Formalin-fixed Paraffin-embedded Tumor Biopsies Research Articles

Construction of a risk stratification model integrating ctDNA to predict response and survival in neoadjuvant-treated breast cancer

BackgroundThe pathological complete response (pCR) to neoadjuvant chemotherapy (NAC) of breast cancer is closely related to a better prognosis. However, there are no reliable indicators to accurately identify which patients will achieve pCR before surgery, and a model for predicting pCR to NAC is required.MethodsA total of 269 breast cancer patients in Shandong Cancer Hospital and Liaocheng People’s Hospital receiving anthracycline and taxane-based NAC were prospectively enrolled. Expression profiling using a 457 cancer-related gene sequencing panel (DNA sequencing) covering genes recurrently mutated in breast cancer was carried out on 243 formalin-fixed paraffin-embedded tumor biopsies samples before NAC from 243 patients. The unique personalized panel of nine individual somatic mutation genes from the constructed model was used to detect and analyze ctDNA on 216 blood samples. Blood samples were collected at indicated time points including before chemotherapy initiation, after the 1st NAC and before the 2nd NAC cycle, during intermediate evaluation, and prior to surgery. In this study, we characterized the value of gene profile mutation and circulating tumor DNA (ctDNA) in combination with clinical characteristics in the prediction of pCR before surgery and investigated the prognostic prediction. The median follow-up time for survival analysis was 898 days.ResultsFirstly, we constructed a predictive NAC response model including five single nucleotide variant (SNV) mutations (TP53, SETBP1, PIK3CA, NOTCH4 and MSH2) and four copy number variation (CNV) mutations (FOXP1-gain, EGFR-gain, IL7R-gain, and NFKB1A-gain) in the breast tumor, combined with three clinical factors (luminal A, Her2 and Ki67 status). The tumor prediction model showed good discrimination of chemotherapy sensitivity for pCR and non-pCR with an AUC of 0.871 (95% CI, 0.797–0.927) in the training set, 0.771 (95% CI, 0.649–0.883) in the test set, and 0.726 (95% CI, 0.556–0.865) in an extra test set. This tumor prediction model can also effectively predict the prognosis of disease-free survival (DFS) with an AUC of 0.749 at 1 year and 0.830 at 3 years. We further screened the genes from the tumor prediction model to establish a unique personalized panel consisting of 9 individual somatic mutation genes to detect and analyze ctDNA. It was found that ctDNA positivity decreased with the passage of time during NAC, and ctDNA status can predict NAC response and metastasis recurrence. Finally, we constructed the chemotherapy prediction model combined with the tumor prediction model and pretreatment ctDNA levels, which has a better prediction effect of pCR with the AUC value of 0.961.ConclusionsIn this study, we established a chemotherapy predictive model with a non-invasive tool that is built based on genomic features, ctDNA status, as well as clinical characteristics for predicting pCR to recognize the responders and non-responders to NAC, and also predicting prognosis for DFS in breast cancer. Adding pretreatment ctDNA levels to a model containing gene profile mutation and clinical characteristics significantly improves stratification over the clinical variables alone.

Mass spectrometry quantifies target engagement for a KRASG12C inhibitor in FFPE tumor tissue

BackgroundQuantification of drug-target binding is critical for confirming that drugs reach their intended protein targets, understanding the mechanism of action, and interpreting dose-response relationships. For covalent inhibitors, target engagement can be inferred by free target levels before and after treatment. Targeted mass spectrometry assays offer precise protein quantification in complex biological samples and have been routinely applied in pre-clinical studies to quantify target engagement in frozen tumor tissues for oncology drug development. However, frozen tissues are often not available from clinical trials so it is critical that assays are applicable to formalin-fixed, paraffin-embedded (FFPE) tissues in order to extend mass spectrometry-based target engagement studies into clinical settings.MethodsWild-type RAS and RASG12C was quantified in FFPE tissues by a highly optimized targeted mass spectrometry assay that couples high-field asymmetric waveform ion mobility spectrometry (FAIMS) and parallel reaction monitoring (PRM) with internal standards. In a subset of samples, technical reproducibility was evaluated by analyzing consecutive tissue sections from the same tumor block and biological variation was accessed among adjacent tumor regions in the same tissue section.ResultsWild-type RAS protein was measured in 32 clinical non-small cell lung cancer tumors (622–2525 amol/µg) as measured by FAIMS-PRM mass spectrometry. Tumors with a known KRASG12C mutation (n = 17) expressed a wide range of RASG12C mutant protein (127–2012 amol/µg). The variation in wild-type RAS and RASG12C measurements ranged 0–18% CV across consecutive tissue sections and 5–20% CV among adjacent tissue regions. Quantitative target engagement was then demonstrated in FFPE tissues from 2 xenograft models (MIA PaCa-2 and NCI-H2122) treated with a RASG12C inhibitor (AZD4625).ConclusionsThis work illustrates the potential to expand mass spectrometry-based proteomics in preclinical and clinical oncology drug development through analysis of FFPE tumor biopsies.

High Prevalence of Epigenetic Mutations in Histiocytic and Dendritic Cell Disorders: Results from Molecular Analysis of a Large Cohort from Histiocytosis Working Group

Introduction Over the last decade, recurrent oncogenic alterations in the MAPK/ERK pathway have been discovered in various histiocytic and dendritic cell disorders, especially Erdheim-Chester disease (ECD), Langerhans cell histiocytosis (LCH), and Rosai-Dorfman disease (RDD). These discoveries have led to the successful use of targeted BRAF- and MEK-inhibitors. However, most data are derived from small institutional studies mostly focusing on ECD and LCH, which leaves room for a comprehensive evaluation of the molecular underpinnings in a large cohort. Methods Formalin-fixed paraffin embedded tumor biopsies from various histiocytic/dendritic cell disorders seen at a tertiary institution from 2001-2021 were subjected to whole exome sequencing (WES) using one of two assays: in-house SureSelectXT AllExon V5+UTR or commercially available Tempus xT assay. Identified variants were filtered to remove polymorphic, low coding impact and low read coverage variants. A variant was then selected for further analysis if it met one of the following criteria: 1) observed as somatic in ≥1 tumor-germline pairs or COSMIC database or 2) observed in at least three samples with an allele frequency of ≥10% in ≥1 sample. Cases were clustered into mutually exclusive groups based on their primary driver mutational profile using the following funneling order: BRAF-V600E (BRAF-V600E mutations), BRAF*/NRAS/KRAS (BRAF non-V600E, KRAS, or NRAS mutations), CSF1R (CSF1R mutations), MAPK (MAPK/MEK mutations), and PI3K (AKT3, INPP4B, MTOR, PIK3CA, PIK3R1, PIK3R2, PIK3R3, PPP2R1A, MTR mutations). A cancer specific pathway enrichment was performed on the unknown cases, which identified epigenetic mutations as major pathway alterations observed in the cohort. We also correlated the presence of epigenetic mutations with clinical features and outcomes. Results We included 114 cases with histiocytic and dendritic cell disorders in the study. Median age at diagnosis was 51y (IQR 37-64). The disease distribution was as follows: 46 LCH, 20 ECD, 17 RDD, 8 follicular dendritic cell sarcoma (FDCS), 1 adult xanthogranuloma, 12 malignant histiocytosis (MH; including histiocytic and Langerhans cell sarcoma), and 10 multicentric reticulohistiocytosis (MRH). Based on the known driver mutations in histiocytoses, the following 5 molecular groups were identified: BRAF-V600E (n= 19, 17%), BRAF*/KRAS/NRAS (n=17, 15%), MAPK (n=14, 12%), PI3K (n=8, 7%), CSF1R (n=2, 2%) (Figure 1). A large proportion (n=48, 42%) did not have any of the known driver mutations, while 6 cases (5%) had no alterations identified. Cancer-specific enrichment analysis in the entire cohort revealed mutations in epigenetic genes in 47 (41%) cases, including 30 (26%) cases with identifiable driver mutations and 17 (15%) cases without known driver mutations. Median number of epigenetic mutations was 1 (range 1-11); 21% cases had 2 and 10% cases had 3 epigenetic mutations, respectively. The most common epigenetic mutations included those in KMT2C (34%), KMT2D (23%), and ARID1A (19%). The mutational groups with highest prevalence of epigenetic mutations were CSF1R (100%), MAPK (71%), and PI3K (63%), while lowest prevalence of these alterations was found in BRAF-V600E group (16%; p=0.0003). The epigenetic gene mutations were seen more commonly in MRH (90%) and MH (50%) compared with ECD (45%), RDD (41%), FDCS (38%), and LCH (33%; p=0.037). There was no association between epigenetic gene mutations and age at diagnosis or organ involvement. The median overall survival (OS) was not significantly different between those with epigenetic mutations and those without such alterations (22y vs. 17y, p=0.89). Conclusions In this large study of patients with diverse histiocytic and dendritic cell disorders, we found a high prevalence of mutations in epigenetic genes (41%). These mutations were specifically enriched in tumors with non-BRAF-V600E mutations, and in MRH/MH. These data shed light on additional mechanisms of disease pathogenesis, and with available epigenetic therapies, warrant further investigation. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

The addition of FAIMS increases targeted proteomics sensitivity from FFPE tumor biopsies

Mass spectrometry-based targeted proteomics allows objective protein quantitation of clinical biomarkers from a single section of formalin-fixed, paraffin-embedded (FFPE) tumor tissue biopsies. We combined high-field asymmetric waveform ion mobility spectrometry (FAIMS) and parallel reaction monitoring (PRM) to increase assay sensitivity. The modular nature of the FAIMS source allowed direct comparison of the performance of FAIMS-PRM to PRM. Limits of quantitation were determined by spiking synthetic peptides into a human spleen matrix. In addition, 20 clinical samples were analyzed using FAIMS-PRM and the quantitation of HER2 was compared with that obtained with the Ventana immunohistochemistry assay. FAIMS-PRM improved the overall signal-to-noise ratio over that from PRM and increased assay sensitivity in FFPE tissue analysis for four (HER2, EGFR, cMET, and KRAS) of five proteins of clinical interest. FAIMS-PRM enabled sensitive quantitation of basal HER2 expression in breast cancer samples classified as HER2 negative by immunohistochemistry. Furthermore, we determined the degree of FAIMS-dependent background reduction and showed that this correlated with an improved lower limit of quantitation with FAIMS. FAIMS-PRM is anticipated to benefit clinical trials in which multiple biomarker questions must be addressed and the availability of tumor biopsy samples is limited.

Identification of proteomic markers for prediction of the response to 5-Fluorouracil based neoadjuvant chemoradiotherapy in locally advanced rectal cancer patients

BackgroundNeoadjuvant chemoradiotherapy (nCRT) prior to surgery is the standard treatment for patients with locally advanced rectal cancer (LARC), while parts of them show poor therapeutic response accompanied by therapy adverse effects. Predictive biomarkers for nCRT response could facilitate the guidance on treatment decisions but are still insufficient until now, which limits the clinical applications of nCRT in LARC patients.MethodsIn our study, 37 formalin-fixed paraffin-embedded tumor biopsies were obtained from patients with LARC before receiving 5-fluorouracil based nCRT. Proteomics analyses were conducted to identify the differentially expressed proteins (DEPs) between total responders (TR) and poor responders (PR). The DEPs were validated via ROC plotter web tool and their predictive performance was evaluated by receiver operating characteristic analysis. Functional enrichment analyses were performed to further explore the potential mechanisms underlying nCRT response.ResultsAmong 3,998 total proteins, 91 DEPs between TR and PR were screened out. HSPA4, NIPSNAP1, and SPTB all with areas under the curve (AUC) ~ 0.8 in the internal discovery cohort were independently validated by the external mRNA datasets (AUC ~ 0.7), and their protein levels were linearly correlated with the graded responses to nCRT in the internal cohort. The combination of HSPA4 and SPTB could distinctly discriminate the TR and PR groups (AUC = 0.980, p < 0.0001). Moreover, multiple combinations of the three proteins realized increased specificity and/or sensitivity, while achieving favorable predictive value when moderate responders were introduced into the ROC analysis. Pathways including DNA damage repair, cell cycle, and epithelial mesenchymal transition were involved in nCRT response according to the enrichment analysis results.ConclusionsHSPA4, SPTB and NIPSNAP1 in tumor biopsies and/or their optional combinations might be potential predictive markers for nCRT response in patients with LARC. The DEPs and their related functions have implications for the potential mechanisms of treatment response to nCRT in patients with LARC.

Mutational Characterization and Potential Prognostic Biomarkers of Chinese Patients with Esophageal Squamous Cell Carcinoma.

PurposeEsophageal squamous cell carcinoma (ESCC) is the most common type of esophageal cancer in China and the 5-year mortality rate is up to 70%. Studies on the ESCC genetic landscape are needed to further explore clinical therapeutic strategies. In this study, we evaluated the genetic landscape of ESCC to aid the search for clinical therapeutic strategies.Patients and MethodsA total of 225 ESCC patients were enrolled in this study. Deep sequencing of 450 cancer genes was performed on formalin-fixed paraffin-embedded tumor biopsies and matched blood samples from patients. Tumor mutational burden (TMB) was calculated using an algorithm developed in-house.ResultsOur results showed that the most commonly mutated genes in ESCC were TP53 (96%), CCND1 (46%), FGF4 (44%), FGF19 (44%), FGF3 (44%), CDKN2A (31%), PIK3CA (26%), NOTCH1 (24%), KMT2D (18%), FAT1 (16%), and LRP1B (16%). We found that TMB correlated with patient drinking status. We identified mutations associated with sex, early ESCC, high TMB, and metastasis lymph nodes. KMT2D mutations associated with sex (P = 0.035), tumor stage (P = 0.016), high TMB (P = 0.0072), and overall survival of patients (P = 0.0026). SPEN mutations associated with high TMB (P = 0.0016) and metastasis-positive lymph nodes (P = 0.027). These results suggested that SPEN and KMT2D could be potential prognosis biomarkers for Chinese patients with ESCC. We also found that the number of positive lymph nodes was associated with disease-free survival. Clinical target gene analysis indicated that nearly half of Chinese ESCC patients might benefit from treatment with gene-specific target drugs.ConclusionOur study revealed the ESCC mutational landscape in 225 Chinese patients and uncovered the potential prognosis biomarker for Chinese patients with ESCC.

Shallow whole-genome sequencing of plasma cell-free DNA accurately differentiates small from non-small cell lung carcinoma

BackgroundAccurate lung cancer classification is crucial to guide therapeutic decisions. However, histological subtyping by pathologists requires tumor tissue—a necessity that is often intrinsically associated with procedural difficulties. The analysis of circulating tumor DNA present in minimal-invasive blood samples, referred to as liquid biopsies, could therefore emerge as an attractive alternative.MethodsConcerning adenocarcinoma, squamous cell carcinoma, and small cell carcinoma, our proof of concept study investigates the potential of liquid biopsy-derived copy number alterations, derived from single-end shallow whole-genome sequencing (coverage 0.1–0.5×), across 51 advanced stage lung cancer patients.ResultsGenomic abnormality testing reveals anomalies in 86.3% of the liquid biopsies (16/20 for adenocarcinoma, 13/16 for squamous cell, and 15/15 for small cell carcinoma). We demonstrate that copy number profiles from formalin-fixed paraffin-embedded tumor biopsies are well represented by their liquid equivalent. This is especially valid within the small cell carcinoma group, where paired profiles have an average Pearson correlation of 0.86 (95% CI 0.79–0.93). A predictive model trained with public data, derived from 843 tissue biopsies, shows that liquid biopsies exhibit multiple deviations that reflect histological classification. Most notably, distinguishing small from non-small cell lung cancer is characterized by an area under the curve of 0.98 during receiver operating characteristic analysis. Additionally, we investigated how deeper paired-end sequencing, which will eventually become feasible for routine diagnosis, empowers tumor read enrichment by insert size filtering: for all of the 29 resequenced liquid biopsies, the tumor fraction could be increased in silico, thereby “rescuing” three out of five cases with previously undetectable alterations.ConclusionsCopy number profiling of cell-free DNA enables histological classification. Since shallow whole-genome sequencing is inexpensive and often fully operational at routine molecular laboratories, this finding has current diagnostic potential, especially for patients with lesions that are difficult to reach.

Abstract 2896: Impact of intratumoral heterogeneity on gene expression profiles and known biomarkers in FFPE tumor biopsies from early stage breast cancer patients

Abstract Background: Cancer is a dynamic disease that becomes more heterogeneous during its course, with cellular subpopulations with distinct molecular profiles across different regions. Because this intratumoral heterogeneity may result in varying levels of gene expression patterns and sensitivity to treatment across different regions of the tumor, it is important to understand patterns of heterogeneity to develop effective biomarkers. We conducted multiregion sampling within single breast tumor biopsies to evaluate the effect of intratumoral heterogeneity on expression of known and novel biomarkers. Methods: We sampled three sequential regions of breast tumor biopsies to compare gene expression patterns determined by a novel biomarker panel (BA355), 5 known biomarkers (ERS1, ERBB2, PGR, AR, and MKi67) and a novel predictive classifier (BA100). RNA expression levels were measured using a proprietary custom designed nanoString nCounter assay. We investigated formalin-fixed paraffin-embedded (FFPE) breast tumor biopsies from a cohort of 40 early stage breast cancer patients (T1 N0 M0) collected prospectively. RNA was extracted from upper, middle, and lower regions of each biopsy, resulting in a total of 120 samples quantified and QC using standard procedures. Each layer consisted of 40 micrometers, with an interval of 50 micrometers of tissue. BA100 linear regression models were used to generate prediction scores. Analyses were done in R. Results: Eighteen of the 120 samples had poor data quality (sample median = 0 or inter quantile range &amp;gt; mean + 2*sd). Further analyses only included 102 good quality samples corresponding to 34 FFPE biopsies. Among 33 of 34 samples, gene expression patterns between the three regions of the biopsies were highly correlated, with correlation coefficients ranging from 0.90-0.99 (p-values ranged from 2.07e-19 to 8.45e-319). Among 31 of 34 samples, the three regions were clustered together in cluster analysis. The coefficient of variation (CV) for ESR1 ranged from 0.37%-15.5%, with a majority (32 of 34) of the samples having CV &amp;lt;10%. While the CVs of ERBB2, PGR, and AR were within the same range as ESR1, MKI67 had much larger variation (CV ranged from 0.99%-173%). BA100 scores yielded consistent predictions for the three regions in 33 of 34 samples. Conclusion: Examination of gene expression patterns across multiple regions of breast tumor biopsies did not reveal a significant impact of intratumoral heterogeneity on expression patterns using a novel gene signature panel. However, at the individual gene level, while a number of genes including ESR1, ERBB2, PGR, and ARdisplayed consistent gene expression, genes such as MKI67 showed significant variation, indicating that tumor heterogeneity may impact individual gene biomarkers. Further studies are ongoing to evaluate intratumoral heterogeneity. Citation Format: Lucas Delmonico, Joan Chen, Trushna Desai, Lisa Barnett, Aladdin Shadyab, John Obenauer, Gilda Alves Brown, Mauricio Magalhaes Costa, Marcia Fournier. Impact of intratumoral heterogeneity on gene expression profiles and known biomarkers in FFPE tumor biopsies from early stage breast cancer patients [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 2896.

Abstract 1801: MicroRNA expression profiling predicts clinical outcome in patients with locally advanced larynx and hypopharynx cancer treated with chemoradiotherapy

Abstract Background For patients with locally advanced larynx and hypopharynx cancer, organ preservation chemoradiotherapy (CRT) protocols are frequently applied. However, some tumors are resistant to CRT or will recur within a short time frame. For these patients upfront surgical resection may result in an improved cure rate. Currently, there are no molecular markers that can predict CRT response. In this study, tumor microRNA (miRNA) profiling was performed to predict clinical outcome in patients with locally advanced larynx or hypopharynx cancer treated with CRT. Methods First, matched fresh frozen (FF) and formalin-fixed paraffin-embedded (FFPE) tumor biopsies from 12 patients were profiled by Next Generation Sequencing (NGS) to ensure valid miRNA results from FFPE. Next, miRNA profiles of FFPE tumor biopsies from 37 patients with larynx or hypopharynx cancer, treated with primary CRT, were determined. Clinicopathological data, including response to CRT, tumor stage and nodal stage, was collected of all patients. Differential expression analysis was performed to compare patients with poor clinical outcome, i.e residual tumor after treatment or progression free survival (PFS) &amp;lt; 2 years versus patients with good clinical outcome, i.e. complete response and PFS &amp;gt; 2 years. The median follow up of this cohort was 60 months (range 6 - 118). The Cancer Genome Atlas (TCGA) head and neck miRNA expression data of larynx and hypopharynx tumors and normal squamous cell tissues were used to select tumor specific miRNAs. By group-regularized penalized logistic ridge regression analysis, a model to predict clinical outcome was generated using tumor specific miRNAs combined with clinical covariates. Results In total, 324 novel candidate and 2,262 mature miRNAs were detected. A comparative analysis of miRNA profiles generated using matched FF and FFPE tumor biopsies showed a high correlation (mean r2=0.75). Using only FFPE samples, 13 miRNAs were significantly differentially expressed between patients with good versus poor clinical outcome (p&amp;lt;0.05, FDR&amp;lt;0.1). Class prediction resulted in a model of three miRNAs combined with AJCC stage, able to predict recurrence with an area under the curve of 0.83. Further validation of the model is ongoing in an additional cohort of patients with larynx or hypopharynx cancer treated with CRT. Conclusion NGS miRNA profiling using FFPE tumor biopsies from patients with head and neck cancer is technically feasible, resulting in high quality miRNA expression data comparable to fresh frozen tissue. Based on miRNA expression profiling, candidate miRNAs were identified predictive for clinical outcome following treatment with CRT. A validation study is ongoing. Citation Format: Dennis Poel, Francois Rustenburg, Daoud Sie, Hendrik F. van Essen, Paul P. Eijk, Elisabeth Bloemena, Teresita E. Benites, Bauke Ylstra, Brakenhoff H. Ruud, René C. Leemans, Tineke E. Buffart, Henk M. Verheul, Jens Voortman. MicroRNA expression profiling predicts clinical outcome in patients with locally advanced larynx and hypopharynx cancer treated with chemoradiotherapy [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 1801.

Baseline Tumor Transcriptome Characteristics Associated with the Response of Relapsed/Refractory Hodgkin Lymphoma Patients to Brentuximab Vedotin in Combination with Nivolumab

Introduction: In recently-reported interim results from a phase 1/2, open-label, multicenter study (NCT02572167), we showed that 61% of relapsed/refractory Hodgkin Lymphoma (HL) patients achieved complete responses (CRs) after treatment with Brentuximab Vedotin (BV) in combination with Nivolumab (Nivo) (Herrera et al., 2018). To gain insight into tumor microenvironment driven disease characteristics associated with CRs after combined BV+Nivo therapy, we performed RNA sequencing (RNA-Seq) transcriptome analysis of formalin-fixed paraffin-embedded tumor biopsies obtained from 37 study participants (23 CR, 14 PR/SD/PD) prior to the start of treatment.Results: Within a set of 132 candidate genes comprised of BV+Nivo therapeutic targets (CD30, PD-1, PD-L1), the markers for immune cells, inflammatory response factors, and potential resistance mechanism factors within the tumor microenvironment, the gene most strongly associated with CRs was the BV target CD30 (TNFRSF8) (p=4x10-4, FDR=5%). Baseline tumor expression of CD30 above 16 counts per million mapped reads (CPM) was strongly associated with achieving CR after BV+Nivo (65% sensitivity and 93% specificity). CRs were associated with several other factors, including higher baseline tumor expression of macrophage markers CD14 and CD163 and lower expression of the Nivo target PD-1 (PDCD1) and multidrug resistance-associated protein 2 (MRP2/ABCC2) in the tumor microenvironment. Both PD-1 and MRP2 were significantly complementary to CD30 for discriminating CRs from the other patients on the basis of baseline gene expression profiling in tumor microenvironment (p<0.01, FDR<20%), and a simple baseline gene expression ratio of CD30/PD-1 achieved high accuracy (area under the receiver operating characteristic curve = 0.93). Interestingly, baseline expression of PD-L1 (CD274) was not significantly associated with development of CRs after BV+Nivo, either by itself (p=0.13) or in combination with CD30 (p=0.22), which may be expected given that consistently high PD-L1 expression has been reported for HL. To gain further insight into the role of the tumor microenvironment (TME) in determining HL responsiveness to BV+Nivo, we performed mathematical deconvolution of the baseline HL biopsy transcriptomes. This analysis suggested that both macrophages and T cells were major leukocyte constituents of the HL biopsy TMEs, with higher baseline macrophage abundance and lower baseline T cell abundance being associated with higher CR rates after BV+Nivo treatment.Conclusions: The CR rate for combination therapy of BV+Nivo in relapsed/refractory HL is nearly double the rate reported for this patient population for either agent individually, suggesting the possibility of complementary mechanisms between BV-targeted and Nivo-targeted pathways. By performing RNA-Seq analysis of patient baseline tumor biopsies from the BV+Nivo trial (NCT02572167), we were able to identify specific tumor characteristics that are associated with favorable responses to treatment. In particular, the gene expression ratio of CD30/PD-1 was an accurate discriminator for patients that would achieve CRs. This result suggests that BV+Nivo efficacy may be observed in tumors in which the potential for BV-driven immunogenic cell death (ICD) - as indicated by CD30 expression - exceeds a threshold set by the immunosuppressive potential of the tumor (as indicated by PD-1 expression). This hypothesis is supported by the deconvolution analysis results that implicated a favorable role for macrophages (which may become activated in response to BV-induced ICD) in the tumor and a detrimental role for immunosuppressive tumor T cells. Immunohistochemical analyses of biopsies from the study are underway to quantify target abundance and to further characterize the TMEs. These data will be integrated with the transcriptomics to validate and refine the model for BV+Nivo efficacy in relapsed/refractory HL. DisclosuresZak:Seattle Genetics: Employment, Equity Ownership. Ogden:Seattle Genetics: Employment, Equity Ownership. Herrera:Pharmacyclics: Consultancy, Research Funding; Seattle Genetics: Research Funding; Immune Design: Research Funding; Bristol-Myers Squibb: Consultancy, Research Funding; Merck, Inc.: Consultancy, Research Funding; AstraZeneca: Research Funding; Genentech: Consultancy, Research Funding; KiTE Pharma: Consultancy, Research Funding; Gilead Sciences: Research Funding. Sacchi:Bristol-Myers Squibb: Employment, Equity Ownership. Onsum:Seattle Genetics: Employment, Equity Ownership.

Increased expression of long non-coding RNA XIST predicts favorable prognosis of cervical squamous cell carcinoma subsequent to definitive chemoradiation therapy.

The present retrospective study aimed to examine the association between the expression of long non-protein-coding RNAs (lncRNAs) and clinical prognosis in the pretreatment formalin-fixed, paraffin-embedded (FFPE) tissue samples of cervical squamous cell carcinoma patients that underwent platinum-based chemoradiation therapy. Between 2001 and 2013, 49 consecutive patients with squamous cell cervical carcinoma were selected for the present study (median follow-up period, 44.1 months). The patients possessed an International Federation of Gynecology and Obstetrics stage of IB1/IIA1 (with pelvic lymph node metastasis), IB2 or IIA2-IVA, and had been treated with definitive chemoradiation therapy. The pretreatment FFPE tumor biopsies of the patients obtained diagnosis were used for analysis. Total RNAs were extracted from the FFPE tumor tissues and reverse transcription-quantitative polymerase chain reaction was performed to examine the expression level of lncRNAs. The expression level of X inactive-specific transcript (XIST) demonstrated a significant association with the overall survival rate (P=0.014). The 4-year overall survival rates were 87.1 and 54.4% in the high and low XIST expression groups, respectively. Since the expression of XIST is associated with the overall survival rate, this lncRNA has the potential to become a predictor for the prognosis of cervical squamous cell carcinoma patients that are treated with chemoradiation therapy. Additional studies are required to investigate the underlying mechanisms of XIST that are associated with prognosis.

MicroRNA Profiling of Primary Cutaneous Large B-Cell Lymphomas

Aberrant expression of microRNAs is widely accepted to be pathogenetically involved in nodal diffuse large B-cell lymphomas (DLBCLs). However, the microRNAs profiles of primary cutaneous large B-cell lymphomas (PCLBCLs) are not yet described. Its two main subtypes, i.e., primary cutaneous diffuse large B-cell lymphoma, leg type (PCLBCL-LT) and primary cutaneous follicle center lymphoma (PCFCL) are characterized by an activated B-cell (ABC)-genotype and a germinal center B-cell (GCB)-genotype, respectively. We performed high-throughput sequencing analysis on frozen tumor biopsies from 19 cases of PCFCL and PCLBCL-LT to establish microRNA profiles. Cluster analysis of the complete microRNome could not distinguish between the two subtypes, but 16 single microRNAs were found to be differentially expressed. Single microRNA RT-qPCR was conducted on formalin-fixed paraffin-embedded tumor biopsies of 20 additional cases, confirming higher expression of miR-9-5p, miR-31-5p, miR-129-2-3p and miR-214-3p in PCFCL as compared to PCLBCL-LT. MicroRNAs previously described to be higher expressed in ABC-type as compared to GCB-type nodal DLBCL were not differentially expressed between PCFCL and PCLBCL-LT. In conclusion, PCFCL and PCLBCL-LT differ in their microRNA profiles. In contrast to their gene expression profile, they only show slight resemblance with the microRNA profiles found in GCB- and ABC-type nodal DLBCL.

Abstract 4498: Prognostic significance of nuclear factor kappa B expression in human papillomavirus-associated oropharyngeal cancer

Abstract Nuclear factor kappa B (NFκB) has been implicated in the development of head and neck squamous cell carcinoma (HNSCC) from premalignant changes to tumor progression and recurrence. However, the prognostic significance of cytoplasmic or nuclear localization of NFκB expression in human papillomavirus (HPV)-associated HNSCC has not been investigated largely due to limited patient samples. We hypothesized that cytoplasmic and/or nuclear expression of NFκB may be a poor prognostic factor for tumor recurrence and overall survival (OS) regardless of p16 status. We conducted a retrospective study with archival formalin-fixed paraffin-embedded tumor biopsies collected from 188 patients with oropharyngeal SCC between October 1994 and February 2008 from Emory University Department of Pathology. Median follow-up was 21.7 and 58.0 months for p16- and p16+ patients, correspondingly. Tumors were evaluated for p16 expression (a surrogate marker for HPV) by immunohistochemistry (IHC). Cytoplasmic expression of NFκBp65 by IHC or total expression combining cytoplasmic and nuclear localization were significantly greater in p16- patients (N=58) than p16+ patients (N=130), p&amp;lt;0.05. High cytoplasmic expression of NFκBp65 was significantly associated with poor OS (p=0.0276) and the total expression was associated with recurrence (p=0.0081) by log rank test, whereas nuclear NFκBp65 displayed weak association. Positive nuclear expression significantly correlated with smoking history (p=0.025). In conclusion, p16+ and low total NFκBp65 expression is associated with the best while p16- and high total NFκBp65 is associated with the worst OS in this patient cohort (Table I). (The study was supported by Head and Neck SPORE Grant P50CA128613). Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 4498. doi:1538-7445.AM2012-4498

Abstract 4905: Quantitative high resolution detection of O6-MGMT methylation in routinely processed human glioblastoma specimens

Abstract Hypermethylation of the DNA repair gene 06-Methylguanosinemethyltransferase (06-MGMT) is an important prognostic and predictive marker for the management of glioblastoma patients. However, assessment of O6-MGMT gene methylation is not yet standardized and a comprehensive comparison of different methods described in the literature is missing. Therefore, the methylation status of the MGMT gene was analyzed in a large series of glioblastoma specimens, WHO grade IV (n = 92) using nested conventional methylation-specific PCR (MSP), real-time PCR using either SybrGreen or a sequence specific probe for detection and pyrosequencing. All specimens were routinely processed formalin-fixed paraffin-embedded tumor biopsies. For a subset of samples (n = 30) three different protocols for the treatment of genomic DNA with bisulfite were compared. As a control, 20 brain biopsies from 10 patients without glioblastoma were analyzed. MGMT mRNA level was measured using quantitative real-time RT-PCR in n = 39 samples and the mRNA expression levels were correlated to the methylation status of these samples. The three different bisulfite protocols did not show statistically significant differences concerning quality and yield of recovered DNA. A region in the MGMT promoter in which many published primers bind, displayed constitutively elevated methylation levels also in healthy brain tissue samples, thereby leading to false-positivity under many circumstances. Nested MSP and pyrosequencing correlate fairly well in scoring samples as “methylated” versus “unmethylated”, but a considerable number of samples scored “methylated” by nested MSP (n = 7) displayed only background methylation when analyzed by pyrosequencing. Only methylation levels measured by pyrosequencing correlated with mRNA expression levels (r2 = 0.32, p = 0.0002). Also, threshold definition was much more straightforward and easier to standardize using pyrosequencing. Real-time PCR based methods did not correlate very well with nested PCR or pyrosequencing. In conclusion, pyrosequencing is the most reliable methodology for methylation analysis of O6-MGMT gene methylation analysis in routinely processed glioblastoma specimens. Note: This abstract was not presented at the AACR 101st Annual Meeting 2010 because the presenter was unable to attend. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 101st Annual Meeting of the American Association for Cancer Research; 2010 Apr 17-21; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2010;70(8 Suppl):Abstract nr 4905.

ERCC1 mRNA Expression Is Not Associated with Response and Survival after Platinum-Based Chemotherapy Regimens in Advanced Non-Small Cell Lung Cancer

Platinum-based therapy is pivotal to the treatment of advanced non-small cell lung Cancer (NSCLC). Excision repair cross-complementation group 1 (ERCC1) is a key component of the platinum-DNA repair machinery responsible for nucleotide excision repair. We sought to determine the influence of ERCC1 mRNA expression in advanced NSCLC on chemotherapy response, toxicity, and survival after platinum-based chemotherapy. Patients randomized to a phase III trial of platinum-based chemotherapy were eligible for inclusion. Formalin-fixed paraffin-embedded tumor biopsies were retrieved for mRNA extraction and purification before quantitative real-time polymerase chain reaction analysis using Taqman technology. Expression data were correlated with treatment response, toxicity, and overall survival. Sixty-six patients were enrolled. No statistically significant relationship existed between ERCC1 mRNA expression and response to chemotherapy (p = 0.794) or hematological toxicity. No statistically significant difference in median survival was demonstrated according to ERCC1 expression (high expression, 415 days, 95% confidence interval [95%CI]: 197-633 days; low expression, 327 days [95%CI: 211-433 days]; p = 0.801). High ERCC1 mRNA expression was associated with a hazard ratio for death of 0.96 (95% CI 0.919-1.004; p = 0.08). In contrast to recent publications, ERCC1 mRNA expression in our study did not favor a prognostically better outcome after platinum-based chemotherapy in advanced NSCLC. We explore potential reasons for this, including the need for cautious interpretation of mRNA expression data from archival materials and highlight the need for additional translational research linking gene expression with a promising ERCC1 polymorphism.