A Hopeful Leap Forward by Multicentric Cooperation for Precision-Based Therapy for Very Resistant, Relapsed, or Refractory Childhood Leukemia

International cooperation and evidence-based trial development were key for the fantastic progress with the treatment of acute leukemias that for decades has been based on risk-stratified treatment intensification (1). Clinical trials driven by the necessity to reduce treatment-related mortality in more challenging socioeconomic or geographic conditions showcase the importance of risk stratification also to reduce treatment intensity for a relevant proportion of patients (2). New opportunities arise both to address resistant disease and to reduce treatment toxicity for all patients with immunotherapy, but evidence is building that similar risk factors may apply to identify leukemias that remain resistant to current CD19-directed approaches in B-cell precursor acute lymphoblastic leukemia (B-ALL). Due to the lack of immunotherapeutic approaches in T-cell precursor ALL (T-ALL), relapsed T-ALL will require alternative approaches. Modern oncology builds for decades on the concept of personalized medicine to guide treatment choices based on multiple parameters. With the advances of genomics and systems biology, a more basic concept of precision medicine emerged. In contrast to personalized medicine, a definition for precision medicine is to be “more closely related to disease-specific features and molecular or cellular interactions and to explore how these features impact the patient's diagnosis, prognosis, and outcome in general” (3). The focus thus far was mainly directed on genomic approaches, but translation into clinically actionable targets remains challenging. In Europe, the pediatric INFORM study constitutes a pillar of the development of precision oncology. Likewise, the Australian Zero Childhood Cancer program most recently showed that genomic profiling could identify potentially targetable alterations with evidence of a clinical response for 11 patients from 134 who received molecular tumor board recommendations (4).

In this issue of Cancer Discovery, the LEAP consortium reports that targeted next-generation sequencing identified actionable lesions in 116 patients, although the majority of these were not high-confidence matches (5). With an expert panel and individualized interpretation, 17 patients received treatment with available targeted drugs, mostly off-label, one patient in a clinical study, providing several signals of clinical activity. Most of these targetable lesions would not have been identified with the current standard-of-care diagnostic evaluation performed in the participating institutions. This study provides additional very important points to be considered in the field. In particular for patients with ambiguous-lineage leukemia, the mutational profile may provide relevant information to guide the choice of standard chemotherapy protocol. A typical example of an early T-ALL phenotype with myeloid lineage features is provided. As precise molecular classification becomes increasingly possible (6), new approaches will be needed in clinical trial design to develop benefits of this potential. Most of the high-confidence recommendations for targetable lesions that resulted in interventions encompassed tyrosine kinase inhibitors (TKI). This limitation reflects the difficulty to predict targets based on static genomic information. Interestingly, this study includes selected ex vivo testing of drug sensitivity. Indeed, heterogeneous responses to selumetinib and other inhibitors of the RAS pathway are reported within patients with mutations in that pathway, albeit with no correlation to the variant allele frequency. Similarly, ex vivo sensitivities of leukemias with activating mutations in FLT3 to four targeted TKIs were also heterogeneous, raising the question of whether inclusion of functional assays could contribute to predict on-target activity. An impressive example is provided with the inclusion of a patient with TCF3-HLF–positive ALL, a subtype previously identified to be very dependent on BCL2 and highly responsive to venetoclax in combination with vincristine and dexamethasone in patient-derived xenografts (7). The combination of venetoclax with four drug reinduction resulted in a complete molecular response after multiple lines of therapy for this refractory leukemia, allowing the patient to bridge to CAR-T therapy. Although this patient relapsed after CAR-T therapy, the combination of venetoclax and chemotherapy after second relapse induced remission with low minimal residual disease. Based on a rationale for synergy and the detection of a KRAS mutation in TCF3-HLF ALL, experimental evidence could be obtained in patient-derived xenografts for the combination of venetoclax and selumetinib, a principle of broader relevance provided that patients who are dependent on these pathways can be identified. For this index patient, clinical translation was not yet possible given the context and the lack of safety data. The sum of such efforts will make a difference for patients.

As the genomic landscape of leukemia is now tangible in real time in the clinic, much attention is directed at functional dependencies in cancer, with new opportunities to understand the bottlenecks that enable cancer cells to evade chemotherapy (8). An active community of researchers engages in scientific working groups of the European Hematology Association and the Society for Functional Precision Medicine to develop ex vivo drug-response profiling and other assays to capture the relevant information related to drug activity. Major genomics research institutes join forces to develop a systematic map of cancer vulnerabilities by functional approaches in patient-derived models (9). In a way, we are facing the situation of the pioneers of modern oncology more than 50 years ago. As we honor the memory of a pioneer founder of modern leukemia therapy, Emil Freireich (10), imagine the courage needed to initiate the first trials, based on information from ex vivo assays and from a syngeneic leukemia transplantation mouse model by Skinner at the NCI. The next generation of precision medicine trials is in preparation. Agile platform-based trials are needed to address the challenges and opportunities of the spectrum of vulnerabilities that will be discovered. The LEAP consortium and similar networks have demonstrated that effective multicentric cooperation for precision medicine trials is possible. Even more so, given the complexity and diversity of phenotypes to handle, cooperation between various precision programs worldwide will be needed to accelerate the development of more personalized leukemia treatment. In this respect, the LEAP program does represent a large step forward and lives up to its name. This does come with huge challenges: multiple stakeholders and regulators will have to agree on more agile trial design and new ways to validate clinical observations in subsets with very small numbers of patients. Our thinking around trial design, the ways to postulate hypotheses and to measure outcome will need to be revisited. A direct connection to strong preclinical programs will be needed more than ever, as outlined by the LEAP consortium project. With the progress in patient-derived preclinical modeling and integration of new dimensions such as the Cancer Dependency Map and other systems biology initiatives, new solutions will emerge for patients who are currently not curable. The possibilities are within reach; systematic translation will shape the new treatment types. Funding of this preclinical and clinical trial space by different agencies, industry, and philanthropy is crucially needed to enable faster and more effective development. The field is eager to take on the challenge.

No disclosures were reported.