Methods: Epithelial-type hepatoblastoma responses to neoadjuvant chemotherapy were analyzed among patients (n=23) treated at two children’s hospitals between 1996 and 2010. Using digital imaging data, ellipsoid and point-based models were created to measure tumor volume regression and respective distances from tumor borders nearest to the COHV, PVB, and IVC. Changes over time in tumor volumes and distances were evaluated using mixed models analysis of variance and monotonic regression. Benchmark 1cm vascular margins were compared before and after neoadjuvant chemotherapy using McNemar’s test. Features predictive of achieving a 1cm margin were determined using logistic regression.
Results: Hepatoblastoma volumes regressed with incremental neoadjuvant chemotherapy cycles (p<0.001). Although tumor borders regressed away from the COHV (p=0.008), on average only 1.1mm was gained. No change from tumor borders to the PVB was detected (p=0.102). Distances from tumor borders to the IVC remained stable at one hospital (p=0.612) but increased only 0.15mm every 10 days of therapy at the other (p=0.002). For the COHV, 11 of 23 (47.8%) patients had a vascular margin >1cm at baseline and 17 (73.9%) patients had a margin >1cm at conclusion of neoadjuvant chemotherapy (p=0.058). For the PVB, 11 of 23 (47.8%) patients had a margin >1cm at baseline, and 15 (65.2%) had a margin >1 cm at conclusion of neoadjuvant chemotherapy (p=0.157). For the IVC, 4 of 23 (17.4%) patients had a margin >1cm at baseline and 10 (43.5%) had a margin >1cm at conclusion of neoadjuvant chemotherapy (p=0.034). Using logistic regression, we detected no differences in age at diagnosis, in tumor volumes or margins at baseline, or in COG and PRETEXT stage between patients who achieved a 1cm margin at the conclusion of chemotherapy and those who did not for all three vascular margins analyzed.
Conclusion: Hepatoblastoma volumes regress significantly with increasing neoadjuvant chemotherapy cycles, however, tumors often remain anchored to the major hepatic vascular structures and show marginal improvement in resectability criteria. These findings support the COG aims to facilitate earlier surgical decision-making and to reduce treatment toxicities by limiting the number of neoadjuvant chemotherapy cycles.