From February 1997 to December 2007, all patients who were newly diagnosed with neuroblastoma at the Samsung Medical Center were retrospectively reviewed for the present study. A diagnosis of neuroblastoma was made based on either histological examination of the tumor specimens or bone marrow infiltration by neuroblastoma cells and elevated urine catecholamine levels. The N-myc copy number was determined using Southern blot analysis or the quantitative reverse transcriptase-polymerase chain reaction. The patients were staged according to the international staging system (INSS) reported by Brodeur et al. (23) and the extent of the disease was evaluated using computerized tomography, a technetium-99 bone scan, bilateral bone marrow aspirates and biopsy specimens and an iodine-131 or 123-metaiodobenzylguanidine scan. Measurement of serum LDH, ferritin, NSE and urine vanillylmandelic acid (VMA) was part of a routine evaluation at diagnosis.

Table 1 summarizes the treatment plans. Patients were treated according to age-, stage- and N-myc-related protocols. Briefly, for patients with low-risk tumors, the primary treatment was surgery with (stage 2) or without (stage 1) postoperative chemotherapy. For patients with intermediate-risk tumors, preoperative chemotherapy followed by surgery and postoperative chemotherapy was the treatment regimen. While local radiation therapy to the primary site was given only if gross total tumor removal was not possible in the early period of the study (diagnosis until December 2003), all patients received local radiation therapy regardless of resectability later in the study (diagnosis from January 2004). High-dose chemotherapy and autologous stem cell rescue (HDCT/ASCR) was provided when gross tumor persisted despite the above-mentioned conventional therapies. For patients with high-risk tumor, HDCT/ASCR was routinely given as consolidation therapy after preoperative chemotherapy, surgery and postoperative chemotherapy (24). Thirteen-cis-retinoic acid was usually administered for 10 cycles (125 mg/m²/day for 14 days per every 4 weeks) to treat possible residual tumor cells for intermediate- and high-risk patients after completion of the scheduled chemotherapy. For high-risk patients, immunotherapy using interleukin-2 was given (2×10⁶ U/m²/day, subcutaneous injection for 5 consecutive days per every 4 weeks) with or without preceding induction therapy (3×10⁶ U/m²/day, continuous infusion for 5 consecutive days per week for 2 weeks) after HDCT/ASCR. IL-2 immunotherapy was usually initiated when the platelet count exceeded 50×10⁹/L without a transfusion after HDCT/ASCR. The first dose of IL-2 was begun on the first day of the 13-cis-retinoic acid treatment cycle and was applied until 1 yr after HDCT/ASCR (24).

The clinical and biological characteristics were compared according to the site of tumor origin (extra-abdominal versus abdominal). Regarding the clinical prognostic factors, age, sex and INSS stage were compared between the two groups. For the biological prognostic factors, the Shimada classification, N-myc amplification (≥3 copies), serum LDH, ferritin, NSE and urine VMA were compared between the two groups. Differences in the frequencies of the clinical and biological factors were compared using the Pearson chi-square test. Differences in the levels of tumor markers were compared using the Mann-Whitney U test. Levels of tumor markers are presented as the median values along with the ranges. Overall survival (OS) and event-free survival (EFS) rates along with 95% confidence interval (CI) were estimated using the Kaplan-Meier method. An event was defined as the occurrence of relapse/progression or treatment-related death. The differences in the survival rates between the two groups according to the clinical and biological factors were compared using the log-rank test. Statistical significance was accepted when the P values were less than 0.05.

From February 1997 to December 2007, 143 consecutive patients were newly diagnosed with neuroblastoma. Among 143 neuroblastomas, 115 (80.4%) tumors originated from the abdomen, 26 (18.2%) from extra-abdominal sites (thorax in 21, neck in 4 and orbit in 1) and two (1.4%) from unknown primary sites. Table 2 lists the clinical and biological characteristics of the 141 patients, except for the two patients with tumors from unknown primary sites. The median age at diagnosis of the 141 patients (81 boys and 60 girls) was 25 months (range 1-129), and 101 (71.6%) patients were over 12 months of age at diagnosis. Sixty-three (44.7%) tumors were localized (stage 1 in 16 patients, stage 2 in 12 and stage 3 in 35) and 78 (55.3%) tumors were disseminated (stage 4 in 77 patients and stage 4S in 1). Fifty-one (37.8%) of 135 tumors analyzed were N-myc amplified (≥3 copies), and 72 (53.7%) of 134 tumors analyzed had unfavorable Shimada pathology. The number of low-, intermediate- and high-risk patients were 28 (19.9%), 32 (22.7%), and 81 (57.4%), respectively. One hundred-eight, 120, 133, and 132 patients had serum LDH, ferritin, NSE and 24-hr urine VMA measured preoperatively at diagnosis. The values measured postoperatively were not included in the analysis. The levels of the serum LDH, ferritin, NSE and 24-hr urine VMA were a median of 1,352 (range 176-9,944) IU/L, 142 (10-6,690) ng/mL, 68 (7-1,081) ng/mL and 5.6 (range 0.1-101.8) mg/day, respectively.

All 16 patients with stage 1 disease received surgery only and all 12 patients with stage 2 disease received surgery followed by postoperative chemotherapy. Among 32 intermediate-risk patients, 10 received single or double HDCT/ASCR because of persistent gross tumor or tumor relapse. All 81 high-risk patients except five patients (progression in 2 patients, early death in 2 and refusal of HDCT/ASCR in 1) received single or double HDCT/ASCR, or are scheduled to receive HDCT/ASCR.

Table 3 shows the clinical and biological characteristics of the extra-abdominal and abdominal neuroblastomas. As for sex, there was no difference between the two groups. The median age at diagnosis in the abdominal and extra-abdominal group was 28 (range 1-129) months and 22.5 (range 7-74) months, respectively (P=0.484). In addition, there was no difference in the number of patients older than 1 yr of age at diagnosis between the two groups. The frequency of disseminated tumors (stage 4 and 4S) was lower in the extra-abdominal group than in the abdominal group (34.6% vs. 60.0%, P=0.019). The frequency of N-myc amplified tumors was also lower in the extra-abdominal group than in the abdominal group (4.2% vs. 45.0%, P<0.001). Accordingly, the proportion of high-risk tumors was lower in the extra-abdominal group than in the abdominal group (34.6% vs. 62.6%, P=0.026). Furthermore, the frequency of unfavorable Shimada pathology was lower in the extra-abdominal group than in the abdominal group, albeit this difference was not significant. The levels of serum LDH, ferritin, NSE and urine VMA were lower in the extra-abdominal group than in the abdominal group (median 679 vs. 1,391 IU/L, P<0.001, Fig. 1A; median 36 vs. 171 ng/mL, P=0.042, Fig. 1B; median 17 vs. 103 ng/mL, P<0.001, Fig. 1C; median 3.9 vs. 6.2 mg/day, P=0.056, Fig. 1D). The characteristics of the five extra-thoracic and extra-abdominal tumors were very similar to those of thoracic tumors. All five tumors were localized (stage 1 in 1 patient, stage 2 in 2 and stage 3 in 2). All of four both extra-thoracic and extra-abdominal tumors analyzed had non-amplified N-myc results. The levels of serum LDH, ferritin and NSE were very low (median 504 IU/L, 24 ng/mL and 12 ng/mL, respectively).

Tumors relapsed or progressed in 16 patients and treatment-related mortality occurred in 14 patients. Therefore, 111 out of 141 patients remained event free with a median follow-up of 47 months (range 1-131) from diagnosis. The probability of 5-yr OS and EFS rate (±95% CI) for all 141 patients was 76.3±8.5% and 73.9±8.1%, respectively. The probability of 5-yr EFS rates and the results of the unstratified log-rank test with respect to the clinical and biological characteristics are listed in Table 2. The probability of 5-yr EFS rate in patients with extra-abdominal tumors was higher than in patients with abdominal tumors (94.4±10.6% vs. 69.4±9.4%, P=0.026). Fig. 2A shows the EFS curves for patients with extra-abdominal and abdominal tumors. All but one patient in the extra-abdominal group are currently event free. When the analysis was confined to only the patients older than 1 yr of age at diagnosis, the probability of 5-yr EFS rate was higher in patients with extra-abdominal tumors than in those with abdominal tumors (92.3±14.5% vs. 65.3±11.7%, P=0.047, Fig. 2B). Similarly, when the analysis was confined to only the patients with stage 4 tumors, the probability of 5-yr EFS rate was higher in patients with extra-abdominal tumors than in those with abdominal tumors (87.5±22.9% vs. 59.2±12.8%, P=0.133, Fig. 2C), albeit this difference was not significant. All five patients with both extra-thoracic and extra-abdominal tumors remained event free with a median follow-up of 123 months (range 10-131) from diagnosis.