Clinical Validation of Liquid Biopsy for Faster Molecular Diagnosis of EBV-Positive Burkitt Lymphoma

Burkitt Lymphoma (BL) is common in sub-Saharan Africa, but its diagnosis is often delayed. Here, we comprehensively evaluate blood-based liquid biopsies from 377 children and young adults with clinically suspected lymphoma for diagnostic accuracy, yield and turnaround time (TAT). Following extensive pathology capacity building at participating study pathology sites to streamline and quality-ensure the histopathology assessment, diagnoses were reached from best local pathology consisting of tissue morphology, a previously validated limited immunohistochemistry (IHC) panel and review by at least two histopathologists, together called the “gold standard”. Next, we used a combination of clinical characteristics and circulating tumour DNA attributes ( MYC mutations, MYC-IG translocations, and EBV fragmentomics) to train five different diagnostic models based on penalised logistic regression and 10-fold cross-validation on 212 children and young adults with clinically suspected lymphoma. External real-world validation was conducted on the best-performing model on a prospective cohort of 56 children and young adults with clinically suspected lymphoma. Diagnostic accuracy, yield and TAT for the liquid biopsy test were compared head-to-head against the gold-standard in 58 children and young adults with clinically suspected lymphoma. All five models showed good discrimination for BL (AUC ≥ 0.8). The comprehensive model had the best overall performance with AUC of 0.94 (95% CI: 0.895–0.974; p = 4.86×10⁻⁴), sensitivity of 0.83, and specificity of 0.95 confirmed by external validation (AUC of 0.98 (95% CI: 0.942–1)). Liquid biopsy was the only diagnostic result available in the multi-disciplinary team meeting in 42% (26/61) of participants and reduced the median diagnostic TAT from 46.8 days to 6.5 days (p = 4.42e-10) compared to gold standard. This study demonstrates that our liquid biopsy model enables fast, highly accurate molecular diagnosis of EBV-positive BL. Its future integration into the diagnostic pathway may increase diagnostic accuracy and minimise treatment delays in resource-limited settings.