Low Cellularity Tumor Research Articles

Abstract 5345: Complementary NGS approaches on digitally sorted pure tumor cells reveal hidden molecular characteristics in low tumor content FFPE specimens

Abstract Introduction: Precise characterization of tumor cell populations is an essential requirement for guiding the cancer care, allowing patients to receive personalized therapies. Poor biopsies with low-tumor content represent a significant barrier for sample enrollment in clinical trials. Here we describe a multi-level approach for precisely characterizing the genetic mutation landscape of pure tumor cell populations sorted by the DEPArray™ technology from low-cellularity FFPE samples. Methods: 50μm sections of FFPE from breast infiltrating ductal adenocarcinoma with 10% tumor content were processed by DEPArray sorting protocol. Illumina-compatible libraries were prepared from sorted stromal (n=497) and tumor (n=419) cell pools, and from the unsorted sample. An aliquot of these libraries was processed using SeqCap EZ MedExome enrichment kit (Roche) for whole-exome sequencing (WES) with a HiSeq 2500, reaching a mean coverage of 27x for tumor and 25x for stromal libraries. A second aliquot was used for low-pass (≈1M fragments per sample) whole-genome sequencing (WGS) on a MiSeq to analyze copy-number alterations (CNA). Other cell lysates from stromal (n=104, n=112) and pure tumor (n=75) populations were used in DEPArray OncoSeek amplicon-based assay for focused analysis of clinically relevant somatic variants and copy-number alterations. Results: In sorted pure tumor populations, WES analysis of B-allele frequencies of germline heterozygous SNPs clearly outlined an aberrant profile, precisely revealing several Loss-of-Heterozygosity (LOH) and copy-altered regions. Conversely, unsorted gDNA showed a flat profile non-distinguishable from sorted stromal cells, as expected for a low-cellularity tumor sample. Similar results were obtained by low-pass WGS, where the huge number of copy number aberrations (≈1.2 Gbp) in tumor is contrasted by lack of gains and losses in stromal cells and unsorted gDNA. Noteworthy, WES, low-pass and targeted sequencing by OncoSeek panel highlighted a focal amplification of ERBB2 gene (>13 copies), which was just barely detectable as a 1-copy gain in bulk gDNA. Genetic analyses showed a high concordance between WES and targeted panel data, with two non-synonymous homozygous somatic mutations found in TP53 (p.L111R) and ERBB2 (p.D769Y). In the unsorted sample, the TP53 mutation was missed because allelic frequency was below the limit of detection due to normal-cell dilution, while the ERBB2 mutation was still detectable only because of the high-level amplification. Conclusions: DEPArray sorting technology is an indispensable tool for accurately investigating cancer genomes, enabling multi-level applications for obtaining a fine-grain characterization of copy-numbers, LOH, and tumor-specific variants, independent of original tumor content. Citation Format: Claudio Forcato, Alberto Ferrarini, Genny Buson, Cassie Schumacher, Chiara Bolognesi, Paola Tononi, Valentina del Monaco, Chiara Mangano, Francesca Fontana, Gianni Medoro, Timothy Harkins, Nicolò Manaresi. Complementary NGS approaches on digitally sorted pure tumor cells reveal hidden molecular characteristics in low tumor content FFPE specimens [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 5345. doi:10.1158/1538-7445.AM2017-5345

Abstract P2-05-07: Comparison of the Xpert breast cancer stratifier mRNA assay with central ER, PR, HER2, and Ki67 immunohistochemistry (IHC) for rapid biomarker analysis in developing countries

Abstract Breast cancer care in the developing world is limited by access to quality ER and HER2 IHC diagnostic assays needed to justify hormone and HER2 therapeutics. Shipping pathology specimens to a central testing site often out of country delays therapy and is costly. The Xpert Breast Cancer Stratifier assay makes quantitative measurements of ESR1, PGR, ERBB2, and MKi67 mRNAs from FFPE specimens in <2 hours on an easy-to-use automated diagnostic platform, the GeneXpert (GX). 10,000 GX machines are currently in use in 182 countries offering the possibility of a point-of-care solution. We compared concordance in tumor samples between IHC and mRNA intending to challenge the limits of the GX mRNA assay. 83 breast tumor samples were chosen including those with low cellularity, small volume disease, unusual subtypes, ER- tumors with surrounding benign epithelium, and low level HER2+ tumors. mRNA, IHC and FISH assays were performed. Slides were tested following macrodissection of invasive carcinoma and as non-macrodissected whole sections. GX measurements for Ki67 were compared with mitotic rate as an alternative to Ki67 IHC. Overall percent agreement following macrodissection was 95% for ER, 89% for HER2, 76% for PR, and 80% for Ki67 (>20% positive cut), and using whole section, 99% for ER, 80% for PR, 92% for HER2, and 73% for Ki67. Concordance was 92% for both macrodissection and whole section using mitotic rate to assess proliferation. Ignoring HER2 2+ calls which represented low level amplified tumors by FISH, the concordance rates were 95% for macrodissection and 99% for whole section. Discordance when testing long-term stored 4μm sections was resolved in a number of cases by using a fresh cut from the FFPE block. Half the ER discrepancies were in very small volume tumors ≤25mm2 and 75% were classified as ER-ve by IHC, and positive by Stratifier. 80% of ER IHC- cases were appropriately identified as ER- by the Stratifier in the presence of benign breast epithelium. HER2+ DCIS adjacent to HER2- invasive tumor resulted in a discrepant HER2 mRNA result even with macrodissection. No ER or HER2 discrepancies occurred in low cellularity tumors (≤30% cellularity) nor in lobular and mucinous subtypes. In a study intended to challenge an mRNA breast biomarker assay, concordance between mRNA results and IHC was high for ER and HER2, the two most important prognostic markers needed for therapeutic decision making. Use of whole sections rather than tumor macrodissection did not decrease concordance. Discrepant ER cases were more prevalent when analyzing low volumes of tumor and in this setting were seen in ER IHC- tumors surrounded by ER+ normal epithelium, or with weak IHC expression, highlighting predictable limitations of the assay. Concordance was better between Ki67 mRNA and mitotic rate than with IHC. Re-test data suggested that a fresh cut of the FFPE block yields the best results by GX, perhaps due to mRNA degradation in stored 4μm sections. The Xpert Breast Cancer Stratifier may provide a rapid, cost-effective solution to the problem of obtaining accurate diagnostic results at the point-of-care in low resource settings, and deserves further evaluation in developing countries. Citation Format: Brock JE, Milner DA, Ho K, Natalie W, Victor C, Annaliza R, Teresa B, Kathryn G-F, Edwin LW, Jodi W, Wendy W, Michael B. Comparison of the Xpert breast cancer stratifier mRNA assay with central ER, PR, HER2, and Ki67 immunohistochemistry (IHC) for rapid biomarker analysis in developing countries [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P2-05-07.

SCNAphase: using haplotype resolved read depth to genotype somatic copy number alterations from low cellularity aneuploid tumors.

Accurate identification of copy number alterations is an essential step in understanding the events driving tumor progression. While a variety of algorithms have been developed to use high-throughput sequencing data to profile copy number changes, no tool is able to reliably characterize ploidy and genotype absolute copy number from tumor samples that contain less than 40% tumor cells. To increase our power to resolve the copy number profile from low-cellularity tumor samples, we developed a novel approach that pre-phases heterozygote germline single nucleotide polymorphisms (SNPs) in order to replace the commonly used ‘B-allele frequency’ with a more powerful ‘parental-haplotype frequency’. We apply our tool—sCNAphase—to characterize the copy number and loss-of-heterozygosity profiles of four publicly available breast cancer cell-lines. Comparisons to previous spectral karyotyping and microarray studies revealed that sCNAphase reliably identified overall ploidy as well as the individual copy number mutations from each cell-line. Analysis of artificial cell-line mixtures demonstrated the capacity of this method to determine the level of tumor cellularity, consistently identify sCNAs and characterize ploidy in samples with as little as 10% tumor cells. This novel methodology has the potential to bring sCNA profiling to low-cellularity tumors, a form of cancer unable to be accurately studied by current methods.

Toward the molecular dissection of peritoneal pseudomyxoma

Outcome of pseudomyxoma peritonei (PMP) after cytoreductive surgery (CRS) and hypertermic intraperitoneal chemotherapy (HIPEC) is heterogeneous even after adjusting for clinico-pathological prognostic variables. The identification of additional prognostic or even predictive biomarkers is an unmet clinical need. Forty patients with mucinous appendiceal tumors and PMP were clinically eligible and had evaluable tumor samples obtained after CRS and HIPEC. We carried out next-generations sequencing (NGS) of 50 gene's hotspot regions contained in the Hotspot Cancer Panel v2 using the Ion Torrent Personal Genome Machine platform (Life Technologies). KRAS and GNAS mutations were found in 72% and 52%, and their allelic frequency was below 10% in 55% and 43% of samples, respectively. KRAS and GNAS mutations were associated with worse progression-free survival (PFS) at univariate analysis (P = 0.006 and 0.011, respectively). At multivariate analysis, only KRAS mutations were independently associated with PFS (P = 0.012); GNAS mutations were not-being significantly associated with other poor prognostic features such as incomplete cytoreduction or KRAS mutations. Validation of results was carried out in an independent bi-institutional cohort of 25 patients and the prognostic effect of KRAS mutations was again confirmed in the multivariate model (P = 0.029). NGS approach allowed the discovery of other potentially druggable mutations such as those in PI3K, AKT, LKB1, FGFR3 and PDGFRA. Given the homogeneity of this series and the sensitivity of NGS in this low-cellularity tumor, we demonstrated for the first time a poor prognostic role of KRAS mutations.

Abstract 3617: Digital sorting rescues low-cellularity tumor FFPE samples for genome-wide copy-number profiling

Abstract Introduction: The analysis of formalin-fixed paraffin embedded (FFPE) cancer specimens is particularly challenging due to minute amount of tissue, low-tumor cellularity and heterogeneity, associated with low quality DNA. This results in an imprecise characterization of somatic variants and copy-number alterations (CNA). Here we show how digital sorting combined with low-pass whole genome sequencing (WGS) can resolve genome-wide copy-number profiling even for the more challenging FFPE samples. Methods: An archival FFPE sample from one pancreatic cancer patient, rejected by the International Cancer Genome Consortium (ICGC) criteria due to low tumor content (<25%), was dissociated into a single cell suspension. Using the DEPArray™ digital sorter, tumor and normal stromal cell subpopulations (range = 70-77) were recovered using Keratin/Vimentin immunofluorescence and DNA ploidy, from which 0.4-0.5ng DNA was obtained. After lysis, Illumina®-compatible libraries were prepared from one pool of pseudo-diploid sorted tumor cells, one pool of stromal cells and one unsorted sample (1500 cells, i.e. about 10ng gDNA) using the Accel-NGS® 2S DNA Library Kit from Swift Biosciences. Libraries were used for low-pass whole genome sequencing on Illumina® MiSeq. Paired-end (2×100) reads were aligned to hg19 human reference using BWA software and copy-number profiles were generated with Control-FREEC. Results: A total of 11M paired-end reads and a mean coverage of 0.12x were obtained from low-pass WGS, enabling the detection of CNAs at good resolution. The resulting copy-number profiles clearly show the difference between the stromal and tumor populations, with the first characterized by a flat profile (0 gains, 0 losses) and the second presenting several somatic copy-number alterations (6 gains, 22 losses). As expected due to dilution by normal cells, CNA profiles of unsorted samples missed most of the gains and losses (only 5/28 = 18% of aberrant genomic regions were detected). At a single base level, the difference between unsorted and sorted samples was more significant as the unsorted fraction missed >99% of CNA bases found in the sorted tumor population. Conclusions: DEPArray™ sorting combined with Accel-NGS® 2S kits and Illumina® low-pass whole genome sequencing enables high quality genome-wide profiling of pure tumor and stromal populations. The capacity of this method to deal with low DNA input (0.5ng) from archival FFPE samples characterized by low cancer cell content (<25%) makes it a valuable and easily accessible tool for studying tumor CNA profiles. Citation Format: Claudio Forcato, Julie Lalibertè, Chiara Bolognesi, Ivana Cataldo, Genny Buson, Chiara Mangano, Cinzia Cantù, Francesca Fontana, Paola Tononi, Gianni Medoro, Tim Harkins, Rita T. Lawlor, Aldo Scarpa, Nicolò Manaresi. Digital sorting rescues low-cellularity tumor FFPE samples for genome-wide copy-number profiling. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 3617.

Abstract 619: Identification of clinically actionable genomic alterations in the tumor and circulation of pancreatic cancer patients

Abstract Pancreatic adenocarcinoma has the worst overall mortality of any solid tumor, with only 6% of patients surviving after 5 years. To evaluate the clinical implications of genomic alterations in this low cellularity tumor type, we deeply sequenced the genomes of 134 enriched pancreatic adenocarcinomas from patients who underwent potentially curative resections. Given the low neoplastic cellularity of pancreatic cancers, we enriched for neoplastic cells either by macrodissection of primary tumors or by flow-sorting of tumor nuclei, and performed deep sequencing (high coverage) of these enriched samples using next-generation sequencing approaches. We obtained a total of >1Tb of sequence data, resulting in an average coverage within the target regions of >200-fold for each tumor analyzed by whole-exome sequencing and >750-fold for each tumor analyzed by targeted cancer gene sequencing. These approaches allowed us to identify sequence changes, including single base and small insertion or deletion mutations, as well as copy number alterations in >20,000 genes in the whole-exome analyses and in 116 specific genes in the targeted analyses. These analyses revealed that somatic mutation of chromatin remodeling genes were associated with improved progression-free and overall survival. Alterations in genes with potential clinical utility were observed in a majority of cases and included alterations of AKT1, AKT2, BRCA2, ERBB2, KIT, and PIK3CA. Non-invasive liquid biopsy analyses were performed before and after surgery to evaluate the presence of circulating tumor DNA in the plasma of 83 patients. Through these approaches, we were able to diagnose early stage pancreatic tumors in the majority of patients and to detect the presence of circulating tumor DNA prior to clinical relapse. These observations provide genetic markers of clinical outcome in pancreatic cancer and suggest new avenues for personalized therapy. Citation Format: Mark Sausen, Jillian Phallen, Vilmos Adleff, Siân Jones, Rebecca J. Leary, Karli Lytle, Sonya Parpart-Li, Derek Murphy, Michael T. Barrett, David C. Linehan, Anirban Maitra, Ralph Hruban, Daniel D. Von Hoff, Julia S. Johansen, Luis A. Diaz, Jeffrey A. Drebin, Victor E. Velculescu. Identification of clinically actionable genomic alterations in the tumor and circulation of pancreatic cancer patients. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 619. doi:10.1158/1538-7445.AM2015-619

Rapid Intraoperative Molecular Characterization of Glioma

Conclusive intraoperative pathologic confirmation of diffuse infiltrative glioma guides the decision to pursue definitive neurosurgical resection. Establishing the intraoperative diagnosis by histologic analysis can be difficult in low-cellularity infiltrative gliomas. Therefore, we developed a rapid and sensitive genotyping assay to detect somatic single-nucleotide variants in the telomerase reverse transcriptase (TERT) promoter and isocitrate dehydrogenase 1 (IDH1). This assay was applied to tissue samples from 190 patients with diffuse gliomas, including archived fixed and frozen specimens and tissue obtained intraoperatively. Results demonstrated 96% sensitivity (95% CI, 90%-99%) and 100% specificity (95% CI, 95%-100%) for World Health Organization grades II and III gliomas. In a series of live cases, glioma-defining mutations could be identified within 60 minutes, which could facilitate the diagnosis in an intraoperative timeframe. The genotyping method described herein can establish the diagnosis of low-cellularity tumors like glioma and could be adapted to the point-of-care diagnosis of other lesions that are similarly defined by highly recurrent somatic mutations.

Abstract A82: Virtual microdissection reveals tumor specific heterogeneity in pancreatic cancer

Abstract Using matrix factorization, we have removed confounding normal tissue gene expression from profiles of primary and metastatic tumors, facilitating the study of underlying tumor biology without the need for laser capture microdissection. Understanding molecular mechanisms of metastasis in pancreatic cancer has the potential to yield rationally designed therapies. Previous work has used gene expression arrays and next generation sequencing to look for potential targets, focusing on identifying differences between metastatic lesions and primary tumors. However, such analyses are hampered by the low cellularity of malignant epithelium in patient samples. Since profiles of primary and metastatic tumors often contain contributions from adjacent normal tissues from the tumor site, comparison of these samples are confounded by the differences between normal pancreas and normal distant organs. We seek to avoid this pitfall by leveraging the blind source separation technique, nonnegative matrix factorization (NMF)[1], to the analysis of gene expression data. Microarray data were obtained from matched primary (n=8) and metastatic tumors (n=31) as well as adjacent normal tissues (n=85) for both local and distant sites of patients who died of metastatic pancreatic cancer[2]. Pathological analysis showed that tumor cellularity varied with a mean of 48% and a standard deviation of 31% across all sites. We applied NMF to our cohort, thus computationally microdissecting tumors into the source tissues composing our samples. 200 iterations of 5-fold resampling were performed to achieve stable NMF results. Genes with expression ranked in the top 50 of any factor together were recorded in a consensus matrix. This consensus matrix was then used as the basis of a hierarchical clustering as to yield k gene clusters. These k gene-clusters were used to seed a supervised NMF using the MATLAB NMF multiplicative update solver until completion. Using an unbiased approach, we identified distinct molecular signatures associated with adjacent normal and tumor tissue in primary and metastatic pancreatic cancer samples. Specifically, we identified confounding liver, lung, lymph, and muscle tissue gene expression from our metastatic samples as well as both endocrine and exocrine pancreas gene expression from our primary tumors. Furthermore, we estimated the relative composition of each of our samples in terms of a weighted sum of these tissue-specific signatures. Both primary tumors and metastases appeared to be admixtures of pancreatic tumor epithelial expression and adjacent normal tissue. For example, in liver tissues histological assessment of tumor content correlated well with NMF predictions of tumor content, (n=39, R2=0.72, p<0.01). Prior to NMF, unsupervised clustering of all data caused samples to cluster by site of harvest. After digital subtraction of confounding tissue factors identified by NMF, samples grouped by patient of origin, suggesting intrinsic similarity among tumor sites within a patient. We have demonstrated a method to virtually microdissect tissues, thereby identifying tumor-specific gene expression data in pancreatic cancer. By applying a fresh computational approach to a large cohort of data, we can generate new insight into the complex nature of low-cellularity tumors such as pancreatic cancer, and facilitate the study of inter- and intra-patient heterogeneity. This approach may be further leveraged to study the role of tumor and stroma signatures in pancreatic cancer. 1. Brunet, J.P., et al., Metagenes and molecular pattern discovery using matrix factorization. Proceedings of the National Academy of Sciences, 2004. 101(12): p. 4164-4169. 2. Stratford, J.K., et al., A six-gene signature predicts survival of patients with localized pancreatic ductal adenocarcinoma. PLoS medicine, 2010. 7(7): p. e1000307. Citation Format: Richard A. Moffitt, Keith A. Volmar, Judy M. Anderson, Michael A. Hollingsworth, Jen Jen Yeh. Virtual microdissection reveals tumor specific heterogeneity in pancreatic cancer. [abstract]. In: Proceedings of the AACR Special Conference on Pancreatic Cancer: Innovations in Research and Treatment; May 18-21, 2014; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2015;75(13 Suppl):Abstract nr A82.

Somatic point mutation calling in low cellularity tumors.

Somatic mutation calling from next-generation sequencing data remains a challenge due to the difficulties of distinguishing true somatic events from artifacts arising from PCR, sequencing errors or mis-mapping. Tumor cellularity or purity, sub-clonality and copy number changes also confound the identification of true somatic events against a background of germline variants. We have developed a heuristic strategy and software (http://www.qcmg.org/bioinformatics/qsnp/) for somatic mutation calling in samples with low tumor content and we show the superior sensitivity and precision of our approach using a previously sequenced cell line, a series of tumor/normal admixtures, and 3,253 putative somatic SNVs verified on an orthogonal platform.

Clinical and Histopathologic Features of Chordomas in Children and Young Adults

Chordoma is a tumor of the axial skeleton that is distinctly uncommon in children and adolescents. Previous observations have suggested that chordomas in pediatric patients differ from adult chordomas in presentation, morphology, and behavior. This study examines the clinical and histologic features of chordomas in young patients (< or =25 years old). All cases from the Mayo Clinic files were graphed according to age and a bimodal distribution was observed. The 35 cases representing the youngest population were selected for review. Histopathology ranged from low cellularity tumors with lobulated architecture and abundant myxoid matrix (conventional chordoma), to those with varying amounts of chondroid matrix (chondroid chordoma), to more cellular tumors (atypical chordoma), and finally to neoplasms with high-grade spindle-cell differentiation (dedifferentiated chordoma). Over an average follow-up period of 129 months (range 1 to 501 months), there were 13 deaths (37%) and 3 patients with metastasis. This survival rate was slightly better than the reported mortality rate for adults with chordoma, but there was a subset of young patients (those with atypical chordoma) that had a significantly worse survival rate, suggesting that histologic subtyping may be predictive of prognosis.

A CASE OF STROMAL TUMOR OF THE DUODENUM

A 50-year-old woman was seen at the hospital because of hematemesis and tarry stool. Emergency endscopic examination revealed a submucosal tumor with central easy bleeding ulcer at the anal side of the Papilla Vateri of the duodenum. Preoperative examination showed that the tumor which was 5cm in diameter, solid, well-delineated, and hypervascular, developed extramurally from the second portion of the duodenum to the head of pancreas. It was diagnosed as leiomyoma or leiomyosarcoma and a pancreatoduodenectomy was performed. Histopathologically, this tumor was shown to be composed of proliferated spindle cells which arranged in fascicles and were complicated. Amyloid protein deposition was recognized in this tumor. Immunohistochemical study showed that the tumor was negative for α-smooth muscle actin, desmin, S-100 protein and NSE, but was positive for vimentin only. Therefore, it was diagnosed as duodenal stromal tumor whose histogenesis was neither leiomyogenic nor neurogenic. Although mitosis was not seen so frequently in this low cellularity tumor, it was supposed to be possibly malignant because the diameter of this tumor was 5.5cm. We would have to follow up the patient carefully.