Pediatr Surg Int (2012) 28:887891 DOI 10.

1007/s00383-012-3139-x

ORIGINAL ARTICLE

Can we predict the prognosis of resectable hepatoblastoma from serum alpha-fetoprotein response during preoperative chemotherapy?
Hiroaki Fukuzawa Naoto Urushihara Koji Fukumoto Maki Mitsunaga Kentaro Watanabe Takeshi Aoba Shinya Yamoto Hiromu Miyake Shiro Hasegawa

Published online: 4 August 2012 Springer-Verlag 2012

Abstract Purpose The objective of this study was to clarify whether the alpha-fetoprotein (AFP) reduction rate during preoperative chemotherapy represents a prognostic factor for hepatoblastoma. Method We divided 14 hepatoblastoma patients who underwent preoperative chemotherapy and curative resection into Group A (no recurrence; n = 10) and Group B (recurrence; n = 4). We then compared AFP levels before and after preoperative chemotherapy between groups. Result Mean AFP level after completing the rst cycle of chemotherapy was reduced to 7.28 % (range 1.236.8 %) in Group A and 17.05 % (range 12.020.5 %) in Group B (p \ 0.05). Mean AFP after total preoperative chemotherapy was reduced to 1.42 % (range 0.078.5 %) in Group A and 7.55 % (range 3.412.4 %) in Group B (p \ 0.02). Eight patients in whom AFP levels decreased [1 log after the rst cycle of preoperative chemotherapy survived without recurrence. Conclusion A large, early decrease in AFP level during preoperative chemotherapy may offer a strong indicator of survival. Patients in whom AFP levels do not decrease easily during preoperative chemotherapy may have increased risk of recurrence and should be followed very closely.

Keywords Hepatoblastoma Prognostic factor AFP Chemotherapy

Introduction Hepatoblastoma is a rare disease, accounting for around 1 % of malignancies in children. Outcomes for hepatoblastoma have been improving with the development of more efcient chemotherapy regimens. The Japanese Study Group for Pediatric Liver Tumor (JPLT) reported a 5-year overall survival rate of 80.9 % [1]. Complete resection is necessary to achieve disease-free survival. However, some cases show recurrence even after curative resection. Prognostic factors for identifying patients at increased risk of residual disease are thus needed. Alpha-fetoprotein (AFP) levels at diagnosis have been reported as a prognostic factor, with initial AFP level \100 ng/ml or [1,000,000 ng/ml associated with worse outcomes [2, 3]. However, the abilities of AFP levels at specic time points and of serial changes in AFP levels to predict outcomes have not been established and have been described in detail in only a few studies [46]. Koh et al. [6] recently reported that AFP response to preoperative chemotherapy may offer a useful prognostic factor. They also noted that an initial favorable AFP response, dened as a [1 log decline in serum AFP level after the rst cycle of chemotherapy was signicantly associated with survival outcome. Similarly, Van Tornout et al. [4] reported that patients in whom AFP levels fail to decrease by at least 2 log during preoperative chemotherapy may show a greater risk of recurrence. The objective of this study was thus to clarify whether the AFP reduction rate during preoperative chemotherapy offers a prognostic factor for hepatoblastoma in our institution.

H. Fukuzawa (&) N. Urushihara K. Fukumoto M. Mitsunaga K. Watanabe T. Aoba S. Yamoto H. Miyake S. Hasegawa Department of Pediatric Surgery, Shizuoka Childrens Hospital, 860 Urushiyama, Aoi-ku, Shizuoka 420-8660, Japan e-mail: fukuzawa1203@yahoo.co.jp

123

888

Pediatr Surg Int (2012) 28:887891

Materials and methods Materials Between January 1991 and August 2011, 19 patients with hepatoblastoma were treated at our institution. Of these, ve cases were excluded from this study. In two cases, primary surgery had been performed without preoperative chemotherapy. In another two cases, metastases remained when hepatectomy was performed. In the nal case, the dosage of preoperative chemotherapy had to be reduced due to severe renal failure. The 14 remaining patients underwent curative surgery after preoperative chemotherapy according to the JPLT protocol. Liver transplantation was not needed at curative operation in any of these cases. We then retrospectively reviewed the medical records for all 14 cases.

used in the JPLT-2 protocol [1]. These regimens were broadly similar. The JPLT91B2 regimen consisted of combination chemotherapy including cisplatin (CDDP) at 80 mg/m2 and tetrahydropyranyl (THP)-adriamycin (ADR) at 30 mg/m2 for 2 days. CITA was a modication of the JPLT91B2 regimen. THR-ADR was administered as a 1-h infusion for 2 days in CITA, whereas a 48-h infusion of THP-ADR was performed for the JPLT91B2 regimen. The precise protocol has been described in the JPLT [1, 7]. Although various regimens were used in individual cases during preoperative treatment, the rst cycle of chemotherapy comprised JPLT91B2 or CITA in all cases.

Results Clinical outcomes

Methods We divided the 14 patients into two groups: Group A, no recurrence (n = 10); and Group B, recurrence (n = 4). To investigate possible correlations between AFP responses to preoperative chemotherapy and outcomes, we reviewed AFP levels in each group at diagnosis after the rst cycle of chemotherapy and after all cycles of preoperative chemotherapy. Responses of AFP levels to preoperative chemotherapy were compared between groups. Next, we divided the 14 patients according to AFP levels showing a decrease of [1 log or B1 log after the rst cycle of chemotherapy. We also divided the 14 patients according to AFP levels showing a decrease of [2 log or B2 log after all cycles of chemotherapy. We then reevaluated whether changes in AFP level before curative operation could represent a prognostic factor. For statistical analysis, the MannWhitney U test and Fishers test were used for group comparisons. Values of p \ 0.05 were considered statistically signicant. We also measured the diameter of these liver tumors bidirectionally before initiating preoperative chemotherapy and after two courses of preoperative chemotherapy. Reduction rates in tumor size were calculated and compared between groups. All study protocols were approved by institutional review board at our hospital. Treatment Chemotherapy was performed based on the JPLT protocol. In the JPLT-1 protocol, the JPLT91B2 regimen was used for preoperative chemotherapy [7]. The CITA regimen was

Of the four patients showing recurrence after curative resection (Group B), lesions were in the lungs in two patients, in the liver in one patient, and in both lungs and liver in one patient. Only one of these four patients survived, with resection of a lung metastasis. The other three patients died despite various treatments. Patient characteristics Clinical characteristics of both groups are summarized in Table 1. Group A comprised seven boys and three girls, while Group B comprised three boys and one girl. In Group A, nine patients had unifocal tumors and one patient had a multifocal tumor. In Group B, three patients had unifocal tumors and one patient had a multifocal tumor. Mean age at the time of diagnosis was 9.9 months in Group A and 13.0 months in Group B. According to the PRETEXT grouping system, one patient was classied as PRETEXT I, ve as PRETEXT II, one as PRETEXT III, and three as PRETEXT IV in Group A. In Group B, two patients were classied as PRETEXT II, and the other two as PRETEXT III or IV. One patient in Group B displayed multiple lung metastases at diagnosis. These metastases totally disappeared during preoperative chemotherapy and the patient was able to undergo curative resection. Mean reduction rates in lesion size were 40.2 % in Group A and 58.7 % in Group B. No signicant differences were seen between groups. The total number of courses of preoperative chemotherapy including JPLT91B2, CITA and others was 3.0 (range 25) in Group A and 3.5 (range 25) in Group B. No signicant difference in the number of courses of preoperative chemotherapy was seen between groups. All patients achieved curative resection histologically.

123

Pediatr Surg Int (2012) 28:887891 Table 1 Background characteristics of study participants Group A Sex Male Female Age at diagnosis (months) Unifocal/Multifocal PRETEXT I II III IV Metastasis at diagnosis Reduction rate of tumour after two courses of chemotherapy (%) Total number of courses of preoperative chemtherapies (JPLT91B1/CITA etc.) 1 5 1 3 0 40.2 (12.4106.2) 0 2 1 1 1 (lung) 58.7 (24.4100.2) 7 3 9.9 (119) 9/1 3 1 13.0 (520) 3/1 Group B

889

Fig. 2 Decreases in AFP level after the rst cycle of preoperative chemotherapy 3.0 (25) 3.5 (25)

Prognostic signicance of AFP Median levels of AFP at diagnosis were 240,000 ng/ml (range 15,900688,000 ng/ml) in Group A and 375,000 ng/ml (range 95,100959,700 ng/ml) in Group B. No signicant difference was seen between groups (Fig. 1). Mean AFP level after completion of the rst cycle of chemotherapy decreased to 7.28 % (range 1.236.8 %) in Group A and 17.05 % (range 12.020.5 %) in Group B, showing a signicant difference between groups (p = 0.0237) (Fig. 2). Mean AFP level after all courses of preoperative chemotherapy decreased to 1.42 % (range 0.078.5 %) in Group A and 7.55 % (range 3.412.4 %) in Group B.

Fig. 3 Decreases in AFP level after all cycles of preoperative chemotherapy

Fig. 1 Serum AFP levels at diagnosis

Again, a signicant difference was identied between groups (p = 0.0088) (Fig. 3). The eight patients in whom AFP levels decreased [1 log after the rst cycle of preoperative chemotherapy all survived without recurrence. Conversely, among the six patients in whom AFP levels decreased B1 log after the rst course of chemotherapy, four patients experienced recurrence (Fig. 4). A signicant difference was apparent between these groups (p = 0.006). The seven patients in whom AFP levels decreased [2 log after all courses of preoperative chemotherapy all survived without recurrence. On the other hand, among the seven patients in whom AFP levels decreased B2 log after all courses of preoperative chemotherapy, four patients

123

890

Pediatr Surg Int (2012) 28:887891

Fig. 4 Magnitude of decreases in AFP level after the rst cycle of preoperative chemotherapy

Fig. 5 Magnitude of decreases in AFP level after all cycles of chemotherapy

experienced recurrence (Fig. 5). Again, a signicant difference was apparent between groups (p = 0.020).

initial AFP levels \100 ng/ml or [1,000,000 ng/ml appear to be associated with poor outcomes [2, 3, 13, 14]. Koh et al. [6] also suggested that the initial AFP level may offer a predictor of disease-free survival, after observing a signicant difference in this outcome measure between patients with initial AFP level above and below the median. However, the present study was unable to identify initial AFP level as a prognostic factor in this series. Other reports [4, 15, 16] have indicated that no strong relationship exists between initial AFP level at diagnosis and outcome. Whether initial AFP level can be used as a prognostic factor thus remains unclear. However, the abilities of AFP levels at specic time points and of serial changes in AFP levels to predict outcomes have not been established and have been described in detail in only a few studies, suggesting that serum AFP response during preoperative chemotherapy may represent a good indicator of prognosis [46]. We conrmed a large early response of AFP to treatment as a strong predictor of good outcome. Patients in whom AFP levels decreased [1 log after the rst cycle of chemotherapy survived without recurrence. In our study, patients received various types of chemotherapy before hepatectomy, but the rst cycle was almost identical in all patients. We can therefore reliably predict better outcomes in patients showing a decrease in AFP of [1 log just after completing the rst cycle of chemotherapy. Even though only a small number of reports have included data regarding outcomes and serial changes in AFP levels before hepatectomy, each has indicated that a large early decrease in AFP level during preoperative chemotherapy appears to offer a strong indicator of survival. In other words, patients showing no substantial decrease in AFP levels after the rst cycle of chemotherapy may have an increased risk of recurrence and should thus be followed very closely.
Conict of interest of interest. The authors declare that they have no conict

Discussion References Recently, outcomes of hepatoblastoma have improved thanks to cisplatin-based chemotherapy and advances in surgical treatment, including liver transplantation. Curative resection is necessary for cure and the outcomes for patients with unresectable and/or metastatic tumor remain poor [811]. Although curative resection was achieved for all cases of hepatoblastoma in this series, some patients developed recurrence. Clarication of factors contributing to an increased risk of recurrence is thus needed. Some reports have identied various prognostic factors. The Childrens Oncology Group has shown small-cell undifferentiated histological subtype as a prognostic factor signicantly associated with increased risk of death [12]. Some reports have noted that
1. Hisiki T, Matsunaga T, Sasaki F et al (2011) Outcome of hepatoblastomas treated using the Japanese Study Group for Pediatric Liver Tumor (JPLT) protocol-2: report from the JPLT. Pediatr Surg Int 27:18 2. Meyers RL, Rowland JR, Krailo M et al (2009) Predictive power of pretreatment prognostic factors in children with hepatoblastoma: a report from the Childrens Oncology Group. Pediatr Blood Cancer 53:10161022 3. Fuchs J, Rydzynski J, Von Schweinitz D et al (2002) Pretreatment prognostic factors and treatment results in children with hepatoblastoma: a report from the German Cooperative Pediatric Liver Tumor Study HB94. Cancer 95:172182 4. Van Tornout JM, Buckley JD, Quinn JJ et al (1997) Timing and magnitude of decline in alpha-fetoprotein levels in treated children with unresectable or metastatic hepatoblastoma are

123

Pediatr Surg Int (2012) 28:887891 predictors of outcome: a report from the Childrens Cancer Group. J Clin Oncol 15:11901197 Lovvorn HN 3rd, Ayers D, Zhao Z et al (2010) Dening hepatoblastoma responsiveness to induction therapy as measured by tumor volume and serum alpha-fetoprotein kinetics. J Pediatr Surg 45:121128 Koh KN, Park M, Kim BE et al (2011) Prognostic implication of serum alpha-fetoprotein response during treatment of hepatoblastoma. Pediatr Blood Cancer 57:554560 Sasaki F, Matsunaga T, Iwafuchi M et al (2002) Outcome of hepatoblastoma treated with the JPLT-1 (Japanese Study Group for Pediatric Liver Tumor) Protcol-1: a report from the Japanese Study Group for Pediatric Liver tumor. J Pediatr Surg 37:851856 Perilongo G, Shafford E, Maibach R et al (2004) Risk-adapted treatment for childhood hepatoblastoma: nal report of the second study of the International Society of Paediatric OncologySIOPEL 2. Eur J Cancer 40:411421 Katzenstein HM, London WB, Douglass EC et al (2002) Treatment of unresectable and metastatic hepatoblastoma: a Paediatric Oncology Group Phase II Study. J Clin Oncol 20: 34383444 Matsunaga T, Sasaki F, Ohira M et al (2003) Analysis of treatment outcome for children with recurrent or metastatic hepatoblastoma. Pediatr Surg Int 19:142146

891 11. Malogolowkin MH, Katzenstein H, Krailo MD et al (2006) Intensied platinum therapy is an ineffective strategy for improving outcome in pediatric patients with advanced hepatoblastoma. J Clin Oncol 24:28792884 12. Meyers LM, Rowland JR, Krailo M et al (2009) Predictive power of pretreatment prognostic factors in children with hepatoblastoma: A report from the childrens oncology group. Pediatr Blood Cancer 53:10161022 13. von Schweinitz D, Byrd DJ, Hecker H et al (1997) Efciency and toxicity of ifosfamide, cisplatin and doxorubicin in the treatment of childhood hepatoblastoma. Study Committee of the Cooperative Paediatric Liver Tumour Study HB89 of the German Society for Paediatric Oncology and Haematology. Eur J Cancer 33:12431249 14. De Ioris M, Brugieres L, Zimmermann A et al (2008) Hepatoblastoma with low serum alpha-fetoprotein level at diagnosis: the SIOPEL group experience. Eur J Cancer 44:545550 15. Browo J, Perilongo G, Shafford E et al (2000) Pretreatment prognostic factors for children with hepatoblastomaresult from the International Society of Paediatric Oncology (SIOP) study SIOPEL1. Eur J Cancer 36:14181425 16. Jung SE, Kim KH, Kim MY et al (2001) Clinical characteristics and prognosis of patients with hepatoblastoma. World J Surg 25:126130

5.

6.

7.

8.

9.

10.

123