Magnetic resonance imaging (MRI) is emerging as a critical diagnostic tool for identifying pediatric acute lymphoblastic leukemia (ALL) in children who present with persistent musculoskeletal pain but maintain normal blood counts. A retrospective observational study conducted between 2006 and 2024 at a single institution analyzed 424 newly diagnosed pediatric ALL cases and found that 20 patients (4.7%) initially presented with suspicious MRI findings that triggered the subsequent hematologic evaluation. In this diagnostically challenging subgroup, more than one-third of patients, specifically 7 out of 20, displayed a completely normal complete blood count (CBC) at the time of presentation, highlighting the potential for conventional laboratory screening to produce falsely reassuring results in early-stage disease.
The study sought to clarify the role of MRI in the diagnostic pathway for childhood ALL, particularly when clinical symptoms lack hematological indicators. Patients included in the cohort predominantly presented with persistent bone pain, which served as the leading symptom preceding imaging in 18 of the 20 cases. Because these musculoskeletal complaints often overlap with inflammatory conditions such as osteomyelitis or juvenile idiopathic arthritis, many patients experienced significant diagnostic delays. The mean interval from the onset of symptoms to the initial MRI was 20.8 days, with some cases experiencing delays of up to 65 days. These intervals illustrate the clinical difficulty of distinguishing leukemic infiltration from common orthopedic or developmental issues in pediatric populations.
Radiological re-evaluation revealed that diffuse bone marrow signal alterations were the most consistent findings. Every patient in the musculoskeletal imaging group demonstrated T1-weighted (T1w) bone marrow signal loss. Furthermore, 16 of the 17 patients in this group showed proton-density-weighted (PDw) or T2-weighted short tau inversion recovery (T2w-STIR) hyperintensity. In cases where trunk imaging was performed, diffusion-weighted imaging (DWI) consistently revealed high signal intensity with low apparent diffusion coefficient (ADC) values, providing additional radiological markers of marrow infiltration. Standard T2-weighted turbo spin-echo sequences, however, contributed less to the identification of the underlying malignancy compared to these other protocols.
The diagnostic interval from MRI to the eventual hematologic confirmation varied substantially, with a mean of 18.4 days and a range extending up to 180 days. This variability was frequently attributed to the misinterpretation of initial MRI scans, which were sometimes incorrectly labeled as osteomyelitis or even interpreted as normal. In some instances, patients received unnecessary antibiotic or anti-inflammatory treatments while the malignancy remained undiagnosed. The researchers noted that one patient with persistent back pain underwent an MRI that showed diffuse marrow signal abnormalities, yet due to the absence of peripheral blood blasts and negative flow cytometry, the leukemia diagnosis was only established six months later after a second scan prompted by worsening symptoms.
The study underscored the importance of recognizing specific “red flag” imaging patterns, such as diffuse or confluent T1w marrow hypointensity, especially when paired with diffusion restriction or multifocal lesions. Such findings should prompt urgent referral to pediatric oncology, even when initial blood work does not indicate leukemia. The researchers suggest that targeted sampling of bone marrow at MRI-abnormal sites may be necessary in cases where initial iliac crest aspiration is negative but suspicion remains high. This approach was exemplified by a patient whose iliac aspirate showed no blasts, but an MRI-guided biopsy from a site identified as abnormal by imaging successfully confirmed B-cell precursor ALL.
Genetic analysis revealed that all 20 patients in the cohort had B-cell precursor ALL, with no identified cases of T-cell ALL or acute myeloid leukemia presenting via these specific MRI abnormalities. Standardized genetic testing identified common alterations including ETV6::RUNX1 fusion, hyperdiploidy, and various other fusions such as TCF3::PBX1 and PAX5::SOX5. The researchers observed that there was no association between a specific genetic subtype and the clinical manifestation of musculoskeletal symptoms, nor did the presence of radiological marrow alterations serve as a predictive marker for treatment response or the incidence of treatment-related complications such as osteonecrosis.
Limitations of the study include its retrospective, single-center design and the small sample size, which restricts the generalizability of the findings. Because the study did not include a control group of patients with inflammatory or orthopedic conditions who did not have leukemia, the researchers emphasized that the reliability of these MRI findings in distinguishing ALL from other marrow alterations cannot be determined through this analysis alone. Nevertheless, the study provides evidence that radiological markers of marrow infiltration can appear early in the course of pediatric ALL, often predating hematological markers by weeks or even months.
Future prospective multicenter research is needed to validate these MRI patterns and refine the proposed clinical diagnostic pathway. Such studies could help establish age-appropriate quantitative ADC thresholds that would reduce observer-dependent misinterpretation and provide more objective evidence of leukemic infiltration. By standardizing MRI descriptors and scoring systems, clinicians may be better equipped to detect leukemia earlier in children with ambiguous musculoskeletal presentations, potentially shortening the duration of the diagnostic process and ensuring that urgent oncologic assessment is not delayed by normal blood counts.
The findings ultimately suggest that a normal CBC should not be the sole determinant for delaying an oncologic evaluation when clinical presentation and MRI results are suspicious for leukemia. In children with persistent, unexplained musculoskeletal pain, the presence of diffuse T1w bone marrow signal loss should be considered a potential early indicator of systemic malignancy. Incorporating MRI into the clinical algorithm as a complementary tool when blood counts are inconclusive may allow for more rapid intervention, potentially improving the management of children who fall into this diagnostically challenging category of ALL presentation.
Beyond the overall patient count, the demographic breakdown and imaging distribution provide further insight into the clinical makeup of the cohort. Among the 20 evaluated children, 12 were male and 8 were female, reflecting a median age at diagnosis of 7.8 years across an overall age range of 1 to 17 years. The diagnostic imaging was not confined to a single scanning center; while 8 patients were imaged directly at the treating institution, 12 underwent scans across five referring hospitals and two outpatient radiology practices prior to tertiary admission. Across the entire cohort, clinicians evaluated a total of 24 separate MRI examinations. To maintain diagnostic rigor across these disparate settings, all imaging studies were retrospectively re-evaluated by a pediatric radiologist possessing over three decades of clinical experience in interpreting pediatric magnetic resonance examinations. None of the children in the cohort had been referred for imaging with a clinical suspicion of an underlying leukemia or hematological malignancy, underscoring how non-oncologic initial presentations routinely prompted these scans.
Detailed laboratory assessments gathered prior to imaging illustrate why hematologic diagnoses proved elusive in these patients. Although 7 children had an entirely normal complete blood count, 13 of the 20 patients exhibited normal white blood cell counts, and an isolated mild anemia was the sole hematologic anomaly identified in another 6 cases. Across the entire cohort, the mean leukocyte count was 8.54 × 10⁹/L with a range of 1.2 to 33 × 10⁹/L, while the mean hemoglobin level was 10.1 g/dL with a range of 3.4 to 13 g/dL. Platelet counts averaged 270 × 10⁹/L, spanning from 31 to 812 × 10⁹/L. Inflammatory markers demonstrated mild to moderate elevation, with mean C-reactive protein measuring 6.14 mg/dL and mean lactate dehydrogenase reaching 462 U/L. Peripheral blood blasts were detected in only 8 of the 20 cases, yielding a cohort mean of 6.2 percent, whereas subsequent bone marrow blast counts averaged 59.4 percent and reached 91 percent in aspirates, with one patient diagnosed solely via a bone biopsy showing 75 percent blast infiltration.
The diagnostic complexity is prominently exemplified by the medical trajectory of a four-year-old boy with Down syndrome designated as Patient 12. Presenting with pain and swelling localized to the right elbow joint, he was initially treated for presumed osteomyelitis with clindamycin and cefuroxime after initial imaging revealed arthritis and synovitis. Because antibiotic therapy provided no clinical improvement, re-evaluation of the imaging approximately one week later revealed extensive T1-weighted marrow hypointensity and T2-weighted hyperintensity, which prompted urgent bone marrow sampling. This patient had actually undergone an aspirate and biopsy four months earlier to investigate transient bicytopenia, showing leukocytes of 3.2 × 10⁹/L and platelets of 50 × 10⁹/L. At that earlier juncture, marrow samples revealed no leukemic blasts, and mild signs of dysplasia and hypocellularity had led clinicians to suspect a myelodysplastic syndrome instead, showing how rapidly marrow infiltration can evolve.
Technical sequence comparisons also provided valuable methodological insights regarding which imaging acquisitions yield diagnostic clarity. In the musculoskeletal imaging group, protocols reliably combined standard T1-weighted turbo spin-echo sequences with proton-density-weighted or T2-weighted short tau inversion recovery acquisitions. In contrast, diffusion-weighted imaging was incorporated in only three musculoskeletal studies but was deployed in all three chest and abdominal evaluations. In those trunk examinations, diffusion-weighted imaging demonstrated high signal intensity with low apparent diffusion coefficient values across affected compartments, while specialized T2-weighted sequences such as HASTE and TrueFISP delineated multifocal tumor masses within the peritoneal cavity and kidneys. Importantly, the researchers noted that the diagnostic contribution of additional contrast-enhanced T1-weighted sequences, which were administered in nine patients, was minimal, as gadolinium contrast did not uncover any additional findings relevant to establishing the diagnosis or planning clinical treatment.
Subject of Research: Diagnostic role of MRI in pediatric acute lymphoblastic leukemia for patients with non-diagnostic blood counts
Article Title: MRI paving the way to diagnosis of acute lymphoblastic leukemia in children and adolescents – a retrospective observational study
Article References: Lipp, J., Schaper, J., Rummel, F., Kural, M., Borkhardt, A., Babor, F., & Brozou, T. (2026). MRI paving the way to diagnosis of acute lymphoblastic leukemia in children and adolescents – a retrospective observational study. Molecular and Cellular Pediatrics, 13(1), Article 41. https://doi.org/10.1186/s40348-026-00253-0
Image Credits: AI Generated
DOI: 10.1186/s40348-026-00253-0
Keywords: pediatric oncology, acute lymphoblastic leukemia, MRI diagnosis, musculoskeletal pain, bone marrow infiltration, diagnostic delay, T1w hypointensity, diffusion-weighted imaging, apparent diffusion coefficient, B-cell precursor ALL, retrospective study, clinical pathways, radiological findings, hematology, normal blood counts, diagnostic challenges
News Source: Nathaniel Bowman. (November 18, 2026). MRI Patterns May Detect Early Childhood Leukemia Without Abnormal Blood Counts. Scienmag.



