Surgical view of the treatment of patients with hepatoblastoma

Results from the first prospective trial of the International Society of Pediatric Oncology Liver Tumor Study Group (SIOPEL-1)

Authors


Abstract

BACKGROUND

Surgical resection is the cornerstone of treatment for patients with hepatoblastoma (HB). The Society of Pediatric Oncology Liver Tumor Study Group launched its first prospective trial (SIOPEL-1) with the intention to treat all patients with preoperative chemotherapy and delayed surgical resection. The objective of this article was to assess the assumed surgical advantages of primary chemotherapy.

METHODS

Between 1990 and 1994, 154 patients age < 16 years with HB were registered on SIOPEL-1. The pretreatment extent of disease was assessed, and, after undergoing biopsy, patients were treated with cisplatin 80 mg/m2 intravenously over 24 hours and doxorubicin 60 mg/m2 intravenously over 48 hours by continuous infusion (PLADO). Generally, tumors were resected after four of a total of six courses of PLADO.

RESULTS

One hundred twenty eight patients underwent surgical resection (13 patients underwent primary surgery, and 115 patients underwent delayed surgery after PLADO). A pretreatment surgical biopsy was performed in 96 of 128 patients (75%). Biopsy complications occurred in 7 of 96 patients (7%). Twenty-two patients showed pulmonary metastases at the time of diagnosis, and 7 patients underwent thoracotomy. Operative morbidity and mortality were 18% and 5%, respectively. Complete macroscopic surgical resection was achieved in 106 patients (92%), including 6 patients who underwent orthotopic liver transplantation. The actuarial 5-year event free survival (EFS) rate for all 154 patients in the study was 66%, and the overall survival (OS) rate was 75%. For the 115 patients who were included in the surgical analysis that followed the exact protocol, the EFS and OS rates were 75% and 85%, respectively.

CONCLUSIONS

Biopsy is a safe procedure and should be performed routinely. Preoperative chemotherapy seems to make tumor resection easier. Reresection of a positive resection margin does not necessarily have to be performed, because postoperative chemotherapy showed good results. Resection of lung metastases can be curative if there is local control of the primary tumor; however, results showed that the patient's prognosis was worse. Surgical morbidity or mortality rates were not necessarily higher in large multicenter studies. More importantly, countries of lesser economic status also can contribute effectively to these trials. Cancer 2002;94:1111–20. © 2002 American Cancer Society.

DOI 10.1002/cncr.10282

There is general agreement that complete surgical resection is the cornerstone of treatment for patients with hepatoblastoma (HB) and hepatocellular carcinoma (HCC) and the only way for eventual cure. HB occurs most frequently in the first few years of life, whereas HCC usually is observed in older children and adolescents.1 This articles deals with HB only.

Historically, only 30% of patients with HB were amenable to primary surgical resection. Currently, with the help of more sophisticated imaging and surgical techniques, the rate is probably closer to 50%. This means that 50% of the tumors are still considered unresectable at the time of diagnosis. Half of these neoplasms can be made resectable with modern preoperative chemotherapy. This has been due mainly to the good tumor response of systemic cisplatin (CDDP)-based chemotherapy, which is capable of reducing tumor volume. In other words, eventually, 75% of all tumors can be resected completely.2–11

The International Society of Pediatric Oncology Liver Tumor Study Group launched its first prospective trial (SIOPEL-1) with the intention of treating all patients with preoperative chemotherapy for the following reasons: 1) The experience of individual surgeons. Compared with the resection of HBs at the time of diagnosis, most surgeons agreed that operating on tumors that had become smaller after chemotherapy was easier, and, hence, safer, because the tumor became better defined, less friable, and less prone to bleeding. 2) Visible metastases were present at the time of diagnosis in 20% of patients; these and micrometastases would be exposed to chemotherapy earlier. In other studies, these patients usually, but not consistently, were given preoperative chemotherapy as well. 3) It was desirable to establish a multidisciplinary approach at the onset with the objective of standardizing the selection and clinical grouping of patients, including use of a uniform staging system. For this reason, a pretreatment extent of disease (PRETEXT) grouping system was designed specifically for patients with liver tumors (the predictive value of this system will be discussed in more detail elsewhere). Standardization is particularly important in an international and multi-institutional study of such a rare pediatric tumor. The published overall results of this completed study support the strategy chosen.12–14

The primary objective of this article was to assess whether the assumed surgical advantages of primary chemotherapy stand up to more detailed scrutiny. Ease and safety of surgery cannot be defined scientifically, so surrogates are needed. These included 1) complications of biopsies, which are prerequisite for preoperative chemotherapy; 2) resectability rate at first attempt; 3) microscopic residual disease; 4) local recurrence rate; and 5) local and systemic complications at surgery and within the first postoperative month.

MATERIALS AND METHODS

The SIOPEL-1 study was open to patient registration between January 1990 and February 1994. Ninety-one centers in 33 different countries registered 154 patients with HB age < 16 years. After biopsies had been taken, patients were treated with preoperative chemotherapy. In the protocol, a biopsy was recommended in face of unequivocal clinical findings and was mandatory in patients age < 6 months and > 3 years because of the increased prevalence of other tumor types in these age groups. Biopsy techniques and their complications were analyzed. Small, localized tumors could be treated with primary resection followed by chemotherapy.

Chemotherapy consisted of cisplatin 80 mg/m2 intravenously over 24 hours on Day 1 and doxorubicin 60 mg/m2 intravenously over 48 hours by continuous infusion (PLADO) on Days 2 and 3. After four courses, tumor resectability was assessed, and definitive surgery was performed if it was considered feasible. Tumor resection was then followed by two more courses of PLADO. If the tumor was still judged to be unresectable after four courses of PLADO and was still responsive to chemotherapy, then two more courses of PLADO were administered, and resectability was assessed again.

Imaging

The extent of the primary tumor was assessed by abdominal ultrasound with or without Doppler, computed tomography scan, and (optionally) magnetic resonance imaging or hepatic angiography. The original radiologic findings were reviewed centrally.

Grouping

A new prechemotherapy and presurgery system based on imaging findings and using the four main liver sectors was used to evaluate tumor extension. It was named the PRETEXT system (Fig. 1). The left lobe of the liver consists of a lateral sector (Couinaud segments 2 and 3) and a medial sector (Couinaud segment 4), and the right lobe is divided in an anterior sector (Couinaud segments 5 and 8) and a posterior sector (Couinaud segments 6 and 7).15 The number of affected liver sector(s) determined the PRETEXT category (Fig. 1). Extrahepatic extension was indicated by adding one or more letters as follows: involvement of the hepatic and/or caval vein (V), involvement of the portal vein (P), extrahepatic tumor extension (E), and presence of distant metastases (M). The system was used to assess tumor extent and response to chemotherapy and to determine the optimal timing and type of surgical resection. Its ultimate goal was to ascertain preoperatively whether complete resection of the tumor would be possible.

Figure 1.

The pretreatment extent of disease grouping system used for the first prospective trial of the International Society of Pediatric Oncology Liver Tumor Study Group.

Surgery

The following surgical guidelines were recommended: The pretreatment biopsy was a needle biopsy or a wedge resection through a small laparotomy. At least three cores of tissue were taken from different sites. Laparotomy allowed the surgeon to collect more material, to sample different areas, and to control any tumor bleeding. For accurate histologic diagnosis, a generous sample of tumor was taken. Fine-needle biopsy was not accepted.

No specific surgical guidelines were provided for tumor resection, because it would have been futile and misleading given the use of so many techniques. However, to minimize the risk of perioperative deaths, evaluation of the nutritional status of the patients was recommended before surgery, as described by Fan et al.,16 and patients generally were treated in centers with a fairly large experience in liver surgery and anesthesia. Primary surgery was recommended only for those patients with tumors confined to the left lateral sector or the right posterior sector (PRETEXT Group I tumors). Usually, after four courses of chemotherapy, partial liver resection was performed if local resectability was feasible, also in the presence of lung metastases. Thus, as soon as local tumor control had been achieved, metastasectomy through either thoracotomy or sternotomy was performed. In patients with unresectable tumors who had lung metastases, chemotherapy was continued to gain maximal tumor response. If reduction of the lung metastasis did not occur, then they were removed surgically to make the patient eligible for transplantation. A specific guideline for the acceptable amount of tumor free margin was not given as long as no tumor cells were found in the plane of resection.

Orthotopic liver transplantation (OLT) was considered in patients with HB completely confined to the liver that, despite response to adequate first line chemotherapy, remained unresectable. Results of OLT in SIOPEL-1 will be presented in a separate article. In patients with recurrent disease or with persistently elevated α-fetoprotein (AFP) levels, the center had to decide which treatment strategy was best suited for the particular patient.

RESULTS

Patients Characteristics

Of 154 patients who entered the study, 128 patients were considered eligible for analysis. Twenty-six patients were excluded for the following reasons: Six patients died within 16 weeks from the time of diagnosis before surgical resection could be performed; in 16 patients, tumors remained unresectable after PLADO; and, in 4 patients, a complete data set was not available. Of the 128 patients analyzed, 13 patients underwent primary surgery for various reasons but, on the intention-to-treat principle, are included in this analysis. The clinical characteristics are summarized in Table 1.

Table 1. International Society of Pediatric Oncology Liver Tumor Study Group Trial 1: Characteristics of 128 Patients with Hepatoblastoma who Underwent Surgical Resection
CharacteristicNo.%
  1. CT: computed tomography; PRETEXT: pretreatment extent of disease; VPEM: involvement of the hepatic/caval vein (v), portal vein (p), extrahepatic tumor extension, and presence of distant metastases; PLADO: cisplatin 80 mg/m2 over 24 hours on Day 1 and doxorobicin 60 mg/m2 over 48 hours by continuous, intravenous infusion on Days 2 and 3; SIOPEL-1. the first prospective trial designed by the International Society of Pediatric Oncology Liver Tumor Study Group.

Gender
 Male8264
 Female4636
Age (yrs)
 Median1
 Range0–13
Serum α-fetoprotein (ng/L)
 Median172,714
 Range = 39–40 × 106
 Platelet count > 500 × 109/L7760
Solitary tumor10280
Size at imaging (cm)
 Median6.8
 Range = 3–17
Pulmonary metastases (chest X-ray or lung CT scan)2217
PRETEXT group (at diagnosis)
 I86
 II5745
 III4636
 IV1411
 Missing32
 VPEM positive2116
PRETEXT group (at surgery)
 I65
 II5543
 III4737
 IV65
 Missing1310
 Impossible10
Surgery
 Primary surgery1310
 Delayed surgery (post-PLADO)11590
5-yr overall survival
 All 154 patients included in SIOPEL-111675
 The 115 patients who received delayed surgical resection (post-PLADO)9885

SURGICAL PROCEDURES

Biopsy (n = 96 patients)

Of 128 patients who underwent surgical resection, 96 patients (75%) underwent a pretreatment surgical biopsy. Sixty-three patients underwent closed biopsy (needle biopsy in 20 patients, Tru-Cut biopsy in 37 patients, and other in 6 patients), 30 patients underwent open biopsy, and data about the biopsy technique used was missing in 3 patients. Eighty percent of biopsies were reviewed centrally, with an even spread across all centers, and central review was broadly in agreement with the center's own diagnosis of the primary pathologist. Thirty-two patients did not have a preoperative histologic diagnosis, because clinicians relied on clinical and biochemical characteristics. Complications of biopsy occurred in 7 of 96 patients (7%) (Table 2): bleeding from the biopsy site in four patients (one open biopsy and three closed biopsies), abdominal pain in two patients (one open biopsy and one closed biopsy), and a wound infection developed in one child who had an open biopsy. All seven patients recovered completely within hours or a few days.

Table 2. International Society of Pediatric Oncology Liver Tumor Study Group Trial 1: Surgical Complications During or Shortly After Surgery
ComplicationBiopsy (n = 96 patients)Primary resection (n = 13 patients)Delayed resection (n = 115 patients)
  • a

    These two patients experienced abdominal pain.

  • b

    One patient had transient hyperbilirubinemia; one patient had tubular dysfunction characterized by high urine volume and mild congestive cardiac failure; one patient had duodenal ulceration with perforation; one patient had anesthesia leading to hypothermia and metabolic acidosis; one patient had inflammation and small discharge from wound only; one patient had impaired biliary drainage and portoenterostomy; one patient had hepatomegaly 2 months postresection with clinical jaundice and abnormal liver function tests (biopsy showed portal tract fibrosis, bile duct proliferation, and inflammation; jaundice disappeared, but liver function tests remained abnormal); one patient died 18 hours postoperatively of possible myocardial damage, (anthracycline-related), because there were no technical problems during the operation; one patient had acute liver graft rejection (treated successfully); and one patient had biliary fistula and subphrenic collection later drained percutaneously under computed tomography control.

7 (7%)3 (23%)27 (23%)
Bleeding43
Infection129
Death015
Other2a10b

Primary surgery (n = 13 patients)

Five girls and eight boys with a median age of 16 months (range, 6 months to 11 years) underwent primary surgery (PRETEXT Group I tumors in four patients, Group II tumors in six patients, Group III tumors in two patients, Group IV tumors in no patients, and an unknown group tumor in one patient). The reasons for primary surgery were an emergency procedure for bleeding in two patients, a pedunculated tumor in one patient, and for a small tumor according to protocol guidelines in one patient. In the remaining nine patients, primary surgery was performed by an individual clinical decision despite protocol guidelines.

The median time from the date of diagnosis to resection was 5 days (range, 0–45 days). A left lateral segmentectomy was performed in one patient, five patients underwent hemihepatectomy (two right and three left), and four patients underwent right-sided trisegmentectomy. Nonanatomic resections were performed in two patients (one in Couinaud segments 4 and 5 and one in Couinaud segments 5 and 6). In eight patients, the intraoperative blood loss was < 500 cc (62%); in one patient, blood loss was between 500 cc and 1000 cc (8%); and, in two patients (15%), blood loss was > 1000 cc. In two patients, the amount of intraoperative bleeding was not retrievable from the medical record. Surgery could not be completed in one patient because of massive intraoperative bleeding. This patient ultimately died of postsurgical shock the day after surgery. The other 12 patients underwent complete surgical resection. Two of them developed a wound abscess (Table 2).

The median follow-up of 13 patients who underwent primary surgery was 5.5 years (range, 3–8 years). Twelve patients are alive with no evidence of disease.

Delayed surgery post-PLADO (n = 115 patients)

Forty-one girls and 74 boys with a median age of 21 months (range, 4 months to 5 years) underwent surgery after pretreatment with chemotherapy. The median time between the date of diagnosis and the resection was 4 months (range, 2–6 months). Thirty-two of 115 patients (28%) were down-staged after receiving preoperative PLADO. Nine patients initially had Group IV (unresectable) tumors and became resectable according to the PRETEXT tumor grouping system. Four patients were up-staged (3%). In one patient, a resectable tumor (initially, a Group II tumor) became unresectable (finally, a Group IV tumor). This patient was salvaged with OLT and is in complete remission 58 months after transplantation (the predictive value of the PRETEXT grouping system is reviewed in detail elsewhere).

A left lateral segmentectomy was performed in 5 patients, 26 patients underwent right hemihepatectomy (5 including Couinaud segment 1), 12 patients underwent left hemihepatectomy (1 with Couinaud segment 1 and 1 with a part of Couinaud segment 5), 26 patients underwent right trisegmentectomy (11 with Couinaud segment 1), and 9 patients underwent left trisegmentectomy (4 with Couinaud segment 1). In 15 patients, an extra-anatomic resection was performed (5 with Couinaud segment 1, 6 with Couinaud segment2, and 4 with Couinaud segment 3). Six patients (5%) underwent primary transplantation (OLT; these patients are reviewed in detail elsewhere). Complete data sets could not be obtained in 16 of 115 patients (14%) who underwent delayed surgery post-PLADO. Intraoperative bleeding was < 500 cc in 52 patients (45%), between 500 cc and 1000 cc in 26 patients (23%), and > 1000 cc in 11 patients (10%). In 26 patients (23%), the amount of intraoperative bleeding was not recorded. There was wide variety in the techniques used for surgical resection. In 40 patients, finger fracture was applied; in 31 patients, ultrasonic dissection with the Cavitron Ultrasonic Surgical Aspirator (CUSA) was used; and, in 9 patients, resection was performed with (Bovie) electrical coagulation. One patient underwent resection with vascular bypass. In 34 patients, the technique was not documented. Twenty-seven of 115 patients (23%) developed a surgical complication (Table 2). There was no clear link between advanced PRETEXT category and surgical morbidity/mortality (data not shown).

Five patients died perioperatively (Table 3), and, macroscopically, four tumors were resected incompletely; thus, successful, complete surgical resection was achieved in 106 patients (92%), including 6 patients who underwent OLT. Microscopic assessment revealed tumor present at margins of surgical resection in 11 patients, and 4 patients showed tumor present in the main vein (for details, see Table 4). In 9 of 11 patients with microscopic tumor present at the margins of surgical resection, the resection had been performed using the CUSA. Two of 15 patients with residual disease died (1 patient with progressive disease and 1 patient because of Budd–Chiari syndrome), 1 patient was treated with local radiotherapy and survived without evidence of disease (follow-up 6 years), 1 patient with tumor present in the main vein was lost to follow-up, and all other patients had no evidence of disease at a mean follow-up of 5.5 years (range, 2–8 years). Nine patients were treated with postoperative chemotherapy. In one patient with microscopic residual disease, data regarding postsurgical treatment were missing.

Table 3. International Society of Pediatric Oncology Liver Tumor Study Group Trial 1: Characteristics of Six Patients who Died of Surgery-Related Causes
GenderAge (months)PRETEXT groupLMCRResectionMajor techniqueCause of death
  1. PRETEXT: pretreatment extent of disease; LM: lung metastes; CR: complete resection; M: male; FF: finger fraction; F: female; R: right; hemi: hemihepatectomy; m: presence of distant metastases; L: left; OLT: orthotopic liver transplantation; vpe: involvement of the hepatic/caval vein (v), portal vein (p), and extrahepatic tumor extension (e); trisegm: trisegmentectomy.

PRIMARY RESECTION (n = 13 patients)
 M12III+Not completedFFPostsurgical shock
DELAYED RESECTION (n = 115 patients)
 F27II+R hemiFFCardiac arrest at operation
 M19II+UnknownUnknownBleeding
 F8III (m)++L hemiFFBleeding postsurgery; cardiac arrest at second look
 F31II (m)++R hemiFFKinking of hepatic artery with total necrosis of OLT 6 days after OLT
 M24III (vpe)R trisegmFF and diathermyBleeding during surgery; died shortly after surgery
Table 4. International Society of Pediatric Oncology Liver Tumor Study Group Trial 1: Characteristics of 15 Patients with Microscopic Residual Disease who Underwent Delayed Surgical Resection
GenderAge (months)PRETEXT groupLMResectionMajor techniqueResection marginMain vein/tumor presentFollow-up
  1. PRETEXT: pretreatment extent of disease; LM: lung metastases; F: female; FF: finger fraction; NED: no evidence of disease; M: male; L: left; hemi: hemihepactectomy; R: right; trisegm: trisegmentectomy; CUSA: Cavitron ultrasonic surgical aspirator; v: involvement of the hepatic/caval vein; p: involvement of the portal vein; m: distant metastases; e: extrahepatic tumor extension.

F13IIILocal excisionFF and diathermy+NED
M19IIIL hemiFF+NED
F4IIIR trisegmCUSA+NED
M8III (vpm)+L hemiCUSA+NED
F8III (y)R trisegmFF and diathermy+NED
M34IIIL trisegmCUSA+Died
M155IIIR trisegmCUSA+NED
F16III (vpe)R hemiCUSA and diathermy+Died
M0IIL hemiFF and CUSA+NED
F46II (pm)+R hemiCUSA and diathermy+Lost
M16IVR hemiCUSA and diathermy+NED
F91III (vp)L trisegmFF+NED
M4IIIL trisegmUnknown+NED
M11IILocal excisionOther+NED
F9IIL hemiCUSA+NED

The median follow-up of all patients who underwent tumor resection after PLADO was 4.5 years (range, 1–8 years). Three patients were lost to follow-up. The 5-year overall survival (OS) rate and the 5-year event free survival (EFS) rate for all 154 patients who were included in the study were 75% (95% confidence interval [95% CI], 68–82%) and 66% (95% CI, 59–74%), respectively.14 The OS and EFS rates (with 95% CI) for the 115 patients included in this surgical analysis who followed the exact protocol were 85% (95% CI, 78–92%) and 75% (95% CI, 67–83%), respectively (Figs. 2, 3).

Figure 2.

The 5-year overall survival of patients who underwent delayed surgery (n = 115 patients) after receiving chemotherapy with cisplatin 80 mg/m2 intravenously over 24 hours and doxorubicin 60 mg/m2 intravenously over 48 hours by continuous infusion.

Figure 3.

The 5-year event free survival of patients who underwent delayed surgery (n = 115 patients) after receiving chemotherapy with cisplatin 80 mg/m2 intravenously over 24 hours and doxorubicin 60 mg/m2 intravenously over 48 hours by continuous infusion.

Mortality and Recurrence

Sixteen patients died (13%) (Table 3). One patient with a PRETEXT Group III tumor died during primary surgery, and five patients died during or after delayed surgery (two patients with Group II tumors and three patients with Group III tumors); thus, the overall surgical mortality rate was 5% (6 of 128 patients). Five of these patients underwent hepatic resection using the finger-fracture technique (one patient with diathermy). Data on the operative technique used in the sixth patient were not available. The remaining 10 patients died of disease-related causes.

Only 2 of 16 patients who died (13%) had microscopic residual disease after surgery. One patient died 7 months postoperatively of Budd–Chiari syndrome after local recurrence (tumor-related death), because, after undergoing a right hemihepatectomy, the left hepatic vein became occluded by tumor. AFP levels remained elevated in the other patient, but the exact time between the diagnosis and the date of death could not been traced, because he was lost to follow-up.

Ten patients (8%) experienced disease recurrence (PRETEXT Group II tumors in three patients, Group III tumors in five patients, and Group IV tumors in two patients) after undergoing apparently complete resection: Five patients developed a local recurrence, and five patients developed a distant recurrence in the lung. Of the five children with a local recurrence, two children had lung metastases at the time of diagnosis. Of the five children who had a recurrence in the lung, three children had lung metastases at the time of diagnosis. None of these children had positive resection margins at first resection. One of them was treated with primary surgery. The primary tumor resections were performed with finger fracture (n = 3 patients), electrocoagulation (n = 3 patients), or CUSA (n = 2 patients). In two patients, details of the technique could not be found. The five patients with local recurrences underwent surgery: All patients were reresected completely, two patients underwent surgery combined with chemotherapy, and one patient underwent combined surgery with chemotherapy and radiotherapy. Two of those patients showed no evidence of disease at the most recent follow-up, 24 months and 84 months after the recurrence, respectively. One patient died during surgery due to uncontrollable bleeding. One patient developed a second recurrence then a third recurrence, underwent transplantation, and finally died. Another patient experienced four recurrences. The first three recurrences were treated with resection, and the last recurrence was treated with chemotherapy. Follow-up at 48 months after the first resection showed evidence of disease, and the patient underwent OLT 4 months later. The patient was alive without evidence of disease at last follow-up, 5.5 years from the time of diagnosis.

The five patients who developed recurrent disease in the lungs underwent metastasectomy followed by chemotherapy for local control (n = 2 patients; see Thoracotomies, below) or with chemotherapy alone (n = 3 patients). A second recurrence occurred in two patients. The first patient was treated with chemotherapy alone but developed recurrent disease 12 months postoperatively and was treated again with chemotherapy alone. The second patient developed a recurrence 13 months postoperatively and was treated with surgery and chemotherapy. Both children were in complete remission at follow-up 3 years and 5 years after their recurrence. Of the other three children, two children showed no evidence of disease at 3 years and 5 years of follow-up, respectively. One child died 6 years after undergoing OLT.

Thoracotomies (n = 7 procedures in 4 patients)

Of 22 patients with lung metastases at the time of diagnosis, 2 patients underwent single metastasectomy. Complications did not occur, and neither patient showed evidence of disease at follow-up 2.5 years and 7.0 years after resection. Two patients underwent multiple thoracotomies for recurrence in the lungs (two metastasectomies in one patient and three metastasectomies in another patient) without complications. These patients also were free of disease with relatively long follow-up (3 years and 5 years after resection). Of the remaining 18 patients, 5 patients who were treated with chemotherapy only also were long-term survivors (median follow-up, 6 years; range, 3–7 years). Thus, in total, 9 of 22 patients (41%) were long-term survivors, and 4 of them underwent metastasectomies as well as chemotherapy.13

DISCUSSION

This article reports the surgical details of patients with HB who were treated according to the SIOPEL-1 protocol based on preoperative PLADO chemotherapy. The demographic characteristics of the patients in SIOPEL-1 are comparable with other studies on treating HB. The tumor occurred mostly in younger children, with a male predominance, and, in most patients, serum AFP levels and platelet counts were highly elevated.1, 2, 6, 17 The vast majority occurred as a solitary tumor in the right lobe of the liver, and, in 22 of 128 patients (17%), there was pulmonary metastatic spread.

Biopsy appeared to be a safe procedure with a low complication rate. Only 7% of patients developed a minor complication, a rate comparable with data in the literature.10 Tumor spill or implantation metastases, which have been reported after biopsies for patients with HCC,18, 19 did not occur. The need for biopsy when imaging findings are characteristic and serum AFP levels are elevated remains a controversial matter among surgeons, at least in Europe. We conclude that the low complication rate in SIOPEL-1 justifies the use of biopsies. Malignant primary germ cell tumors of the liver and HCC can be excluded and, in the future, unfavorable histologic features, like undifferentiated HB, may be used to direct treatment to a high-risk regimen. Furthermore, biopsy material (untreated tumor tissue) is becoming more and more important in molecular biologic studies of tumor markers and for studying tumor biology.

Effect of Preoperative Chemotherapy on Surgery

This study has shown the advantages of treating patients with PLADO before they undergo tumor resection. First and foremost, excluding the patients who underwent OLT, it was possible to resect the tumor completely in 100 of 115 patients (87%) who were treated with chemotherapy. This emphasizes the fact that modern treatment strategies based on effective chemotherapy regimens have dramatically improved the complete resection rate if we keep in mind that, historically, only 30% of patients were eligible for complete tumor resection. Furthermore, the OS rate has improved from 35% in the early 1970s to the current rate of 70–75%.2, 11, 14

It often has been commented by surgeons that tumor resection is easier after chemotherapy. The neoplasm becomes more solid, less prone to bleeding, and more demarcated from the surrounding healthy liver parenchyma. In comparing the data on intraoperative blood loss between patients who underwent primary surgical resection and patients who underwent delayed surgical resection, this opinion could not be confirmed: The data were more or less equal. However, 28% percent of patients were down-staged after receiving preoperative PLADO. In addition, nine patients with initially unresectable (Group IV) tumors became resectable according to the PRETEXT system. In these patients, a smaller and, thus, easier resection could be performed as a result of the preoperative chemotherapy. The disadvantage of the delay in resection due to preoperative chemotherapy also was shown, although in a smaller group of patients. Four patients (3%) were up-staged, and, in one patient, a resectable tumor (initially, PRETEXT Group II) even became unresectable (finally, PRETEXT Group IV).

Surgical procedures showed an overall morbidity rate of 18% (24 of 128 patients) and a surgical mortality rate of 5% (6 of 128 patients). Thus, although preoperative chemotherapy may make more tumors resectable, hepatic surgery remains difficult, with definitive morbidity and mortality. An experienced surgical team, therefore, should perform hepatic resections in children. Furthermore, results may be influenced by the design of the study, in which centers with different levels of experience with liver surgery that were located in countries with different economic status participated. Patient selection was not equal in all centers, and this eventual bias may have contributed to the differences found.

Positive Resection Margins

The favorable outcome for these patients was a notable finding in this series. Of 11 patients with positive margins, only 2 died, but neither of those 2 patients had a local recurrence. None of the 11 patients underwent second resection, and all but one patient (radiotherapy) were treated only with postoperative chemotherapy. There were no recurrences, and all survivors were in complete remission at their last follow-up (mean, 5.5 years). This shows that reresection of the positive margin may not necessarily have to be performed. These favorable results may be explained in most patients by the use of the CUSA. The resection margin is vacuum cleaned and, thus, ablated. Hypothetically, it is possible that, despite the positive margin at the specimen site, there may have been a negative margin at the patient's site. Alternatively or in addition, residual tumor may not be viable because of lethal damage from preoperative chemotherapy.

Surgical resection of pulmonary metastases seems a good treatment option if local control of the liver tumor has been accomplished. All 4 of the 22 patients with pulmonary metastases at the time of diagnosis who underwent metastasectomy survived without residual disease. Even when there was a recurrence, surgical treatment was a curative procedure as long as there was local control of the primary tumor. Micrometastases at the time of diagnosis obviously are treated by the preoperative chemotherapy.

In summary, our data show that biopsy is a safe procedure with a low complication rate, and we still recommend it in our current trials. Not only does it provide an exact histologic diagnosis; in a treatment strategy based on the use of preoperative chemotherapy, it also is the only way to learn more about the histologic variants of hepatoblastoma and their biologic characteristics. If this issue is not addressed systematically within the context of a well-defined study, then the risk will be run that potential crucial information will be lost that may rectify and improve the actual treatment results. The use of a treatment strategy based on preoperative chemotherapy and delayed surgery for patients with such chemosensitive tumors as hepatoblastoma, whose response to therapy can be monitored easily, is at least as effective as strategies based on primary surgery. From a surgical perspective, it seems that preoperative treatment makes resection easier without any relevant drawback for the patients. Surgical morbidity and mortality rates of 18% and 5%, respectively, are comparable with the rates cited in the literature and support the chosen strategy.6, 17, 20 If a positive resection margin is noted by the pathologist, then reresection does not necessarily have to be performed. The treatment with postoperative chemotherapy or radiotherapy showed good results, and these patients did not experience local disease recurrence. Although the overall prognosis for patients with lung metastases is worse, resection of lung metastases may be curative if local tumor control has been achieved. Finally, the results show that large, multicenter studies do not necessarily lead to higher surgical morbidity or mortality rates. More importantly, the results have shown that countries of lesser economic status can contribute crucially to these trials.

Acknowledgements

The authors thank A. Philips and J. Brown from the Northern and Yorkshire Clinical Trial and Research Unit, Leeds, United Kingdom, for collecting and managing the data and emphasize that this study could be conducted only with the participation of the following centers: Argentina: Italian Hospital of Buenos Aires (Buenos Aires); Australia: Adelaide Children's Hospital (Adelaide), Royal Children's Hospital (Brisbane), Royal Children's Hospital (Melbourne), The New Children's Hospital (Paramatta), and Westmead Hospital (Westmead); Belgium: Cliniques Universitaires Saint-Luc (Brussels), Hopital Universitaire des Enfants (Brussels), Gent, University Hospital/Kliniek voor Kinderziekten (Gent), University Hospital Gasthuisburg (Leuven), and Clinique de Montegnee (Montegnee); Brazil: A. C. Camargo Hospital (Sao Paulo), Amico Hospital (Sao Paulo), Centro Infantil (Sao Paulo), Hospital Servidor Publico Estadual (Sao Paulo), and Santa Casa (Sao Paulo); Croatia: Children's Clinic Salata (Zagreb); Pediatric Oncological Center/Regional Hospital (Banska Bystrica) and Clinic of Childrens Oncology (Prague); Denmark: University Hospital (Copenhagen) and Odense University Hospital (Odense); Egypt: University of Alexandria (Alexandria); Finland: Children's Hospital (Helsinki); France: Centre Oscar Lambret (Lille), Centre Leon Berard (Lyon), Hopital d'Enfants (Nancy), and Institut Curie (Paris); Germany: University of Tubingen/Eberhard Karls Universitat (Tubingen); Greece: Children's Hospital (Athens) and Ippokation Hospital (Thessaloniki); Hungary: Semmelweis University Medical School (Budapest) and Miskolc Medical School (Miskolc); Ireland: Our Lady's Hospital for Sick Children (Dublin); Israel: Rambam Medical Center (Haifa); Italy: Policlinico Universita Bari (Bari), Giannina Gaslini Children's Hospital (Genova), Chirurgica Pediatrica (Padova), and Ospedale Regina Margherita (Torino); Japan: Sapporo National Hospital (Saporro); Malaysia: Hospital Universiti Sains Malaysia (Kelantan); Netherlands: Emma Children's Hospital AMC (Amsterdam), Vrije Universiteit Medical Center (Amsterdam), University Hospital of Leiden (Leiden), and University Hospital Nijmegen (Nijmegen); New Zealand: Starship Children's Hospital (Auckland) and Wellington School of Medicine (Wellington South); Northern Ireland: Royal Hospital for Sick Children (Belfast); Norway: University Hospital (Bergen) and Rikshospitalet (Oslo); Poland: Pomeranian Medical Academy (Szczecin), Research Institute of Mother and Child (Warsaw), and Medical Academy (Wroclaw); Portugal: Hospital St. Antonio (Porto); Slovenia: University Pediatric Hospital (Ljubljana); South Africa: Cape town, Red Cross Children's Hospital (Cape Town), Baragwanath Hospital (Johannesburg), Medunsa, Kalatonge Hospital (Pretoria), and Tygerberg (Tygerberg); Spain: Hospital Infantil Valle Hebron (Barcelona), Hospital Infantil de Cruces (Bilbao), Hospital Materno-Infantil (Malaga), and Hopital “La Fe” (Valencia); Sweden: University of Goteborg (Goteborg), University Hospital (Lund), and Karolinska Hospital (Stockholm); Switzerland: Universitäts-Kinderklinik (Bern), University Hospital (CHUV) (Lausanne), and Children's University Clinic (Zurich); Taiwan Republic of China: National Taiwan University Hospital (Taipei); Turkey: Hacattepe University (Ankara); United Kingdom: Royal Aberdeen Children's Hospital (Aberdeen), The Children's Hospital (Birmingham), Royal Hospital for Sick Children (Bristol), Royal Hospital for Sick Children (Edinburgh), Royal Hospital for Sick Children (Glasgow), St. James' University Hospital (Leeds), Leicester Royal Infirmary (Leicester), Royal Liverpool Children's Hospital (Liverpool), Hospital for Sick Children (London), King's College Hospital (London), Middlesex Hospital (London), Royal Marsden Hospital (London), St. Bartholomew's Hospital (London), Royal Manchester Children's Hospital (Manchester), Royal Victoria Infirmary (Newcastle), Queen's Medical Center (Nottingham), John Radcliff Hospital (Oxford), Children's Hospital (Sheffield), and Southampton General Hospital (Southampton); Uruguay: Hospital Pereira Rossell (Montevideo).

Ancillary