PRETEXT Annotation factors

C caudate lobe involvement

E extrahepatic abdominal disease

F tumor focality

H tumor rupture or intraperitoneal hemorrhage

M distant metastasis

N lymph node metastasis

P portal vein involvement

V involvement of the inferior vena cava and/or hepatic veins

1 INTRODUCTION

Hepatoblastoma (HB) is a rare hepatic malignant tumor in children that accounts for approximately 1% of pediatric malignant neoplasmas.¹ In 1989, the Japanese Study Group for Pediatric Liver Tumor (JPLT) was launched, and clinical trials (JPLT1-4 studies) for pediatric hepatic malignant tumors have been sequentially carried out since that time.^2-8 In this report, we present the cure rates and toxicities of CDDP (cisplatin, cis-diamminedichloroplatinum) monotherapy as a less intense therapy for patients under 3 years of age in the standard-risk group (resectable tumors, without any annotation factors in the PRETreatment EXTent of disease [PRETEXT] classification). Patients were evaluated using central radiological and pathological reviews.

In the JPLT2 trial, patients with resectable PRETEXT I, II, and III HB were mainly treated with a combination of CDDP and pirarubicin (CITA regimen), and the cure rate (event-free survival) was satisfactory in these standard-risk HB patients. However, the incidence of late complications remains not a few.^{3, 4} The JPLT3 study was conducted using a new risk stratification based on the survival analyses of the JPLT2 cohort.

In the JPLT3-S trial, we aimed to evaluate the effectiveness of CDDP monotherapy in HB patients under 3 years of age, considering the risk due to age at diagnosis. As a new initiative of the JPLT3, central radiological and pathological reviews were conducted to accurately define eligibility and evaluate the response in all cases. The Société Internationale d'Oncologie Pédiatrique Epithelial Liver Tumor (SIOPEL) group conducted the SIOPEL 3 trial using regimens mainly based on PLADO (CDDP + doxorubicin) and CDDP monotherapy for standard-risk patients, and showed the non-inferiority of CDDP monotherapy for standard-risk HB in comparison to PLADO,⁹ but 19 of 126 cases treated with CDDP monotherapy had relapse or disease progression, indicating that some additional prognostic factors might be correlated in this cohort. Considering that the age at the diagnosis might be correlated with the malignant grade of HB, which was elucidated later,^{10, 11} the JPLT3-S study focused on younger patients who were diagnosed under 3 years of age.

Moreover, to ensure diagnostic accuracy, computed tomography (CT)/magnetic resonance imaging (MRI) images and pathological specimens (tumor biopsy specimens) of all patients were centrally diagnosed by the Japan Children's Cancer Group (JCCG) radiological and pathological committees, and the Response Evaluation Criteria in Solid Tumors (RECIST) classification after two and four courses of neoadjuvant chemotherapy were also evaluated based on imaging by central radiologists.

2 METHODS

The trial period for JPLT3 was from 2012 to 2017, followed by a 3-year observation period. This study was approved by the Ethics Committee of Hiroshima University (Ethics Committee No: E-2198) and the Certified Review Board (CRB) for special clinical research (jRCTs061180086). The cases were collected from 116 JPLT participating facilities in Japan and were approved by the institutional ethics committee. Written informed consent was obtained from all parents.

2.4 Treatment flowchart

A treatment flowchart for JPLT3-S is presented in Figure 1. After performing a biopsy to confirm the diagnosis by central pathological review, four courses of preoperative chemotherapy with CDDP alone were administered, followed by surgery. Two courses of CDDP monotherapy were administered postoperatively. In patients who initially underwent upfront surgery, which was allowed only for those with PRETEXT I disease, four courses of CDDP treatment were administered postoperatively.

The dose of CDDP in each course was 80 mg/m² for those weighing more than 10 kg, 2.7 mg/kg for those weighing 5−10 kg, and 1.8 mg/kg for those weighing less than 5 kg, over 24 hours.

3 RESULTS

A total of 43 patients were enrolled in this study between 2012 and 2017. Of the 43 patients enrolled, five were deemed ineligible for the following reasons: outside the target age, registered before CRB approval, late registration, misdiagnosis (adrenocortical tumor) by central pathological review, or staging changed after the central imaging review (the decision was made as PRETEXT annotation factor N1(+)). Finally, 38 patients were eligible for analysis in JPLT3-S (Figure 2).

Among these 38 patients, two were diagnosed with progressive disease (PD) during preoperative chemotherapy. In one of these cases, the tumor was progressive and invaded three hepatic veins, leading to a change in treatment. Liver transplantation was performed and disease-free survival was achieved. Early emergent surgery was performed in one patient because of intratumoral bleeding after one course of chemotherapy. Five other cases deviated from the protocol at the facility's discretion. In two cases, the tumor showed stable disease (SD) on imaging, but the attending physician determined that the treatment was ineffective and changed to second-line treatment. Treatment (sixth course of CDDP) was delayed in one patient due to postoperative bile leakage. In one case, the institutional surgeon determined that consultation with a transplant facility was necessary after four cycles, and deviated from the protocol. In one case, treatment was discontinued and early surgery was performed because of hyperkalemia and hypertension. Thirty-one patients (81.6%) completed the study.

3.4 Primary endpoint

The 3-year PFS rate, which was the primary endpoint of this study, is shown in Figure 3A. Among the 38 patients, two had PD, and there were five deviated cases. The 3-year PFS rate was 93.9% (95% confidence interval [CI]: 86.4%–100%).

3.5 Secondary endpoints

3.5.1 Overall survival

All 38 patients survived (median follow-up period: 63 months; range: 38–98 months), and the OS rate was 100% (CI: 100) (Figure 3B).

3.5.2 Treatment and response

Preoperative AFP levels after four courses of CDDP were obtained from 33 patients, except for five cases, who underwent surgery at different times. The median preoperative AFP level was 59,142 (22.0–116,743) ng/mL. In 32 cases, except for one in which imaging was unavailable, the responses according to the RECIST criteria based on central imaging were as follows: PD, n = 1; SD, n = 11; and partial remission (PR), n = 20. The relationship between the RECIST criteria and median AFP decline ratio is shown in Figure 3. A preoperative AFP decrease was observed in the PR group (median preoperative AFP/AFP ratio at diagnosis = 0.009), and in the SD group (median preoperative AFP/AFP ratio at diagnosis = 0.034), but the difference was not significant (p = .226). Of the 32 patients, 25 had at least a 1-log decline in AFP at the completion of four courses of CDDP. The PD patient showed less than a 1-log decrease in AFP levels (Figure 4).

3.5.3 Surgery (n = 34)

Detailed surgical records were available for 34 of the 38 patients. Except for one liver transplantation, 16 patients underwent right or left hemihepatectomy, five underwent right or left trisegmentectomy, two underwent lateral segmentectomy, and 10 underwent other methods, including sectorectomy and partial resection. Initial upfront surgery (partial resection) was performed in one patient. There were no cases with positive microscopic margins, and complete resection was achieved in all cases, including the PD case (complete surgical resection rate: 100%).

3.5.4 Serious adverse events (n = 38)

One case of intraoperative cardiac arrest due to massive intraoperative bleeding during partial resection of segments 4 and 5 recovered without sequelae (grade 4). In one patient with postoperative bile leakage, Cycle 6 chemotherapy was delayed (grade 3). Regarding SAEs other than surgical complications, one case of asymptomatic hypomagnesemia and one case of hypercholesterolemia were observed during the sixth course of CDDP (grade 4). One patient required emergency surgery because of intratumoral hemorrhage after one course of treatment (grade 3). One patient required 6 days of glucose-insulin therapy because of hyperkalemia after one course of treatment (grade 3).

3.5.5 Late complications

Regarding late complications, high-frequency hearing loss was detected in 18 of the 38 patients (47.4%). The diagnosis of hearing loss was made according to the institutional criteria. Two patients had mild renal dysfunctions. One patient with ELBWI had short stature, and one patient with trisomy 18 had epilepsy. There were no cases of cardiac toxicity, and none of the patients developed secondary malignant neoplasms.

4 DISCUSSION

In this JPLT3-S study, the 3-year PFS was 93.9% (CI: 86.4%–100%) and OS was 38/38 (100%), which was the expected result. The primary endpoint, the 3-year PFS, was significantly higher than the 80% threshold. From this result, we can conclude that CDDP monotherapy is a suitable first-line treatment for standard-risk group HB cases diagnosed at less than 3 years old.

In the JPLT2 trial, performed from 1999 to 2012 in Japan, patients with a standard risk of PRETEXT I, II, and III were mainly treated with CITA and/or low-CITA therapy.³ Of the 212 patients, 163 were retrospectively re-categorized into the same standard-risk group according to JPLT3-S risk stratification. The 3-year OS of these 163 patients was 93.4% (CI: 40%–100%). The current JPLT3-S results showed that the OS was 100% with a median follow-up period of 63 months (range: 38–98 months), which was not significantly different from that of the previous JPLT 2 trial (p = .74). Therefore, CDDP monotherapy is an effective regimen for standard-risk JPLT3 patients.

Perilongo et al.⁹ reported the results of SIOPEL3, a randomized controlled trial that compared CDDP monotherapy and PLADO therapy. This study demonstrated the non-inferiority of CDDP monotherapy in the standard group. In their study, 126 patients received CDDP monotherapy. Table S1 shows a comparison of the results of SIOPEL3 and JPLT3-S. In that report, the 3-year event-free survical (EFS) rate was 83% (CI: 77%–90%) and the 3-year OS rate was 95% (CI: 91%–99%).⁹ Compared to the results of SIOPEL3, in JPLT3-S results, the 3-year PFS rate was 93.9% and the 3-year OS rate was 100%, which might be more satisfactory, and CDDP monotherapy is considered to be more feasible in these young localized HB cases (JPLT3 standard-risk patients).

In the JPLT3 trial, we conducted a central review of CT images to accurately diagnose and precisely judge the response according to the RECIST criteria. As Miyazaki et al.⁶ reported, a central review was useful not only in the JPLT-3 standard-risk group but also in the intermediate- and high-risk groups. In fact, the diagnosis changed in only two cases in the standard-risk group after the central review, but we believe that the proper risk classification of such cases contributed to the favorable PFS in this study.

The AFP decline ratios before surgery were compared according to RECIST criteria. Several reports have shown that AFP decline ratios are correlated with the prognosis of HB.^13-15 In this report, the average AFP decline in the PR group was greater than 2 before surgery; however, there was no significant difference between the PR and SD groups (p = .226) (Figure 4). Response to chemotherapy may be evaluated by a combination of imaging findings and AFP decline.

In one case, the local surgeon discussed the necessity of liver transplantation after the four scheduled preoperative courses of chemotherapy had been completed. The patient was subsequently referred to a transplant center, but as the referral process was time-consuming, another course of chemotherapy was administered preoperatively before the patient was transferred; thus, this case deviated from the protocol. In the JPLT3 study, we introduced a surgical review system wherein experts from the pediatric liver surgery and liver transplantation fields formed an expert panel to provide surgical suggestions based on imaging studies upon institutional requests.⁷ The abovementioned case never had the chance to be evaluated by the panel. It is challenging for some low-volume local centers to accurately determine the potential need for liver transplantation, and professional advice is warranted in such cases to facilitate early referral to transplant centers. In the current JPLT4 study, all cases underwent real-time central review by an expert surgical panel periodically during treatment after the diagnosis.

In the CDDP monotherapy protocol, no long-term treatment-related late complications were observed, except high-frequency hearing loss. Sodium thiosulfate (STS) was not used in any of the patients enrolled in this study, because STS was unavailable in Japan at the time of the study. To reduce ototoxicity, it is necessary to consider reducing the amount of CDDP or the combined use of STS in the future.^16-18 Although 26/38 patients (68%) in this study were classified as PRETEXT I (n = 8) or II (n = 18), upfront surgery was performed in only one patient. Upfront surgery may decrease the exposure to cisplatin and the long-term risk of ototoxicity.

In this study, none of the patients had cardiac toxicity or secondary primary malignancies due to chemotherapy. In contrast, of the 163 patients in the standard-risk group in JPLT2, secondary malignancies developed in five cases, three of which occurred within 3 years,⁹ and cardiac toxicity was observed in two cases. CDDP monotherapy may have contributed to the reduction in secondary malignancies and cardiotoxicities due to the omission of anthracycline.

AUTHOR CONTRIBUTIONS

All authors have contributed significantly to the study and have reviewed and agreed upon the manuscript content.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.