Search

International audience; Abstract Ki67 has potential clinical importance in breast cancer but has yet to see broad acceptance due to inter-laboratory variability. Here we tested an open source and calibrated automated digital image analysis (DIA) platform to: (i) investigate the comparability of Ki67 measurement across corresponding core biopsy and resection specimen cases, and (ii) assess section to section differences in Ki67 scoring. Two sets of 60 previously stained slides containing 30 core-cut biopsy and 30 corresponding resection specimens from 30 estrogen receptor-positive breast cancer patients were sent to 17 participating labs for automated assessment of average Ki67 expression. The blocks were centrally cut and immunohistochemically (IHC) stained for Ki67 (MIB-1 antibody). The QuPath platform was used to evaluate tumoral Ki67 expression. Calibration of the DIA method was performed as in published studies. A guideline for building an automated Ki67 scoring algorithm was sent to participating labs. Very high correlation and no systematic error ( p = 0.08) was found between consecutive Ki67 IHC sections. Ki67 scores were higher for core biopsy slides compared to paired whole sections from resections ( p ≤ 0.001; median difference: 5.31%). The systematic discrepancy between core biopsy and corresponding whole sections was likely due to pre-analytical factors (tissue handling, fixation). Therefore, Ki67 IHC should be tested on core biopsy samples to best reflect the biological status of the tumor.; v2.5, 25 March 2021.

The nuclear proliferation biomarker Ki67 has potential prognostic, predictive, and monitoring roles in breast cancer. Unacceptable between-laboratory variability has limited its clinical value. The International Ki67 in Breast Cancer Working Group investigated whether Ki67 immunohistochemistry can be analytically validated and standardized across laboratories using automated machine-based scoring. Sets of pre-stained core-cut biopsy sections of 30 breast tumors were circulated to 14 laboratories for scanning and automated assessment of the average and maximum percentage of tumor cells positive for Ki67. Seven unique scanners and 10 software platforms were involved in this study. Pre-specified analyses included evaluation of reproducibility between all laboratories (primary) as well as among those using scanners from a single vendor (secondary). The primary reproducibility metric was intraclass correlation coefficient between laboratories, with success considered to be intraclass correlation coefficient >0.80. Intraclass correlation coefficient for automated average scores across 16 operators was 0.83 (95% credible interval: 0.73–0.91) and intraclass correlation coefficient for maximum scores across 10 operators was 0.63 (95% credible interval: 0.44–0.80). For the laboratories using scanners from a single vendor (8 score sets), intraclass correlation coefficient for average automated scores was 0.89 (95% credible interval: 0.81–0.96), which was similar to the intraclass correlation coefficient of 0.87 (95% credible interval: 0.81–0.93) achieved using these same slides in a prior visual-reading reproducibility study. Automated machine assessment of average Ki67 has the potential to achieve between-laboratory reproducibility similar to that for a rigorously standardized pathologist-based visual assessment of Ki67. The observed intraclass correlation coefficient was worse for maximum compared to average scoring methods, suggesting that maximum score methods may be suboptimal for consistent measurement of proliferation. Automated average scoring methods show promise for assessment of Ki67 scoring, but requires further standardization and subsequent clinical validation.

Introduction: Chimeric antigen receptor T (CAR T) cells are an important emerging therapy for hematologic malignancies and are currently being investigated for other cancer types as well. CAR-T cells require a retrovirus or lentivirus to introduce the chimeric antigen receptor gene into the activated T cells. Safety measures are put in place by design to eliminate any chance to create replication competent virus. Multiple plasmids carrying specific components of the virus are used to reduce the chances of creating a replication competent lentivirus. Despite these measures, safety concerns have been posed by the FDA regarding use of a lentivirus for this treatment due to the potential for replication competent lentiviruses (RCL) to arise. FDA requires monitoring of both CAR T products as well as patients undergoing therapy. It has proven to be difficult to achieve acceptable detection sensitivity of the VSV-G sequence in a background of gDNA due to the high level of homology to many areas in gDNA. Methods: We have developed a new rapid and sensitive qPCR assay targeting the vesicular stomatitis virus G glycoprotein (VSV-G), an envelope protein on lentiviral vectors. Incorporation of the envelope gene would be required to generate a replication competent virus. The intended use of the validated assay is to detect and monitor for RCL presence in human blood samples collected during clinical monitoring. A validated reference assay targeting RPPH1 will be used to quantify the amount of amplifiable genomic DNA in each extracted DNA sample. The size of the amplicons are 66 and 87 base pairs for VSV-G and RPPH1 respectively. The VSV-G assay test includes one positive and one negative control per assay run. The VSV-G positive control is composed of plasmid containing the VSV-G target sequence spiked into a background of VSV-G negative human gDNA. The negative control is composed of negative gDNA without VSV-G plasmid. Results: Improved sensitivity without non-specific amplification was achieved by optimized primer and probe selection. Additionally sensitivity and specificity were enhanced by employing a step-down PCR protocol whereby the annealing and extension temperature is lowered in stepwise fashion over the first five cycles of PCR. Genomic DNA inputs up to 200ng per reaction showed high efficiency of amplification from 10 to 1e6 copies target sequence per reaction. Conclusion: This assay will provide a rapid and sensitive method for safety monitoring of patients undergoing CAR-T therapy with therapeutic cells generated using lentiviral vectors. Citation Format: David Burke, Kristine Drafahl, Clark Fjeld, Chad Galderisi, Cindy Spittle. Enhanced sensitivity detection of replication competent lentivirus by qPCR [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 896.

PD-L1 expression and MSI status are useful predictive biomarkers for checkpoint inhibitors (CIs) in some tumor types. However, studies have demonstrated that the identification of all potential patients who may benefit from CIs will require the analysis of both tumor cell and tumor microenvironment (TME) biomarkers. For example, recent data suggests that tumor mutation burden (TMB) and tumor mutation signatures as well as inflammatory signatures and T cell clonality in the tumor microenvironment may serve as independent predictive biomarkers of CI response. The need for evaluation of multiple biomarkers using IHC, PCR and/or NGS using limited amounts of tissue poses a challenge in the clinical setting. Here we conduct a study to evaluate whether a panel of tumor and TME biomarkers can be accurately analyzed using low DNA and RNA input from FFPE specimens or using plasma-derived cfDNA. Multiple methods for assessing TMB, MSI status and mutation signatures were compared. The impact of sample quality on biomarker analysis was also evaluated. A set of late stage CRC and matched plasma samples were used in this study. DNA and RNA were co-extracted from FFPE sections using RecoverAll™ kit. cfDNA was extracted from plasma using the MagMAX™ Kit. DNA and RNA yields from FFPE ranged from 855-3875ng and 477-5200ng, respectively. cfDNA yields ranged from 8-38ng. Based on the KAPA DNA QC results, the FFPE samples were determined to be high quality. TMB and mutation signature analysis of FFPE samples were performed using Oncomine Tumor Mutation Load (TML) Assay, the Oncomine Comprehensive Assay v3 (OCAv3) and WES. MSI status in FFPE tissue and cfDNA was determined using the Promega MSI Analysis System. The mutation spectrum in the matched cfDNA samples was determined using the Oncomine PanCancer cfTNA Assay. Inflammatory (IFNG) signatures were assessed using the targeted RNASeq panel, Oncomine Immune Response Research Assay (OIRRA). The clonality of TILs was analyzed using NGS. TMB results generated from the targeted panels using only 20ng DNA input were comparable to WES results. Differences in TMB counts between the NGS assays are likely due to differences in genomic coverage and data filtering. MSI-H status and/or mutations in critical DNA repair or proofreading genes such as MSH6 and POLE correlated with high TMB. Gene expression profiling using OIRRA was successful for all FFPE RNA samples. In summary, this study demonstrates that comprehensive genomic analysis of predictive biomarkers for CI therapy can be performed from limited FFPE samples using co-extraction methods and a combination of targeted NGS and PCR-based assays. DNA and RNA quality have a significant impact on the accuracy of TMB analysis and gene expression profiling. In cases where no tumor tissue is available, the analysis of biomarkers such as MSI and mutation signatures may be possible using cfDNA. Citation Format: Xiaodong Wang, Zhenyu Yan, Quyen Vu, David Smith, Chad Galderisi, Cynthia S. Spittle, Jin Li. Comprehensive genomic analysis of predictive biomarkers for checkpoint inhibitor therapy from a limited amount of FFPE tissue and cfDNA [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 1702.

In Ph+ ALL, the absence of detectable disease has shown prognostic value for a reduced risk of relapse and improved survival. However, as the level of undetectable disease is determined by the lower limit of detection of the test in use, standardization of such an endpoint for a drug regulatory submission is critical. To date, no tyrosine kinase inhibitor (TKI) has received approval for the newly diagnosed Ph+ ALL adult patient population in the US. Takeda (Millennium Pharmaceuticals, Inc.) is conducting a phase 3, randomized, open-label, multicenter efficacy study comparing ponatinib versus imatinib, administered in combination with reduced-intensity chemotherapy, in participants with newly diagnosed Ph+ ALL (NCT03589326). The primary endpoint for this study is minimal residual disease (MRD)-negative complete remission (CR), where MRD-negativity is defined as a BCR-ABL:ABL raw ratio of ≤0.01% (MR4.0) in bone marrow aspirate samples at the end of induction. Patients who achieve post-induction ponatinib or imatinib maintained MRD-negative CR will potentially delay or avoid stem cell transplantation. Previously, for the purposes of initiating and monitoring treatment free remission or discontinuation of TKI therapy in chronic phase CML patients, we developed and validated the MRDx® BCR-ABL Test which is an FDA authorized test for the quantitative detection of BCR-ABL e13a2 or e14a2 transcripts. This test will be used in this study and reports BCR-ABL:ABL levels on the International Scale (IS) with traceability to the World Health Organization (WHO) first International Genetic Reference Panel and with a limit of detection below 0.0032% (i.e., MR4.5). Similarily, for assessment of the e1a2 (p190) BCR-ABL:ABL transcripts, we developed and validated a one-step reverse transcription, quantitative polymerase chain reaction (RT-qPCR) test in order to accurately and precisely assess all clinical decision points and disease levels for this study. Because of the lack of available reference material for e1a2, a droplet digital PCR (ddPCR) based test was co-developed to quantify e1a2 BCR-ABL copy numbers in bone marrow aspirates, as well as in peripheral blood samples (to allow assessment of concordance). e1a2 in vitro transcribed RNA calibrators assign copy numbers to determine the e1a2 BCR-ABL:ABL raw % ratios of unknown samples. The e1a2 RT-qPCR test exceeded an analytical sensitivity of MR4.5 (0.0032% raw ratio of BCR-ABL:ABL) with a dynamic linear range from MR4.5 to MR1.0. The test also includes cell line derived RNA assay controls formulated to 10%, 0.1% and 0.01% BCR-ABL:ABL, necessary for decision points in the clinical trial. Validation studies included limit of blank, limit of detection (LOD), limit of quantification, assay range, analytical specificity, repeatability, reproducibility (multi-day, multi-operator, and multi-instrument), and accuracy by comparison to a reference method (ddPCR). The validation of the e1a2 RT-qPCR test with bone marrow aspirate samples was conducted with 1 µg RNA inputs per well and LOD was also verified with 0.5 µg RNA input per well. In conclusion, the validated e1a2 RT-qPCR test allows for accurate standardization of BCR-ABL:ABL measurement across multiple centers in an international Phase 3 study. The e1a2 RT-qPCR test data will be used to assess the primary endpoint in the first registrational trial to be conducted in newly diagnosed Ph+ ALL adult patients. Disclosures Drafahl: MolecularMD, Corp: Employment. Smith:MolecularMD, Corp: Employment. Graham:MolecularMD, Corp: Employment. Glynn:MolecularMD, Corp: Employment. Spittle:MolecularMD, Corp: Employment. Verrow:Takeda (Millennium Pharmaceuticals, Inc.): Employment. Rivera:Takeda (Millennium Pharmaceuticals, Inc.): Consultancy. Srivastava:Takeda (Millennium Pharmaceuticals, Inc.): Employment. Hawkins:MolecularMD, Corp: Employment. Galderisi:MolecularMD, Corp: Employment, Equity Ownership.

FLT3 ITD, a prognostic marker and drug target, is present in ~19% of AML patients. FLT3 ITD is a tandem repeat of the entire or partial exon 13-15 region. FLT3 ITDs range in size from 3bp to above 300bp and impact RTK signaling. Recent publications have shown that FLT3 ITD can be detected at single nucleotide resolution by NGS panels. However, detection of large FLT3 ITD and accurate reporting of ITD frequency remain challenging for NGS methods. False negative ITD results or inaccurate mutant/wild type allele ratio measurements could negatively alter treatment decisions for AML patients. The Illumina TruSight Myeloid panel (TMP) is a commercial NGS assay that covers 54 genes involved in myeloid malignancies. Although the TMP assay can detect SBS and small indel, medium to large size FLT3 ITD cannot be detected. Here we present an enhanced workflow to overcome these challenges. Positive controls, dynamic range, accuracy, and inter-run reproducibility were validated for the detection of SBS, indel, and FLT3 ITD using reference standards and DNA extracted from cell lines and AML patient EDTA blood and bone marrow aspirate samples. NGS libraries were prepared following Illumina's standard protocol. 275bp X2 sequencing instead of the standard 150bp X2 was run on MiSeq flowcell v3. Data analysis was performed using a MolecularMD custom pipeline. The analysis of FLT3 ITD was conducted using three methods including two off-the-shelf programs, Pindel and ITDseek, as well as Cigar, a MolecularMD developed ITD detection algorithm. FLT3 ITDs at 126 bp and 21 bp in PL-21 and MOLM-13 cell lines, respectively, were detected by TMP assay using the three algorithms. In contrast to ITDseek which only reports ITD sequence without frequency information, Pindel and Cigar report both ITD sequence and frequency. Compared to Pindel, ITD frequency reported by the custom algorithm is closer to that determined by the fragment analysis. Serial dilution of the 126bp ITD into wild type DNA showed the 0.55% ITD (from fragment analysis) can be reproducibly detected at ~0.6% by the Cigar algorithm. FLT3 ITDs ranging in size from 6bp to 300bp were detected in clinical AML samples using the enhanced TMP workflow and confirmed by a fragment analysis method. Limit of detection, reproducibility and accuracy of SBS and indel detection in other key genes by the TMP assay have also been validated. Low-level mutations, i.e. NRAS G12D at 1.7%, NRAS G13D at 2%, and IDH2 R140Q at 1.4%, were detected by TMP panel from AML samples and further confirmed with an independent assay. A long deletion (52bp) in exon 9 of CALR was also detected by the custom pipeline. In summary, the enhanced TMP workflow allows for accurate detection and reporting of SBS, indel and FLT3 ITDs including large duplications. Optimized detection of FLT3 ITDs will improve AML patient selection for targeted therapy. Citation Format: Xiaodong Wang, Zhenyu Yan, Peng Fang, Weihua Liu, Scott Glynn, Jennifer Biroschak, Chad Galderisi, Cindy Spittle, Jin Li. An optimized NGS workflow for detection of FLT3 internal tandem duplication (ITD) in AML samples [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 4522.

Checkpoint inhibitors have been approved for the treatment of solid tumor and hematological malignancies. While significant responses have been observed in a subset of patients, outcomes are variable and there is a need to identify additional predictive biomarkers beyond PD-L1 levels as measured by IHC. Tumor mutation burden (TMB) has been correlated with response to checkpoint inhibitors and is emerging as a key biomarker for predicting checkpoint inhibitor response. So far, the methods used to assess tumor mutation burden have included exome sequencing and multiple laboratory-developed targeted NGS panels (e.g., FoundationOne and MSK-IMPACT). In order to fully determine the value of TMB as a predictive biomarker for immunotherapy, a standardized panel, workflow and data analysis pipeline for TMB assessment are needed. In this study we evaluated the performance of a commercially available targeted NGS panel and workflow for TMB analysis. A set of 30 FFPE tumor samples including colon, renal, gastric, endometrial, and lung tissues was analyzed. DNA and RNA were extracted using the RecoverAll Total Nucleic Acid Isolation Kit. DNA quantity and quality were assessed using Qubit and qPCR, respectively. Samples were analyzed with the ThermoFisher Oncomine™ Mutation Load Research Assay (TML), which evaluates tumor mutation load (mutations/Mb) by interrogating 409 cancer-related genes, spanning ~1.7 megabases of the genome. TMB was measured by counting somatic single-base substitutions per Mb at ≥10% allele frequency in single, non-matched, tumor DNA samples. The impact of DNA quality on the TMB score was evaluated. Deamination errors (i.e. G>A and C>T) in poor quality FFPE samples was found to cause the overestimation of TMB. Therefore, a delta Ct cutoff was established to qualify DNA samples for TMB analysis. 12 of the 30 samples were also analyzed using a comparator NGS panel covering ~1.25 megabases. The correlation of TMB results between the two panels was 0.87. Overall, TMB was lowest in RCC (9-17/Mb) compared to NSCLC and CRC (16-37/Mb). MSI status was determined using the Promega MSI Analysis System v1.2. A correlation was observed between TMB and MSI status in a subset of samples. Reproducibility of the assay was also evaluated. To identify clinically relevant mutations and genetic alteration associated with high mutation burden, the Oncomine Comprehensive assay v3 (OCAv3) was also used to analyze the sample set. Mutations in genes involved in several DNA repair pathways were found to correlate with TMB. This study demonstrated the feasibility of utilizing a commercial targeted NGS panel and data analysis pipeline for TMB evaluation in clinical FFPE tumor samples. Standardization of TMB analysis will enable the clinical validation of TMB as a predictive biomarker for therapy selection. Citation Format: Peng Fang, Zhenyu Yan, Quyen Vu, David Smith, Chad Galderisi, Cynthia S. Spittle, Jin Li. Evaluation of a commercial targeted NGS panel for tumor mutation burden assessment in FFPE tissue [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 3614.

Introduction: Analysis of cfDNA has potential applications in therapy selection and disease monitoring in the clinical management of NSCLC. Here we use EGFR ddPCR and the Oncomine Lung cfDNA Assay to evaluate the mutation profile and allele burden in a set of plasma samples collected from NSCLC patients. Methods: Plasma samples were collected at Dana-Farber Cancer Institute (DFCI). cfDNA was extracted using the QIAamp DSP circulating NA kit. EGFR ddPCR assays were designed by DFCI. One third of cfDNA extracted from 1 ml plasma was used for each EGFR ddPCR assay (L858R, T790M and ex19del). The L858R and T790M ddPCR assays detect mutations using mutant specific probes. The ex19del assay allows for the detection of a broad range of exon 19 deletions via the loss of wild-type exon 19 signal. Twenty-two plasma samples from patients with EGFR WT tumors were analyzed to assess the background noise (Limit of Blank). The other 32 plasma samples were analyzed in a blinded fashion. ddPCR analysis was performed on the QX200 system and data was analyzed using QuantaSoft software. A subset of 12 plasma samples had sufficient cfDNA for additional analysis using the Oncomine Lung cfDNA Assay on the Ion S5. This NGS panel is designed to detect 150 hotspot mutations in 11 lung cancer-related genes. Torrent Suite Software 5.2 was used for NGS data analysis. The HorizonDx cfDNA Multiplex Reference Standard (AF 5%, 1%, 0.1%, and WT) was also included in the ddPCR and NGS studies. Results: Based on the Limit of Blank data, the positive mutation cut-off for each EGFR ddPCR assay was set at 3 mutant copies with minimum 3000 wild type copies. Studies using the HorizonDx cfDNA standard demonstrated that the analytical sensitivity of the ddPCR assays increases with cfDNA input: the lowest detected allele frequency (AF) is 1%, 0.5%, and 0.25% for 3, 6, and 12ng input, respectively. In plasma samples, EGFR mutations were detected as low as 0.08% AF using ddPCR (4.5 L858R mutant copies/5827 wild type copies). L858R and exon 19 deletions were detected in 13/54 and 21/54 plasma samples, respectively. T790M mutations were detected in 19/54 plasma samples, and coexisted with L858R or ex19del in all cases. The Oncomine™ Lung cfDNA Assay showed 100% concordance with EGFR ddPCR results for the 12 plasma samples analyzed by both methods. Mutations were detected as low as 0.18% AF with 8.5ng plasma cfDNA input in the NGS assay. Using the HorizonDx cfDNA multiplex standard mutations were detected as low as 0.1% AF using 20ng input. In addition, PIK3CA mutations were found to coexist with EGFR ex19 deletions in 3/12 plasma samples using NGS. The AF of PIK3CA mutations and EGFR mutations was comparable in the same samples (TRL23: 24.3% for EGFR del19 and 10.4% T790M vs 20.2% for PIK3CA H1047R; TRL48: 7.8% for EGFR del19 vs 3.4% for PIK3CA E545K; TRL51: 0.85% for EGFR del19 vs 1.02% for PIK3CA E545K). Conclusions: Both ddPCR and NGS were able to detect mutations at 0.1% AF with sufficient cfDNA input. The Oncomine panel confirmed EGFR ddPCR results and detected coexisting PIK3CA mutations. EGFR and PIK3CA mutations were present at a similar AF, suggesting that EGFR and PIK3CA mutations may develop at the same stage of tumorigenesis in these patients. Our study results demonstrate that comprehensive profiling of cfDNA using NGS could be critical for the selection of appropriate therapies for NSCLC patients while ddPCR is an ideal platform for monitoring disease burden during the course of therapy. Citation Format: WeiHua Liu, Peng Fang, Chad Galderisi, Cindy Spittle, Jin Li. Liquid biopsy analysis of NSCLC patient plasma using ddPCR and NGS [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference: Molecular Targets and Cancer Therapeutics; 2017 Oct 26-30; Philadelphia, PA. Philadelphia (PA): AACR; Mol Cancer Ther 2018;17(1 Suppl):Abstract nr B072.

Haemophilus influenzae HMC-C with high-level macrolide resistance after multi-step selection by clarithromycin reverted spontaneously and became hypersusceptible to macrolides.Determination of macrolide resistance mechanism(s) in hypersusceptible and hyperresistant strains.The presence of macrolide efflux in the strains was studied by radioactive erythromycin accumulation. Ribosomal mutations were investigated by sequencing. The possible role of acrAB clusters in macrolide resistance was studied by sequencing and expression analysis.The parent strain had no ribosomal alteration, but both high-level resistant and hypersusceptible strains had R88P mutations in ribosomal protein L22. Radioactive macrolide accumulation studies pointed to the presence of macrolide efflux in the high-level resistant and parent strains, but not in the hypersusceptible derivative. Transformation of hypersusceptible strains using total DNA from the parent strain restored the macrolide efflux system in the hypersusceptible strain, which was confirmed by MIC levels and radioactive erythromycin accumulation similar to that of the mutant resistant strain. Analysis of sequence and transcription of acrAB gene clusters showed no significant differences between resistant and hypersusceptible derivatives.Mutation in ribosomal protein L22 alone does not confer high-level macrolide resistance unless efflux is present.

The recent identification of activating fibroblast growth factor receptor 2 (FGFR2) mutations in endometrial cancer has generated an opportunity for a novel target-based therapy. Here, we explore the therapeutic potential of 2 FGFR inhibitors, the multikinase inhibitor dovitinib (TKI258) and the more selective FGFR inhibitor NVP-BGJ398 for the treatment of endometrial cancer. We examined the effects of both inhibitors on tumor cell growth, FGFR2 signaling, cell cycle, and apoptosis using a panel of 20 molecularly characterized human endometrial cancer cell lines. Anchorage-independent growth was studied using soft agar assays. In vivo studies were conducted using endometrial cancer xenograft models. Cell lines with activating FGFR2 mutations (S252W, N550K) were more sensitive to dovitinib or NVP-BGJ398 when compared with their FGFR2 wild-type counterparts (P = 0.073 and P = 0.021, respectively). Both agents inhibited FGFR2 signaling, induced cell-cycle arrest, and significantly increased apoptosis in FGFR2-mutant lines. In vitro, dovitinib and NVP-BGJ398 were both potent at inhibiting cell growth of FGFR2-mutant endometrial cancer cells, but the activity of dovitinib was less restricted to FGFR2-mutant lines when compared with NVP-BGJ398. In vivo, dovitinib and NVP-BGJ398 significantly inhibited the growth of FGFR2-mutated endometrial cancer xenograft models. In addition, dovitinib showed significant antitumor activity in FGFR2 wild-type endometrial cancer xenograft models including complete tumor regressions in a long-term in vivo study. Dovitinib and NVP-BGJ398 warrant further clinical evaluation in patients with FGFR2-mutated endometrial cancer. Dovitinib may have antitumor activity in endometrial cancer beyond FGFR2-mutated cases and may permit greater flexibility in patient selection. Mol Cancer Ther; 12(5); 632–42. ©2013 AACR.

Introduction: The Oncomine™ Focus Assay (OFA) is aimed to simultaneously detect and report hotspot mutations, Copy Number Variants (CNVs) and gene fusions in 52 oncogenes clinically relevant to solid tumors. With minimal DNA/RNA input, OFA employs AmpliSeq™ library preparation chemistry to enrich target regions for Ion Torrent™ Next Generation Sequencing. Variants are annotated for actionability by Oncomine® Knowledgebase. Here we report an analytical validation of OFA. Methods: DNA and RNA, extracted from the FFPE processed GM12878, was used as the negative control to evaluate the specificity of OFA. A RNA sample containing multiple oncogenic gene fusions and a DNA sample containing multiple hotspot SNV and indels were used as the positive controls. Fresh DNA from cancer cell lines (HCC1143 or NCI-H2122) along with DNA from matched normal cell lines, HorizonDx NGS standards TruQ-1, TruQ-2, several engineered FFPE samples with gene fusions or copy number changes, and 43 clinical FFPE samples of a variety of solid tumor types (non-small cell lung cancer, colorectal cancer, gastrointestinal stromal tumor, ovarian cancer, pancreatic cancer, and melanoma) were used to evaluate the OFA performance. The analysis of the sequencing data was primarily performed with the OFA pipeline integrated with the Oncomine® Knowledgebase. Only the genetic alterations with clinical utility were selected as the final output from the data pipeline. The SBS and indels detected by OFA were confirmed by Sanger sequencing. Any CNVs detected were confirmed by FISH, if possible, and detected fusions were confirmed by RT-PCR or FISH. Results: No clinically relevant genetic alterations were detected from the negative control FFPE-GM12878, indicating the high specificity of the OFA. The LOD for SBS and short indel detection was 5%, as assessed by TruQ-1 and TruQ-2, each containing 9-10 variants. The assay can detect gene fusion RNA present as 0.1%-1% of the total RNA, and gene amplifications are detected with a minimum of 50% tumor cell content. A known 12X CCND1 amplifications in HCC1143 and a 13X MYC amplification in NCI-H2122 were detected. In addition, a 13X EGFR amplification and 11X CDK4 amplification were detected in a lung cancer FFPE sample and confirmed by FISH. The analytical specificity for detection of SBS and indels was 100% (25/25). The analytical specificity for detection of CNV gain and gene fusions, and the assay sensitivity are currently under assessment. Conclusions: These results demonstrate the high sensitivity and specificity of OFA. With as little as 10 ng RNA and 10 ng DNA, OFA provides biomarker analysis of patient FFPE tumor specimens focusing on the detection of genetic alterations that have been targeted by oncology drugs or supported by published clinical evidence. Citation Format: Peng Fang, Zhenyu Yan, Paul Labrousse, Weihua Liu, Jennifer Biroschak, Jennifer Wright, Selby Weeks, Cindy Spittle, Chad Galderisi, Jin Li. Oncomine focus assay: Simultaneous detection of clinically relevant hotspot mutations, CNVs, and gene fusions in 52 oncogenes relevant to solid tumors. [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 1397.

Introduction: Analysis of ctDNA has potential applications in targeted therapy selection and disease monitoring in the clinical management of colorectal cancer. Here we use an amplicon based target enrichment method, AmpliSeq v2 (CHP2), to study a set of tumor/plasma pairs (n = 13) collected from stage IV colon cancer patients to determine whether amplicon based NGS can be used to profile mutations in ctDNA from CRC patients. Methods: The CHP2 is designed to survey 2800 mutations in 50 cancer-related genes. Matched tumor and plasma samples were purchased from Indivumed. DNA was extracted from 3 or 5 ml plasma using QIAamp DSP circulating NA kit and eluted in 10 μl buffer per ml plasma. The sample input used in the CHP2 reactions was between 3.8 - 55 ng ctDNA. The background noise level of hotspot nt positions was calculated from 10 healthy donor plasma samples. A custom data analysis program was developed to detect variants clearly different from background noise. ddPCR mutation assays were purchased from BioRad and performed using the QX200. ddPCR data was analyzed using QuantaSoft software. Results: The ctDNA samples can be divided into two groups based on the DNA concentration measured by Qubit (range of 0.16-13 ng/ul): 6 with >1 ng/ul and 7 with 10% (N = 7), < 1% (N = 11) and between 1-10% (N = 10). Co-existing mutations in the same tumor tend to be detected simultaneously in the matched plasma at the same level of allele frequency. Two exceptions of this were that only driver mutations (NRAS Q60L and PIK3CA E545G) were detected in two ctDNA samples, while mutations in other genes TP53, APC and GNAQ were found only in the matched tumors, probably due to differential roles of these genes in cancer progression and metastasis. The sequencing read depth of these mutations ranged from 1528-17364. Four mutations with low allele frequency between 0.06-0.2% were reproducibly detected in an independent run. The background error rate at these nt positions was zero. The somatic mutations with >10% allele frequency were detected in 3 ctDNA samples (ctDNA concentration > 1ng/ul) that have co-existing mutations in KRAS and PIK3CA and/or APC gene, suggesting polyclonal evolution in multiple oncogenic pathways. Several Germline SNPs were detected in multiple genes and excluded from concordance calculation between tumor and plasma. A small subset of ctDNA samples, including 1 negative and 2 positive samples (KRAS G12D 1.4% and PIK3CA E545G 0.9%) was also analyzed by ddPCR and results were consistent with those obtained using CHP2. Conclusions: AmpliSeq CHP2 cancer panel achieved 0.06-0.2% allele frequency sensitivity and 85% (11/13) tumor/plasma concordance, which are comparable with previous reports using alternative panels. We also demonstrate the efficiency of target enrichment by AmpliSeq based multiplex PCR, which can detect as low as two mutant copies from wild type background. This high sensitivity method is ideal for clinical sample testing where the amount of plasma-derived DNA is limited. Citation Format: WeiHua Liu, Zhenyu Yan, Paul Labrousse, Peng Fang, Jennifer Biroschak, Cindy Spittle, Chad Galderisi, Jin Li. Mutation profiling of colorectal cancer ctDNA using AmpliSeq CHP2 cancer panel. [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference: Molecular Targets and Cancer Therapeutics; 2015 Nov 5-9; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2015;14(12 Suppl 2):Abstract nr A29.

Background. Angiosarcoma arising after radiation is described in breast cancer but occurs elsewhere. Here, we present an angiosarcoma of the neck occurring after curative chemoradiation. Methods. This is a case of angiosarcoma developing 5 years after curative therapy for T3N0 squamous cell carcinoma of the supraglottic larynx. Therapy consisted of 3 cycles of induction cis- platin/5-fluorouracil chemotherapy followed by radiotherapy. The patient did well until developing a rapidly progressive lesion of the left neck. Examination showed a 7-cm mass and biopsy revealed angiosarcoma. CT scans assessed the local extent of the tumor and ruled out metastatic disease prior to initiating therapy. Results. Therapy consisted of 4 cycles of paclitaxel chemo- therapy. At completion, examination revealed mild induration of the neck with near-complete resolution of the mass. Imaging confirmed this excellent response. Conclusion. This rare therapy-related second malignancy developed after curative larynx-preserving treatment. Paclitaxel was an effective therapy in this setting. V C 2008 Wiley Periodi- cals, Inc. Head Neck 31: 829-832, 2009

Introduction: The Oncomine® Cancer Panel (OCP) is designed to simultaneously detect and report hotspot mutations, Copy Number Variants (CNVs) and gene fusions in 143 genes with minimal DNA/RNA sample input. The OCP utilizes the AmpliSeq™ library preparation chemistry for the Ion Torrent™ Next Generation Sequencing (NGS) platform, in combination with data annotations by the Oncomine® Knowledgebase. Here we report analytical validation of OCP. Methods: DNA and RNA, extracted from the FFPE processed GM12878, was used as the negative control to evaluate the specificity of OCP. A RNA sample containing multiple oncogenic gene fusions, and a DNA sample containing multiple hotspot SNV and indels were used as the positive control. Fresh DNA from cancer cell lines (HCC1143 or NCI-H2122) along with DNA from matched normal cell lines, HorizonDx NGS standards TruQ-1, TruQ-2, several engineered FFPE samples with gene fusions or copy number changes, and 26 clinical FFPE samples of a variety of solid tumor types (lung, breast, colon, ovary, stomach, uterus and larynx) were used to evaluate the OCP performance. The analysis of the sequencing data was primarily performed with the OCP pipeline integrated with Oncomine® Knowledgebase from Life Technologies, supplemented by the MolecularMD proprietary pipeline. Only the genetic alterations with clinical utility were selected as the final output from the data pipeline. Of these detected by OCP, the SBS and indels were confirmed by the Illumina TruSeq, Ion Torrent AmpliSeq commercial cancer panels, or by Sanger sequencing. Any CNVs detected were confirmed by FISH, if possible, and detected fusions were confirmed by RT-PCR or FISH. Results: No clinically relevant genetic alterations were detected from the negative control FFPE-GM12878, indicating the high specificity of the OCP. High specificity was achieved, in part, using stringent filters, which removed error prone regions from analysis. The LOD for SBS and short indel detection was 5%, as assessed by TruQ-1 and TruQ-2, each containing 15 variants. While the exact LOD for gene fusion detection is currently under evaluation, the 100%, 50% and 20% fusion were detected in the RNA samples from serially diluted EML4-ALK or SL34A2-ROS1 fusion into the FFPE-GM12878 RNA. Four known gene amplifications (MYC, CCND1, MDM2, AKT1) with CNV from 5.4X to 14.5X in HCC1143, and a 15.7X MYC amplification in NCI-H2122 were detected. In addition, an EML4-ALK.E13A20 fusion was detected in a lung cancer FFPE sample, and confirmed by FISH. The analytical sensitivity and the specificity of the OCP are currently assessed with 26 clinical samples and three engineered FFPE samples. Conclusions: These results demonstrate the high sensitivity and specificity of OCP. With as little as 10 ng RNA and 20 ng DNA, OCP provides comprehensive screening of patient FFPE tumor specimens for the detection of a broad spectrum of clinically relevant genetic alterations. Citation Format: Peng Fang, Zhenyu Yan, Weihua Liu, Jennifer Biroschak, Paul Labrousse, Jennifer Wright, Cindy Spittle, Chad Galderisi, Li Jin. Oncomine® Cancer Panel: simultaneous detection of clinically relevant hotspot mutations, CNVs, and gene fusions in solid tumors. [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 4261. doi:10.1158/1538-7445.AM2015-4261

Introduction: Plasma cell-free circulating tumor DNA (ctDNA) has the potential to serve as a noninvasive surrogate to determine tumor genotype and monitor treatment response. Here we report the analytical validation of the Ion AmpliSeq v2 Cancer Panel (ACP) in plasma ctDNA samples. Methods: Mutations were assessed using the ACP, designed to survey 2800 mutations in 50 cancer-related genes. TruQ reference standards were used to determine LOD. Healthy donor and cancer patient plasma samples were used for other studies. Eight libraries were barcoded and sequenced with PGM sequencing 200 kit v2 on a 318 v2 chip. The sequencing data were analyzed with a combination of Torrent Suite 3.4.2, VarScan, our proprietary analysis pipeline and IGV. The background noise level of hotspot nt positions was calculated from multiple runs of known wild type DNA. In order to increase the sensitivity of the assay to detect low-level mutation, a custom program was developed to detect variants clearly different from background noise (95% confidence) and not reported by the two variant callers. PCR methods were used for concordance testing. Results: The performance characteristics of the AmpliSeq v2 Cancer Panel were demonstrated in 1) Minimum DNA input and LOD: With 10 ng DNA input, 43/43 and 42/43 expected mutations were detected at 2.5% or 1%, respectively. With 1 ng DNA input, 42/43 and 40/43 expected mutations were detected at 2.5% and 1%, respectively. 2) Coverage: The mean library coverage for 40 plasma libraries was 2912x with only 5 non-critical regions below 250x coverage. 3) Baseline error rate: 95% of 2290 hot-spots for SBS showed Conclusions: The ACP offers a useful tool for comprehensive somatic mutation profiling in ctDNA. The limit of detection is 1-2% mutation frequency for most mutations with slight variation across different nucleotide positions. The low concentration of ctDNA obtained from plasma samples can present a challenge. Less than 10 ng DNA input may result in allele drop out for low level mutations due to the inefficient amplification of multiplex PCR compared with PCR methods using a short length single amplicon. Higher DNA input to rescue allele drop-out is under evaluation. Citation Format: WeiHua Liu, Zhenyu Yan, Candice L. Horn, Fabio Nunes, Steven M. Bray, Philip J. Ebert, Peng Fang, Jennifer Biroschak, Cindy Spittle, Chad Galderisi, Jin Li. Liquid biopsy using the ion AmpliSeq v2 cancer panel. [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 4925. doi:10.1158/1538-7445.AM2015-4925

Treatment of chronic myeloid leukemia (CML) patients with tyrosine kinase inhibitors (TKI) has led to progressively lower levels of disease burden and higher rates of complete molecular responses. Very few assays with an analytical sensitivity of 4.5-logs (0.00316% BCR-ABL/ABL) have been validated, and assay standardization at these levels has been problematic (EUTOS, 2007). In order to have an International Scale (IS) standardized assay that can precisely and accurately detect 4.5 log reductions in BCR-ABL/ABL ratios, the test system design must be tightly controlled to ensure this critical performance metric can be reproducibly met. The MolecularMD MRDxTM BCR-ABL Test was developed and validated to be an accurate, reproducible, and highly sensitive IS-standardized solution to measure minimal residual disease (MRD) in CML patients using either PAXgene Blood RNA or EDTA blood collection tubes. The MRDx Test is a one-step quantitative real-time polymerase chain reaction (RT-qPCR) test that allows for quantitation of BCR-ABL e13/14a2 (b2/3a2) transcripts (covering ≥ 95% of CML patients) and ABL transcripts in RNA extracted from peripheral blood. To achieve a sensitivity of 4.5 logs, several one step enzyme systems were evaluated against RNA samples that spanned the potential dynamic range of the assay (MR1.0 to MR5.0: MR1.0=1 log molecular response (MR), or 10% IS; MR5.0=5 log MR, or 0.001% IS). The one step enzyme system with the best sensitivity and precision was chosen based on the performance of three distinct lots. In addition, in order to better control for the variation of the reverse transcription, in vitro transcribed RNA calibrators were developed to create a standard curve for copy number determination rather than the more commonly used plasmid calibrators that may not accurately reflect the same processing as the patient sample RNA. The limit of detection (LOD) for blood drawn into PAXgene Blood RNA tubes was validated by testing a dilution series created from a baseline CML patient blood sample diluted into non-diseased subject blood. Creating this type of dilution series allows for the determination of the LOD in the actual clinical sample matrix as compared using cell line dilution series that have much higher BCR-ABL and ABL copy numbers than routine patient specimens. Based on the analysis of each level of sample tested over a multi-day, multi-operator, and multi-instrument study, the LOD of the MRDx BCR-ABL Test was determined to be MR4.7 (Mean=MR4.9, 95% CI=MR5.2 to MR4.7) using the more conservative upper bound of the 95% confidence interval. The precision was evaluated based on the standard deviation of the log10 BCR-ABL/ABL ratio and was found to be ≤ 0.20 SD at MR5.0 and above. The LOD for blood drawn into EDTA blood collection tubes was validated by creation of a dilution series from a baseline CML patient RNA sample diluted into RNA isolated from blood of non-diseased subjects. Based on the number of samples with detectable BCR-ABL in at least 95% of replicates over a multi-day, multi-operator, and multi-instrument study, the LOD of the MRDx BCR-ABL Test was determined to be MR4.9 (Mean=MR5.0, 95% CI=MR5.1 to MR4.9) using the more conservative upper bound of the 95% confidence interval with precision being ≤ 0.22 SD at MR5.0 and above. The accuracy of the BCR-ABL copy numbers, ABL copy numbers, and BCR-ABL/ABL ratio using the MRDx BCR-ABL Test was evaluated for ten patient samples using one step droplet digital PCR (ddPCR) as a reference method. For the BCR-ABL copy numbers, ABL copy numbers, and the BCR-ABL/ABL ratio, the bias of the MRDx Test relative to ddPCR by Bland-Altman method was 0.053, 0.030, and 0.023 respectively. The MRDx BCR-ABL Test reports on the IS with a conversion factor of 1.0 by use of the WHO International Standard 1st WHO International Genetic Reference Panel for quantitation of BCR-ABL translocation by RQ-PCR; secondary standards were created for bi-annual monitoring of the assay conversion factor to ensure accurate reporting of patient results on the IS. Confirmation of the conversion factor was done by an independent laboratory using different real-time PCR instruments and multiple technicians. Based on the validation data, the MRDx BCR-ABL Test is an accurate, reproducible, and highly sensitive IS-standardized solution to the growing need for a reliable and robust quantitative BCR-ABL assay that can be used for the monitoring of minimal residual disease in CML patients. Disclosures: Toplin: MolecularMD: Employment. Drafahl:MolecularMD: Employment. Eibl:MolecularMD: Employment. Fjeld:MolecularMD: Employment. Yager:MolecularMD: Employment. McAdams:MolecularMD: Employment. Hawkins:MolecularMD: Employment. Galderisi:MolecularMD: Employment.

Background: Mucin-1 (MUC1, epithelial membrane antigen, EMA) is a transmembrane glycoporotein that is expressed on the apical membrane of epithelial cells of many tissues including breast, prostate, lung, pancreas, stomach, ovaries, intestines, and kidneys. In tumor cells, MUC1 is often overexpressed and aberrantly glycosylated, revealing new epitopes that can trigger a cytotoxic T-cell response. Several approaches are currently being pursued for targeting MUC1 in cancer therapy, mainly focused on vaccines targeting MUC1 antigens. We have developed an immunohistochemistry (IHC) assay to measure MUC1 expression in formalin fixed paraffin-embedded (FFPE) tissues as a potential biomarker assay for MUC1-targeted therapies. Methods: The IHC assays were developed using the Ventana Benchmark™. The MUC1 antibody was obtained from Ventana. The assay was validated using the recommended protocol supplied by the vendor. The assay was optimized using cell pellets generated from cell lines overexpressing MUC1 cDNA as well as an FFPE breast tumor tissue array which consists of 16 cases of breast cancer paired with adjacent normal tissues. In addition, tissue specimens obtained from remnant material from surgery or autopsy were used for the assay optimization and validation. Results: Specificity of the staining was demonstrated by positive staining with T47D breast cancer cell line as well as a B16 cell line overexpressing MUC1 and no staining with HMEC human mammary epithelial cells and the control B16 cell line. Thirty-six FFPE tumor tissues from lung, ovary, breast, colon, kidney, liver, thyroid, and prostate were evaluated for MUC1 expression. Tumor positive specimens were scored based on expression in at least 25% of tumor cells. In addition, staining intensities as well as patterns i.e. apical or membrane polarization and cytoplasmic patterns were noted. 29/36 specimens had positive MUC1 staining; negative staining was reported in thyroid, liver, colon, and prostate tumor tissues. In addition, evaluation of the array of paired normal/tumor breast tissues showed specific staining of tumor tissues in 14/16 specimens with no or little staining in the matched normal tissue. Interestingly, one of the tumor tissues exhibited very distinct apical staining pattern that was unique in the set of 16 breast cancer tissues. Conclusions: We have developed a robust IHC assay for MUC1 that can clearly distinguish MUC1 expression in normal versus tumor tissue as well as demonstrated variable MUC1 expression in a variety of tumor tissues. We are currently exploring the utility of the assay as a biomarker for ONT-10, a novel liposomal cancer vaccine that is currently in clinical trials. Citation Information: Mol Cancer Ther 2013;12(11 Suppl):B297. Citation Format: Sabita Sankar, Jennifer Wright, Kevin Klucher, Scott Peterson, Chad Galderisi. Clinical assessment of MUC1 protein expression in FFPE tissue: Developent and validation of an immunohistochemistry assay as a predictive assay for response to MUC1 vaccines. [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference: Molecular Targets and Cancer Therapeutics; 2013 Oct 19-23; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2013;12(11 Suppl):Abstract nr B297.

Background: Although epidermal growth factor receptor (EGFR) T790M mutation accounts for majority of the resistance to the first generation EGFR tyrosine kinase inhibitors (TKIs), activation of the MET pathway through increased expression of hepatocyte growth factor (HGF) or amplification of MET (the HGF receptor) has also been identified as a resistance mechanism in patients. Effective treatment of patients that have progressed on first-generation TKIs requires diagnostic assays that detect the major resistance mechanisms and direct therapy selection. We have developed immunohistochemistry (IHC) assays to detect both HGF and MET protein expression in formalin-fixed paraffin-embedded (FFPE) tissues, and evaluated expression levels in patients previously treated with EGFR TKIs or chemotherapy. Methods: The IHC assays were developed using the Ventana Benchmark automated staining system. The MET antibody (SP44) was obtained from Ventana. Five different HGF antibodies from several vendors were evaluated. The assay was established using cell pellets generated from cell lines overexpressing HGF as well as a FFPE metastatic cancer tissue array. In addition, tissue specimens obtained from remnant material from surgery or autopsy were used for the assay optimization. Results: The MET (SP44) assay was validated using the recommended protocol supplied by Ventana. Robust staining was observed in multiple FFPE samples, and both cytoplasmic and membranous staining were defined as positive. The expression was scored on the basis of intensity and fraction of positive cells. In contrast, among the 5 HGF antibodies tested for specificity, only the IBL antibody demonstrated HGF-specific staining in cell lines and the metastatic tissue array. The selected MET and HGF antibodies were used to evaluate a set of 25 NSCLC patient FFPE tissues. 19/25 specimens tested showed positive MET expression. Of these, 13 were 3+ whereas 6 were 2+ in terms of staining intensity. Additionally, 2 of the specimens with 2+ score exhibited predominantly cytoplasmic staining. In contrast, HGF expression was more variable with 5/24 with strong, 9/24 with weak and 10/24 with negative staining (one specimen was not evaluable for HGF). There was no clear correlation between MET and HGF staining patterns. In this sample set, there was not a clear correlation between MET or HGF expression and patient status as relapsed versus newly diagnosed. Conclusions: We have developed robust assays for evaluation of MET and HGF expression in FFPE sections and applied these assays in analyzing 25 NSCLC patient specimens, documenting a higher than expected proportion of the samples as MET or HGF positive. A more expansive dataset, potentially evaluating factors beyond protein expression levels, will be needed to elucidate a relationship between MET pathway activation and response to EGFR TKIs. Citation Information: Mol Cancer Ther 2013;12(11 Suppl):A42. Citation Format: Sabita Sankar, Jennifer Wright, Chad Galderisi. Evaluation of HGF and MET protein expression in NSCLC tumor specimens from patients previously treated with targeted or chemotherapies. [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference: Molecular Targets and Cancer Therapeutics; 2013 Oct 19-23; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2013;12(11 Suppl):Abstract nr A42.

Introduction: LKB1 (liver kinase B1)/STK11 (serine-threonine kinase 11) is a tumor suppressor that encodes a serine/threonine kinase which negatively regulates the mTOR (mammalian target of rapamycin) signaling pathway. Somatic mutations in LKB1 occur most notably in lung adenocarcinoma, cervical cancers and melanoma. Inactivation of LKB1 is caused by point mutations, homozygous deletions or promoter methylation. LKB1 mutation status is a predictive marker for responsiveness to both MEK and PI3K inhibitors. We have developed next-generation sequencing (NGS), Sanger sequencing and immunohistochemistry (IHC) assays to assess LKB1 mutation and protein expression status in a set of FFPE patient specimens. Methods: For NGS, we used the Ion Torrent platform to sequence a 4.4kb region of interest (ROI) around LKB1. A combination of TorrentSuite and in-house tools were used for alignment, variant calling and prioritization of variants. Sanger sequencing was used to validate the presence and absence of variants with a frequency of at least 10% that were identified by NGS. An IHC assay was developed using a commercially available antibody. These assays were validated using a combination of cell lines, a genetically engineered mouse model and FFPE tissue for specificity, sensitivity, reproducibility and concordance. Results: 35 patient FFPE tissues from a variety of tumors including cervix, endometrium, pancreas, skin and lung were evaluated for LKB1 mutation status using NGS and IHC methods. About 96% of the coding region of LKB1 had sufficient sequencing coverage (read depth >500x) to reliably call low frequency somatic variants. Analytical validation using cell line dilutions demonstrated that our NGS assay is able to detect single base changes with frequencies as low as 2-5%, and small indels with frequencies as low as 5-10%. We identified 9 coding and splice site variants in our preliminary analysis of 18/35 of the FFPE specimens. We compared the variants in 7 of these FFPE samples to the variants identified in these samples by Sanger sequencing, and found 100% concordance between the variant calls, for NGS variants with frequencies greater than 20%. Sanger sequencing in the remaining FFPE samples is ongoing, and lower frequency NGS variants will be confirmed by other methods. We compared the validated NGS variants with the IHC results to examine the correlation of sequencing results with protein expression analysis. In our preliminary analysis of the first 18 FFPE samples, all of the samples with an IHC score of zero or one (IHC negative) had either a nonsense (n=1) or missense (n=4) variant in STK11. Conclusions: Our data suggest that accurate assessment of LKB1 status may require complementary methods including NGS and IHC methods. Citation Format: Kimberly Pelak, Jennifer Wright, Zhenyu Yan, Agus Darwanto, Weihua Liu, Peng Fang, Jin Li, Sabita Sankar, Chad Galderisi. Complementary analysis of LKB1/STK11 mutation and protein expression status using next-generation sequencing, Sanger sequencing and immunohistochemistry. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 3509. doi:10.1158/1538-7445.AM2013-3509

Introduction: The TruSeq Amplicon Cancer Panel (TSACP), a highly multiplexed targeted resequencing assay for use on the Illumina MiSeq platform, is designed for detecting the hotspot mutations in 212 Regions of Interest (ROI) from 48 cancer related genes. Here we report our validation study on the reproducibility, sensitivity and the specificity of detecting single base substitutions and small indels by the TSACP. Methods: We used well characterized cancer cell lines harboring clinically relevant variants as positive controls and HapMap samples NA12878 and NA19240 as wild type control samples. DNA from 8 cancer cell lines was serially diluted into the control DNA NA12878 for validating the Limit of Detection (LOD) of the TSACP assay. A total of 41 FFPE patient specimens representing a variety of cancer types were analyzed in a blinded fashion to evaluate the analytical sensitivity and specificity. DNA quality was assessed using a qPCR assay. With no gold standard available as a reference method to detect mutations with comparable sensitivity, concordance testing was performed using the Ion AmpliSeq Cancer Panel. Variants detected by both panels were considered as true positives. Variants that were only covered by one of the two panels were confirmed by a third method, either Sanger sequencing for variants with frequencies above 10% or a custom Ion TargetSeq Assay for variants with frequencies below 10%. Data were analyzed using MiSeq Reporter software and our proprietary analysis pipeline. In addition to reporting hotspots mutations, we also report “Critical Variants” such as non-synonymous coding mutations and splicing site mutations that fall within the ROI. Results: 95% of our ROIs were sequenced at minimum of 0.2X normalized coverage. A cell line dilution study showed that the LOD of confirmed variants is 5%. DNA extracted from 4 of the 41 FFPE specimens failed the template QC by qPCR and failed in the subsequent sequencing run. A total of 124 unique critical variants, including single base substitution, single- or multi- base (up to 21bp) deletion, one- or two- base insertion, were identified in the cancer cell lines and 37 qualified FFPE samples. The intra-assay and inter-assay reproducibility was ≥96%. Using our proprietary analysis pipeline, the analytical sensitivity and specificity for the FFPE samples were both 99%. One false negative of TSACP was identified by the Ion AmpliSeq Cancer Panel and was further confirmed by Sanger sequencing. This allele drop-off occurred as a result of the capture probe falling on a SNP. Homopolymer indels in KIT and STK11 were accurately identified with the Illumina sequencing chemistry. Conclusions: These studies demonstrate that the TSACP assay is highly specific and sensitive and is suitable for screening patient FFPE tumor specimens for a spectrum of clinically relevant somatic mutations. Citation Format: Peng Fang, Zhenyu Yan, Weihua Liu, Agus Darwanto, Kim Pelak, Kim Anoe, Cynthia Spittle, Sabita Sankar, Chad Galderisi, Jin Li. Validation of Illumina TruSeq Amplicon Cancer Panel with concordance testing using Ion AmpliSeq Cancer Panel and other methods. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 3229. doi:10.1158/1538-7445.AM2013-3229

Introduction: EGFR T790M mutation leads to treatment resistance in ∼50% of NSCLC patient undergoing TKI treatment. Early detection of the emergence of this resistance mutation allows for tailoring of the treatment regimen. Detection of EGFR T790M requires higher sensitivity techniques than those needed to detect EGFR activating mutations, and quantitative measurement of T790M may also provide value in monitoring disease progression. Here we demonstrate that our proprietary amplicon next-generation sequencing (NGS) on Ion Torrent PGM provides an attractive solution for T790M detection with the advantage of high sensitivity, specificity, and quantification capability. Methods: We developed a proprietary library prep method for amplicon sequencing on the Ion Torrent PGM. Ion Torrent barcode sequences were refined in our validation to minimize barcode cross contamination. This optimization of barcodes allows for accurate mutation quantification, especially critical for variants with a frequency below 1%. This also added the benefit of higher throughput and lower cost. We employ this non-enrichment method to quantitatively measure T790M. DNA from the NCI-H1975 cell line was serially diluted into wild-type NA19240 DNA to determine the limit of detection (LOD), sensitivity, specificity, accuracy and reproducibility of the method. We also used reference FFPE DNA from HorizonDx to validate the accuracy and robustness of the assay. We further validated the test by comparing the performance of our NGS method with ddPCR using the same sample set (serially diluted H1975 into wild-type DNA controls). Results: Our data demonstrated that the LOD for T790M is 0.2%. And the method could reproducibly identify variants at these frequencies. With an input of 10ng of the HorizonDx FFPE DNA, we measured T790M frequency at 6.7%, which is concordant with the 6.5% mutation frequency reported by HorizonDx. We cross-validated our NGS assay with ddPCR, and the mutation frequencies detected by both platforms are nearly identical, and have a regression coefficient of 0.9995, with comparable LOD of 0.2%. Furthermore, our NGS assay allows us to detect other variants located in this amplicon, e.g. SNP Q787Q, with LOD at 1%. The phasing status of compound mutations can also be determined by this assay. To further improve the sensitivity, an enrichment method was developed, and the enriched T790M was sequenced by PGM. This method can semi-quantitatively measure T790M frequency with LOD of 0.03%. Conclusion: Our EGFR T790M NGS assay provides a unique option to reliably quantify T790M down to 0.2% with low DNA input. This highly sensitive and specific detection capability may enable earlier detection of emerging therapeutic resistance, particularly if FFPE results reported here extend to a more accessible specimen type, i.e. plasma circulating DNA, amenable to periodic patient monitoring. Citation Format: Agus Darwanto, Peng Fang, Zhenyu Yan, Weihua Liu, Kim Pelak, Jessica Kristof, Sabita Sankar, Cynthia Spittle, Chad Galderisi, Jin Li. Highly sensitive detection of EGFR T790M on Ion Torrent PGM. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 3490. doi:10.1158/1538-7445.AM2013-3490

Background: PTEN is a tumor suppressor that negatively regulates the PI3K signaling pathway by dephosphorylating PIP3, converting it to PIP2. PTEN loss of function results in activation of Akt and downstream signaling pathways, and has been associated with numerous cancers, most commonly endometrial cancer, glioblastoma, melanoma and prostate cancer. PTEN loss can be a result of multiple different aberrations which include mutations in the coding region leading to frame-shift or stop codons, genomic deletions as well as promoter methylation leading to loss of PTEN protein. Determination of PTEN status, both at the DNA and protein levels, may be important for clinical decision-making because loss of PTEN is associated with resistance to various targeted therapies, while inactivating mutations may confer sensitivity to therapeutics targeting PI3K pathway members. We have developed Sanger sequencing, immunohistochemistry (IHC) and chromogenic in situ hybridization (CISH) assays to assess PTEN mutation status in a set of FFPE patient specimens. Methods: Our Sanger sequencing assay for PTEN was developed to detect known hotspot mutations in exons 5-8, which occur in the majority of PTEN mutations reported in the COSMIC database. The IHC and CISH assays were developed using commercially available reagents. All assays were validated using a combination of plasmids, cell lines and FFPE tissues for specificity, sensitivity, reproducibility and concordance. Results: 22 patient FFPE tissue specimens from a variety of tumor types including colon, lung, pancreas, uterus and prostate were evaluated for PTEN mutation and expression status by both Sanger sequencing and IHC methods. Sequencing results were: 9 wild-type, 12 heterozygous mutant, and 1 hemizygous mutant. 7 of the 9 specimens reported as wild-type by sequencing exhibited normal protein expression levels by IHC, while 2 were PTEN negative by IHC, suggesting that mechanisms other than mutations in exons 5-8 contributed to the loss of expression of PTEN protein. These 2 specimens were further analyzed by CISH to determine copy number. Of the 13 specimens reported as mutant by sequencing, IHC protein expression results indicated that 6 were PTEN positive and 7 were PTEN negative, underscoring the observation that exon 5-8 mutation status alone is not sufficient to predict PTEN expression levels. Data on CISH analysis of the specimen that was hemizygous mutant by sequencing will be reported to confirm the copy loss. Conclusion: The data suggest that additional mutations outside the exon 5-8 region as well as promoter methylation may result in loss of PTEN protein, and that accurate clinical assessment of PTEN mutation and protein expression status in FFPE tissues requires complimentary methods including Sanger sequencing, IHC and CISH methodologies. Citation Format: Kellen Sakrison, Jennifer Wright, Eric Bruening, Stephane Wong, Cynthia Spittle, Sabita Sankar, Chad Galderisi. Clinical assessment of PTEN mutation in FFPE tissue: comparison of Sanger sequencing, immunohistochemistry and chromogenic in situ hybridization methods. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 39. doi:10.1158/1538-7445.AM2013-39

Abstract 2549 Background: BCR-ABL1 mutation testing is recommended for CML and Ph+ ALL patients who fail first line tyrosine kinase inhibitor (TKI) therapy or who have a suboptimal response to therapy. BCR-ABL1 mutations in the kinase domain (KD) of ABL1 account for at least 40–50% of all TKI resistant cases. Rare mutations such as E123Q and T212R in the regulatory domain of ABL upstream of the kinase domain have also been reported to lead to resistance to imatinib. The current gold standard for BCR-ABL1 mutation detection is Sanger sequencing, which has an analytical sensitivity of ∼10–30%. Based on recent findings that mass spectrometry can identify low level BCR-ABL1 mutations that confer clinical resistance in patients sooner than Sanger sequencing, it is likely useful to have a significantly more sensitive BCR-ABL1 test than Sanger sequencing. Although commercial NGS cancer panels have included ABL1 in the region of interest, ABL1 resistance mutations should be sequenced from BCR-ABL1 fusion transcripts instead of being sequenced from genomic DNA as in the commercial panels. Here we developed a fusion transcript based BCR-ABL1 mutation assay on the scalable and cost-effective Ion Torrent platform that has 1–5% sensitivity and comprehensive coverage of the kinase domain, regulatory domain, and the SH2/SH3 domains. The assay was designed to detect both the major and minor BCR-ABL1 fusion gene products and can also detect the micro BCR-ABL1 fusion product accounting for over 99% of all CML and Ph+ ALL patients. Methods: RT and long range PCR was performed to amplify BCR-ABL1 e1, e13, and e14 fusion transcripts and the PCR products were enzymatically fragmented and ligated with Ion Torrent sequencing adaptors. Size-selected libraries were quantified, pooled, amplified with OneTouch system and sequenced with Ion Torrent PGM. Sequencing data was analyzed with Torrent Suite 2.2 and the associated variant caller with variant frequency cutoff adjusted to 1%. Results: Initial work with cell lines harboring the T315I mutation in both e1 and e14 BCR-ABL1 transcript types diluted into wild type cell line demonstrated that Ion Torrent NGS can detect T315I at least down to 1%. In a set of 17 blinded clinical samples, Ion Torrent NGS not only identified all the mutations found by Sanger sequencing but additionally identified rare imatinib resistant mutations such as K357E (MMD-3) present at 7.6%; this mutation was previously reported in patient CD34+CD38-stem cells (Table 1). This patient also expressed the E255K resistance mutation; in patients, the presence of multiple resistance mutations has been shown to be an important predictor of poor response. Understanding whether compound mutations are present in cis or in trans may be important in understanding therapy resistance. For patient MMD-9, although both the predominant mutation G250E (79%) and subclone E255V(12.6%) were identified by both Sanger sequencing and NGS, only Ion Torrent was able to show that the two mutations were on different reads, indicating that the mutations are on different alleles. The BCR-ABL1 Ion Torrent based assay reads DNA fragments between 150–200bps in size and can identify cis and trans mutations from individual fragments with this read length, which is not possible by Sanger sequencing. In MMD-9, another low frequency subclone F359C (3%) was detected by only NGS and may have important implications due to its reported sensitivity to dasatinib and not imatinib or nilotinib. Similarly, in MMD-10 the E255V predominant mutation was identified by both NGS and Sanger sequencing, while the F317L mutation with a low frequency of 1.70% was solely detected by NGS. In this patient, a combination of dasatinib and nilotinib treatment may be required to eliminate the dasatinib sensitive E255V dominant clone and F317L nilotinib sensitive subclone. Ion Torrent NGS was also capable of identifying and calling the 35 base pair 475 insertion from 4 samples (MMD-2 and 8–10), only two of which were detected by Sanger sequencing. In our ongoing study, the low-level mutations not detected in Sanger sequencing will be confirmed with analyses on the MiSeq platform and mutation enrichment methods. Additional patient samples representing 30 CML patients will be analyzed and presented. Disclosures: Wong: MolecularMD: Equity Ownership.

Introduction: Major Molecular Response (MMR) is defined as a three-log reduction from a standardized baseline of BCR-ABL/control gene transcript ratio in CML patients at diagnosis. MMR has prognostic significance for progression-free survival for patients on Imatinib® therapy. Day-to-day monitoring of the MMR value in clinical laboratories is challenging due to the absence of a commercially available standardized MMR control RNA. To improve the reliability of BCR-ABL quantitation, MolecularMD has evaluated the feasibility of a single MMR control RNA valid for blood samples drawn in EDTA or PAXgene™ tubes. Material and Methods: Patient sample RNAs were interchanged between our laboratory and an International Randomized Interferon versus STI571 study (IRIS) laboratory, which had established an MMR value and international scale reporting. This exchange enabled our laboratory to establish an MMR value and reporting on an international scale using a validated conversion factor. A serial dilution of a BCR-ABL positive cell line into a human BCR-ABL negative cell line was prepared. These dilutions were tested in IRIS laboratories with established MMR value and international scale reporting and at our laboratory by QRT-PCR to determine the BCR-ABL/control gene ratio using respectively BCR and ABL control genes. We compared the BCR-ABL/ABL ratio in 104 paired PB CML patient samples drawn either in EDTA and PAXgene tubes and the BCR-ABL/BCR ratio in 32 patient samples. Stability studies were performed to evaluate the degradation of liquid and dried forms of the MMR RNA. Results: We established a conversion factor (CF) of 0.81 with an MMR value of 0.123%. Using this CF and MMR value, we created appropriate RNA dilutions that matched the MMR value using ABL as a control gene. Repeated analyzes of this MMR control RNA confirmed the accuracy of the sample with a median value of 0.124%, very close to the MMR value defined previously (0.123%). Stability studies demonstrated that the dried RNA samples could be stored several days at 37°C and freeze-thawed up-to 10 times without significant degradation. These RNA samples once reconstituted with water could also be used several times for BCR-ABL monitoring without any significant degradation. Comparison of BCR-ABL/ABL ratio between EDTA and PAXgene tubes revealed differences unlikely to have clinical impact on disease management suggesting that the MMR RNA created would be suitable under both EDTA and PAXgene extraction methodologies. Conclusions: We produced a stable MMR control RNA in large quantity for accurate monitoring of the MMR value. This MMR control RNA is now be tested in several laboratories to confirm the stability and reliability of this reagent. The MMR control RNA will be an important tool for standardizing MMR value in laboratories, and an integral part of a BCR-ABL QRT-PCR diagnostic kit.