Extensive molecular analysis revealed genetic alterations related to relapse such as mutations of CREEB, MSH2, or NT5C2 in B-cell precursor acute lymphoblastic leukemia (BCP-ALL). Recently Li B, et al. have established relationship between relapse-specific somatic alterations and timing of relapse (Blood 2020;135:41-55). They identified close link between alterations of DNA mismatch repair genes, including MSH2, MSH6, and PMS2 and early relapse of BCP-ALL as a result of alteration of thiopurine response and resistance to treatment. However, it remains to be clarified which subtype of BCP-ALL is prone to acquisition of particular type of relapse-specific molecular alteration. To further elucidate mechanism of recurrence we have analyzed childhood BCP-ALL cases, particularly focusing on DNA mismatch repair pathway. We analyzed diagnosis-relapse pair samples of recurrent 16 BCP-ALL cases, who had been treated in our institution to find single nucleotide variant (SNV), small Indel, and copy number variation in the coding exons, particularly focusing on mismatch repair pathway using the data captured by Ion AmpliSeq Exome kit and Ion Proton (Thermo Fisher Scientific, MA, USA). The identified variants were confirmed by Sanger sequence. Additionally, we performed RNA-seq using SMART-Seq Ultra Low Input RNA Kit (Clontech Laboratories, Inc, CA, USA), Ion Plus Fragment Library Kit, and Ion Proton, and in vitro cell culture of the leukemic blasts for several cases. Of several DNA mismatch repair pathway genes, we have identified somatic SNV of MLH1 in a case. The index case, three years old male had diploid BCP-ALL with t(7;9) and PAX5 alteration at diagnosis, who developed early relapse (11 months from diagnosis) and died of the disease. From the sample at 1st and 2nd relapse we have identified somatic MLH1 variant (NM_000249.4:c.901C>T;p.Gln301*). This SNV was detected in small fraction of the diagnostic sample. Furthermore, we have established permanently growing cell line, ICH-BCP-1 from the sample obtained at 2nd relapse. The doubling time is approximately 37 hours and the karyotype was, 46,XY. The same MLH1 variant was identified in the cell line. Of note is that the number of detected SNV increased rapidly at 1st relapse and 2nd relapse as suggested by function of MLH1 product. No other alteration of DNA mismatch repair pathway was observed in the cohort. As previously discovered relapse-specific alterations, we identified somatic SNV of NT5C2 in one hyperdiploid BCP-ALL case and somatic SNV of CREBBP in one hyperdiploid BCP-ALL case. The former case with SNV of NT5C2 gained deletion of IKZF1 and formation of P2RY8-CRLF2 fusion gene at recurrence. Throughout the cohort, hyperdiploid BCP-ALL cases had tendency to have RAS pathway somatic SNVs (KRAS, NRAS, FLT3, and PTPN11). In addition, one low hypodiploid BCP-ALL case had germline small Indel of TP53 and somatic SNV of RB. We add MLH1 alteration to the list of DNA repair pathway relapse-specific somatic alterations, further supporting the particular significance of DNA repair pathway as mechanism of BCP-ALL recurrence, probably related to massive acquisition of complex genetic alteration as a result of loss of DNA repair. In our cohort, the prevalence of previously reported relapse-specific mutation is relatively low, which may be caused by detection method and different ethnicity. Our novel cell line is useful staff for investigation to identify the role of DNA mismatch repair pathway in BCP-ALL leukemogenesis. Disclosures No relevant conflicts of interest to declare.