Enspyre MRD：一种低测序深度下的超灵敏试剂盒式MRD解决方案的验证

Detection of MRD post surgery and monitoring of ctDNA during treatment are gaining increasing traction in patient care and drug development. First-generation assays have considerable clinical datasets, but suffer from limited sensitivity due to low numbers of mutations tracked. Recent ‘ultra-sensitive' solutions detect ctDNA at <100 parts per million (ppm), but suffer from high sequencing requirements and costs, limiting their use to large central labs. Enspyre is a novel sample preparation technique that enriches mutated DNA prior to sequencing, significantly improving signal to noise and reducing sequencing requirements. We present results from two independent sites demonstrating the analytical validation of Enspyre MRD, a kit-based MRD solution utilizing Enspyre. Specificity was tested with >180 replicates from >50 healthy blood donors at two different DNA input levels. LoD and LoQ were determined through >280 replicates using contrived samples in a 2-phase approach. Assay linearity was assessed using an additional 24 replicates spanning 5 orders of magnitude of ctDNA concentration. Analytical accuracy and precision were assessed using cell free DNA (cfDNA) derived from patients with NSCLC and serially diluted in cfDNA from healthy donors, as well as DNA extracted from cancer cell lines, fragmented to mimic cfDNA. Different blood collection tubes were also assessed using healthy donors. The effect of variant panel size on performance was assessed through the design of panels from 500 to 2,000 variants and in-silico downsampling. Enspyre MRD produced consistent results between sites, with sensitivity, specificity and quantitation all exceeding target specifications and meeting or exceeding that achieved by current ultra-sensitive tests. Specificity was 100% at all input levels, while LoD was <10ppm for 2,000 variant panels at 20 ng input. Sensitivity was positively correlated with panel size, with potential to trade off cost vs performance for different applications/indications, and was inversely correlated with input mass, with an LoD <40 ppm for 2,000 variant panels at 5 ng input. Results were replicable and precise across >6 operators, >3 reagent lots, >3 runs, and 2 sites. Consistent results were obtained using low- and mid-throughput sequencing platforms. Importantly, sequencing requirements were reduced from ~500M read pairs using standard techniques to just 5M, while retaining ultra-sensitive ctDNA detection and negating the need for molecular barcoding. These results demonstrate the potential of Enspyre as an ultra-sensitive MRD solution, in a format that is much more amenable to adoption in mid- and low-throughput laboratories than existing approaches. Through mutated DNA enrichment, Enspyre enables ctDNA detection at <10ppm with a >98% reduction in NGS requirements, facilitating significant cost reductions and decentralization of MRD testing.