Background: Large next-generation sequencing (NGS) panels (> 300 genes) offer multiple potential therapeutic options for patients with metastatic cancer. However a large portion of the population in developing countries is unable to avail the benefits of such testing due to limited availability of many drugs or suitable clinical trials, coupled with the high cost of these tests. There is an urgent need to offer a compact, affordable and robust testing solution which can offer expanded but feasible therapeutic options. Hence we decided to develop a custom mid-sized panel to fulfil these unmet requirements.
Methods: A targeted solid tumor (DNA + RNA) panel comprising 74 genes (SA74) was designed to cover all genes with a Tier 1 drug recommendation for therapy, evaluating single nucleotide variants (SNV), indels, copy number variants (CNV) and gene fusions (GF). We also included certain Tier2 genes for prognosis or added clinical impact. GFs were evaluated by RNA. Inferior sample quality often results in poor quality data, so we added a DNA component for tiling select intronic regions to identify GF which to be used when RNA could not be analyzed. The analytical sensitivity was > 99% for SNV/Indels with a 5% limit of detection, > 99% for CNV and GF. The clinical sensitivity was 100% for SNV, 95% for indels, 84.2% for CNV, 100% for RNA GF and 70% for DNA GF.
Results: 239 formalin-fixed paraffin embedded tumor samples were evaluated using SA74 and the data was scored for actionability across tumor types. This was compared with data from 706 samples across multiple cancers analyzed using the Illumina TSO500 panel. The average number of actionable alterations was 1.1 in SA74 and 1.6 in TSO500. The overall actionability (cases with at least one Tier1 or Tier2 actionable variant) of SA74 was 63.4%; while that of TSO500 was 78.3%. Of this, the overlap with SA74 was 73.7%. The actionability in the remainder was due to Tier2 genes with lesser evidence altering the same pathway as an approved drug target or targeting investigational drugs. The actionability for each cancer type was calculated and found to be: colon (34.1%; n = 44), non-small cell lung (NSCLC) (78.4%; n = 37), breast (91.7%; n = 24), carcinoma of unknown primary (CUP) (45%; n = 20), uterine (53.3%; n = 15), gallbladder (75%; n = 12) and sarcoma (40%; n = 10), among others. Smaller panels often do not include CNV and GF. Addition of these variant types increased actionability across cancers. The lack of CNV and GF would have decreased total actionability from 62.9% to 54.5%. NSCLC and CUP were most impacted with a difference in actionability of 16.2% and 15% respectively. GF increased actionability mainly in NSCLC, while CNV contributed to increases across all cancers.
Conclusions: SA74 demonstrated high actionability across cancers. It therefore presents a practical alternative to large panel testing by optimizing actionability and affordability, useful in a cost-sensitive setting.
Click this link to access the complete article.