Background: Use of biomarkers for patient treatment stratification is an increasingly important topic with the recent FDA approval of Pembrolizumab and Nivolumab for microsatellite instability (MSI) high/mismatch repair (MMR) deficient cancers. This ground-breaking approval allows physicians to make patient treatment decisions based on molecular biomarker status. Thus, performance of detection methods for these biomarkers is of great importance. The gold standard for MSI analysis in a research setting is PCR followed by fragment analysis to resolve DNA size. Recently, PCR followed by melt curve analysis has been presented as an alternative approach. In this set of experiments these two methods are compared using a subset of human colorectal cancer samples. Methods: A cohort of matched human colorectal cancer and adjacent normal FFPE samples were used for this comparison. Eight pairs were selected that contained subtle shifts, low tumor volume or heterozygosity at microsatellite loci. These samples were then tested for MSI by fragment or melt curve analysis. Samples were then classified as MSI-H, MSI-L or microsatellite stable (MSS) according to manufacturer specifications for melt curve analysis or by an interpretive software for fragment analysis. Results: 40 mononucleotide markers were examined for each method. Melt curve analysis was concordant with fragment analysis in 70% of markers tested (28/40). Of discordant markers, 92% (11/12) were identified as unstable with fragment analysis but were stable by melt curve analysis. Additionally, two of the samples identified as MSI-H by fragment analysis were called MSS or MSI-L using the melt curve analysis method. These preliminary results suggest there may be differences in sensitivity when using challenging samples with melt curve compared to the gold standard method, fragment analysis. Conclusions: The standard method for determining MSI status utilizes PCR and fragment analysis. Recently, melt curve analysis has been proposed as an alternative method. We present preliminary findings of decreased sensitivity with melt curve compared to fragment analysis. Additional studies with a larger, more diverse sample set will be required to define this relationship further.