Research ArticleMinimal residual disease

High-Throughput Sequencing Detects Minimal Residual Disease in Acute T Lymphoblastic Leukemia

See allHide authors and affiliations

Science Translational Medicine  16 May 2012:
Vol. 4, Issue 134, pp. 134ra63
DOI: 10.1126/scitranslmed.3003656


High-throughput sequencing (HTS) of lymphoid receptor genes is an emerging technology that can comprehensively assess the diversity of the immune system. Here, we applied HTS to the diagnosis of T-lineage acute lymphoblastic leukemia/lymphoma. Using 43 paired patient samples, we then assessed minimal residual disease (MRD) at day 29 after treatment. The variable regions of TCRB and TCRG were sequenced using an Illumina HiSeq platform after performance of multiplexed polymerase chain reaction, which targeted all potential V-J rearrangement combinations. Pretreatment samples were used to define clonal T cell receptor (TCR) complementarity-determining region 3 (CDR3) sequences, and paired posttreatment samples were evaluated for MRD. Abnormal T lymphoblast identification by multiparametric flow cytometry was concurrently performed for comparison. We found that TCRB and TCRG HTS not only identified clonality at diagnosis in most cases (31 of 43 for TCRB and 27 of 43 for TCRG) but also detected subsequent MRD. As expected, HTS of TCRB and TCRG identified MRD that was not detected by flow cytometry in a subset of cases (25 of 35 HTS compared with 13 of 35, respectively), which highlights the potential of this technology to define lower detection thresholds for MRD that could affect clinical treatment decisions. Thus, next-generation sequencing of lymphoid receptor gene repertoire may improve clinical diagnosis and subsequent MRD monitoring of lymphoproliferative disorders.


T-lineage acute lymphoblastic leukemia/lymphoma (T-ALL) is an aggressive, immature, malignant T cell neoplasm that affects both adult and pediatric patients. Although there has been progress in treating T-ALL patients, with some achieving durable responses, a subset of these patients are inadequately treated and have disease relapse, and others may be overtreated because of the inability to sufficiently individualize clinical treatment. Several studies have confirmed the importance of assessing minimal residual disease (MRD) to predict clinical outcomes of patients (14). For example, patients who demonstrate an early and sustained response to therapy with no evidence of MRD fare significantly better than those who do not (3).

Current clinical strategies to assess MRD include multiparametric flow cytometry (mpFC) and quantitative polymerase chain reaction (PCR)–based methods using patient-specific primers (5, 6). mpFC typically permits detection of recurrent/persistent disease with a sensitivity on the order of 1 cell in 104; however, data interpretation is operator- and laboratory-dependent, and consequently poorly standardized. Furthermore, variable expression of leukemic antigens in the posttherapy setting confounds MRD detection by mpFC (7). Molecular-based methods for detection of MRD can achieve increased sensitivity, on the order of 1 cell in 105 (8, 9). However, the implementation of these molecular assays—principally real-time quantitative PCR–based assays using patient-specific primers that target variable junctional sequences or patient-specific translocations—is expensive, labor-intensive, and difficult to implement in a uniform manner (8, 10).

High-throughput sequencing (HTS) is an emerging technology that can provide insight into the complexity of the adaptive immune response through the analysis of lymphoid receptor gene rearrangement (11). Recent studies using this technology have challenged our understanding of the extent of lymphocyte diversity occurring within and shared by individuals (11, 12) and have provided mechanistic insight into the early molecular genetic events critical for T cell lineage maturation (13). Recently, HTS of lymphoid receptor genes has been used to monitor lymphocyte diversity after adoptive immunotherapy with chimeric antigen receptor–modified T cells for the treatment of chemotherapy-refractory chronic lymphocytic leukemia (14) as well as to monitor disease in B lymphoproliferative disorders (15). We have used HTS to identify rare T cell clones (1 T cell in 100,000) with high accuracy and reproducibility (16).

To determine whether HTS could contribute to the clinical management in ALL, we used HTS to diagnose and detect MRD in patients with T-ALL. We directly compared HTS to mpFC for MRD assessment and found that HTS of T cell receptor (TCR) gene loci may equal or surpass mpFC in the detection of MRD. In addition, our findings suggest that HTS may aid in the identification of T-ALL cases with early thymic precursor (ETP) immunophenotype (17, 18) because these cases in our cohort lacked a complete clonal TCRB gene rearrangement at diagnosis.


Forty-three matched pairs of T-ALL samples were collected sequentially at the University of Washington’s Hematopathology Laboratory from patients enrolled in the Children’s Oncology Group (COG) AALL0434 trial. The complementarity-determining region 3 (CDR3) of TCRB and TCRG was amplified and sequenced from 67,000 cells for pretreatment samples and 200,000 cells for samples taken 29 days after treatment. More DNA was sequenced for posttreatment samples because the proportion of neoplastic lymphoblasts was expected to be markedly diminished after chemotherapy. In two posttreatment samples (cases 49 and 63), there was only enough input material to sequence ~5000 and 16,000 cells, respectively, which limited the sensitivity of this assay for MRD detection for these patients. To ensure accurate detection of MRD, we sequenced all samples with at least fivefold coverage to account for PCR and sequencing errors, as previously described (11).

The pretreatment analysis of TCRB and TCRG sequences permitted us to identify for most patients the unique, recombined TCR gene sequences representing the patient’s clonal, neoplastic T lymphoblasts (Table 1). The frequency of these clonal sequences represented at least 19% of the total T cell repertoire before treatment, a proportion that is greater than four SDs (3.7%) from the mean frequency of the largest reactive T cell clone (3.8%) seen in normal individuals (fig. S1). The proportion of T lymphoblasts in these pretreatment samples was uniformly high, averaging 83.7 ± 17.2% of total leukocytes (range, 35.4 to 98.7%) by mpFC (table S1). For pretreatment samples, a clonal sequence of TCRB was readily identified in 31 of 43 cases (Table 1 and Fig. 1, top panel). One of these samples, case 58, had two TCRB sequences with similar high frequency, consistent with biallelic rearrangement. By contrast, 12 patient samples did not have an identifiable, complete, clonal TCRB sequence at diagnosis. Of these 12 patients, four had a demonstrable pretreatment, clonal TCRG rearrangement on subsequent analysis (ETP, cases 54 and 55; near-ETP, case 56; and non-ETP, case 11). Of the remaining eight cases, these patients could not be studied further for MRD detection by HTS because no pretreatment complete clonal TCRB or TCRG sequence was identified. For our cohort then, HTS identified at least one clonal TCRB or TCRG sequence in 35 of 43 samples (81.3%).

Table 1

Pretreatment, clonal TCRB CDR3 sequences. Clonal TCRB CDR3 sequences were identified in most of the patients (31 of 43). Cases designated by mpFC as ETP (five of five) and a subset designated as near-ETP (six of nine) are not included here because these did not have a pretreatment, clonal TCRB gene rearrangement. One non-ETP case, case 11, is also not included because this similarly did not have a clonal TCRB sequence. Nucleotide insertions (underlined) and D gene (italicized) regions are shown. The V and J gene family (number) or gene segment usage is indicated.

View this table:
Fig. 1

T cell clonality as detected by HTS for TCRB and mpFC. Pretreatment (top) and posttreatment day 29 (bottom) clonal T lymphoblast populations were identified by HTS (red) or by mpFC (blue). HTS data are reported as the frequency of the clonal sequence of total rearranged T cell sequences; mpFC data are reported as the T-ALL frequency of total T cells including all CD7+ T/NK events by mpFC. Twelve cases (5 of 5 ETP, 6 of 9 near-ETP, and 1 of 31 non-ETP) without a complete clonal TCRB gene rearrangement in pretreatment samples are not shown, despite each having a high proportion of immunophenotypically abnormal T lymphoblasts as identified by mpFC. On the basis of MRD detection, three groups of cases are identified in posttreatment samples (left to right): group 1, MRD not detected by either HTS or mpFC; group 2, MRD detected by HTS, but not by mpFC; and group 3, MRD detected by both HTS and mpFC. No cases with a pretreatment, clonal TCRB rearrangement were mpFC-positive and HTS-negative for MRD at day 29.

With knowledge of the specific TCRB CDR3 sequence of the patient’s clonal population in the pretreatment samples, we next examined the potential for HTS to identify the same clonal sequence in the posttreatment day 29 sample and compared these findings to results obtained by mpFC performed as part of the current protocol in the COG AALL0434 trial (Fig. 1, bottom panel). HTS to detect the patient’s original clonal TCRB sequences at day 29 revealed three subgroups of patients in this cohort: (i) those for whom MRD was not detected by either HTS or mpFC, 9 cases; (ii) those for whom MRD was detected only by HTS but not mpFC, 10 cases; and (iii) those for whom MRD was detected by both HTS and mpFC, 12 cases (Fig. 1, bottom panel). Of the 10 cases for which HTS did and mpFC did not detect MRD, the MRD was 10- to 100-fold lower than for the 12 cases for which both mpFC and HTS detected MRD (Fig. 1, bottom panel). There were no cases in which MRD was detected by mpFC but not by HTS for TCRB. For 4 of the 12 cases with a clonal TCRG but not a clonal TCRB (ETP, cases 54 and 55; near-ETP, case 56; and non-ETP, case 11), MRD was detected by HTS of TCRG in 3 of 4 cases at day 29. MRD for case 56 was missed by HTS for TCRG, but detected by mpFC.

Immunophenotypically, by mpFC, 5 of the 43 cases had a pretreatment immunophenotype compatible with ETP T-ALL (17), a subtype of T-ALL recognized to have more aggressive clinical behavior with higher propensity for disease relapse with standard chemotherapy. Of the other 38 cases, 9 had an ETP-like immunophenotype that we designate as “near-ETP,” in which all but the CD5 expression requirement of ETP tumors is met. We designate the remaining 29 cases that did not have an ETP or near-ETP T-ALL immunophenotype as “non-ETP” (Fig. 2 and table S1).

Fig. 2

Immunophenotype of representative T-ALL cases by mpFC. Pretreatment cases were classified as ETP using reported criteria (17). Near-ETP cases met all but the CD5 expression requirement of ETP cases. All other cases were otherwise designated as non-ETP T-ALL. Representative cases are shown.

Further evaluation of the cases for which no complete clonal TCRB sequence was identified in pretreatment samples showed that these patients had substantial MRD at day 29 detectable by mpFC (Fig. 3). Review of the pretreatment immunophenotype of these 12 cases showed that 5 had an immunophenotype compatible with an ETP subtype (patients 3, 12, 54, 55, and 57), whereas 6 had a near-ETP immunophenotype (patients 13, 42, 45, 56, 60, and 68). One case (patient 11) did not have either an ETP or a near-ETP immunophenotype, but had a V9JP clonal TCRG sequence and is likely a γδ-differentiated T-ALL. In contrast, of the 31 cases in which a clonal TCRB rearrangement was identified at diagnosis, no cases were classified as ETP, 3 were near-ETP, and the remaining 28 had a non-ETP T-ALL immunophenotype. In this cohort, the absence of a complete clonal TCRB sequence in the pretreatment sample was associated with either an ETP or a near-ETP immunophenotype by mpFC (P < 0.0001, Fisher’s exact test; fig. S2). Moreover, the relatively higher amount of MRD by mpFC in ETP patients at day 29 compared to typical T-ALL patients (13.1 ± 20.8% ETP compared with 0.46 ± 1.00% non-ETP T-ALL lymphoblasts of total mononuclear cells, P = 0.001; fig. S3) is compatible with a previous report identifying the ETP subtype of T-ALL as having a more aggressive clinical course with increased propensity for disease relapse (17). On the basis of these data, our findings suggest that HTS may be used to segregate non-ETP T-ALL from ETP and near-ETP T-ALL. Further, these data suggest that the near-ETP group that we describe may behave similarly to ETP subtypes of T-ALL with increased propensity for MRD after treatment (fig. S3); however, clinical outcome data will be required to address this issue and are not yet available for this patient cohort.

Fig. 3

Posttreatment day 29 MRD by mpFC for cases without a complete clonal TCRB rearrangement in the corresponding pretreatment sample. (A and B) mpFC data are shown as both (A) frequency of T lymphoblast clone of total CD7+ T/NK cells for comparison with HTS and (B) frequency of total mononuclear cells for comparison with standard clinical enumeration. ETP cases are highlighted in yellow, whereas near-ETP cases are highlighted in green. Case 11 represents a probable γδ-differentiated T-ALL, lacking a clonal TCRB sequence, but having a pretreatment, clonal TCRG sequence.

To evaluate the specificity of HTS for assessment of MRD, we evaluated the frequency of identifying the specific clonal sequence from each MRD-positive patient in the other 42 patient samples at day 29 (table S2). This analysis demonstrated 17 “false positives” in 1512 TCRG comparisons and 3 “false positives” in 1344 comparisons for TCRB. This false-positive rate reflects not the background frequency of a clonal sequence in a given individual but the proportion of individuals who by chance might have at least one T cell with an identical sequence of TCRB or TCRG. The level of these clones was generally much lower by about one order of magnitude on average compared to true MRD positives (table S2).


Clinical management of patients with T-lineage acute lymphoblastic leukemia is dependent on accurate risk stratification. Recent data from the AIEOP-BFM-ALL 2000 study, a multi-institutional, prospective clinical trial involving 464 patients with T-ALL, confirmed the importance of MRD to define standard, intermediate, and high-risk disease on the basis of MRD at two time points after therapy (2). Patients with detectable MRD greater than 10−3 at the second time point (TP2) (day 78) fared poorly, and persistent MRD at TP2 was concluded to be an important predictive factor for T-ALL relapse. The HTS technology described here, which can detect MRD with a frequency below 10−5, may further improve MRD-dependent risk stratification.

Our data support TCRB sequencing for MRD assessment due to the relatively greater germline diversity and therefore specificity versus TCRG. This finding is supported by previous studies comparing analysis of TCR gene rearrangements at diagnosis and relapse for T-ALL (19, 20). TCRG clones were less specific because cross-patient, coincidental sequences were more commonly seen versus TCRB, but TCRG HTS could still be of value, particularly in cases in which the cancer clone has a TCRG rearrangement, but not a TCRB rearrangement.

In our analyses of day 29 samples, clonal TCRB sequences were found to be highly specific, with coincidental identification of cross-patient sequences being generally rare. It is important to recognize that current molecular protocols using patient-specific primers do not routinely test for cross-patient specificity (8, 9), and it is likely that such studies would also identify rare, coincidental overlap of TCR gene sequences from one patient to another. Notably, the level of these clones detectable in the false-positive, cross-patient comparisons was lower in frequency on average than seen in true patient MRD cases (table S2). For T-ALL MRD, TCR gene rearrangements appear generally stable with limited evidence for clonal evolution at relapse (20, 21). Consequently, HTS for identification of MRD should be feasible and specific for most of the patients with T-ALL.

Our findings suggest that HTS may also contribute to T-ALL diagnosis (17, 18) by identifying those cases with a high proportion of T lymphoblasts in the absence of a clonal, complete TCRB gene rearrangement. In our cohort, these cases tended to have either an ETP immunophenotype (17) or a similar immunophenotype that we describe as being near-ETP. These cases also tended to have increased, detectable MRD at day 29 as measured by mpFC compared to non-ETP T-ALLs.

The ETP subtype of T-ALL has been proposed to be derived from an early precursor subset of T cells that have not yet undergone TCR gene rearrangement, although a recent study suggests that a subset may also represent reversion from the double-positive, immature T cell state to a more immature immunophenotype (22). The presence of TCRG gene rearrangements in a subset of these tumors (17), as seen here (patients 54 and 57), would seem to support the reversion hypothesis for some patients. Although we did not sequence TCRD in this study and cannot exclude the possibility of incomplete rearrangements (D-to-J rearrangement without V-to-DJ) in TCRB with these data, comparison with flow cytometry results of MRD at day 29 shows that the 12 cases for which no clonal TCRB sequence was identified at pretreatment had increased levels of MRD at day 29 compared to non-ETP cases (fig. S3), consistent with the more aggressive clinical course reported for these ETP tumors. The similarity of increased levels of MRD at day 29 of these near-ETP tumors suggests a possible, expanded definition of the ETP category, but definite conclusions will require correlation with clinical outcome data that are not available at present.

The use of HTS for sequencing of lymphoid genes in T-ALL should be directly extendable to routine clinical monitoring of acute B lymphoblastic leukemia/lymphomas (B-ALLs) because MRD by molecular and mpFC has previously also been shown to be important for patient prognosis (3). B-ALL commonly shows cross-lineage rearrangements of TCRB and/or TCRG, so the approach used in this work should be directly applicable to a subset of cases (1). Similarly, HTS of immunoglobulin heavy chain gene rearrangements can readily be performed, and studies of a separate B-ALL cohort are currently under way. We also anticipate that HTS of lymphoid genes could be informative for the diagnosis and MRD surveillance of mature B and T cell lymphomas.

Although our findings provide evidence that HTS of TCR gene receptors could potentially affect clinical detection of MRD, our approach is suitable only for those patients whose tumors have a detectable clonal rearrangement at diagnosis in either TCRB or TCRG genes. As configured, our TCRB assay detects only complete clonal TCRB gene rearrangement and does not detect incomplete D-to-J rearrangements. This may in part explain the decreased proportion of cases that we report having a clonal TCRB rearrangement (31 of 43, 72%) compared to that in the literature—about 85% for complete TCRB rearrangements (23). Further, as currently designed, our TCRG primers likely miss a subset of clonal rearrangements because we do not sequence pseudogenes that account for about one-third of possible TCRG gene rearrangements. To create a clinically viable assay for TCRG, we will need to include primers to amplify these pseudogenes in our assay. This may explain in part the relatively poorer performance of the TCRG assay (27 of 43 cases, 62.7%) versus that generally accepted in the literature at about 80 to 90% (1). Notably, our TCRG assay currently cannot detect the absence of biallelic TCRG deletion reported recently to be associated with poor prognosis in a subset of T-ALL patients (18).

Application of this HTS for MRD detection is additionally limited by the amount of sample material available for sequencing analysis. To enhance our sensitivity for detecting a malignant clone in posttreatment samples, we sequenced an increased number of cells compared to pretreatment samples. In two patients, cases 49 and 63, there was insufficient material for us to sequence 200,000 cells. Accordingly, our a priori ability to detect MRD was limited for these patients. Nevertheless, for patient 49, MRD was still detected by HTS at day 29. For patient 63, MRD was not detected by HTS, but this case was also negative by concurrent mpFC analysis. To ensure adequate specificity and to exclude sequencing errors, we oversampled to ensure that clonal sequences representing posttreatment MRD were identified in at least 10 sequencing reads. Thus, as with other quantitative MRD assays, the limit of detection of our assay is determined by input DNA, whereas the precision of our measurement is dependent on the redundancy of sequencing reads (fold coverage).

Lastly, the ability to generalize our finding of the absence of a complete clonal TCRB gene rearrangement at diagnosis with ETP and near-ETP immunophenotype and increased propensity for MRD requires further study. In our cohort, we identified a proportion of ETP cases similar to that reported previously (our cohort had 5 of 43, 11.6%, versus previous report showing 30 of 239, 12.6%) (17). Whether the absence of a complete clonal TCRB rearrangement was a prominent feature in the previous report is unclear based on the published data. Similarly, although our findings raise consideration of near-ETP immunophenotype being potentially related to ETP tumors, having increased MRD at day 29 (see fig. S3), additional study is needed to confirm this finding.

Despite these limitations, our findings suggest the potential of HTS of lymphoid receptor genes to enhance detection of MRD in many cases compared to mpFC, the current clinical standard of care used in the United States for MRD assessment in T-ALL. In addition, the use of HTS for MRD detection has many advantages over more classical patient-specific PCR-based approaches. HTS can contribute to detection of clonality without the laborious requirement of individualized PCR or allele-specific oligonucleotide design, validation, and implementation. HTS is readily scalable and less organizationally complex and should be adaptable to any clinical laboratory environment once clinically focused HTS instruments become available. For example, both the Illumina MiSeq and the Life Technologies Ion Torrent can sequence a few million TCR reads in a matter of hours. HTS could therefore prove to be a compelling alternative for the patient-specific, molecular assessment of MRD of lymphoid neoplasms compared to the complex, oftentimes multi-institutional infrastructure currently required for today’s molecular monitoring of MRD. Additional studies, however, will need to be performed to make this determination and to ensure that this technology meets the high clinical standards of current-generation molecular diagnostic tests.

Materials and Methods

Sample preparation

Pretreatment (peripheral blood or bone marrow aspirates) and day 29 posttreatment (bone marrow aspirate) samples were from patients diagnosed with T-ALL. Samples were submitted for flow cytometry analysis at the University of Washington as part of the AALL0434 COG trial protocol. Residual samples for 43 patients were collected consecutively after mpFC analysis and were further de-identified and submitted for HTS. All patients provided informed consent as part of the COG trial for the use of their residual samples.

Flow cytometry

mpFC was performed at the University of Washington as part of the routine evaluation for MRD (24). Flow cytometry data were reviewed by B. Wood who further classified samples as ETP subtype as previously described (17). In addition, cases in which the immunophenotype met all but the CD5 expression criteria for ETP were designated as near-ETP. All samples were labeled with each of two antibody combinations: (i) CD8 BV421, CD2 FITC (fluorescein isothiocyanate), CD5 PE (phycoerythrin), CD34 PE-TR (Texas Red), CD16 + CD56 PE-Cy5, CD3 PE-Cy7, CD4 A594, CD7 APC (allophycocyanin), and CD45 APC-H7, and (ii) CD16 V450, cytoCD3 FITC, CD7 PE, CD56 PE-Cy5, CD5 PE-Cy7, CD38 A594, CD3 APC, and CD45 APC-H7. Samples were processed with standard methods for surface (tube 1; NH4Cl + 0.25% formaldehyde) or surface and cytoplasmic [tube 2; Fix and Perm (Invitrogen)] antigens, and 750,000 events were acquired on a Becton-Dickinson LSRII. Clusters of events that differed from normal T cell maturation were designated MRD and quantified relative to total mononuclear cells and CD7+ T/natural killer (NK) cells. Data were analyzed with Woodlist software version 2.7. Equivalent software is publicly available.

High-throughput sequencing

Sequencing CDR3 regions. TCRG and TCRB CDR3 regions were amplified and sequenced from 400 ng of pretreatment and 1200 ng of day 29 posttreatment samples, or in a subset of cases with all available extracted DNA. Amplification and sequencing of TCRB CDR3 regions were performed on the ImmunoSEQ platform at Adaptive Biotechnologies as previously described (11), and amplification and sequencing of TCRG CDR3 regions were performed as previously described (13). The sequences for both the TCRB and the TCRG CDR3 regions were delineated according to the definition established by the International ImMunoGeneTics collaboration (25). Sequences that did not match CDR3 sequences were removed from the analysis. A standard algorithm was used to identify which V, D, and J segments contributed to each TCRB CDR3 sequence and which V and J segments contributed to each TCRG CDR3 sequence (25).

Identifying CDR3 sequences. Ten milliliters of blood was isolated from six healthy control individuals to define a normal proportion of reactive T cells. Peripheral blood mononuclear cells (PBMCs) were isolated and DNA-extracted, and 1200 ng of extracted DNA was used to amplify TCRB and TCRG sequences with the same protocol as the day 29 posttreatment T-ALL samples. The frequency of the most common αβ T cell–reactive clones in PBMCs from these healthy individuals averaged 3.8% of the total repertoire with an SD of 3.5%. For tumor cases, because αβ T cells carry both rearranged TCRG and TCRB CDR3 chains (13), the frequencies of both the highest copy TCRB and the TCRG CDR3 chains were assumed to represent the neoplastic T lymphoblasts. For pretreatment samples, TCRB CDR3 sequences that constituted a proportion greater than 19% of all sequences [equivalent to four SDs above the mean (3.8%) of reactive T cell clones in normal individuals] were considered by our definition to be a clonal population. For samples in which the two most common TCR sequences were of comparable frequency, both sequences were considered as being the probable cancer clone. Posttreatment MRD at day 29 was identified by searching for CDR3 sequences that identically matched the clonal sequence derived from analysis of paired, pretreatment samples, requiring an exact 60–base pair match. Both the presence and the frequency of the MRD clone relative to that of the total TCR repertoire were noted.

To determine whether the clonal CDR3 sequences identified were specific for the patient’s neoplastic clone, we screened all day 29 posttreatment samples derived from other patients for the presence and frequency of all identified clonal TCR CDR3 sequences.

Statistical methods

We used a Fisher’s exact test in fig. S2 to test the null hypothesis that the presence of a rearranged TCRB gene is independent of ETP and near-ETP phenotype.

Supplementary Materials

Fig. S1. Frequency of reactive T cells by TCRB sequencing in normal individuals.

Fig. S2. Association of TCRB gene rearrangement with T-ALL subtype by mpFC.

Fig. S3. MRD assessment by mpFC at day 29 posttreatment.

Table S1. Pretreatment T lymphoblast immunophenotype and extent of disease involvement and MRD at day 29 posttreatment by mpFC.

Table S2. Cross-patient identification of TCRB and TCRG clones in other posttreatment samples.

References and Notes

  1. Acknowledgments: We thank M. G. M. Wilson for excellent technical support. Funding: Supported by the Department of Laboratory Medicine, University of Washington (D.W., J.R.F., H.A.G., D.E.S., and B.L.W.); Becton, Dickinson and Company, NJ, for flow cytometry reagents (B.L.W.); and Children’s Oncology Group grant U10 CA98543 (B.L.W., S.S.W., K.P.D., and M.L.L.). Author contributions: D.W., A.S., J.R.F., H.A.G., D.E.S., B.L.W., and H.R. designed the experiments and analyzed the data; S.S.W., K.P.D., and M.L.L. contributed cases; D.W., A.S., B.L.W., and H.R. drafted the manuscript; all authors reviewed and edited the manuscript. Competing interests: H.R. has consultancy, equity ownership, patents, and royalties with Adaptive Biotechnologies; A.S. has employment and equity ownership with Adaptive Biotechnologies; D.W., H.R., A.S., and Adaptive Biotechnologies have applied for research funding, pending review. Adaptive Biotechnologies has submitted for patent application on this technology, entitled “Diagnosis of lymphoid malignancies and minimal residual disease detection,” application 61569118. The other authors declare that they have no competing interests.
View Abstract

Navigate This Article