Abstract
Background: We aimed to retrospectively investigate the role of neoadjuvant chemotherapy in low-risk patients with hepatoblastoma (HB) who underwent curative resection between February 2009 and December 2017. We also verified the feasibility of the risk stratification system to select the optimal patients for upfront resection.
Procedure: We compared 5-year overall survival (OS) and event-free survival (EFS) between the upfront surgery (US) (n=26) and neoadjuvant chemotherapy (NC) (n=104) groups at three oncology centers in Beijing, China. To reduce the effect of covariate imbalances, propensity score matching (PSM) was used. We explored whether preoperative chemotherapy affected surgical outcomes and identified the risk factors for events and death, including resection margin status, PRETreatment EXTent of disease stages, age, sex, pathology classification, and α-fetoprotein levels.
Results: The median follow-up period was 64 months (interquartile range 60–72). After PSM, 22 pairs of patients were identified and the patient characteristics were similar for all variables included in propensity score matching. In the US group, the 5-year EFS and OS rates were 81.8% and 86.3%, respectively. In the NC group, 5-year EFS and OS rates were 81.8% and 90.9%, respectively. No significant differences in EFS or OS were observed between the groups. Pathological classification was the only risk factor for death and disease progression, tumor recurrence, diagnosis of other malignant neoplasms, and death from any cause (p =0.007 and p =0.032, respectively).
Conclusion: Upfront resection can achieve long-term disease control in low-risk patients with resectable HB, thus reducing the cumulative toxicity of platinum-based chemotherapy drugs.
Introduction
Hepatoblastoma (HB) is a rare disease with an incidence of 1.6/million children. However, it is the most common malignant pediatric liver cancer, which usually develops in patients aged <3 years, and its incidence has increased by >5% annually1,2. With improvements in surgical techniques and the use of adjuvant chemotherapy, the 5-year survival rate of HB has increased from approximately 35% (50 years ago) to 80%–90% (currently) 3.
The primary treatment modality for HB is complete surgical resection; however, 60%–70% patients are inoperable during diagnosis because of a high tumor bulk volume or major blood vessel invasion4. Systemic cisplatin-based chemotherapy is effective for reducing tumor volume in patients with HB and can convert most unresectable tumors into resectable tumors. Therefore, the International Childhood Liver Tumors Strategy Group (SIOPEL) prefers the use of neoadjuvant chemotherapy (NACT) and delayed resection to facilitate tumor resection 4. However, the Children’s Oncology Group (COG) trial AHEP0731 and Japanese Study Group for Pediatric Liver Tumors (JPLT) study indicated that upfront resection in selected HB cases achieved excellent outcomes 5,6. Unnecessary NACT may result in potential exposure to chemotherapy and treatment-related toxicity, and an increased number of chemotherapy cycles are associated with chemotherapy resistance7-9. However, the value of upfront resection has not been established, and there is no clear consensus regarding which patients with HB benefit the most from this form of treatment.
Due to the lack of a uniform staging system, it is difficult to interpret and compare the results reported by different research groups, and it creates difficulties in terms of optimal patient selection. Four major cooperative trial groups (SIOPEL, COG, the German Society for Pediatric Oncology and Hematology, and JPLT) formed the Children’s Hepatic Tumors International Collaboration (CHIC) to define a common hepatoblastoma stratification (CHIC-HS), which makes the heterogenous results of prior research more comparable 10.
In this retrospective, multicenter study, we compared event-free survival (EFS) and overall survival (OS) in CHIC-HS low-risk patients with HB who underwent surgical resection and NACT and verified the feasibility of the risk stratification systems used to select the optimal treatment for patients with HB.
Methods
The study was performed in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology guidelines. This study complied with the tenets of the Declaration of Helsinki and the Capital Institute of Pediatrics’ (CIP) ethics committee approved this study (SHERLL2022047). Due to the retrospective nature of the study, the need for informed consent was waived.
Patient selection
The medical records of patients with HB who underwent curative resection with or without NACT at CIP (Beijing, China) and two other centers between February 2009 and December 2017 were collected. We classified patients into upfront surgery (US) and NACT (NC) groups, based on whether they had received NACT or not. The CHIC-HS system assigned patients to four risk groups based on age, serum α-fetoprotein (AFP) levels, and PRETreatment EXTent of disease stages (PRETEXT) stage and its annotation factors 11. Eligible patients were stratified into a very low- or low-risk group (Table 1), were <8 years old, and had a histopathologic diagnosis of HB. The tumors were estimated to be resectable at the time of diagnosis, patients had complete clinical and follow-up data, and liver and kidney function were normal. Patients with other tumors or serious medical diseases, who refused surgery, and who refused postoperative chemotherapy were excluded.
Outcomes
The primary outcomes were 5-year OS and EFS. We defined EFS as the time from surgery to tumor recurrence, diagnosis of other tumors, death from any cause, or last follow-up without the occurrence of any of these events.
The secondary objectives were to 1) explore whether preoperative chemotherapy affected surgical outcomes and 2) identify risk factors for events and death, e.g., resection margin status, PRETEXT stages, age, sex, pathology classification, and AFP level at diagnosis. R0 resection was defined as a microscopically negative margin, and R1 resection as macroscopically complete resection with positive microscopic margins.
Statistical analysis
The non-parametric Mann–Whitney U test was used to compare non-normal data between groups. All tests were two-sided, and ap -value of <0.05 was considered statistically significant.
To reduce the confounding effects of imbalances in the study covariates, propensity score matching (PSM) was performed. The propensity score (PS) was estimated using a logistic regression model, in which the treatment modality was regressed onto sex, age at surgery, AFP levels at diagnosis, and PRETEXT stage as potential covariates. The US group was PS-matched to the NC group in a 1:1 ratio, using maximum distance (caliper) of 0.15 between matched participants based on their propensity score. The balance in covariates between the groups before and after PSM was evaluated using standardized mean differences (SMDs). SMD <0.2 was deemed to be the ideal balance.
The Kaplan–Meier method was performed to estimate OS and EFS, and a log-rank test was conducted to compare these results among the patient groups.
The relationships between resection margin status, PRETEXT stages, age, sex, pathology classification, AFP levels, and outcome events (events and deaths) were analyzed using logistic regression, since we only studied the effects of variables on death and events, but not on the length of survival. Statistical significance was set at two-sided p -value <0.05.