ALBERT

Description: Abstract 755 Second hematologic malignancies in non-syndromic children without a pronounced family history for cancer may be mistaken for relapses or therapy-related malignancies. Recently, we characterized diagnosis and presumed relapse samples of 22 patients with very late disease recurrences (〉2.5 years), and identified 8 patients with leukemic presentations that were fully discordant at the level of TCR-rearrangements and DNA copy number aberrations (J Clin Oncol 2011; 29:1643-9). One of these patients showed a germline deletion comprising the recombination activating genes RAG1 and RAG2, and regulatory sequences of LMO2, genes frequently affected somatically in T-ALL, suggesting a genetic predisposition to leukemia. In the current study, we performed exome sequencing to assess whether consecutive leukemic presentations in such patients are indeed fully discordant, also at the sequence level, and to identify candidate pathogenic germline variants that point at a genetic predisposition. We sequenced the exomes in samples obtained from 2 consecutive leukemic presentations, and intermittent complete remissions, from 2 patients with very late disease recurrences (〉2.5 years) and discordant leukemic presentations. We found on average 26,600 variants per exome. Recurrent variants recorded in the dbSNP and/or 1000 Genomes databases, or those present in our in-house database (〉300 exomes) were excluded, resulting in an average of 989 private variants per exome. We divided these variants into 3 groups (i) somatic variants shared between the consecutive leukemic samples but not detected in remission (ii) somatic variants present in only one of the leukemic samples and (iii) germline variants present in the remission samples of the patients. All candidate somatic variants shared between two consecutive leukemic samples were re-sequenced by Sanger sequencing and were shown to be either present in all three samples, and thus originally missed in the remission sample, or falsely detected in one or more leukemic samples. Therefore, we conclude that in both patients no somatic variants were shared between the first and second leukemic presentations, which confirms that these patients suffered from clonally unrelated second T-ALLs. From all somatic variants present in only one of the leukemic samples, we focused on variants in exons or splice junction sites. We found 4 nonsense mutations, 9 frame-shift mutations, 12 in-frame in/dels and 7 non-synonymous missense variants with a high interspecies conservation score (PhyloP〉3.0), mostly affecting genes implicated in oncogenesis like PTEN, TET3, CDKN2C, CD109, and GLRX2. Each leukemic sample harbored 2–11 of these putative deleterious variants. In the germline of the two patients, we identified 314 and 190 non-synonymous unknown variants in exons or splice junction sites, respectively. Among these were 12 nonsense mutations, 7 canonical splice-site mutations, 20 frame-shift mutations, 11 in-frame in/dels and 143 non-synonymous missense variants at highly conserved positions (PhyloP〉3.0). Filtering of these variants for known T-ALL associated genes resulted in several interesting novel candidate predisposing genes such as, among others, RANBP17 and HOXC13. Sequencing of the entire RANBP17 open reading frame in a cohort of 24 sporadic T-ALL samples revealed that this gene was somatically affected in one of them. In conclusion, we confirmed by exome sequencing that consecutive leukemic presentations in patients with late T-ALL recurrences may be fully discordant and thus represent independent leukemia occurrences, most likely caused by predisposing germline abnormalities. Disclosures: No relevant conflicts of interest to declare.

Description: Resistance to glucocorticoids (GCs) is a major clinical problem in the treatment of acute lymphoblastic leukemia (ALL), but the underlying mechanisms are not well understood. Although mutations in the glucocorticoid receptor (GR) gene can give rise to therapy resistance in vitro, acquired somatic mutations in the GR are rarely encountered in patients. Here we report that the protein encoded by the BTG1 gene, which is frequently deleted in (pediatric) ALL, is a key determinant of GC responsiveness. Using RNA interference, we show that loss of BTG1 expression causes GC resistance both by decimating GR expression and by controlling GR-mediated transcription. Conversely, reexpression of BTG1 restores GC sensitivity by potentiating GC-induced GR expression, a phenomenon known as GR autoinduction. In addition, the arginine methyltransferase PRMT1, a BTG1-binding partner and transcriptional coactivator, is recruited to the GR gene promoter in a BTG1-dependent manner. These results implicate the BTG1/PRMT1 complex in GR-mediated gene expression and reveal that deregulation of a nuclear receptor coactivator complex can give rise to GC resistance. Further characterization of this complex as part of the GR regulatory circuitry could offer novel opportunities for improving the efficacy of GC-based therapies in ALL and other hematologic malignancies.

Description: Abstract 1104 Poster Board I-126 Relapse is the most common cause of treatment failure in pediatric acute lymphoblastic leukemia (ALL), and is difficult to predict from information at diagnosis in the majority of cases. To explore the prognostic impact of recurrent copy number abnormalities on relapse in children diagnosed with precursor-B cell ALL, we performed genome-wide copy number profiling of 34 paired diagnosis-relapse samples. Lesions detected at diagnosis were often absent at relapse, including recurrent targets in precursor-B ALL like PAX5 (not preserved in 2 out of 7 cases with deletions at diagnosis), CDKN2A (not preserved in 1 out of 15 cases), and EBF (not preserved in 2 out of 5 cases), which illustrates that these lesions are often secondary events that are not present in the therapy-resistant progenitor that causes relapse. In contrast, deletions and nonsense mutations in IKZF1, which encodes the lymphoid differentiation factor IKAROS, were highly frequent (38%) and always preserved at time of relapse. Locus-specific copy number screening of IKZF1 in an additional cohort of diagnosis samples from children enrolled in the Dutch treatment protocol DCOG-ALL9 with (n=40) or without (n=51) relapse revealed that IKZF1 deletions were significantly enriched in relapse-prone cases (22.5% vs 3.9%; P=0.007). An independent and unbiased validation cohort of 150 DCOG-ALL9 cases was used to confirm these findings, which established that 28.6% of the cases with IKZF1 deletion at diagnosis developed a relapse. Together, we conclude that deletions of IKZF1 in DCOG-ALL9 treated pediatric precursor-B ALL patients provide a strong prognostic marker for relapse. Disclosures No relevant conflicts of interest to declare.

Description: Background Treatment outcome in acute lymphoblastic leukemia (ALL) has improved over the past 30 years, with overall survival rates of ∼45% in adults and ∼85% in children. Gross cytogenetic abnormalities, including numerical changes and chromosomal translocations, are of considerable prognostic value in both pediatric and adult ALL. In addition, we and others have recently identified novel molecular markers associated with a poor outcome in ALL, including deletions of the lymphoid transcription factor IKZF1. In order to identify downstream signaling events associated with these genetic alterations, we performed an integrated analysis of genomic abnormalities, including copy number alterations, sequence mutations and chromosomal translocations, with alterations in protein expression and modification. Methods A cohort of 91 precursor B-ALL cases treated at M.D. Anderson Cancer Center in Houston, USA, including 82 newly diagnosed cases and 5 diagnosis-relapse pairs was used for this study. The cohort consisted of 6 children (age 1-6), 30 young adults (age 15-39) and 45 adults (age〉39), and 20 patients carried a BCR-ABL1 chromosomal translocation. Copy number alterations in eight genes frequently deleted in ALL (IKZF1, PAX5, EBF1, RB1, CRLF2, CDKN2A/2B, BTG1, and ETV6) were determined by multiplex ligation-dependent probe amplification analysis. IKZF1 deletions were associated with relapse (Pearson's chi-square test, p=0.009), and the presence of BCR-ABL1 translocation (p=0.032). Protein expression and modification levels were determined by probing Reverse Phase Protein Arrays (RPPA) containing protein lysates of all above samples with 128 rigorously validated antibodies including 34 phospho-specific antibodies. Hierarchical clustering analysis was used to determine which (phospho)proteins are differently expressed in genetic subsets of ALL. The significance of correlations was determined using two-sample t-test, with correction for multiple testing (Beta-Uniform Mixture model). Results We identified clustering of cases with a BCR-ABL1 chromosomal translocation (p=0.01; false discovery rate (FDR)=0.1), IKZF1-deletions (p=0.01, FDR=0.072), RB1-deletions (p=0.03, FDR=0.43) and EBF1 deletions (p=0.05, FDR=0.63). As expected RB1 deletion positive cases were characterized by decreased levels of (phospho)-RB1 and increased levels of cyclin E, illustrating the validity of our approach. EBF1-deleted cases showed relatively high levels of SHIP1, SSBP2 and phospho-STAT5, and lower levels of FAK and LYN. The protein signatures of BCR-ABL1-positive cases and IKZF1-deletion positive cases largely overlapped, and were characterized by elevated levels of (phospho)PKCα, SMAD1, phospho-STAT3, and phospho-STAT5 and lower levels of LYN and cyclinD3 (Figure 1). In total 70% of the BCR-ABL1-positive cases carried an IKZF1 deletion and several BCR-ABL1-negative cases with similar RPPA signature could be identified, all of which were IKZF1-deletion positive. These cases may represent the “BCR-ABL1-like” cases that were previously identified using gene expression signatures (Mullighan et al. 2009, NEJM 360:470-480; Den Boer et al. 2009, Lancet Oncol. 10:125-134), and could reflect activation of cAbl or other cellular tyrosine kinases. Together, we conclude that integrated analysis of genetic and proteomic aberrations identified protein signatures downstream of recurrent mutational events in ALL, a strategy that promises to facilitate the discovery of novel therapeutic targets in ALL and may aid in the identification of (high risk) patients that would benefit from tyrosine kinase inhibition. Disclosures: No relevant conflicts of interest to declare.

Description: B cell precursor acute lymphoblastic leukemia (BCP-ALL) is one of the most common malignancies in children. In the period 1991-2013, the Dutch Childhood Oncology Group (DCOG) has completed three treatment trials for childhood ALL: ALL8, 9 and 10, each protocol with stratifications into risk-groups (details: www.skion.nl). Although the cure rates increased in these subsequent trials, relapses still occurred in a significant number of children. Since consecutive upfront treatment protocols usually change at multiple levels, genomic alterations that are associated with relapse may also be variable, which could provide insight into the biology underlying therapy failure and relapse. In this study, we characterized the genetic architecture of relapsed BCP-ALL patients within the context of these three Dutch upfront protocols. We identified 3 patient groups based on upfront treatment as follows: Group-1: patients treated upfront with high-amounts of corticosteroids (CS) and relatively mild additional chemotherapy (ALL9 NHR/HR); Group-2: patients treated with high-amounts of CS and intensive additional chemotherapy (ALL10 MR); Group-3: patients treated with low-amounts of CS and moderately-intensive additional chemotherapy (ALL8 SR/MR, ALL10 SR). The number of high-risk patients that relapsed after ALL8 HR and ALL10 HR chemotherapy courses was too low to be included for analysis. We determined, at relapse, the presence of copy number alterations and sequence mutations in 21 recurrently affected genes involved in B-cell development, cell cycle regulation and RAS signaling, in 123 patients that relapsed after treatment in group-1 (n=56), group-2 (n=20) and group-3 (n=47). The number of CREBBP mutations in patients that relapsed after treatment according to group-1 (ALL9) was significantly lower compared to the other two groups, whereas B-cell development alterations were most common in patients that relapsed after treatment according to group-1, mainly due to a higher number of IKZF1 alterations (Figure 1). The high number of relapsed patients with leukemic clones carrying IKZF1 alterations in patients treated with high-amounts of CS and relatively mild additional chemotherapy is in line with our recent finding that IKZF1 is a key determinant of GC-induced apoptosis in normal and leukemic B-cells, and that loss of IKZF1 function confers resistance to dexamethasone, the major treatment component in group-1 (Marke et al., submitted). Additionally, in the group-2 patients treated with high-amounts of CS and highly intensive additional chemotherapy, a lower percentage IKZF1-deleted clones was detected at relapse, indicating that more GC-resistant, IKZF1-deleted clones are killed by the intense chemotherapy given in addition to CS in group-2 patients. Similarly, in the group-3 patients relapsing after treatment with lower amounts of CS and moderately-intensive additional chemotherapy, the percentage of surviving IKZF1-deleted clones was lower than in patients treated with high-amounts of CS. Taken together, our data indicate that the genetic architecture of relapsed BCP-ALL patients depends on the upfront treatment and, in addition, that the poor-prognostic feature of IKZF1-deletions may be more prominent in upfront treatment with high-amounts of CS and relatively mild additional chemotherapy. Figure 1. The frequency of genetic alterations in studied genes in patients that relapsed after treatment according to group-1, 2 and 3. Genes were grouped by their corresponding pathways. Group-1: patients treated upfront with high-amounts of CS and relatively mild additional chemotherapy (ALL9 NHR/HR); Group-2: patients treated with high-amounts of CS and intensive additional chemotherapy (ALL10 MR); Group-3: patients treated with low-amounts of CS and moderately-intensive additional chemotherapy (ALL8 SR/MR, ALL10 SR). Asterisk showed significant difference between upfront treatment groups, **p

Description: Abstract 126 Relapse is the major cause of treatment failure in pediatric acute lymphoblastic leukemia (ALL). Recent studies have shown a complex, dynamic architecture of clonal diversity in ALL and other leukemia subtypes, both at diagnosis and relapse. This multiclonal diversity follows a Darwinian model of evolution, and likely contributes to the selective outgrowth of therapy-resistant leukemic cells during or after chemotherapy treatment, resulting in relapse. In order to gain more insight into the multiclonal architecture of ALL, we selected two cytogenetically normal B-cell precursor ALL patients treated according to the DCOG-ALL9 protocol who, based on a previous study (Kuiper et al, Leukemia 2010, 24:1258–64), developed relapses with only minor genomic alterations as compared to diagnosis. Patient 1 had an IKZF1 deletion (exons 4 to 8) and a 23-nt insertion in IKZF1 exon 4 at diagnosis and developed a relapse after 12 months, whereas patient 2 was IKZF1 wild-type, and developed two relapses at 32 months and 4.5 years after diagnosis, respectively. Comparison of clonal rearrangements in the Ig and TCR genes between diagnosis and relapse(s) revealed that both leukemias were genomically stable. Copy number analysis revealed an acquired intragenic PAX5 deletion at relapse in patient 1 and two acquired copy number changes in the second relapse of patient 2 (Kuiper et al., 2010). In the current study, we performed whole exome sequencing on diagnosis, remission and relapse samples of both patients, and identified and confirmed 21 and 7 somatic missense, frameshift or splicesite mutations in the diagnosis and/or relapse samples, respectively. These variants were subsequently selected for amplicon-based ultra-deep sequencing (IonTorrent PGM with 318 chip, Applied Biosystems), using 15 ng of genomic DNA (corresponding to 2,200 haploid genome copies), reaching an average read-depth of 15,000x. All amplicons were mixed at equimolar levels and barcoded per patient sample. In patient 1, 19 mutations, including the 23-nt insertion in IKZF1 exon 4, were detected in 44–52% of the reads at both diagnosis and relapse, thus confirming that this leukemia was genomically stable. Two mutations were present at subclonal levels both at diagnosis and relapse, of which one (FMN1) was detected at 4-fold higher levels in relapse (Table 1). The second patient showed substantially more subclonal variability, revealing a mutation in GHR at diagnosis that was lost at relapse, and three mutations that appeared as novel mutations in the second relapse. These latter mutations thus may have been induced during treatment of the first relapse. Three mutations were detected at subclonal levels already at diagnosis, albeit in very low amounts for RANBP17 (Table 1). Based on these findings, we conclude that i) using paired whole-exome sequencing of diagnosis, remission and relapse samples we have identified novel somatic mutations in childhood ALL, ii) amplicon-based ultra-deep sequencing allows the sensitive detection of relapse-prone subclones at diagnosis, iii) this sequencing effort provides insight into the complex dynamic architecture of clonal diversity in childhood ALL. Table 1. Subclonal mutations in relapsed childhood ALL patients patient gene mutation status at relapse mean read depth per sample % variant reads diagnosis (1st) relapse 2nd relapse 1 FMN1 subclonal outgrowth 20,292 8% 34% - 1 NAT8 preserved subclone 4,249 24% 27% - 2 GHR lost 14,530 38% 0% 0% 2 ARHGEF1 acquired 14,370 0% 0% 44% 2 PNPLA8 acquired 19,301 0% 0% 44% 2 IQGAP1 acquired 19,811 0% 0% 46% 2 RANBP17 subclonal outgrowth 30,260

Description: Abstract 272 The response to therapy as determined by minimal residual disease (MRD) is currently used for stratification in treatment protocols for pediatric acute lymphoblastic leukemia (ALL). Even though MRD classification clearly identifies patients at low or at high risk for relapse, it also results in a large intermediate group (50 to 60% of patients), which still contains approximately half of all relapse cases. To improve risk stratification, we evaluated the added value of the IKZF1 alteration status, recently identified as a prognostic factor, in precursor-B-ALL patients. In an unbiased cohort of 131 uniformly treated precursor-B-ALL patients, we determined MRD levels at 42 and 84 days after treatment initiation using RQ-PCR analysis of Ig/TCR rearrangements. Based on these levels, patients were divided into three groups: MRD-Low (MRD-L), MRD-Medium (MRD-M) and MRD-High (MRD-H). IKZF1 alterations at diagnosis were determined using multiplex ligation-dependent probe amplification and genomic sequencing. We confirmed the strong prognostic significance of MRD classification, which was independent of IKZF1 status. Importantly, in the large MRD-M group (n=81; 62% of patients) containing 46% of the relapsed patients, IKZF1 alteration status identified 8 out of 11 relapsed patients (72%). The 9 year relapse-free survival (RFS) for IKZF1 mutated patients in this MRD-M group was 27% compared to 96% for patients wild-type for IKZF1 (P

Description: Current risk assessments for treatment of children with B-cell precursor acute lymphoblastic leukemia (BCP-ALL) are based on several clinical and biological criteria, including genomic alterations. Genomic profiling of BCP-ALL in the last few years has substantially extended the number of risk factors that can be used for risk stratification, including a novel entity known as BCR-ABL1 -like or Ph-like ALL. This subgroup of BCR-ABL1-like cases is characterized by the high recurrence of a diverse repertoire of novel gene fusions and mutations which frequently result in enhanced tyrosine kinase and cytokine receptor signaling [Roberts et al., NEJM 2014;371:1005-15]. Leukemia's with these alterations could potentially be targeted with appropriate tyrosine kinase inhibitors. Clinical trials with newly-diagnosed patients carrying these alterations are therefore required, but the large genomic diversity within this group of patients currently provides a major bottleneck. Here, we describe the use of Targeted Locus Amplification (TLA), combined with deep-sequencing to detect fusion genes and sequence mutations relevant for stratification of BCP-ALL. TLA involves a strategy to selectively amplify and sequence regions 〉100 kb around a preselected primer pair by crosslinking of physically proximal genomic sequences [de Vree et al., Nat Biotechnol. 2014;32:1019-25]. Since TLA results in the amplification of all sequences at either end of the primer pair, TLA is highly effective in picking up structural variations including novel fusion partners. Furthermore, breakpoints can be identified from the TLA sequencing data from which targets for detection of minimal residual disease can be directly designed. A total of 31 primer sets targeting 19 recurrently affected genes were designed and multiplexed, including the 'classical' players MLL, RUNX1, TCF3, and IKZF1, the tyrosine kinase genes ABL1, ABL2, PDGFRB, CSF1R, JAK1, JAK2, JAK3, FLT3, and TYK2, and the cytokine signaling genes CRLF2, EPOR, IL7R, TSLP, SH2B3, and IL2RB. Primer sets were chosen such that the most relevant regions were sufficiently covered. As a pilot, viable cells from 46 selected BCP-ALL samples were analysed, including 26 cases with a BCR-ABL1 -like expression profile [Den Boer et al., Lancet Oncol. 2009;10:125-34], of which 6 had a known kinase fusion. 7 Gb of aligned sequence data was obtained for each patient sample. All 21 rearrangements known to be present in these samples were detected by TLA, including rearrangements in ETV6-RUNX1 (n=5), MLL (n=4), TCF3-PBX1 (n=3), CRLF2 (n=4), EBF1-PDGFRB (n=2), BCR-ABL1 (n=1), RCSD1-ABL2 (n=1), and SSBP2 -CSF1R (n=1). For 10 of the fusions sequencing depth was sufficient to extract breakpoint-spanning sequences directly. For two cases with known JAK2 fusions with an unknown partner, the fusion gene was identified (TERF2 and BCR), as was the case for an unknown ABL1 fusion (FOXP1). New fusions were identified in 9 cases, including previously described IGH@-EPOR and TCF3-ZNF384 fusions, and novel kinase activating fusions of MAP3K19-TSLP and HDAC9-FLT3. In addition we identified deletion breakpoint fusions in IKZF1, and sequence mutations in JAK1, JAK3,and IL7R. In total, we detected gene fusions or sequence mutations affecting tyrosine kinase or cytokine receptor signaling in 16 of the 26 cases with a BCR-ABL1 -like expression profile. We conclude that TLA is an effective method for the reliable detection of sequence mutations and structural variations that are relevant for disease prognosis and/or could be targeted by approved kinase inhibiton. Disclosures Simonis: Cergentis BV: Employment. Klous:Cergentis BV: Employment. Yilmaz:Cergentis BV: Employment. Splinter:Cergentis BV: Employment.

Description: Chronic myeloid leukemia (CML) is a rare malignancy in children and is mostly diagnosed in the chronic phase (CP). In adults, the five-year overall survival rate is 89% for patients on Imatinib and disease progression occurs in 1-3% per year (Druker 2006). Once a blast crisis (BC) has occurred, treatment options are limited with a median survival of only a few months (Cortes 2008). Therefore, early recognition of patients at risk for developing a BC is desirable. Besides the translocation t(9;22)(q34;q11), IKZF1, PAX5, and CDKN2A deletions have been reported in CML lymphoid blast crisis (LyBC) of both adult and pediatric patients (Mullighan 2008, Alpár 2012). The aim of this study was to investigate the presence of IKZF1 deletions and other copy number alterations (CNAs) by MLPA analysis in a large cohort of pediatric CML patients at time of diagnosis in order to determine whether CNAs commonly found in pediatric ALL might predict disease progression and / or treatment response. Between October 1991 and October 2012 a total of 86 children with newly diagnosed CML were included. The median follow up was 31 months. Among the 86 patients, 82 patients were diagnosed in CP, 2 patients in accelerated phase (AP), and 2 patients in LyBC. Six patients experienced progression to a BC respectively a myeloid blast crisis (MyBC) (N=2) and LyBC (N=4). At time of diagnosis, an IKZF1 deletion was detected in one patient diagnosed with CML-AP (Table A, patient no 58). IKZF1 and EBF1 deletions were detected in one patient diagnosed with CML-LyBC (Table A, patient no 22). No CNAs were detected in the 82 patients diagnosed with CML-CP. At time of disease progression, new CNAs were detected at time of the LyBC (Table A, patient no 62, 64, and 67). Due to the absence of material no CNAs could be detected in both patients experiencing a MyBC. In conclusion, we were able to detect CNAs in progressive CML disease (CML-AP and CML-LyBC) and not in the samples at the time of chronic phase in this large pediatric cohort of CML patients. Therefore, the investigated CNAs could not be used to predict disease progression at time of diagnosis. The CNAs detected in patients with progressive CML were similar to specific CNAs detected in pediatric B-cell precursor ALL, indicating a similar disease development (Kuiper 2010). Additionally, our results are in accordance with existing literature, suggesting that mechanisms of disease progression in pediatric and adult CML might be similar (Brazma, 2007). Disclosures: No relevant conflicts of interest to declare.

Description: Introduction: Recent work has demonstrated that relapses in pediatric acute lymphoblastic leukemia (ALL) can arise from minor subclones present at diagnosis. Several genes have been associated with therapy resistance in these subclones, including the Ras pathway genes KRAS, NRAS, and PTPN11, the H3K36 methyltransferase NSD2 (WHSC1), and the 5'-nucleotidase NT5C2. Retrospective backtracking of these relapse-associated alterations has demonstrated that these alterations are frequently present at time of diagnosis in minor subclones, sometimes in less than a few percent of the cells. The prognostic value of subclonal alterations in these genes at time of diagnosis, however, is less well understood. Prospective screening of subclonal mutations, without prior knowledge of the mutation status, requires extra specificity and sensitivity. Accurate quantification of the subclonal burden of these mutations will provide potential for following the subclonal dynamics during early stages of treatment, and could be informative for adapting therapy. The aim of this study was to develop a targeted next generation sequencing assay to perform quantitative detection of subclonal mutations in the selected genes. We used single molecule molecular inversion probes (smMIPs), an approach that applies single molecule tagging to correct for amplification biases (Hiatt et al., Genome Research. 2013, 23: 843-854), an artifact that becomes relevant in case of low mosaic mutations. Method: We designed a pool of 77 smMIP oligonucleotides targeting the coding sequences of five genes associated with therapy resistance in BCP-ALL, including KRAS, NRAS, PTPN11, NT5C2, and WHSC1. The smMIPs tiled a total of 4124bp of genomic sequence, including hotspot regions of the genes. To demonstrate the potential of this method, we applied this newly designed smMIP panel on 22 BCP-ALL diagnosis samples to retrospective backtrack mutations in KRAS (n=11), NRAS (n=8) and PTPN11 (n=3) that were previously characterized at relapse. We used 100ng of genomic DNA per sample as input, which is the equivalent of 15,000 haploid copies. Sequencing was performed on the Illumina NextSeq platform with pair-end sequencing, data were analyzed by SeqNext v4.2.2. Result: The average read depth obtained varied per gene from 30,081x (NRAS) to 65,749x (PTPN11). Sequencing reads with the same molecular tag were clustered into one tag-defined read group, in which random errors caused by library construction and sequencing were eliminated. These so-called single molecule consensus reads (smc-reads) were comprised of, on average, 139 individual sequencing reads. Using the smMIP approach, 19 out of the 22 Ras pathway mutations identified at relapse were detectable at diagnosis, of which 10 had a low mutant allele frequency (varying from 0.52-8.31%), which is in line with our previous ultra-deep backtracking result. Taking advantage of the known position of the mutations at relapse, we established the noise level in the diagnosis samples by analyzing variant calls outside the hotspot regions. The noise level was varied between samples from 0.03% to 0.24% (average 0.06%). Based on these background settings, we subsequently searched for novel mutations and identified 1 mutation in NT5C2 (p.P534S, 0.38%), 2 hotspot mutations in WHSC1 (p.E1099K, 0.17% and 0.27%), as well as many additional subclonal mutations in KRAS, NRAS and PTPN11. The latter finding suggests the presence of multiple Ras-mutated subclones in individual cases, of which only a subset survive from chemotherapy and grow out in the relapse clone. Conclusions: Taken together, single molecule tagging based smMIP technology allows the accurate detection of low mosaic mutations. These findings illustrate the need for the current ongoing prospective mutation screens in unbiased cohorts of diagnosis samples to determine the prognostic value of subclonal mutations in these five genes on the risk of relapse. Disclosures No relevant conflicts of interest to declare.