High Tumor Content Research Articles

Monitoring hepatocellular carcinoma using tumor content in circulating cell-free DNA.

To evaluate the utility of tumor content in circulating cell-free DNA (ccfDNA) for monitoring hepatocellular carcinoma (HCC) throughout its natural history. We included 67 hepatitis B virus (HBV)-related HCC patients, of whom 17 had paired pre- and post-treatment samples, and 90 controls. Additionally, in a prospective cohort with HBV surface antigen-positive participants recruited in 2012 and followed up biannually with blood sample collections until 2019, we included 270 repeated samples before diagnosis from 63 participants who later developed HCC (pre-HCC samples). Shallow whole-genome sequencing and the ichorCNA method were used to analyze genome-wide copy number and tumor content in ccfDNA. High tumor content was associated with advanced tumor stage (P < 0.001) and a poor survival after HCC diagnosis (HR=12.35; 95% confidence interval [CI]=1.413-107.9; P = 0.023). Tumor content turned negative after surgery (P = 0.027), while remained positive after transarterial chemoembolization treatment (P = 0.578). In non-HCC samples, the mean tumor content (±SD) was 0.011 (±0.007) and had a specificity of 97.8% (95%CI=92.2%-99.7%). In pre-HCC samples, tumor content increased from 0.014 in 4 years before diagnosis to 0.026 in 1 year before diagnosis. The sensitivity of tumor content in detecting HCC increased from 22.7% (95%CI=11.5%-37.8%) within one year before diagnosis to 30.4% (95%CI=13.2%-52.9%) at BCLC stage 0/A, 81.8% (95%CI=59.7%-94.8%) at stage B, and 95.5% (95%CI=77.2%-99.9%) at stage C. The tumor content in ccfDNA is correlated with tumor burden and may help in monitoring HCC one year earlier than clinical diagnosis and in predicting patient prognosis.

Abstract B102: Mutational and copy number-based ctDNA profiles mirror high-grade serous cancer tumors and enable detection of genetic changes appearing at recurrence

Abstract Circulating tumor DNA (ctDNA) is the cancer-derived fraction of extracellular DNA fragments present in the bloodstream as cell-free DNA (cfDNA). Consequently, ctDNA harbors the distinct genomic information specific to the cancer from which it derives. Therefore, ctDNA from liquid biopsies provides a non-invasive, universally available alternative to traditional tumor biopsies. ctDNA amount may provide insight into tumor burden, as its concentration is typically highest during diagnosis and relapse. The non-invasive nature of this method facilitates frequent monitoring, enabling longitudinal tracking of the disease's trajectory and response to treatments. By focusing on the key genomic changes, either short mutations or copy number aberrations (CNA), liquid biopsy allows the investigation of cancer progression and evolution over time, providing invaluable insight into the behavior and adaptability of the cancer. We collected 152 longitudinal plasma samples and 92 fresh tissue samples from 29 patients enrolled in the DECIDER cohort (NCT04846933), after a diagnosis of high-grade serous carcinoma (HGSC), the most common and deadliest histological subtype of ovarian cancer. With ultra-deep targeted sequencing of a gene panel of more than 700 cancer-related genes we were able to comprehensively evaluate the cancer-related genomic changes. The concordance rate between mutations in plasma compared to tumor tissue was 83 % at diagnosis and 90 % at relapse, confirming the reliability of exploiting ctDNA in monitoring genomic alterations. Furthermore, by comparing plasma- and tissue-derived mutations, we demonstrated that comparable levels of ctDNA are released from adnexa, intra-abdominal lesions, and ascites, proposing that a liquid biopsy represents mutations from various locations. Analysis of CNA showed that our approach successfully detected concordant profiles in most plasma samples, even with tumor content as low as 3%. Additionally, highly amplified regions were identified in samples with approximately 1% tumor content. Longitudinal analysis of both mutations and CNA revealed genomic changes in plasma samples at relapse. Twenty-four new exonic mutations were identified in eight patients at relapse; two of these alterations were related to HRD status, supporting PARP-inhibitors as a valid treatment option. Furthermore, mutations in FGFR3, JAK2, FLT3 and CDKN2A could be investigated as target indications for extended use of currently available therapies. Eleven patients had high tumor content in their plasma samples at the time of relapse. We detected altered CNA profiles in seven of these eleven patients, so that five patients harbored only a few novel focal amplifications or losses, while two patients showed significant changes in their copy-number profile during the therapy. In conclusion, our study identified recurrence-specific genomic aberrations, proposing that ctDNA more accurately represents the metastatic tumor and provides possibilities for personalized therapy options. Citation Format: Giovanni Marchi, Mai T.N. Nguyen, Anna Rajavuori, Taru Muranen, Kaisa Huhtinen, Sinikka Oksa, Sakari Hietanen, Johanna Hynninen, Jaana Oikkonen, Sampsa Hautaniemi. Mutational and copy number-based ctDNA profiles mirror high-grade serous cancer tumors and enable detection of genetic changes appearing at recurrence [abstract]. In: Proceedings of the AACR Special Conference on Ovarian Cancer; 2023 Oct 5-7; Boston, Massachusetts. Philadelphia (PA): AACR; Cancer Res 2024;84(5 Suppl_2):Abstract nr B102.

Predicting EGFR mutational status from pathology images using a real-world dataset

Treatment of non-small cell lung cancer is increasingly biomarker driven with multiple genomic alterations, including those in the epidermal growth factor receptor (EGFR) gene, that benefit from targeted therapies. We developed a set of algorithms to assess EGFR status and morphology using a real-world advanced lung adenocarcinoma cohort of 2099 patients with hematoxylin and eosin (H&E) images exhibiting high morphological diversity and low tumor content relative to public datasets. The best performing EGFR algorithm was attention-based and achieved an area under the curve (AUC) of 0.870, a negative predictive value (NPV) of 0.954 and a positive predictive value (PPV) of 0.410 in a validation cohort reflecting the 15% prevalence of EGFR mutations in lung adenocarcinoma. The attention model outperformed a heuristic-based model focused exclusively on tumor regions, and we show that although the attention model also extracts signal primarily from tumor morphology, it extracts additional signal from non-tumor tissue regions. Further analysis of high-attention regions by pathologists showed associations of predicted EGFR negativity with solid growth patterns and higher peritumoral immune presence. This algorithm highlights the potential of deep learning tools to provide instantaneous rule-out screening for biomarker alterations and may help prioritize the use of scarce tissue for biomarker testing.

MO1-6 Characteristics of the tissue obtained in biopsy affecting the success rate of NGS testing in advanced NSCLC patients

Next-generation sequencing (NGS), Oncomine TM DX Target Test (ODxTT) companion diagnostics (CDx) system, has been introduced to screen oncogenic driver mutations in advanced non-small cell lung cancer (NSCLC). It is well known that obtaining tissue with large size and high tumor content is essential for appropriate NGS testing. On the other hands, it remains unclear what kind of the characteristics of tissue obtained by biopsy is related to high success rate in advanced NSCLC patients.

Abstract 197: A systemic model derivation platform for generating 3D neuroendocrine tumor cell spheroids to accelerate cancer research

Abstract The Broad Cancer Cell Line Factory (CCLF) aims to increase the number and representation of in vitro/ex vivo cell models for common and rare cancer types. Neuroendocrine tumor (NET) cell model derivation is one of the CCLFs focus because it lacks well-characterized, publicly available models.Two major barriers existed in deriving NET cell models, including how to collect sufficient patient tumor tissue samples for the model derivation pilot and how to systematically iterate model derivation strategies since there is no prior knowledge for NET cell model generation success. Thus, we partnered with the MD Anderson Cancer Center, the Dana-Farber Cancer Institute, and the Rare Cancer Research Foundation to collect patient tissue samples. All patient’s NET tissues were sequenced with a targeted Pan-Cancer panel to ensure high tumor content. To reduce fibroblast outgrowth, we combined an empirical rich media matrix (HYBRID technology) with a 3D spheroid culture system to initiate one sample in 16-64 conditions. The growing cultures at passage 3-5 were genomically credentialed to ensure the driver events matched with the original patient tissue. So far, we have received more than 70 NET samples. While several derived models are still under culture, we successfully generated 5 genomically verified NET tumor models, including small intestinal, pancreas, and liver subtypes. To phenotypically characterize these NET models, neuroendocrine biomarkers such as chromogranin A, synaptophysin, SSTR2, and VMAT 1/2 were also evaluated using qRT-PCR and ELISA. We observed that these NET spheroid models display long doubling times (2-4 weeks) at later passages which limits their utility for large scale perturbation experiments and model sharing capability with the research community. While we are currently working on several strategies to improve the propagation ability in these models, 1 (out of 5) model has reached passage 15 with a 3 day doubling time. Genomic studies, such as RNAseq, will be performed to address the model transcriptome changes after overcoming growth plateaus. Here we showed that it is feasible to derive NET models from patient biospecimens using our HYBRID strategy. As we expand our NET cohort, we will further refine disease-specific model generation protocols for different NET types. Our goal is to share our model generation experience and make these tumor cell models publicly available to the research community in order to accelerate cancer research. Citation Format: Adel Attari, Madison Liistro, Barbara Van Hare, Jennifer Chan, Emma Coleman, Tim Heffernan, Bianca Amador, Matthew Meyerson, Jesse Boehm, William Sellers, Yuen-Yi (Moony) Tseng. A systemic model derivation platform for generating 3D neuroendocrine tumor cell spheroids to accelerate cancer research [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 197.

Rapid non-destructive volumetric tumor yield assessment in fresh lung core needle biopsies using polarization sensitive optical coherence tomography.

Adequate tumor yield in core-needle biopsy (CNB) specimens is essential in lung cancer for accurate histological diagnosis, molecular testing for therapeutic decision-making, and tumor biobanking for research. Insufficient tumor sampling in CNB is common, primarily due to inadvertent sampling of tumor-associated fibrosis or atelectatic lung, leading to repeat procedures and delayed diagnosis. Currently, there is no method for rapid, non-destructive intraprocedural assessment of CNBs. Polarization-sensitive optical coherence tomography (PS-OCT) is a high-resolution, volumetric imaging technique that has the potential to meet this clinical need. PS-OCT detects endogenous tissue properties, including birefringence from collagen, and degree of polarization uniformity (DOPU) indicative of tissue depolarization. Here, PS-OCT birefringence and DOPU measurements were used to quantify the amount of tumor, fibrosis, and normal lung parenchyma in 42 fresh, intact lung CNB specimens. PS-OCT results were compared to and validated against matched histology in a blinded assessment. Linear regression analysis showed strong correlations between PS-OCT and matched histology for quantification of tumors, fibrosis, and normal lung parenchyma in CNBs. PS-OCT distinguished CNBs with low tumor content from those with higher tumor content with high sensitivity and specificity. This study demonstrates the potential of PS-OCT as a method for rapid, non-destructive, label-free intra-procedural tumor yield assessment.

Abstract 2731: Cell-type deconvolution of African American breast tumors reveals spatial heterogeneity of the immune microenvironment

Abstract Background: As immunotherapy emerges as a possible treatment for triple negative breast cancer (TNBC), it is crucial to understand heterogeneity and the complexities of the tumor microenvironment. GeoMx, a new platform developed by NanoString, allows for gene expression analysis within localized segments of the tumor, and is therefore ideal for determining the spatial distribution of immune cell-types. To test the feasibility of applying spatial transcriptomic to breast tumors from diverse populations, we selected ten paraffin embedded blocks from a cohort of self-reported African American patients with TNBC in the Chicago Multi-Ethnic Breast Cancer Study (ChiMEC). Methods: Tissue sections were immunostained with four antibodies (PanCK, CD3E, CD20 and DAPI) and regions of interest (ROIs) were manually selected based on specific morphology, resulting in 10 to 39 ROIs per slide (with a total of 234 ROIs across the ten slides). Using the GeoMx Cancer Transcriptome Atlas assay, expression levels of 1825 genes were measured within each ROI. For this cell-type deconvolution analysis, we used the Spatial Deconvolution algorithm, developed specifically for GeoMx data. We first normalized read counts across all ROIs of a slide and removed the background expression levels. In order to account for cancer cell types in the deconvolution process, we selected ROIs with high tumor content to predict the gene expression profiles of these tumor cells. We then performed deconvolution using both cancer and immune expression profiles, and clustered ROIs into immune signature groups. Results: As expected, we saw a higher abundance of immune cells in the stroma compared with the tumor, which resulted in lower immune cell-type diversity within the tumor segments. Tumor segments were CD8+ T-cell and neutrophil enriched: 20.3% of immune cells in tumor segments were neutrophils, and 23.7% were CD8+ T-cells, whereas in the stroma, only 3.6% and 13.8% were neutrophils and CD8+ T-cells respectively. The proportion of naïve versus memory cytotoxic T-cells differed between the tumor and the stroma, with 84.1% of CD8+ T-cells in the stroma identified as memory versus only 23.9% for tumor segments. We were also able to compare two slides from the same patient and found similar immune patterns across the two sections of the tumor. Clustering of immune cell-type distribution revealed three major immune signatures within both sections of this tumor: macrophage-rich, B-cell rich, and CD8+ T-cell rich ROIs. Conclusion: Spatial transcriptomics allows for granular analysis of the tumor microenvironment. This level of detail is crucial to understanding the cancer-immune relationship, as our analysis revealed high heterogeneity of the immune landscape within tumors. Future work will compare the tumor microenvironment within and across racial groups. Citation Format: Jean-Baptiste Reynier, Jovian Yu, Anna Biernacka, Galina Khramtsova, Toshio Yoshimatsu, Qun Niu, Anna Woodard, Yonglan Zheng, Jeffrey Mueller, Mengjie Chen, Olufunmilayo I. Olopade. Cell-type deconvolution of African American breast tumors reveals spatial heterogeneity of the immune microenvironment [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 2731.

Innovative Tumor Tissue Dissection Tool for Molecular Oncology Diagnostics

Formalin-fixed, paraffin-embedded (FFPE) tissue is the most commonly used material for tumor molecular profiling, therapy selection, and prognostication. Tumor tissue enrichment by tissue dissection is highly recommended to generate quality data reproducibly for use in downstream assays, such as real-time PCR and next-generation sequencing. The aim of this study was to evaluate the performance of the automated tissue dissection tool AVENIO Millisect System compared with a manual dissection method using 18 FFPE tissue specimens. The study assessed performance of these two methods with paraffinized and deparaffinized sections at 5- and 10-μm thickness as well as at low (5% to 10%) and high (>50%) tumor content. In addition, compatibility with various nucleic acid and protein extraction methods was assessed. Overall, dissection by Millisect resulted in statistically significantly higher yields of nucleic acids and protein compared with manual dissection (P=0.00524). In downstream analysis on a statistically nonpowered sample set, EGFR mutation testing by PCR led to highly concordant results, and next-generation sequencing testing yielded significantly higher allelic frequencies when tissue was dissected by Millisect compared with manual scraping, demonstrating noninferiority of the automated method. In summary, the AVENIO Millisect System may replace manual labor and support automation of FFPE tumor tissue workflows in clinical molecular laboratories with high testing volumes with adequate validation.

Abstract 4990: Flow cytometric characterization of the tumor-infiltrating immune compartment: Significant differences in the tumor immune repertoire versus peripheral blood

Abstract Characterization of the tumor microenvironment (TME) is expected to be critical in informing prognosis and guiding response prediction1,2. Recent studies have classified the TME based on tumor infiltrating leukocytes (TILs), programmed death-ligand 1 (PDL1) expression, and the nature of inflammatory responses1,3. Tumor studies involving transcript profiling and IHC have begun to inform the diversity and complexity of the tumor-infiltrated immune compartment (TIIC) by defining aspects of cellular composition, spatial distribution, and function. The use of flow cytometry allows the elucidation of an array of immune cell subtypes across entire samples. This study characterizes immune cell population subsets and tumor cells in four cancer types (breast, head and neck, colorectal, and gall bladder) from patients using multi-color flow cytometry. A good correlation was observed (r= 0.89, p=0.003) between tumor content, as evaluated by a pathologist (Hematoxylin and Eosin staining), and tumor content (EpCAM/panCK[7/8]+ cells) as estimated by flow cytometry (for all cancers excepting head and neck). Using pathology scoring we divided the samples into high-content (&amp;gt;10% tumor) and low-content (&amp;lt;10% tumor) samples. Samples with high tumor content were further examined by comparing the immune population in the tissue with their matched peripheral blood mononuclear cells (PBMC). We observed significantly higher proportions of T-cells, T-regulatory cells (Tregs) and exhausted T-cells (PD-1+ T-cells) in TIIC in comparison to its matched PBMC, while a lower proportion of B-cells and NK cells was noted. Also, samples with high tumor content had a significantly increased proportion of CD8 T-cells, Tregs and NK cells and a lower proportion of effector CD4 (CD4+Foxp3-) T-cells in comparison to the samples with low tumor content. The results from this study demonstrate the differences in the peripheral immune repertoire from the TIIC, reaffirming the need to robustly characterize the local TME. Additionally, the similarity in Treg and exhausted T-cell profiles across the indications examined may support a broad utility for therapeutics that can successfully target immune exhaustion or suppression.

Abstract 5477: A novel loci-selected method for microsatellite instability detection validated on a large Chinese cohort

Abstract Purpose: Microsatellite instability (MSI) was the FDA-approved companion diagnostic biomarker for using pembrolizumab in solid tumors and the biomarker for Lynch Syndrome. Next-generation sequencing (NGS) based method was recommended by NCCN and ESMO for detecting MSI status due to rich mutation information it could provide. However, current NGS based methods could be discordant on some cases, triggering confusion in clinical practice. To address this concern, we developed a novel and highly accurate MSI detection method, 3D-MSI, with a set of well-selected loci and compared its performance with other 3 public methods (MANTIS, MSISensor2, mSINGS) Materials and Methods: 100 MSI loci were selected from 2539 candidates and then sequenced on 145 FFPE samples as the training dataset. A binomial distribution model was built on each loci to predict PCR-validated MSI statuses of all samples in the training dataset and then all models were ordered by their prediction accuracy. 3D-MSI method determined MSI status by the percentage of top 30 loci models which predict the sample as MSI-H (&amp;gt;40%: MSI-H; otherwise MSS). Performances of 3D-MSI and other 3 methods were compared on 485 MSI-validated sample. Their discordance were further investigated on a large cohort of 10399 Chinese cancer patients. Detection limits of these 4 methods were determined by 4 serial-diluted tumor samples. We also observed concordance between 3D-MSI detected statuses and immunotherapy response on 27 advanced colorectal cancer (CRC) patients. Results: In the dataset of 485 samples, 3D-MSI had the highest accuracy (All: 96.7%, Colorectal cancer CRC: 98.9%, Gastric cancer GC: 97.0%, other cancer: 86.4%), sensitivity (All: 93.2%, CRC: 97.0%, GC: 94.4%, Other: 80.8%) and specificity (All: 98.2%, CRC: 99.5%, GC: 98.8%, Other: 90.0%) among all 4 methods. It could detect correct MSI status under low tumor content down to 5%, but other methods required much higher tumor content to work (MSISensor2: ≥15%, mSINGS: ≥20%, MANTIS: ≥40%). Study on 10399 Chinese cancer patients confirmed superiority of 3D-MSI: compared with other 3 methods, it yielded the fewest discordant and unreliable MSI statuses questioned by PCR (3D-MSI: 1, MSISensor2: 36, mSINGS: 36, MANTIS: 50). Loci selection was the key point for 3D-MSI to outrival other 3 methods, revealed by comparing their scoring distribution. 3D-MSI detection predicted significantly prolonged PFS (log-rank P&amp;lt;0.0001) and OS (log-rank P&amp;lt;0.0001) of the MSI-H group in 27 CRC patients who received immunotherapy, proving its validity in clinical practice. Conclusions: Our study offered an improved and clinically validated method for NGS-based MSI detection and stressed importance of loci selection in MSI detection, which would have crucial implications for future research and clinical management. Citation Format: Jian Li, Zhenghang Wang, Bo Yang, Bin Li, Yunjie Song, Bei Zhang, Xiaochen Zhao, Yifan Zhou, Zhenghua Xie, Xiaofang Li, Hao Chen, Hua Dong, Yuezong Bai, Lei Xiong, Fugen Li, Jing Gao, Lin Shen. A novel loci-selected method for microsatellite instability detection validated on a large Chinese cohort [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 5477.

Decrease in tumor content assessed in biopsies is associated with improved treatment outcome response to pembrolizumab in patients with rare tumors

BackgroundDecreased tumor content (TC) in resection specimens after neoadjuvant therapy is used to predict prognosis. We investigated whether TC assessed in biopsy specimens or the shift in TC from baseline...

Abstract A19: HLA allele-specific loss of heterozygosity detection using augmented exome capture approach

Abstract Loss of human leukocyte antigen (HLA) is of increasing interest as a mechanism of cancer immune evasion and biomarker for cancer immunotherapy response. Each cancer patient has six class I HLA alleles that are capable of presenting a set of tumor-specific neoantigens. However, HLA alleles are often deleted in tumors, resulting in a loss of heterozygosity (LOH). When LOH occurs, neoantigen presentation is significantly impaired, potentially facilitating tumor immune evasion. Given the biologic impact of HLA LOH, there is a need for robust algorithms that can detect allele-specific HLA LOH in tumor samples. Here, we describe a novel computational approach to detect HLA LOH from exome sequencing, demonstrate the robustness of the method, and apply the method to calculate the frequencies of HLA LOH in key cancer types. We performed exome sequencing with augmented HLA region capture on the ImmunoID NeXT platform for tumor and normal samples of 184 patients across several cancer types and identified 430 nonhomozygous HLA genes. Next, we extracted the reads mapping to a custom HLA database and mapped them on to the patient-specific HLA alleles. For each allele, we calculated two key features: 1) the tumor b-allele frequency normalized by the native b-allele frequency and 2) the allele-specific tumor to normal coverage ratio. Using these two features, along with tumor purity and ploidy values, we trained a random forest model on a subset of the HLA genes (n=300). While standard copy-number variant (CNV) tools are unable to detect LOH in the polymorphic HLA genes, they can accurately measure deletions in their flanking regions, which we used to validate the accuracy of our allele-specific HLA LOH algorithm. In our test set (n=130), we found a high concordance between our allele-specific deletion calls and the generic deletion calls (94% accuracy, 0.85 F1 Score). When we constrained our test set to samples with high tumor content (&amp;gt;50%, n=40), we saw even stronger concordance (98% accuracy, 0.95 F1 Score). The only discordant call was a focal deletion within an HLA gene that was detected by our algorithm but missed by the CNV tool. Next, we ran our algorithm on patient samples of different cancer indications. For non-small cell lung cancer, we found a high frequency of patients affected by LOH (35%, 9 of 26), which is similar to frequencies previously reported in the literature. Furthermore, we found a lower frequency of melanoma tumors with LOH (15%, 7 of 48). In conclusion, we developed a novel algorithm to call allele-specific HLA LOH on the ImmunoID NeXT exome sequencing platform that augments coverage in the polymorphic HLA locus and demonstrated overall robust performance. The relatively high frequency of LOH events we detected in the melanoma and lung cancer samples suggests the importance of LOH analysis to inform cancer immunotherapy biomarker studies and personalized cancer therapies that depend on neoantigen presentation. Citation Format: Rachel Marty Pyke, Charles Abbott, Simo V. Zhang, Datta Mellacheruvu, John West, Richard Chen, Sean Michael Boyle. HLA allele-specific loss of heterozygosity detection using augmented exome capture approach [abstract]. In: Proceedings of the AACR Special Conference on Tumor Immunology and Immunotherapy; 2019 Nov 17-20; Boston, MA. Philadelphia (PA): AACR; Cancer Immunol Res 2020;8(3 Suppl):Abstract nr A19.

Discordant and heterogeneous clinically relevant genomic alterations in circulating tumor cells vs plasma DNA from men with metastatic castration resistant prostate cancer.

Circulating tumor cell (CTC) and cell-free (cf) DNA-based genomic alterations are increasingly being used for clinical decision-making in oncology. However, the concordance and discordance between paired CTC and cfDNA genomic profiles remain largely unknown. We performed comparative genomic hybridization (CGH) on CTCs and cfDNA, and low-pass whole genome sequencing (lpWGS) on cfDNA to characterize genomic alterations (CNA) and tumor content in two independent prospective studies of 93 men with mCRPC treated with enzalutamide/abiraterone, or radium-223. Comprehensive analysis of 69 patient CTCs and 72 cfDNA samples from 93 men with mCRPC, including 64 paired samples, identified common concordant gains in FOXA1, AR, and MYC, and losses in BRCA1, PTEN, and RB1 between CTCs and cfDNA. Concordant PTEN loss and discordant BRCA2 gain were associated with significantly worse outcomes in Epic AR-V7 negative men with mCRPC treated with abiraterone/enzalutamide. We identified and externally validated CTC-specific genomic alternations that were discordant in paired cfDNA, even in samples with high tumor content. These CTC/cfDNA-discordant regions included key genomic regulators of lineage plasticity, osteomimicry, and cellular differentiation, including MYCN gain in CTCs (31%) that was rarely detected in cfDNA. CTC MYCN gain was associated with poor clinical outcomes in AR-V7 negative men and small cell transformation. In conclusion, we demonstrated concordance of multiple genomic alterations across CTC and cfDNA platforms; however, some genomic alterations displayed substantial discordance between CTC DNA and cfDNA despite the use of identical copy number analysis methods, suggesting tumor heterogeneity and divergent evolution associated with poor clinical outcomes.

Desmoplastic Infantile Ganglioglioma: A MAPK Pathway-Driven and Microglia/Macrophage-Rich Neuroepithelial Tumor.

MAPK pathway activation has been recurrently observed in desmoplastic infantile ganglioglioma/astrocytoma (DIG/DIA) with reported disproportionally low mutation allele frequencies relative to the apparent high tumor content, suggesting that MAPK pathway alterations may be subclonal. We sought to expand the number of molecularly profiled cases and investigate if tumor cell composition could account for the observed low mutation allele frequencies. Molecular (targeted neuro-oncology next-generation sequencing/RNA sequencing and OncoScan microarray) and immunohistochemical (CD68-PGM1/CD163/CD14/CD11c/lysozyme/CD3/CD20/CD34/PD-L1) studies were performed in 7 DIG. Activating MAPK pathway alterations were identified in 4 (57%) cases: 3 had a BRAF mutation (V600E/V600D/V600_W604delinsDQTDG, at 8%-27% variant allele frequency) and 1 showed a TPM3-NRTK1 fusion. Copy number changes were infrequent and nonrecurrent. All tumors had at least 30% of cells morphologically and immunophenotypically consistent with microglial/macrophage lineage. Two subtotally resected tumors regrew; 1 was re-excised and received adjuvant treatment (chemotherapy/targeted therapy), with clinical response to targeted therapy only. Even with residual tumor, all patients are alive (median follow-up, 83 months; 19-139). This study further supports DIG as another MAPK pathway-driven neuroepithelial tumor, thus expanding potential treatment options for tumors not amenable to surgical cure, and suggests that DIG is a microglia/macrophage-rich neuroepithelial tumor with frequent low driver mutation allele frequencies.

Leveraging Spatial Variation in Tumor Purity for Improved Somatic Variant Calling of Archival Tumor Only Samples.

Archival tumor samples represent a rich resource of annotated specimens for translational genomics research. However, standard variant calling approaches require a matched normal sample from the same individual, which is often not available in the retrospective setting, making it difficult to distinguish between true somatic variants and individual-specific germline variants. Archival sections often contain adjacent normal tissue, but this tissue can include infiltrating tumor cells. As existing comparative somatic variant callers are designed to exclude variants present in the normal sample, a novel approach is required to leverage adjacent normal tissue with infiltrating tumor cells for somatic variant calling. Here we present lumosVar 2.0, a software package designed to jointly analyze multiple samples from the same patient, built upon our previous single sample tumor only variant caller lumosVar 1.0. The approach assumes that the allelic fraction of somatic variants and germline variants follow different patterns as tumor content and copy number state change. lumosVar 2.0 estimates allele specific copy number and tumor sample fractions from the data, and uses a to model to determine expected allelic fractions for somatic and germline variants and to classify variants accordingly. To evaluate the utility of lumosVar 2.0 to jointly call somatic variants with tumor and adjacent normal samples, we used a glioblastoma dataset with matched high and low tumor content and germline whole exome sequencing data (for true somatic variants) available for each patient. Both sensitivity and positive predictive value were improved when analyzing the high tumor and low tumor samples jointly compared to analyzing the samples individually or in-silico pooling of the two samples. Finally, we applied this approach to a set of breast and prostate archival tumor samples for which tumor blocks containing adjacent normal tissue were available for sequencing. Joint analysis using lumosVar 2.0 detected several variants, including known cancer hotspot mutations that were not detected by standard somatic variant calling tools using the adjacent tissue as presumed normal reference. Together, these results demonstrate the utility of leveraging paired tissue samples to improve somatic variant calling when a constitutional sample is not available.

Distinct homologous recombination gene expression profiles after neoadjuvant chemotherapy associated with clinical outcome in patients with ovarian cancer

ObjectiveThe expression of homologous recombination (HR) genes in high grade ovarian cancer (HGOC) samples from debulking surgeries were correlated to outcomes in patients selected for chemotherapy treatment regimens. Study designRNA was extracted from 96 fresh frozen tumor samples from debulking surgeries from chemotherapy naïve patients with HGOC (primary derived surgeries (PDS), n = 55) or following neoadjuvant chemotherapy treatment (NACT), n = 41). The samples were selected for high tumor content by a gynecological pathologist, and cancer cell content was further confirmed using a percent tumor content covariate, and mutation score covariate analysis. Gene expression analysis was performed using a tailored NanoString-based Pancancer Pathway Panel. Cox proportional hazard regression models were used to assess the associations between the expression of 19 HR genes and survival. ResultsIn the PDS group, over-expression of six HR genes (C11orf30, NBN, FANCF, FANCC, FANCB, RAD50) was associated with improved outcome, in contrast to the NACT group where four HR genes (BRCA2, TP53, FANCB, RAD51) were associated with worse outcome. With the adding extent of debulking as a covariate, three HR genes (NBN, FANCF, RAD50), and only one HR gene (RAD51) remained significantly associated with survival in PDS and NACT groups, respectively. ConclusionDistinct HR expression profiles define subgroups associated with overall outcome in patients that are exposed to neoadjuvant chemotherapy and not only chemotherapy-naïve patients.

Abstract LB-007: Leveraging spatial heterogeneity in tumor purity for improved somatic variant calling of archival tumor-only samples

Abstract Archival tumor samples present a rich resource of annotated specimens for genomics research. However, standard variant calling approaches require a matched normal sample from the same individual, which is often not available with archival tissue. Without a matched normal sample, it is very difficult to distinguish between true somatic variants and germline variants that are private to the individual. Archival FFPE sections often contain adjacent normal tissue, but this normal tissue is often contaminated with infiltrating tumor cells. Comparative somatic variant callers are designed to exclude variants present in the normal sample, so a novel approach is required to leverage sequencing of adjacent normal tissue. Here we present LumosVar 2.0, a software package designed to jointly analyze multiple samples from the same patient. The approach is based on the concept that the allelic fraction of somatic variants, but not germline variants, would be diluted by the presence of normal tissue. Given the allele specific copy number state and expected tumor fraction of each sample, one can model the expected allelic fractions of somatic and germline variants. LumosVar estimates allele specific copy number and tumor sample fractions from the data, and uses the model to call somatic and germline variants. LumosVar is also able to detect sub clonal events and spatial heterogeneity between samples. To evaluate this approach, we used a glioblastoma dataset, where enhancing (high tumor content) and non-enhancing (low tumor content) samples were available for each patient, as well as the matched normal sample to define the true somatic variants. We show that both sensitivity and positive predictive value are improved by analyzing the high tumor and low tumor jointly compared to analyzing the samples individually or in silico pooling of the two samples. Finally, we apply this approach to a set of breast and prostate archival FFPE tumor samples where normal samples were not available for germline sequencing. Citation Format: Rebecca F. Halperin, Winnie S. Liang, Erica Tassone, Jonathan Adkins, Megan Russell, Nhan Tran, Nicole Hank, James Newell, Sandeep Gupta, Michael Berens, Ronald Korn, David W. Craig, Sara A. Byron. Leveraging spatial heterogeneity in tumor purity for improved somatic variant calling of archival tumor-only samples [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 LB-007. doi:10.1158/1538-7445.AM2017-LB-007

Feasibility of endobronchial ultrasound fine-needle aspiration for massively parallel next-generation sequencing in cancer patients.

8535 Background: Next generation sequencing (NGS) is an important emerging tool in precision oncology, allowing identification of a growing number of clinically validated and investigational therapeutic molecular targets. A potential limitation is that some NGS assays require more DNA input than more limited molecular assays. Endobronchial ultrasound fine-needle aspiration (EBUS-FNA) is a minimally invasive procedure for sampling mediastinal and pulmonary lesions, but it is unknown if it provides adequate material for NGS. Methods: An IRB approved, retrospective review was performed of patients undergoing EBUS-FNA by thoracic surgeons at our institution 3/1/14 - 9/28/16. NGS was performed using an assay developed at our institution that detects mutations in up to 410 genes (MSK-IMPACT). Samples diagnostic for malignancy and with MSK-IMPACT requested were identified. Pathology and clinical data were drawn from the medical record and MSK-IMPACT results were examined. Results: 784 EBUS-FNA were done in the study period. MSK-IMPACT was requested on 115 positive samples. MSK-IMPACT was successful in 99 samples (86.1%), identifying an average of 12.7 mutations at a mean coverage depth of 806X. In 17 (17.2%) samples, tumor content was suboptimal ( &lt; 20% of nucleated cells), with fewer identified mutations than in cases with higher tumor content (6.8 vs 13.9, p = 0.01). NGS was performed on paraffin-embedded cell blocks in 93 cases (93.9%), and in 6 cases DNA extraction was performed from residual cytological material isolated from supernatant including cell-free DNA. Failures were attributable to low cell content (7), high contamination by benign cells (4) or both (1) and processing issues (4). No difference in surgical or radiologic parameters were identified for failed or suboptimal samples. Conclusions: In our practice, EBUS-FNA has a high rate of success for obtaining adequate tissue for NGS. Ability to utilize cell-free DNA for molecular studies – a new process in our lab – allows increased success of molecular testing in scant samples. Further studies may identify factors contributing to NGS failure and to improving success for samples with minimal cellularity.

Identification of BRCA1-like triple-negative breast cancers by quantitative multiplex-ligation-dependent probe amplification (MLPA) analysis of BRCA1-associated chromosomal regions: a validation study.

BackgroundTriple-negative breast cancer (TNBC) with a BRCA1-like molecular signature has been demonstrated to remarkably respond to platinum-based chemotherapy and might be suited for a future treatment with poly(ADP-ribose)polymerase (PARP) inhibitors. In order to rapidly assess this signature we have previously developed a multiplex-ligation-dependent probe amplification (MLPA)-based assay. Here we present an independent validation of this assay to confirm its important clinical impact.MethodsOne-hundred-forty-four TNBC tumor specimens were analysed by the MLPA-based “BRCA1-like” test. Classification into BRCA1-like vs. non-BRCA1-like samples was performed by our formerly established nearest shrunken centroids classifier. Data were subsequently compared with the BRCA1-mutation/methylation status of the samples. T-lymphocyte infiltration and expression of the main target of PARP inhibitors, PARP1, were assessed on a subset of samples by immunohistochemistry. Data acquisition and interpretation was performed in a blinded manner.ResultsIn the studied TNBC cohort, 63 out of 144 (44 %) tumors were classified into the BRCA1-like category. Among these, the MLPA test correctly predicted 15 out of 18 (83 %) samples with a pathogenic BRCA1-mutation and 20 of 22 (91 %) samples exhibiting BRCA1-promoter methylation. Five false-negative samples were observed. We identified high lymphocyte infiltration as one possible basis for misclassification. However, two falsely classified BRCA1-mutated tumors were also characterized by rather non-BRCA1-associated histopathological features such as borderline ER expression. The BRCA1-like vs. non-BRCA1-like signature was specifically enriched in high-grade (G3) cancers (90 % vs. 58 %, p = 0.0004) and was also frequent in tumors with strong (3+) nuclear PARP1 expression (37 % vs. 16 %; p = 0.087).ConclusionsThis validation study confirmed the good performance of the initial MLPA assay which might thus serve as a valuable tool to select patients for platinum-based chemotherapy regimens. Moreover, frequent PARP1 upregulation in BRCA1-like tumors may also point to susceptibility to treatment with PARP inhibitors. Limitations are the requirement of high tumor content and high-quality DNA.

Digitally guided microdissection aids somatic mutation detection in difficult to dissect tumors

Molecular genetic testing on formalin fixed, paraffin embedded (FFPE) tumors frequently requires dissection of tumor from tissue sections mounted on glass slides. In a process referred to as "manual macrodissection," the pathologist reviews an H&E stained slide at the light microscope and marks areas for dissection, and then the laboratory performs manual dissection from adjacent sections without the aid of a microscope, using the marked reference H&E slide as a guide. Manual macrodissection may be inadequate for tissue sections with low tumor content. We compared manual macrodissection to a new method, digitally guided microdissection, on a series of 32 FFPE pancreatic cancer samples. KRAS hotspot mutation profiling was performed using the Sequenom MassARRAY system (Agena Bioscience). Digitally guided microdissection was performed on multiple smaller areas of high tumor content selected from within the larger areas marked for manual macrodissection. The KRAS mutant allele fraction and estimated neoplastic cellularity were significantly higher in samples obtained by digitally guided microdissection (p < 0.01), and 7 of the 32 samples (22%) showed a detectable mutation only with digitally guided microdissection. DNA quality and yield per cubic millimeter of dissected tissue were similar for both dissection methods. These results indicate a significant improvement in tumor content achievable with digitally guided microdissection.