Module 1: Clinical Cancer NGS & Assay Strategy: Targeted Panels, CGP, WES & Liquid Biopsy
1. The Clinical Paradigm Shift: From Single-Gene Assays to Comprehensive Genomic Profiling
Precision oncology relies fundamentally on identifying genomic alterations that drive tumor proliferation, confer therapeutic sensitivity, or mediate secondary resistance. Historically, clinical diagnostics depended on isolated single-gene assays—such as Sanger sequencing for BRAF V600E, real-time PCR for EGFR exon 19 deletions, or fluorescence in situ hybridization (FISH) for ALK translocations. While specific, sequential single-gene testing rapidly exhausts minute clinical biopsy specimens, prolongs turnaround times, and fails to evaluate multi-gene therapeutic signatures.
Next-Generation Sequencing (NGS) has consolidated these disparate tests into single, unified assays. Modern oncology testing spans four principal computational and assay architectures:
| Assay Strategy | Typical Genomic Target Size | Sequencing Depth & Profile | Primary Clinical Utilities | Diagnostic Trade-offs & Limitations |
|---|---|---|---|---|
| Targeted Hotspot Panels | 10 – 50 genes (20 – 150 kb) | Ultra-deep: $1,000\times - 5,000\times$ | Rapid turnaround (3–5 days), low input DNA ($5-10\text{ ng}$), high sensitivity for common hotspots (KRAS, EGFR, BRAF). | Cannot detect novel structural variants, gene fusions, Tumor Mutational Burden (TMB), or large copy-number deletions. |
| Comprehensive Genomic Profiling (CGP) | 300 – 600 genes (1.0 – 2.5 Mb) | Deep hybrid capture: $500\times - 1,000\times$ | Current Standard of Care: simultaneous detection of SNVs, indels, CNVs (ERBB2, MET), fusions (ALK, ROS1, NTRK), TMB, and MSI. | Requires higher DNA/RNA input ($50-100\text{ ng}$), higher sequencing cost, complex bioinformatics deconvolution. |
| Whole Exome Sequencing (WES) | ~20,000 genes (~35 Mb) | $150\times - 300\times$ (tumor-normal pair) | Research discovery, comprehensive neoantigen prediction, whole-genome copy number and mutational signatures. | Lower clinical depth across difficult GC-rich exons; generates large numbers of Variants of Uncertain Significance (VUS). |
| Cell-Free DNA (cfDNA) Liquid Biopsy | Custom targeted panels (30 – 100 genes) | Ultra-deep with UMIs: $> 10,000\times$ raw depth | Non-invasive monitoring of treatment response, early detection of acquired resistance (EGFR T790M/C797S), Minimal Residual Disease (MRD). | Fragmented DNA (~166 bp), low mutant allele fraction ($ ext{VAF} < 0.1\%$), confounding clonal hematopoiesis of indeterminate potential (CHIP). |
2. Somatic vs. Germline Alterations & Actionability
A crucial conceptual cornerstone in molecular oncology is distinguishing somatic variants from constitutional germline variants:
- Somatic Variants: Acquired mutational events present exclusively within the neoplastic tissue. They dictate immediate targeted therapy selection (e.g. Osimertinib for EGFR L858R) and are not inherited by offspring.
- Germline Variants: Constitutional variations present in all diploid cells of the patient, inherited through the gametes. Identified germline mutations in high-penetrance cancer predisposition genes (e.g. BRCA1, BRCA2, PALB2, MLH1, MSH2) mandate specialized genetic counseling, cascade family testing, and heightened cancer surveillance.
When normal matched tissue (peripheral blood or buccal swab) is unavailable, tumor-only sequencing is performed. In tumor-only testing, variants exhibiting ~50% or ~100% Variant Allele Frequency (VAF) may represent either a true somatic driver that underwent Loss of Heterozygosity (LOH) OR an incidental pathogenic germline finding. Clinical molecular reports must state whether normal tissue was co-sequenced to avoid misinterpreting inherited cancer syndromes.
3. Introducing the Three Longitudinal Case Study Patients
Throughout this training program, you will guide three real-world cancer patients from raw FASTQ reads to final Molecular Tumor Board (MTB) presentation:
| Patient Case | Clinical Demographics & Diagnosis | Sample Specimen Type | Primary Diagnostic Question |
|---|---|---|---|
| Case 1: Lung Adenocarcinoma | 62-year-old female, never-smoker, Stage IV non-small cell lung cancer (NSCLC) with bone metastases. | Core needle pleural biopsy (FFPE), low tumor purity (~30%). | Identify sensitizing EGFR, ALK, ROS1, or BRAF driver alterations for first-line targeted therapy. |
| Case 2: Colorectal Carcinoma | 58-year-old male, metastatic colorectal adenocarcinoma (mCRC) with liver lesions. | Surgical resection specimen (FFPE), high tumor content (~70%). | Determine RAS/RAF status (anti-EGFR therapy eligibility) and MSI/TMB status (immune checkpoint inhibitor eligibility). |
| Case 3: Infiltrating Ductal Breast Cancer | 49-year-old premenopausal female, triple-negative phenotype by IHC, family history of ovarian cancer. | Diagnostic excision biopsy and matched peripheral blood mononuclear cells. | Assess homologous recombination deficiency (HRD), BRCA1/2 germline/somatic status, and PI3K/AKT pathway alterations. |
Common Pitfalls & Troubleshooting Checklist
- Java Heap Exhaustion (GATK / Picard): If processes terminate with
java.lang.OutOfMemoryError, set your environment variable explicitly:export _JAVA_OPTIONS="-Xmx16g -Djava.io.tmpdir=/tmp". - Index Timestamp Mismatch: If tools complain that a BAM index is missing or out of date, regenerate with
samtools index -@ 4 sample.bam. - Conda Channel Conflicts: Avoid ABI incompatibilities between bioconda and conda-forge by setting strict priority:
conda config --set channel_priority strict. - Chromosome Naming Contamination: Always verify whether reference FASTA headers use
chr1or1; divergent naming causes silent alignment drops.