information we considered established gene signatures that distinguish prostate cancer versus noncancerous tissue and high versus low Gleason grade in prostate cancer, and clear cell versus serous adenocarcinoma for the ovarian cancer cohort. The prostate tumor versus normal tissue signature was obtained from Oncomine (Life Technologies, Carlsbad, CA), and was defined by genes that were significantly up- or down-regulated in tumor versus normal comparison in four prostate cancer studies,23–26 and had a median rank of the differential expression of less than 150 across these four studies. An mRNA signature related to Gleason grade was taken from Penney et al.27 For ovarian cancer, the signature was retrieved as the union of three gene signatures describing differences between serous and clear cell carcinomas from GeneSigDb.28–31 Gene signatures used for validation are presented in Supplemental Table S2.

We performed principal components analysis using genes from each signature that were represented on Affymetrix and NanoString platforms (Figure 3). For each comparison we observed meaningful separation of the classes defined by the signatures on the first two principal components.

Figure 3.

Figure 3

Open in a new tab

Genes from the signatures that are represented on each platform were projected onto two principal components and plotted. Colors represent the compared classes and distinct plotting characters represent the technical replicates from the same specimen. A: NuGen + Affymetrix profiles of tumor versus normal signature on prostate samples. Tumor samples are shown in orange and normal samples are shown in blue. B: NanoString profiles of tumor versus normal signature on prostate samples. Tumor samples are shown in orange and normal samples are shown in blue. C: NuGen + Affymetrix profiles of Gleason signature on prostate tumor samples. Gleason 6 samples are shown in blue, Gleason 7 samples are shown in green, and Gleason 8 samples are shown in orange. D: NanoString profiles of Gleason signature on prostate tumor samples. Gleason 6 samples are shown in blue, Gleason 8 samples are shown in orange. E: NuGen + Affymetrix profiles of histology signature on ovarian samples. Clear cell carcinoma samples are shown in blue and serous carcinoma samples are shown in orange. F: NanoString profiles of histology signature on ovarian samples. Clear cell carcinoma samples are shown in blue and serous carcinoma samples are shown in orange. PCA, principal components analysis.

In addition, for the Gleason signature genes, we calculated the log-fold changes in the gene expression values between high and low Gleason grade tumors. We compared our results with the log-fold changes observed in the Gleason signature. By using the NetAffx annotations implemented in the Bioconductor packages oligo and pd.hugene.1.0.st.v1, we mapped 150 of 157 genes from the signature to 157 Affymetrix transcript clusters. The correlation between the log-fold changes from two studies was 0.57. Only 10 genes from the 157-gene Gleason signature were represented on the NanoString Cancer Panel. The correlation between the log-fold changes observed on NanoString and previously published data were 0.89.

Discussion

The reliable use of archival FFPE samples for mRNA signature development will provide enormous opportunities for precision medicine in oncology by unlocking banked tissue collected over decades with mature associated outcomes. Previous work has established that new approaches for RNA expression profiling in FFPE correlate well with corresponding fresh-frozen samples.10,11,32 In this study, we sought to determine the reliability of two approaches for mRNA expression profiling in archival FFPE samples from two representative research cohorts. We selected samples with block ages most typically used for biomarker discovery studies in prostate and ovarian cancer. Prostate cancer has a long natural course and a follow-up period of more than 10 years typically is necessary to study progression and lethality. Ovarian cancer is a much more aggressive disease, and a follow-up period of 5 years is common for lethality and progression studies. The NanoString platform allows assessment using low RNA input quantity and provides expression data on a prespecified set of genes. Whole-genome expression data can be obtained using an array-based approach after amplification.

We found that both the NanoString platform and whole-genome amplification followed by array profiling produced highly concordant results across technical replicates. These findings were independent of FFPE block ages up to 21 years old, and held for the suboptimal RNA input amounts down to 50 ng, highlighting the clinical applicability of this approach. There was good correlation of mRNA abundances between the NanoString and NuGen + Affymetrix expression profiling platforms investigated here. As expected from experience with gene expression profiling from fresh-frozen specimens, reproducibility diminished with the expression levels, and reliable assessment of gene expression remains imperfect for low-abundance transcripts.33 Furthermore, although we did not intend to directly compare results at the sample level with mRNA expression data from fresh-frozen tissue, we did observe biological correlations to existing literature from fresh-frozen samples, suggesting that biologically relevant signatures can be derived from archival FFPE samples.

Concordance between microarray gene expression profiling and the NanoString nCounter platform has been shown previously for frozen specimens.34 To our knowledge, our study is the first to compare the performance of these two platforms on FFPE specimens across two different diseases. We have shown the feasibility of discovery using larger microarray assays and a follow-up validation using the targeted NanoString nCounter approach.

Although the study sample size was relatively modest, no sample failed to produce usable results on either platform. This is in contrast to our prior experience with the Illumina (San Diego, CA) whole-genome DASL HT assay platform, in which filtering out low-quality samples was essential to maintain the reproducibility of the differential gene expression inference.35 This study did not compare the DASL platform with these more recent profiling approaches for FFPE.

Our comparative study and several recent studies34,36 that used NuGen + Affymetrix and NanoString platforms have shown the feasibility of these platforms for clinical applications such as discovery of the gene expression signatures for personalized medicine and its applications from FFPE biopsy material. However, the development of additional statistical methodologies specifically tailored for preprocessing of the gene expression data from FFPE tissues still might be beneficial because the reproducibility between the technical replicates observed in our study allows for further improvement.

With a growing acceptance of the utility of mRNA expression profiles in precision medicine,6,37,38 researchers will require reliable methods for assessing gene expression data from archival specimens. In this study, we found that two modern platforms, one for more targeted profiling and one looking across the genome, can produce reliable, biologically relevant signatures. Importantly, we showed this for two different cancers with very different RNA yields: challenging prostate cancer and ovarian cancer, for which RNA yields usually are abundant. Differences in the RNA extraction techniques between the ovarian and prostate cases reflected previous experience and optimization by the two groups of investigators involved. Also, it previously was shown that both the RNeasy and RecoverAll kits allow extraction of total RNA of similar integrity and suitable for microarray hybridization.9,39 Similar findings in both of our cohorts additionally suggest that extraction protocols did not have much influence on the RNA used for library preparation. From this perspective our analysis provided additional independent validation.

Although additional experience and pilot studies in other cancer types are needed, our data prompt identical conclusions for each cohort that we studied, which suggests that our conclusions are generalizable across cohorts of different diseases and block age ranges. The data presented here should help researchers use FFPE tissue to answer questions central to the care of their patients.

Acknowledgments

We thank Alex Forrest-Hay, Brian Read, and Affymetrix, who provided the 1.0 ST array plates for the prostate cancer study and contributed extensively to the scientific discussion regarding the assay; the late Dr. Edward Fox (Microarray Core Facility, Dana Farber Cancer Institute) for contributions and collaboration; Kathryn Penney, Travis Gerke, Michael Pitt, and Elizabeth Nuttall for assistance on the prostate cancer study; and the participants in the Physicians' Health Study prostate cancer project.

L.A.M., M.B., N.E.M., S.T., G.P., M.J.S., and M.L. conceived and designed the studies; E.C.S., W.W., V.V., M.F., R.T.L., L.W., and M.L. compiled data; S.T. analyzed the data; S.T., N.E.M., G.P., L.A.M., M.B., and L.W. interpreted data; and S.T., N.E.M., G.P., M.B., and L.A.M. wrote the manuscript.

Footnotes

Supported by National Cancer Institute/NIH grants 1RC4CA156551-01 and 5R01CA142832-05 (M.B.), 5R01CA141298 (M.J.S.), and CA136578 and CA090381 (L.A.M.); The Julie Fund; the Prostate Cancer Foundation; and by NIH grants R01CA131945 and P50CA90381 (M.L.).

S.T. and N.E.M. contributed equally to this work.

L.A.M. and M.B. contributed equally to this work as senior authors.

Disclosures: Affymetrix provided the 1.0 ST array plates for the prostate cancer study and contributed extensively to the scientific discussion regarding the assay.

Supplemental Data

Supplemental Figure S1

Boxplots of raw and RMA normalized NuGen + Affymetrix prostate sample data.

mmc1.pdf (9.6KB, pdf)
Supplemental Figure S2

Boxplots of the correlations between technical replicates against the input RNA amount.

mmc2.pdf (5.6KB, pdf)
Supplemental Table S1
mmc3.docx (21KB, docx)
Supplemental Table S2
mmc4.docx (25.6KB, docx)

References

  • 1.Mirnezami R., Nicholson J., Darzi A. Preparing for precision medicine. N Engl J Med. 2012;366:489–491. doi: 10.1056/NEJMp1114866. [DOI] [PubMed] [Google Scholar]
  • 2.Lynch T.J., Bell D.W., Sordella R., Gurubhagavatula S., Okimoto R.A., Brannigan B.W., Harris P.L., Haserlat S.M., Supko J.G., Haluska F.G., Louis D.N., Christiani D.C., Settleman J., Haber D.A. Activating mutations in the epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinib. N Engl J Med. 2004;350:2129–2139. doi: 10.1056/NEJMoa040938. [DOI] [PubMed] [Google Scholar]
  • 3.Slamon D.J., Leyland-Jones B., Shak S., Fuchs H., Paton V., Bajamonde A., Fleming T., Eiermann W., Wolter J., Pegram M., Baselga J., Norton L. Use of chemotherapy plus a monoclonal antibody against HER2 for metastatic breast cancer that overexpresses HER2. N Engl J Med. 2001;344:783–792. doi: 10.1056/NEJM200103153441101. [DOI] [PubMed] [Google Scholar]
  • 4.Paik S., Shak S., Tang G., Kim C., Baker J., Cronin M., Baehner F.L., Walker M.G., Watson D., Park T., Hiller W., Fisher E.R., Wickerham D.L., Bryant J., Wolmark N. A multigene assay to predict recurrence of tamoxifen-treated, node-negative breast cancer. N Engl J Med. 2004;351:2817–2826. doi: 10.1056/NEJMoa041588. [DOI] [PubMed] [Google Scholar]
  • 5.Hoshida Y., Villanueva A., Kobayashi M., Peix J., Chiang D.Y., Camargo A., Gupta S., Moore J., Wrobel M.J., Lerner J., Reich M., Chan J.A., Glickman J.N., Ikeda K., Hashimoto M., Watanabe G., Daidone M.G., Roayaie S., Schwartz M., Thung S., Salvesen H.B., Gabriel S., Mazzaferro V., Bruix J., Friedman S.L., Kumada H., Llovet J.M., Golub T.R. Gene expression in fixed tissues and outcome in hepatocellular carcinoma. N Engl J Med. 2008;359:1995–2004. doi: 10.1056/NEJMoa0804525. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Cuzick J., Swanson G.P., Fisher G., Brothman A.R., Berney D.M., Reid J.E., Mesher D., Speights V., Stankiewicz E., Foster C.S., Møller H., Scardino P., Warren J.D., Park J., Younus A., Flake D.D., Wagner S., Gutin A., Lanchbury J.S., Stone S., Transatlantic Prostate Group Prognostic value of an RNA expression signature derived from cell cycle proliferation genes in patients with prostate cancer: a retrospective study. Lancet Oncol. 2011;12:245–255. doi: 10.1016/S1470-2045(10)70295-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.von Ahlfen S., Missel A., Bendrat K., Schlumpberger M. Determinants of RNA quality from FFPE samples. PLoS One. 2007;2:e1261. doi: 10.1371/journal.pone.0001261. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Waldron L., Simpson P., Parmigiani G., Huttenhower C. Report on emerging technologies for translational bioinformatics: a symposium on gene expression profiling for archival tissues. BMC Cancer. 2012;12:124. doi: 10.1186/1471-2407-12-124. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Linton K., Hey Y., Dibben S., Miller C., Freemont A., Radford J., Pepper S. Methods comparison for high-resolution transcriptional analysis of archival material on Affymetrix Plus 2.0 and Exon 1.0 microarrays. Biotechniques. 2009;47:587–596. doi: 10.2144/000113169. [DOI] [PubMed] [Google Scholar]
  • 10.Thomas M., Poignee-Heger M., Weisser M., Wessner S., Belousov A. An optimized workflow for improved gene expression profiling for formalin-fixed, paraffin-embedded tumor samples. J Clin Bioinforma. 2013;3:10. doi: 10.1186/2043-9113-3-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Reis P.P., Waldron L., Goswami R.S., Xu W., Xuan Y., Perez-Ordonez B., Gullane P., Irish J., Jurisica I., Kamel-Reid S. mRNA transcript quantification in archival samples using multiplexed, color-coded probes. BMC Biotechnol. 2011;11:46. doi: 10.1186/1472-6750-11-46. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Final report on the aspirin component of the ongoing Physicians' Health Study. Steering Committee of the Physicians' Health Study Research Group. N Engl J Med. 1989;321:129–135. doi: 10.1056/NEJM198907203210301. [DOI] [PubMed] [Google Scholar]
  • 13.Hennekens C.H., Buring J.E., Manson J.E., Stampfer M., Rosner B., Cook N.R., Belanger C., LaMotte F., Gaziano J.M., Ridker P.M., Willett W., Peto R. Lack of effect of long-term supplementation with beta carotene on the incidence of malignant neoplasms and cardiovascular disease. N Engl J Med. 1996;334:1145–1149. doi: 10.1056/NEJM199605023341801. [DOI] [PubMed] [Google Scholar]
  • 14.Burger R.A., Brady M.F., Bookman M.A., Fleming G.F., Monk B.J., Huang H., Mannel R.S., Homesley H.D., Fowler J., Greer B.E., Boente M., Birrer M.J., Liang S.X., Gynecologic Oncology Group Incorporation of bevacizumab in the primary treatment of ovarian cancer. N Engl J Med. 2011;365:2473–2483. doi: 10.1056/NEJMoa1104390. [DOI] [PubMed] [Google Scholar]
  • 15.Hall J.S., Leong H.S., Armenoult L.S., Newton G.E., Valentine H.R., Irlam J.J., Moller-Levet C., Sikand K.A., Pepper S.D., Miller C.J., West C.M. Exon-array profiling unlocks clinically and biologically relevant gene signatures from formalin-fixed paraffin-embedded tumour samples. Br J Cancer. 2011;104:971–981. doi: 10.1038/bjc.2011.66. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Kennedy R.D., Bylesjo M., Kerr P., Davison T., Black J.M., Kay E.W. Development and independent validation of a prognostic assay for stage II colon cancer using formalin-fixed paraffin-embedded tissue. J Clin Oncol. 2011;29:4620–4626. doi: 10.1200/JCO.2011.35.4498. [DOI] [PubMed] [Google Scholar]
  • 17.Abdueva D., Wing M., Schaub B., Triche T., Davicioni E. Quantitative expression profiling in formalin-fixed paraffin-embedded samples by affymetrix microarrays. J Mol Diagn. 2010;12:409–417. doi: 10.2353/jmoldx.2010.090155. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Geiss G.K., Bumgarner R.E., Birditt B., Dahl T., Dowidar N., Dunaway D.L., Fell H.P., Ferree S., George R.D., Grogan T., James J.J., Maysuria M., Mitton J.D., Oliveri P., Osborn J.L., Peng T., Ratcliffe A.L., Webster P.J., Davidson E.H., Hood L., Dimitrov K. Direct multiplexed measurement of gene expression with color-coded probe pairs. Nat Biotechnol. 2008;26:317–325. doi: 10.1038/nbt1385. [DOI] [PubMed] [Google Scholar]
  • 19.Irizarry R.A., Hobbs B., Collin F., Beazer-Barclay Y.D., Antonellis K.J., Scherf U., Speed T.P. Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics. 2003;4:249–264. doi: 10.1093/biostatistics/4.2.249. [DOI] [PubMed] [Google Scholar]
  • 20.Irizarry R.A., Bolstad B.M., Collin F., Cope L.M., Hobbs B., Speed T.P. Summaries of Affymetrix GeneChip probe level data. Nucleic Acids Res. 2003;31:e15. doi: 10.1093/nar/gng015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Gentleman R.C., Carey V.J., Bates D.M., Bolstad B., Dettling M., Dudoit S., Ellis B., Gautier L., Ge Y., Gentry J., Hornik K., Hothorn T., Huber W., Iacus S., Irizarry R., Leisch F., Li C., Maechler M., Rossini A.J., Sawitzki G., Smith C., Smyth G., Tierney L., Yang J.Y., Zhang J. Bioconductor: open software development for computational biology and bioinformatics. Genome Biol. 2004;5:R80. doi: 10.1186/gb-2004-5-10-r80. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Faul F., Erdfelder E., Lang A.-G., Buchner A. G*Power 3: a flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behav Res Methods. 2007;39:175–191. doi: 10.3758/bf03193146. [DOI] [PubMed] [Google Scholar]
  • 23.Grasso C.S., Wu Y.M., Robinson D.R., Cao X., Dhanasekaran S.M., Khan A.P., Quist M.J., Jing X., Lonigro R.J., Brenner J.C., Asangani I.A., Ateeq B., Chun S.Y., Siddiqui J., Sam L., Anstett M., Mehra R., Prensner J.R., Palanisamy N., Ryslik G.A., Vandin F., Raphael B.J., Kunju L.P., Rhodes D.R., Pienta K.J., Chinnaiyan A.M., Tomlins S.A. The mutational landscape of lethal castration-resistant prostate cancer. Nature. 2012;487:239–243. doi: 10.1038/nature11125. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Lapointe J., Li C., Higgins J.P., van de Rijn M., Bair E., Montgomery K., Ferrari M., Egevad L., Rayford W., Bergerheim U., Ekman P., DeMarzo A.M., Tibshirani R., Botstein D., Brown P.O., Brooks J.D., Pollack J.R. Gene expression profiling identifies clinically relevant subtypes of prostate cancer. Proc Natl Acad Sci U S A. 2004;101:811–816. doi: 10.1073/pnas.0304146101. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Taylor B.S., Schultz N., Hieronymus H., Gopalan A., Xiao Y., Carver B.S., Arora V.K., Kaushik P., Cerami E., Reva B., Antipin Y., Mitsiades N., Landers T., Dolgalev I., Major J.E., Wilson M., Socci N.D., Lash A.E., Heguy A., Eastham J.A., Scher H.I., Reuter V.E., Scardino P.T., Sander C., Sawyers C.L., Gerald W.L. Integrative genomic profiling of human prostate cancer. Cancer Cell. 2010;18:11–22. doi: 10.1016/j.ccr.2010.05.026. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Tomlins S.A., Mehra R., Rhodes D.R., Cao X., Wang L., Dhanasekaran S.M., Kalyana-Sundaram S., Wei J.T., Rubin M.A., Pienta K.J., Shah R.B., Chinnaiyan A.M. Integrative molecular concept modeling of prostate cancer progression. Nat Genet. 2007;39:41–51. doi: 10.1038/ng1935. [DOI] [PubMed] [Google Scholar]
  • 27.Penney K.L., Sinnott J.A., Fall K., Pawitan Y., Hoshida Y., Kraft P., Stark J.R., Fiorentino M., Perner S., Finn S., Calza S., Flavin R., Freedman M.L., Setlur S., Sesso H.D., Andersson S.O., Martin N., Kantoff P.W., Johansson J.E., Adami H.O., Rubin M.A., Loda M., Golub T.R., Andren O., Stampfer M.J., Mucci L.A. mRNA expression signature of Gleason grade predicts lethal prostate cancer. J Clin Oncol. 2011;29:2391–2396. doi: 10.1200/JCO.2010.32.6421. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Culhane A.C., Schroder M.S., Sultana R., Picard S.C., Martinelli E.N., Kelly C., Haibe-Kains B., Kapushesky M., St Pierre A.A., Flahive W., Picard K.C., Gusenleitner D., Papenhausen G., O'Connor N., Correll M., Quackenbush J. GeneSigDB: a manually curated database and resource for analysis of gene expression signatures. Nucleic Acids Res. 2012;40(Database issue):D1060–D1066. doi: 10.1093/nar/gkr901. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Schwartz D.R., Kardia S.L., Shedden K.A., Kuick R., Michailidis G., Taylor J.M., Misek D.E., Wu R., Zhai Y., Darrah D.M., Reed H., Ellenson L.H., Giordano T.J., Fearon E.R., Hanash S.M., Cho K.R. Gene expression in ovarian cancer reflects both morphology and biological behavior, distinguishing clear cell from other poor-prognosis ovarian carcinomas. Cancer Res. 2002;62:4722–4729. [PubMed] [Google Scholar]
  • 30.Zorn K.K., Bonome T., Gangi L., Chandramouli G.V., Awtrey C.S., Gardner G.J., Barrett J.C., Boyd J., Birrer M.J. Gene expression profiles of serous, endometrioid, and clear cell subtypes of ovarian and endometrial cancer. Clin Cancer Res. 2005;11:6422–6430. doi: 10.1158/1078-0432.CCR-05-0508. [DOI] [PubMed] [Google Scholar]
  • 31.Stany M.P., Vathipadiekal V., Ozbun L., Stone R.L., Mok S.C., Xue H., Kagami T., Wang Y., McAlpine J.N., Bowtell D., Gout P.W., Miller D.M., Gilks C.B., Huntsman D.G., Ellard S.L., Wang Y.Z., Vivas-Mejia P., Lopez-Berestein G., Sood A.K., Birrer M.J. Identification of novel therapeutic targets in microdissected clear cell ovarian cancers. PLoS One. 2011;6:e21121. doi: 10.1371/journal.pone.0021121. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Norton N., Sun Z., Asmann Y.W., Serie D.J., Necela B.M., Bhagwate A., Jen J., Eckloff B.W., Kalari K.R., Thompson K.J., Carr J.M., Kachergus J.M., Geiger X.J., Perez E.A., Thompson E.A. Gene expression, single nucleotide variant and fusion transcript discovery in archival material from breast tumors. PLoS One. 2013;8:e81925. doi: 10.1371/journal.pone.0081925. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Wang D., Wang C., Zhang L., Xiao H., Shen X., Ren L., Zhao W., Hong G., Zhang Y., Zhu J., Zhang M., Yang D., Ma W., Guo Z. Evaluation of cDNA microarray data by multiple clones mapping to the same transcript. OMICS. 2009;13:493–499. doi: 10.1089/omi.2009.0077. [DOI] [PubMed] [Google Scholar]
  • 34.Northcott P.A., Shih D.J., Remke M., Cho Y.J., Kool M., Hawkins C., Eberhart C.G., Dubuc A., Guettouche T., Cardentey Y., Bouffet E., Pomeroy S.L., Marra M., Malkin D., Rutka J.T., Korshunov A., Pfister S., Taylor M.D. Rapid, reliable, and reproducible molecular sub-grouping of clinical medulloblastoma samples. Acta Neuropathol. 2012;123:615–626. doi: 10.1007/s00401-011-0899-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Waldron L., Ogino S., Hoshida Y., Shima K., McCart Reed A.E., Simpson P.T., Baba Y., Nosho K., Segata N., Vargas A.C., Cummings M.C., Lakhani S.R., Kirkner G.J., Giovannucci E., Quackenbush J., Golub T.R., Fuchs C.S., Parmigiani G., Huttenhower C. Expression profiling of archival tumors for long-term health studies. Clin Cancer Res. 2012;18:6136–6146. doi: 10.1158/1078-0432.CCR-12-1915. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Erho N., Crisan A., Vergara I.A., Mitra A.P., Ghadessi M., Buerki C., Bergstralh E.J., Kollmeyer T., Fink S., Haddad Z., Zimmermann B., Sierocinski T., Ballman K.V., Triche T.J., Black P.C., Karnes R.J., Klee G., Davicioni E., Jenkins R.B. Discovery and validation of a prostate cancer genomic classifier that predicts early metastasis following radical prostatectomy. PLoS One. 2013;8:e66855. doi: 10.1371/journal.pone.0066855. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Ward S., Scope A., Rafia R., Pandor A., Harnan S., Evans P., Wyld L. Gene expression profiling and expanded immunohistochemistry tests to guide the use of adjuvant chemotherapy in breast cancer management: a systematic review and cost-effectiveness analysis. Health Technol Assess. 2013;17:1–302. doi: 10.3310/hta17440. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Drukker C.A., van Tinteren H., Schmidt M.K., Rutgers E.J., Bernards R., van de Vijver M.J., Van't Veer L.J. Long-term impact of the 70-gene signature on breast cancer outcome. Breast Cancer Res Treat. 2014;143:587–592. doi: 10.1007/s10549-013-2831-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Doleshal M., Magotra A.A., Choudhury B., Cannon B.D., Labourier E., Szafranska A.E. Evaluation and validation of total RNA extraction methods for microRNA expression analyses in formalin-fixed, paraffin-embedded tissues. J Mol Diagn. 2008;10:203–211. doi: 10.2353/jmoldx.2008.070153. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplemental Figure S1

Boxplots of raw and RMA normalized NuGen + Affymetrix prostate sample data.

mmc1.pdf (9.6KB, pdf)
Supplemental Figure S2

Boxplots of the correlations between technical replicates against the input RNA amount.

mmc2.pdf (5.6KB, pdf)
Supplemental Table S1
mmc3.docx (21KB, docx)
Supplemental Table S2
mmc4.docx (25.6KB, docx)

Articles from The Journal of Molecular Diagnostics : JMD are provided here courtesy of American Society for Investigative Pathology

RESOURCES