Personalized analysis of circulating tumor DNA in childhood malignancies

Abstract

Circulating tumor DNA (ctDNA) analysis has emerged as a powerful technique for precision medicine in pediatric oncology. ctDNA refers to small fragments of tumor-derived genetic material that are released into the bloodstream. ctDNA analysis can enable sensitive monitoring of treatment response, early detection of relapse, and detection of minimal residual disease. Previous studies have used generic cancer gene panels for ctDNA analysis in children which has proven to be challenging in pediatric patients who differ in genotype from adult cancers. In paper I, five patients with relapsed or refractory ALK-driven neuroblastoma were treated with lorlatinib. All patients showed a good response and four were alive three to five years after lorlatinib was started. To investigate if ctDNA analysis could provide a more sensitive method for monitoring residual disease we designed an ALK sequencing panel for sensitive detection of the ALK mutations found in this patient group as well as those mediating resistance to tyrosine kinase inhibitors. ctDNA levels showed a stepwise decline and became negative during lorlatinib treatment in all patients. For two patients with relapsed disease, an increase in ctDNA levels could be seen three and nine months before the clinical relapses, respectively. In paper II and III, we developed personalized sequencing panels for patients diagnosed with neuroblastoma and rhabdomyosarcoma, respectively. We used an ultrasensitive sequencing technique for retrospective longitudinal ctDNA analysis to evaluate the use of ctDNA as a tumor marker. In both studies, ctDNA analysis of patient plasma showed that ctDNA levels at diagnosis correlated with risk group and disease burden. For patients with good response to treatment, ctDNA levels gradually decreased. Follow-up samples from patients who were free from disease were all ctDNA negative. Levels of ctDNA were elevated in all cases of relapse or disease progression. In paper II, one case showed detectable ctDNA 78 days before the clinical relapse. In paper III one patient showed elevated ctDNA levels in several samples five months prior to clinical relapse. In summary, we show that we can accurately and sensitively monitor disease burden using our ALK panel in patients with ALK-driven neuroblastoma and lorlatinib monotherapy shows long-lasting results in this patient cohort. We show that personalized ctDNA analysis is a sensitive and accurate method for longitudinally monitoring of disease burden and detection of relapse in pediatric neuroblastoma and rhabdomyosarcoma, regardless of tumor genetics. ctDNA analysis showed higher sensitivity for relapse detection and treatment monitoring than two serum markers and three urine markers currently used in the clinic for neuroblastoma. As there are no biomarkers in use for rhabdomyosarcoma and the disease is often highly genetically heterogenous showing poor prognosis in metastasized disease, personalized ctDNA analysis could prove a useful complement in disease monitoring and relapse detection also for these patients.