Abstract Precise relapse prediction in cancer patients is urgently needed as conventional serum tumor markers do not predict relapse in a timely manner. Circulating tumor DNA (ctDNA) is a new class of quantitative biomarkers that reflect tumor burden. The ctDNA is usually found as mutated DNA fragments in the blood, which are generally consistent with the mutations found in the concomitant tumor in the body. Variant allele frequency (VAF) of ctDNA is usually very low and ranges from 0.001 to 1% even in advanced cases. To quantify absolute copy number of ctDNA at the very low range, digital PCR (dPCR) is one of the most practical methods. However, specific primers/probes for dPCR must be designed individually as tumor mutation varies between individual patients. The TP53 gene is the most frequently mutated in human cancer and 90% of the mutations occur in its DNA binding domain coding region. According to the COSMIC database, there are 1,284 unique mutations from a total of 25,376 recorded mutations in the DNA binding domain. The majority (77%) of the 1,284 unique mutations are not recurrent and thus 42 unique mutations cover 50%, 132 unique mutations cover 70%, and 426 unique mutations cover 90% of the total mutations. Here we have designed a dPCR probe library for the set of 132 unique mutations to cover 70% of TP53 mutations in the DNA binding domain using Hypercool Primer & ProbeTM technology. With this technology, modified bases, 2-Amino-dA and 5-Methyl-dC, were substituted in adenine and cytosine positions, respectively. This modification increases the Tm, thus allowing for highly specific amplification despite short PCR products of 70-90 bp. The dPCR library allows "off-the-shelf", highly sensitive ctDNA tracking once the TP53 mutation is identified in the target tumor (e.g., surgical specimens). In our pilot cases, we found TP53 mutations from 97% (34/35) of esophagus, 70% (7/10) of stomach, and 93% (13/14) of colorectal primary tumors using next generation sequencer (NGS). There were 40 unique mutations for which all mutations were covered with our TP53 dPCR probe library. We then validated the dPCR library for ctDNA in terms of: (a) the ability to detect the target mutation; (b) tumor genomic heterogeneity; and (c) early detection of relapse. Of >500 DNA samples from primary cancers and patient plasma, 98% of probes tested were confirmed to detect the target mutation properly. Multiregion sequencing of stomach and colon cancers by NGS revealed that 85% (17/20) of tumors exhibited TP53 founder mutations, offering a high chance that the corresponding ctDNA was detected in plasma. Among cases who experienced relapse confirmed by CT scan, ctDNA elevation was seen approximately 6 month prior to the confirmation by CT scan. Overall, the sensitivity of VAF using dPCR with the probes was as low as 0.001%. These results suggest that the TP53 dPCR probe library can potentially be used for a wide range of human cancers as a new class of biomarkers for early relapse prediction. Citation Format: Satoshi S. Nishizuka, Mizunori Yaegashi, Noriyuki Sasaki, Takeshi Iwaya. Digital PCR probe library for TP53 mutations in early relapse prediction [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 1378.