SPANOM: A cost-effective method of detecting MSI in ctDNA.

e24263 Background: Not only has been proved to be an effective biomarker for prognosis and chemotherapies, but also becomes the first pan-cancer biomarker for immunotherapies, MSI/MMR detection has increasing values in clinical applications. The conventional methods, such as PCR and IHC, could merely be applied to tumor tissues. However, tissue biopsy or surgery might not be applicable for some late-stage cancer patients, such as stage IV patients, where most of the immunotherapies are applied. Hence the MSI detection based on ctDNA is crucial. Here we developed SPANOM, a flexible and normal-control-free NGS-based method of comparable cost to the conventional methods, to accurately detect MSI status in tissues of low quality and ctDNA. Methods: A small panel (about 24000 base pairs) was designed directly targeting 100 selected microsatellite loci. The detection model was built on the in silico samples randomly generated from 3 MSI-H and 3 MSS tumor tissues and their corresponding white-blood-cell controls. The model was validated by the clinical samples. The tissues were sequenced at 1000x and the ctDNA was sequenced at 10000x for the validation set. Results: Considering conventional methods (IHC or PCR) on tissues as golden standard, the overall concordance was 92.3% in 27 dMMR/MSI-H and 25 pMMR/MSS tumor tissues whose tumor contents were larger than 50%, and the overall concordance was 86.5% in 17 dMMR/MSI-H and 20 pMMR/MSS ctDNA samples from stage IV treatment-naive patients of colorectal or stomach adenocarcinoma. Conclusions: SPANOM is an accurate method of detecting MSI status in tumor tissues and ctDNA from late-stage patients of colorectal or stomach adenocarcinoma. The sequencing cost of SPANOM could be as low as 3 USD for tumor tissues and 25 USD for ctDNA, which is comparable to the conventional method. The small panel of MSI could be utilized independently or merged with a larger panel, which could benefit more patients in clinical applications.