Model May Identify Non-ETP T-ALL Patients at High Risk of Poor Response to Treatment

An analysis by members of the FDA showed that time to treatment discontinuation may be a viable end point to assess drug efficacy in the postmarketing setting.
An analysis by members of the FDA showed that time to treatment discontinuation may be a viable end point to assess drug efficacy in the postmarketing setting.
The model appears to predict minimal residual disease status.

The following article features coverage from the 2021 American Society of Clinical Oncology (ASCO) Annual Meeting. Click here to read more of MPR‘s conference coverage.

Researchers developed a gene expression classifier that identified a subset of patients with high risk T-cell acute lymphoblastic leukemia (T-ALL), according to a study presented at the 2021 American Society of Clinical Oncology (ASCO) Annual Meeting.1

According to presenter Lauren K. Meyer, an MD/PhD student at the University of California, San Francisco, the genetic heterogeneity of T-ALL has resulted in a lack of biomarkers and limited biology-based risk stratification.

Instead, a number of clinical features are used for risk stratification, one of the most important being minimal residual disease (MRD) at the end of induction therapy, Ms Meyer said. She pointed to data from the Children’s Oncology Group (COG) AALL0434 trial showing that MRD positivity at day 29 was associated with inferior event-free survival and overall survival in non-early T-cell precursor (ETP) T-ALLs.2

Ms Meyer and her colleagues hypothesized that MRD-positive and MRD-negative non-ETP T-ALLs have distinct transcriptomic signatures, and transcriptomic analysis may provide prognostic and biological insights into this group of patients.

The researchers analyzed RNA sequencing data from 146 diagnostic non-ETP samples from the AALL0434 trial. The team identified a set of genes that differentiated MRD-positive samples from MRD-negative samples, and they converted the gene expression pattern into a risk score for the probability of being MRD positive.

The researchers applied their model to the 146 non-ETP samples, 19 ETP samples, and 24 near-ETP samples. The team found that ETP samples have uniformly high risk scores, which is consistent with the high rate of MRD positivity in this group. In the near-ETP and non-ETP samples, MRD-positive samples had higher risk scores relative to MRD-negative samples.

ETP samples had a relatively uniform expression of 119 genes as well as high risk scores. The non-ETP samples had heterogeneous gene expression and risk scores. However, a subset of the non-ETP samples with the highest risk scores had a gene expression pattern that closely resembled the pattern in ETP samples.

The researchers selected a score cut-off of 50 that corresponds with a 50% chance of being MRD positive. The team found that most ETP samples had risk scores above 50, and non-ETP samples with risk scores above 50 were enriched for being MRD positive.

The researchers then validated their score in samples from the COG AALL1231 trial.3 They found the predictive value was independent of induction steroids used (prednisone or dexamethasone).

Finally, the researchers attempted to adapt the classifier to a targeted gene expression platform. They chose the Nanostring platform and developed a custom assay.

The researchers piloted the assay using 96 samples from the ALL0434 trial and applied the same analytical model to the Nanostring model. The Nanostring data confirmed their previous findings.

“Going forward, we anticipate that this model may be used for more rapid identification of patients at high risk for a poor response to treatment and may allow for earlier therapeutic intervention, even prior to completion of the first month of therapy,” Ms Meyer said.

“In addition, our data suggest that these non-ETP samples with high risk scores are biologically distinct from the remaining non-ETP samples, despite having an identical immunophenotype. We anticipate that these biological differences may be exploited therapeutically to improve response to standard therapy in these patients,” she concluded.

Disclosures: This research was supported by the Rally Foundation, Luke Tatsu Johnson Foundation, National Institutes of Health, and UCSF Research Evaluation and Allocation Committee. Some study authors declared affiliations with biotech, pharmaceutical, and/or device companies. Please see the original reference for a full list of authors’ disclosures.

Read more of MPR’s coverage of the 2021 ASCO Annual Meeting by visiting the conference page.


  1. Meyer K, Roy R, Huang BJ, et al. Targeted gene expression classifier identifies pediatric T-cell acute lymphoblastic leukemia (T-ALL) patients at high risk for end induction minimal residual disease positivity. J Clin Oncol. 2021;39:(suppl 15; abstr 10002). doi: 10.1200/JCO.2021.39.15_suppl.10002
  2. Wood BL, Winter SS, Dunsmore KP, et al. T-Lymphoblastic Leukemia (T-ALL) Shows Excellent Outcome, Lack of Significance of the Early Thymic Precursor (ETP) Immunophenotype, and Validation of the Prognostic Value of End-Induction Minimal Residual Disease (MRD) in Children’s Oncology Group (COG) Study AALL0434. Blood. 2014;124 (21):1. doi:
  3. Combination Chemotherapy With or Without Bortezomib in Treating Younger Patients With Newly Diagnosed T-Cell Acute Lymphoblastic Leukemia or Stage II-IV T-Cell Lymphoblastic Lymphoma. Identifier: NCT02112916. Updated March 12, 2021. Accessed June 6, 2021.

This article originally appeared on Cancer Therapy Advisor