A Real World Analysis of Imaging Compared to Biopsy for Diagnosis of Pediatric Liver Tumors.

Purpose: Pediatric liver tumors are challenging and need biopsy to perform when imaging suggests a malignant tumor, findings are equivocal, or confirmation of certain benign lesions is required. This study aims to evaluate the diagnostic accuracy of real-world imaging of liver tumors, using pathology as the gold standard. Methods: Patients less than 18 years old with a liver tumor and imaging obtained before liver tumor biopsy were retrospectively identified. Liver imaging reports (ultrasound, computed tomography, and magnetic resonance imaging) were collected to evaluate specific diagnoses and to differentiate between benign and malignant lesions. A discrepancy was defined as discordance in the benign or malignant classification between imaging and pathological findings. Results: Among 50 patients, imaging and biopsy findings matched between benign and malignant in 40 cases (80.0%). Of these, 22 were benign and 18 were malignant. In 10 (20.0%) discrepant cases, the most frequent misclassification (14.0%, 7/50) involved imaging diagnosis of malignancy found to be benign on pathology. In three (6.0%) cases, imaging suggested a benign lesion (hepatocellular adenoma, regenerative nodule, and indeterminate benign lesion) while biopsy revealed neuroendocrine tumor, hepatocellular carcinoma, and hepatoblastoma, respectively. Conclusion: Overall, imaging and biopsy diagnoses differed in tumor classification (benign vs. malignant) in 20% in this real world series. In a few cases, imaging suggested a benign lesion without specific diagnostic features, but pathology confirmed malignancy; in others, tumors read as malignant on imaging were benign on biopsy. These findings highlight the importance of biopsy or close follow-up where imaging is not diagnostic.