Introduction: Pancreatic adenocarcinoma (PDA) is projected to be the leading cause of cancer deaths by 2030, with a five-year survival rate of 7%. While high patient mortality is driven primarily by late stage diagnosis, patients with localized disease who undergo curative resections still have a five-year survival rate of only 25%. It is hypothesized PDA undergoes dissemination at the earliest stages of tumor formation, driving formation of occult metastases which go undetected using conventional screening methods These occult metastases would be undetectable using conventional screening methods, and may fuel cancer progression in patients clinically staged as having localized disease. A surrogate marker of cancer progression is critical for disease monitoring and early intervention; however, no clinically-validated biomarkers have the required specificity or sensitivity for surveillance of localized PDA. Here, we present the Multiplex Enrichment using Droplet Pre-Amplification (MED-Amp) method for ultrasensitive detection of KRAS mutations, the most commonly mutated oncogene in PDA. This assay preserves ctDNA allelic frequency in the original sample, while increasing the molecular signal over 50-fold. This study demonstrates the feasibility of using our strategy for non-invasive ctDNA detection in low DNA input samples from localized cancer plasma samples. Materials and Methods: Genomic standards were created with controlled mutant KRAS content by diluting a genomic reference standard with KRAS WT cell line DNA. Standards were made to cover allelic frequencies ranging from 10.33% to 0.02% and total DNA content ranging from 50 ng to 5 ng. The standards were then partitioned into microfluidic droplets using ddPCR technology. PCR strips containing emulsified droplets were run in a thermocycler for 9 cycles of preamplification. The droplet suspension was de-emulsified using droplet destabilizer, and the PCR product underwent standard post-PCR cleanup. The PCR product was then re-partitioned into microfluidic droplets containing standard TaqMan probes and the emulsified sample was analyzed using standard ddPCR. Results and Discussion: Overall assay sensitivity was 87%, ranging from 72.2% for 3 mutant DNA fragments to 100% for > 12 fragments. Average mutant DNA signal was increased by 50-fold, and there was high concordance between measured KRAS mutant allele frequency post MED-Amp and calculated fraction across all dilutions tested (r2 = 0.88). We also compared droplet- versus conventional-PCR pre-amplification. Droplet amplification was two-fold more efficient than conventional amplification (p = 0.03) when the known KRAS allelic frequency was lower than 0.2%, near the limit of detection of most ddPCR assays. We next applied our assay to PDA patient plasma samples. We quantified the frequency of KRAS mutations in blinded plasma samples from metastatic PDA patients (n = 7), and age-matched non-PDA patient controls (n = 10). As expected, KRAS mutations were detected in 86% of PDA samples tested. Measured variant allelic frequencies were highly variable, ranging from 0.04% to 13.4% (mean: 3.89%). Based on input DNA concentration, the total number of mutant DNA fragments present in the tested plasma samples ranged from 1.56 to 195 copies.Conclusions: MED-Amp method enables accurate detection of low abundance mutations (< 0.05%) in low DNA input samples by using a droplet preamplification step with a high-fidelity polymerase. Our assay sensitivity for localized cancer is one of the highest reported for a single-gene ddPCR assay in PDA. The MED-Amp assay could also be applied to other tumor types and biospecimens including fine needle aspirates, as well as archived FFPE or frozen tissue samples. Assay sensitivity remained high, even when applied to frozen samples under extended storage conditions (> 2 years), enabling retrospective analysis of stored samples. Furthermore, MED-Amp can be integrated with existing assays, such as mutation detection in circulating tumor cells and next-generation sequencing of samples with limited DNA. These qualities make MED-Amp a potentially versatile tool that could be easily integrated into clinical studies.