Search

Appendicitis is one of the most common surgically treated diseases in the world. CT scans are often over-utilized and ordered before a surgeon has evaluated the patient. Our aim was to develop a tool using machine learning (ML) algorithms that would help determine if there would be benefit in obtaining a CT scan prior to surgeon consultation. Retrospective chart review of 100 randomly selected cases who underwent appendectomy and 100 randomly selected controls was completed. Variables included components of the patient’s history, laboratory values, CT readings, and pathology. Pathology was used as the gold standard for appendicitis diagnosis. All variables were then used to build the ML algorithms. Random Forest (RF), Support Vector Machine (SVM), and Bayesian Network Classifiers (BNC) models with and without CT scan results were trained and compared to CT scan results alone and the Alvarado score using area under the Receiver Operator Curve (ROC), sensitivity, and specificity measures as well as calibration indices from 500 bootstrapped samples. Among the cases that underwent appendectomy, 88% had pathology-confirmed appendicitis. All the ML algorithms had better sensitivity, specificity, and ROC than the Alvarado score. SVM with and without CT had the best indices and could predict if imaging would aid in appendicitis diagnosis. This study demonstrated that SVM with and without CT results can be used for selective imaging in the diagnosis of appendicitis. This study serves as the initial step and proof-of-concept to externally validate these results with larger and more diverse patient population.

Appendicitis is one of the most common surgically treated diseases in the world. CT scans are often over-utilized and ordered before a surgeon has evaluated the patient. Our aim was to develop a tool using machine learning (ML) algorithms that would help determine if there would be benefit in obtaining a CT scan prior to surgeon consultation.Retrospective chart review of 100 randomly selected cases who underwent appendectomy and 100 randomly selected controls was completed. Variables included components of the patient's history, laboratory values, CT readings, and pathology. Pathology was used as the gold standard for appendicitis diagnosis. All variables were then used to build the ML algorithms. Random Forest (RF), Support Vector Machine (SVM), and Bayesian Network Classifiers (BNC) models with and without CT scan results were trained and compared to CT scan results alone and the Alvarado score using area under the Receiver Operator Curve (ROC), sensitivity, and specificity measures as well as calibration indices from 500 bootstrapped samples.Among the cases that underwent appendectomy, 88% had pathology-confirmed appendicitis. All the ML algorithms had better sensitivity, specificity, and ROC than the Alvarado score. SVM with and without CT had the best indices and could predict if imaging would aid in appendicitis diagnosis.This study demonstrated that SVM with and without CT results can be used for selective imaging in the diagnosis of appendicitis. This study serves as the initial step and proof-of-concept to externally validate these results with larger and more diverse patient population.

Background: Liquid biopsy testing has demonstrated the utility of capturing the heterogeneity of solid tumors and relevant actionable mutations using a less invasive sample collection method. The Genexus™ Integrated Sequencer provides a viable pathway to deploy an automated platform for a sensitive pan-cancer liquid biopsy assay to meet the needs of quick turn-around and cost effectiveness for multi-site implementation. The Oncomine™ Precision Assay GX (OPAGX) is an amplicon-based next-generation sequencing (NGS) assay that enables detection of biomarkers across 50 genes from both solid tissue and liquid biopsy samples. OPAGX uses AmpliSeq™ HD technology which includes unique molecular tags for sequencing error correction and increased sensitivity. The assay’s detection of hotspot mutations was evaluated for liquid biopsy samples. Methods: Automated library preparation was completed on the Genexus™ Integrated Sequencer using 20 ng of cfTNA derived from plasma or from well characterized control samples (Acrometrix, Agilent, Horizon, and SeraSeq). After reverse transcription was completed, gene targets of interest were amplified from DNA and cDNA. Amplicons were labeled with dual molecular barcodes during library preparation. Subsequently, libraries were bead normalized, pooled, and sequenced on a GX5™ chip to a targeted depth of 12,000,000 reads per sample. Sensitivity and positive predictive value were examined by evaluating the correctness of calling true positive variants compared to false negative and false positive variant calls, respectively. Precision and limit of detection (LoD) were examined by evaluating the concordance of variants across replicate samples with the range of AF between 0.1% - 35% for SNVs and 0.2% - 15% for insertions and deletions. Specificity and accuracy were examined by evaluating the correctness of calling true negative variants compared to false positive and all other variant calls, respectively. Results: Accuracy was 98% for SNVs and 100% for insertions and deletions above 0.25% allele frequency (AF). Sensitivity was 95% for SNVs and 81% for insertions and deletions above 0.5% AF. Specificity and PPV were 100% for SNVs and insertions and deletions above 0.1% AF. Intra-run precision was 95% for SNVs and 92% for insertions and deletions above 0.5% and 0.25% AF, respectively. Inter-run precision was 93% for SNVs and 76% for insertions and deletions above 0.5% and 0.25% AF, respectively. The estimated AF LoD was 0.6% for SNVs, which is slightly above the advertised 0.5% provided by the vendor. Additionally, CNV gain calls were 100% for ERBB2 above 0.25% AF, MET above 0.5% AF, and FGFR3 at all AFs. Conclusions: The demonstrated performance of OPAGX presents a viable liquid biopsy platform which lays the foundation for global standardization of genomic profiling liquid biopsy samples with minimal nucleic acid input. Citation Format: Lauren Frady, Desiree Unselt, John Pufky, Fernando Torres, Jennifer Sims, Jennifer Mason. Automated liquid biopsy pan-cancer genomic profiling panel demonstrates utility for screening of targeted biomarkers [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 528.

Background: The rapid spread of COVID-19 has resulted in an urgent need for effective diagnostic and therapeutic strategies against SARS-CoV-2. Next-generation sequencing (NGS) is a powerful tool in the identification and characterization of this pathogen and genomic information may aid in understanding the mechanisms of therapeutic resistance, vaccine escape, virulence, and pathogenicity. The Ion AmpliSeq SARS-CoV-2 Research Panel is a targeted NGS solution that facilitates sequence analysis of the SARS-CoV-2 genome. Paired with a bioinformatics assembly and variant calling pipelines, this assay allows for accurate characterization of the dominant SARS-CoV-2 variant. This assay’s performance was analytically validated for the detection of mutations (substitutions, insertions, and deletions) in RNA derived from nasopharyngeal (NP) swabs. Method: The Ion AmpliSeq SARS-CoV-2 Research panel consists of two primer pair pools generating 237 amplicons specific to the SARS-CoV-2 virus. Reverse transcription of the RNA was performed using the SuperScript VILO cDNA Synthesis kit. Library preparation was then completed using the Ion AmpliSeq Library Kit Plus kit. The final library was quantified, normalized, pooled, and sequenced. Raw sequencing data was aligned to the AmpliSeq SARS-CoV-2 Research panel, using the MN908947.3 reference genome. Variants were called using the Torrent Variant Caller and annotated using the COVID19AnnotateSnpEff plugin. The reference-guided iterative assembler IRMA was used to produce a single consensus sequence consisting of the reference genome sequence modified to include sequence variations supported by the reads. The Pangolin COVID-19 lineage assigner software tool was used to assign SARS-CoV-2 lineage. Analytical validation was completed using controls (Twist Biosciences, BEI Resources, ATCC) and RNA derived from NP swabs. Accuracy and specificity were examined by evaluating the correctness of calling true negative variants compared to false positive and all other variant calls, respectively. Precision and limit of detection (LoD) were examined by evaluating the concordance of variants across replicate samples. Limit of Blank (LoB) was calculated as the 95th percentile of reads per amplicon in the negative samples. Results: Accuracy of base calling, specificity, and precision were 100% for SNVs, insertions, and deletions above 25% allele frequency. LoD was determined to be 576 viral copies/mL. LoB was determined to be 202 reads per amplicon. Pangolin lineage assignment was 100% for all samples. Conclusions: This panel accurately characterizes SARS-CoV-2 variants, allowing for accurate consensus sequence generation, mutation annotation, and lineage assignment. Citation Format: Desiree Unselt, Katherine Knudsen, Christopher Rounds, Janet Doolittle-Hall, Fernando Torres, Jennifer Sims, Jennifer Mason. Characterization of SARS-CoV-2 using the Ion AmpliSeq SARS-CoV-2 research panel [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 437.

Background: Genomic profiling of a tumor to characterize clinically relevant variants has become critically important for cancer diagnosis and prognosis, as well as in driving therapeutic decision making. As such, there is a growing need for rapid inline assessment of somatic variants in genes of key interest in oncology. The Oncomine™ Precision Assay on the Ion Torrent Genexus Sequencer (OPAGX) is an amplicon-based automated next-generation sequencing (NGS) assay that enables quick turnaround genomic profiling of biomarkers across 50 genes from solid tissue. The complete workflow, including library prep, templating, sequencing, and final report generation, can be completed in less than 30 hours for up to 32 formalin-fixed, paraffin-embedded (FFPE) samples. This assay uses AmpliSeq™ HD technology which includes unique molecular tags for sequencing error correction and increased sensitivity for low frequency variant detection. The assay’s detection of hotspot mutations (substitutions, insertions, deletions, copy number variants (CNVs), and fusions) was evaluated for FFPE tissue. Method: Fully automated library preparation and sequencing was completed in the Genexus™ Integrated Sequencer using 10 ng of DNA derived from FFPE tissue or from characterized control samples. Following gene specific amplification, AmpliSeq™ HD technology is used to label amplicons with dual molecular barcodes. The resultant libraries are bead normalized, pooled, and sequenced on a GX5™ chip to a targeted depth of 3,000,000 reads per sample. Sensitivity and positive predictive value were examined by evaluating the correctness of calling true positive variants compared to false negative and false positive variant calls, respectively. Precision and limit of detection (LoD) were examined by evaluating the concordance of variants across replicate control samples. Specificity and accuracy were examined by evaluating the correctness of calling true negative variants compared to false positive and all other variant calls, respectively. Results: Accuracy and specificity were 99.9% for SNVs and 100% for insertions and deletions above 1% allele frequency (AF). Sensitivity and intra-run precision were 100% for SNVs and insertions and deletions above 1% and 2% AF, respectively. Inter-run precision and PPV were 100% for SNVs and insertions and deletions above 1% AF. The estimated AF LoD was 3.7% for SNVs. CNV gain calls were 100% for MET and EGFR and loss calls were 100% for PTEN (homozygous and heterozygous) and CDKN2A. Additionally, fusion calling for RET and ROS1 were 100%. Conclusions: OPAGX is a quick turnaround NGS assay enabling precision medicine through automated multi-variant genomic profiling to inform decision making; performance was characterized with analysis of hotspot mutations in FFPE samples in a variety of tumor types, demonstrating utility of implementation in testing of targeted biomarkers across multiple indications. Citation Format: Desiree Unselt, Lauren Frady, John Pufky, Fernando Torres, Jennifer Sims, Jennifer Mason. Quick turnaround NGS assay for the characterization of biomarkers in FFPE tissue [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 2290.

Pepper (Capsicum annuum L.) is an economically important crop with added nutritional value. Production of capsaicin is an important quantitative trait with high environmental variance, so the development of markers regulating capsaicinoid accumulation is important for pepper breeding programs. In this study, we performed association mapping at the gene level to identify single nucleotide polymorphisms (SNPs) associated with capsaicin pathway metabolites in a diverse Capsicum annuum collection during two seasons. The genes Pun1, CCR, KAS and HCT were sequenced and matched with the whole-genome sequence draft of pepper to identify SNP locations and for further characterization. The identified SNPs for each gene underwent candidate gene association mapping. Association mapping results revealed Pun1 as a key regulator of major metabolites in the capsaicin pathway mainly affecting capsaicinoids and precursors for acyl moieties of capsaicinoids. Six different SNPs in the promoter sequence of Pun1 were found associated with capsaicin in plants from both seasons. Our results support that CCR is an important control point for the flux of p-coumaric acid to specific biosynthesis pathways. KAS was found to regulate the major precursors for acyl moieties of capsaicinoids and may play a key role in capsaicinoid production. Candidate gene association mapping of Pun1 suggested that the accumulation of capsaicinoids depends on the expression of Pun1, as revealed by the most important associated SNPs found in the promoter region of Pun1.