Pediatr Blood Cancer 2004;42:574–581

Second Neoplasms After Treatment of Childhood Cancer in Slovenia Janez Jazbec,

1 MD, MSc, *

Patricija Ec´imovic´, MD, MSc,2 and Berta Jereb,

Background. The number of long time survivors of childhood cancer treatment is constantly increasing over the last decades as a result of advances in diagnosis and treatment. The occurrence of second neoplasms is one of most serious late effects observed in cancer survivors. Methods. The risk of secondary neoplasm was studied in a cohort of 1,577 patients treated for childhood cancer registered in the Cancer Registry of Slovenia (CRS) between 1961 and 2000. The time at risk was defined from the date of diagnosis of first malignancy to the time of death or the end of the study. Results. The most frequent primary malignancies were: acute leukemia 28.5%, central nervous system

Key words:

2

(CNS) tumors 21.3%, and lymphomas 16.6%. Median observation time was 7.8 years. Fortyeight patients developed second neoplasms. CNS tumors, acute leukemias, and thyroid carcinoma were most frequent second neoplasms. The cumulative risk for second neoplasm in the entire cohort was 0.06% at 5 years, 5.1% at 15 years, and 12.6% at 25 years after diagnosis of first cancer. The overall survival after second neoplasm was 65% 10 years after the diagnosis of second neoplasm. Conclusions. Patients after treatment of childhood cancer are at special risk for subsequent neoplasms and long-term follow-up is mandatory. Pediatr Blood Cancer 2004;42:574–581. ß 2004 Wiley-Liss, Inc.

childhood cancer; second neoplasm

INTRODUCTION

The number of long time survivors of childhood cancer treatment continues to grow as diagnostic and treatment options improve. With the increasing number of survivors and years of follow-up, the adverse effects of treatment gain more importance. The late effects of treatment on many organ systems, such as cardiovascular, skeletal, endocrinologic, dental, hepatic, pulmonary, and renal as well as psychosocial, educational, and neuropsychological, have been described. Second malignant neoplasms (SMN) are one of the most serious late effects of therapy. As initial treatment results for the primary malignancy improve, the more long-term survival may be compromised by secondary cancers [1]. Reports on childhood cancer survivors estimated the cumulative probability of developing SN vary from 3.3 to 15% at 25 years from diagnosis and the risk of developing another cancer to be 5–15 times greater than in the general population [2]. The wide range of predicted incidences of SN in studies reflects, at least partially, differences in study designs and in observed populations. Most published studies of SMN are single or multicenter studies; population based studies are rare. The organization of treatment of childhood cancer in Slovenia is centralized in a single center and gives an opportunity to conduct a population-based study of SMN risk. The aims of this study were (1) to estimate ß 2004 Wiley-Liss, Inc. DOI 10.1002/pbc.20025

MD, PhD

the risk for SN and (2) to evaluate the survival of patients with SN after treatment of cancer in childhood. MATERIALS AND METHODS Study Population

In Slovenia one pediatric-oncology center at the Department of Pediatrics, University Medical Centre Ljubljana, serves as a national referral center for all pediatric patients with malignant diseases. This center covers a population of approximately two million people. After completion of treatment, all children are followed at the center until they are18 years old and at least 3 years after treatment Afterward, all patients are followed at the outpatient clinic for late effects at the Institute of Oncology, where a multi-disciplinary team, headed by an oncologist known to the patient as a member of pediatric ——————

1

Department of Pediatrics, Division of Oncology and Hematology, University Medical Centre Ljubljana, Ljubljana, Slovenia

2

Institute of Oncology, Ljubljana, Slovenia

*Correspondence to: Janez Jazbec, Department of Pediatrics, Division of Oncology and Hematology, Medical Centre Ljubljana, Vrazov trg 1, Ljubljana, Slovenia. E-mail: [email protected] Received 23 May 2003; Accepted 15 January 2004

Second Neoplasms After Treatment of Childhood Cancer

treatment team, continues follow-up for life [3]. Both the pediatric oncology clinic and the outpatient late effects clinic have procedures in place for ensuring patient compliance with follow-up. The late effects clinic at the Institute of Oncology has ongoing research projects concerning (1) organ deficits such as cardiac, pulmonary, renal, and endocrine deficits; (2) occurrence of secondary tumors; (3) intellectual abilities; (4) psychosocial status in persons treated for cancer in childhood. The Cancer Registry of Slovenia (CRS), founded in 1950, covers the entire population of the Republic and is based on obligatory registration of all cancer patients [4]. Followup of registered cancer patients regarding their vital status is performed annually by linkage with the CRS [5]. Combining data from these three sources, we constructed a database in which only a small number of patients are lost to follow-up and each patient’s data could be validated. Using these resources, we conducted a population-based study of second neoplasm occurrence after cancer treatment in childhood. The study included patients who were younger than 16 years at diagnosis of the first malignancy and registered at the CRS between 1961 and 2000. In all, patients data were collected from the CRS database as well as from the patient’s medical records at the outpatient clinic for late effects and Pediatric clinic. Demographic data, medical diagnoses, type of treatment, data regarding primary cancer recurrence, and occurrence of second neoplasm were abstracted. Treatment exposures were recorded as chemotherapy, radiation therapy, and surgery, without details of agent and dosages used. Treatment methods employed were similar to other pediatric oncology centers [6]. SN cases were diagnosed during follow-up at the Pediatric clinic (J. Jazbec) or at the Clinic for late effects at the Institute of Oncology (B. Jereb). In each SN case original medical documentation and original pathology reports of primary cancer and SN were reviewed. In all cases of death, the cause of death was checked by reviewing the patient’s medical records. SNs were defined as neoplasms on a new location, that were not direct spread or metastases of the primary neoplasm, or the neoplasms on the same location as the primary ones but of different histological type [7]. Cases of non-melanoma skin cancers and meningiomas were also included if they occurred as a second tumor, since occurrence of both types of neoplasms are registered in our database and represent an important number of secondary neoplasms, especially among patients treated for acute lymphoblastic leukemia [8]. Second leukemia and nonHodgkins lymphoma, after treatment of acute leukemia, were included in cases of distinctively different subtype form the first leukemia, to avoid inclusion of cases of late recurrences or extramedullary relapses [9]. Two cases of germinoma, which were non secreting at diagnosis, but were b-HCG positive on recurrence, were not considered

575

to be of different histological type and were therefore not included as second neoplasms [10]. Statistical Methods

The period at risk for second neoplasm was defined as from the time of diagnosis of the first malignancy to the time of occurrence of second neoplasm, the time of death, or the time of the last clinic or registry follow-up. The cumulative incidence of second neoplasms was estimated using the Kaplan–Meier method [11]. The same method was used for estimation of the survival after occurrence of SN and the observation period was defined as the time from diagnosis of SN to the date of the last control at the outpatient clinic or the date of death. For statistical analysis the program package SPSS, version 10.0 (SPSS, Inc., Chicago, IL) was used. The relative risk (standardized incidence ratio
SIR) was calculated as ratio between the number of observed cases of second neoplasms and expected number of cases in the general population. Expected numbers of subsequent neoplasms were estimated by applying age/sex specific rates derived from the National CRS to person years accumulated in the corresponding age/sex categories. In the estimation of relative risk for second neoplasm, the cases of secondary benign CNS neoplasms (meningiomas) were not included, since they are not included in CRS. The cases of nonmelanoma skin cancer were included in the calculation of SIR, since they are included in the CRS. The study was closed in December 2002. RESULTS

Between 1961 and 2000, there were 1,661 patients registered in CRS, who were younger than 16 years at the time of diagnosis of first malignancy. Eighty-four patients (5%) were excluded because they were lost to follow-up. All of them were treated before 1991 and had moved to live outside Slovenia so their follow-up was discontinued. No patients treated after 1992 were lost to follow-up. The demographic characteristics of the remaining 1,577 patients are presented in Table I. The mean age at the diagnosis of first malignancy was 7.2 years. There were 901 boys and 676 girls who accrued a total of 12,304 person-years of follow-up. The mean observation time was 7.8 years, the longest was 43.7 years. We categorized primary neoplasms as: leukemia, Hodgkins disease, nonHodgkin lymphoma, Ewing sarcoma, osteogenic tumors, nephroblastoma, neuroblastoma, hepatoblastoma, rabdomyosarcoma, retinoblastoma, thyroid cancer, germ cell tumors, tumors of central nervous system (CNS), and other tumors. The group of ‘‘other tumors’’ consisted of carcinomas of different organ systems in 41 cases and of 2 melanomas. They were grouped together because each particular group would be too small for meaningful statistical analysis.

576

Jazbec et al. TABLE I. Characteristic of the Study Population—1,577 Persons Treated for Childhood Cancer in Slovenia Between 1st January 1960 and 1st October 2000

Characteristic Mean age at the diagnosis of childhood cancer (years) Sex Male Female Childhood cancer diagnosis Leukemia Central nervous tumors Non-Hodgkin lymphoma Hodgkins disease Soft tissue sarcoma Renal tumors Neuroblastoma Bone tumors Germ cell tumors Ewing/PNET Retinoblastoma Thyroid cancer Hepatic tumors Others Vital status no. (%) (alive) Age at diagnosis of SN (years) Time from diagnosis of childhood cancer to SN (years) Mean time of follow-up after SN (years) (min.–max.)

Overall cohort (n ¼ 1,577)

Cases with second neoplasms (n ¼ 48)

7.2

8.5

901 (57%) 676 (43%) 449 (28.5%) 336 (21.3%) 142 (9%) 120 (7.6%) 120 (7.6%) 87 (5.5%) 81 (5.1%) 60 (3.8%) 44 (2.8%) 27 (1.7%) 35 (2.2%) 18 (1.1%) 15 (1%) 43 (2.7%) 1,014 (64%) N/A* N/A N/A

25 (52%) 23 (48%) 16 (33.3%) 11 (22.9%) 2 (4.1%) 10 (20.8%) 2 (4.1%) 1 (2.1%) 1 (2.1%) 1 (2.1%) 1 (2.1%) 1 (2.1%) 0 1 (2.1%) 0 1 (2.1%) 33 (68.7%) 21.7 13.2 6.2 (0.3–18.4)

*N/A, not applicable.

Forty-eight patients with second neoplasms were identified. There were 23 girls (47.9%) and 25 boys (52.1%). Mean age at the diagnosis of the first malignancy was 8.5 years. Mean time from diagnosis of the first malignancy to occurrence of second neoplasm was 13.2 years (95% confidence interval (CI) 8.5–16.4 years). The shortest intervals were 1.18 years in a girl with soft tissue sarcoma who developed acute lymphoblastic leukemia, and 1.5 years in a boy with osteosarcoma followed by acute myeloblastic leukemia. Distribution of second neoplasms is presented in Tables II and III which presents the number of second neoplasms by type of first malignancy. Three groups of primary malignancies, namely: leukemia, CNS tumors, and lymphomas make for 70% of cases of SN. Acute leukemia (lymphoblastic 14, myeloblastic 2) represents 29.1% and is followed by CNS tumors (11 patients) 22.9%, and Hodgkins disease (10 patients) 20.1%. All patients with acute leukemia and subsequent brain tumor had had lymphoblastic leukemia, and had had prophylactic CNS irradiation with doses ranged from 12 to 24 cGy, depending on the period of treatment. One patient had a CNS relapse of ALL and received also triple inraventricular chemotherapy through Omaya reservoir. Two meningeomas, two glioblastomas, one anaplastic astrocitoma, and one oligodendroglioma occurred as second CNS neoplasma after ALL.

In the group of patients with primary CNS tumors, SN involved the same organ in 8 of 11 cases and histologically differed from the primary tumor. In all but one case, radiotherapy was a part of primary treatment and second neoplasms developed within the irradiated field or on its margin. The mean interval between the first malignancy and SN was 21 years in patients with primary CNS tumors. In this group the shortest period was from the diagnosis of a CNS germinoma to appearance of acute myeloblastic leukemia

TABLE II. Frequency Distribution of Second Neoplasms

Second neoplasm Central nervous system (CNS) tumors Meningioma Acute leukemia Non-Hodgkins lymph Thyroid cancer Breast cancer Hodgkins disease Osteosarcoma Rhabdomyosarcoma Carcinomasa Total a

No. of patients

Percent

16 7 8 4 7 3 2 2 1 5 48

33 14.5 16.7 8.3 14.5 6.2 4.1 4.1 2 10.4 100

Included one case each of carcinoma of parotid gland, carcinoma colli uteri, carcinoma of maxillary sinus, and basal cell carcinoma.

5 Basal cell carcinoma, carcinoma colli uteri, carcinoma of maxillary sinus, and carcinoma of parotid gland. a

3 7 8 16

1 1 1

2 1

4

1

4

2

1

2

2

1

1

1

1

1

1 1

16 11 2 10 2 1 1 1 1 1 1 0 1 0 48 1 1 2 3 2 3 1

449 336 142 120 118 87 81 60 45 44 27 35 18 15 1,577 Leukemia (AML þ ALL) CNS Non-Hodgkins lymphoma Hodgkins disease Soft tissue sarcoma Renal tumors Neuroblastoma Bone tumors Others Germ cell tumors Ewings sarcoma and PNET Retinoblastoma Thyroid cancer Hepatic tumors Total

6 8 1

Non-Hodgkins lymph Thyroid cancer Acute leukemia CNS tumors No. of patients Primary malignancy

TABLE III. Number of Second Neoplasms by Type of Primary Malignancy Second Neoplasms

Breast cancer

Hodgkins disease

Osteosarcoma

Rhabdomyosarcoma

Carcinomasa

Total

Second Neoplasms After Treatment of Childhood Cancer

577

2.75 years later (etoposide was part of postoperative chemotherapy). This is in contrast with second malignancies after acute leukemia, where the mean latency period was 9.3 years. Even after exclusion of non-CNS second neoplasms, the mean time to second CNS neoplasms in leukemia cases was significantly shorter (P ¼ 0.001) (Table IV). The estimated cumulative risk for second neoplasm was 0.06% (95% CI 0.2–1.2%) at 5 years, 3.35% (95% CI 2.3–5%) at 10 years, 5.1% (95% CI 3.2–6.9%) at 15 years, 7.4% (95% CI 5.1–10.7%) at 20 years, and 12.16% (95% CI 7.7–16.4%) at 25 years. There were considerable differences in cumulative risk between different primary tumor groups (Figs. 1 and 2, Table V). In leukemia patients, there was a 1.17% (95% CI 0–2.5%) risk for second neoplasm at 5 years after diagnosis, 7.75% (95% CI 0.3–12.2%) at 10 years, 11.14% (95% CI 5.5–16.8%) at 15 years, 13.03% (95% CI 6–19.6%) at 20 years, and 16.14 (95% CI 6.8–24.6%) at 25 years. In contrast, the cumulative risk was lower in the earlier period after treatment of Hodgkins disease, but increased constantly without reaching a plateau at 25 years. A similar trend was apparent in CNS tumors where the risk of second neoplasm, although lower in the first decade after primary malignancy, increased and almost equaled the risk after leukemia at 25 years. As presented in Table III, SMN occurred relatively rarely in the remaining nine groups of patients. There were two cases of second neoplasms in patients with primary non-Hodgkins lymphoma and soft tissue sarcoma and one SN in each of the remaining seven groups. SMN were not observed in two groups of patients, those with hepatic tumors and retinoblastomas. Patients with hepatic tumors represent a relatively small group of patients with 13 hepatoblastoma patients and 2 hepatocellular carcinoma patients. Contrary to expectation [12] there were no second neoplasms in the 35 retinoblastoma patients, of whom 28 were alive at the time of evaluation. Treatment included radiotherapy in eight cases. There is only one patient with bilateral retinoblastoma in our cohort. He was treated with radiotherapy and died 9.7 years after diagnosis because of recurrence of retinoblastoma. Of the remaining seven irradiated patients with unilateral retinoblastoma, two died of disease 0.15 and 1.3 years after diagnosis and five are alive 12.9–34.9 years (mean 19.4  8.9 years) after diagnosis, without evidence of second neoplasm. The standardized incidence ratio for malignant neoplasms of all types was 8.7 (95% CI 6.23–11.9), thus the incidence of new cancers among patients treated for childhood cancer is significantly greater than the incidence in general population in Slovenia. The relative risk for second neoplasm at varying periods of time from diagnosis of first malignancy is presented in Table VI. The SIR for all secondary malignant neoplasms was 8.71

578

Jazbec et al.

TABLE IV. Mean Time to SN After CNS Tumors and After Leukemia

A. Primary CNS tumors Number of patients with second neoplasm Mean time to second neoplasm (years)

A. Primary CNS tumors with second CNS tumors

11 21.0

(observed 40, expected 4.59; 95% CI 6.23–11.9). The greatest excess number of cases was observed in thyroid carcinoma (SIR 35.9, 95% CI 11.6–83.3), myeloid leukemia (SIR 40.5, 95% CI 13.5–72.2), malignant brain tumors (SIR 9.7, 95% CI 2–28.3), and breast cancer (SIR 9.2, 95% CI 2.5–23.3). The overall survival after SN is presented in Figure 3. Mean time of follow-up after the diagnosis of SMN was 6.2 years, ranging from 0.3 to 18.4 years. The cumulative survival 10 years after diagnosis of SN is 65% (95% CI 51–81%). Fifteen patients died because of progression of SN. Thirteen of these patients died within the first 2 years after diagnosis of SN, 6 of them had CNS tumors, 4 had secondary acute myeloblastic leukemia, 2 had lymphoid malignancies, and 1 had rhabdomyosarcoma. Of special interest is a patient who died of carcinoma of the hypopharynx 37 years after treatment of medulloblastoma. At autopsy thyroid carcinoma was also discovered. He is the only patient with secondary carcinoma who died of disease progression. In other patients with secondary carcinoma (three with breast cancer, seven with thyroid cancer, two with skin cancers, one patient with carcinoma colli uteri, and one patient with carcinoma of maxillary sinus) treatment was successful and they remain in remission.

Fig. 1. Cumulative risk for second neoplasms after the childhood cancer in patients treated is Slovenia between years 1961 and 2000. The cohort consists of 1,577 patients treated for cancer in childhood. Dotted lines indicate 95% confidence interval (CI).

B. Primary leukemia

8

16

22.7

9.3

B. Primary leukemia with second CNS tumors 6 12.7

Two patterns of treatment outcome were noted among patients with secondary CNS neoplasms. One group of patients with astrocytomas grade III, two glioblastomas, and one oligodendrogllioma, all died of disease, most of them (five of six) within 1 year after the diagnosis. The second group of patients with neoplasms of CNS (seven meningiomas, three unclassified tumors) had a good prognosis. Only one patient in these group died 3.5 years after diagnosis of SN. She had anaplastic meningioma. There was no statistically significant difference in the latency period of SMN development between patients with high grade CNS tumors and patients with other neoplasms of CNS (mean time 19.7 years vs. 15.2 years; P ¼ 0.3). The survival of patients with secondary leukemias seemed to be dependent on the type of leukemia. While two of three patients with secondary acute lymphoblastic leukemia are alive, only two of six patients with secondary AML are alive. Both patients with secondary Hodkins lymphoma are alive and so are three of four patients with non-Hodgkins lymphoma. One patient died 1.8 years after the diagnosis of secondary non-Hodgkin lymphoma DISCUSSION

Direct comparison of the results of various studies of SMN after treatment of cancer in childhood is difficult

Fig. 2. Cumulative risk for second neoplasms after treatment of leukemia, CNS tumors, and Hodgkins disease in childhood in the period from 1961 to 2000 in Slovenia.

Second Neoplasms After Treatment of Childhood Cancer

579

TABLE V. Cumulative Risk for Second Neoplasm in Patients Treated in Slovenia Between Years 1961 and 2000 for Childhood Leukemia, Hodgkins Disease, and CNS Tumors Years after first malignancy 5 10 15 20 25

Entire cohort (95% CI)

Leukemia (95% CI)

Hodgkins disease (95% CI)

CNS tumors (95% CI)

0.06 (0.2–1.2) 3.35 (2.3–5) 5.1 (3.2–6.9) 7.4 (5.1–10.7) 12.6 (7.7–16.4)

1.17 (0–2.5) 7.75 (0.3–12.2) 11.14 (5.5–16.8) 13.03 (6–19.6) 16.14 (6.8–24.6)

0.1 (0–3.1) 4.45 (0.1–8.6) 8.21 (1.4–14.5) 15.3 (3.2–25.9) 24.71 (6.9–39.1)

0.58 (0–1.7) 2.43 (0–5.2) 4.17 (0–8.4) 6.63 (0.7–12.8) 13.45 (1.9–23.6)

Cumulative risk for second neoplasm in percentage.

because of differences in length of follow-up, differences in primary cancers, the type of treatment, the treatment era of the first cancer, and differing patient ages. These factors may influence second neoplasms incidence rate estimates [13]. Moreover, different studies use disparate definitions of the risk period for SMN, e.g., from the time of diagnosis of first malignancy, or the time from the end of primary treatment, to 2, 3, or 5 years after the diagnosis of first cancer. In our cohort, there are six patients (12%) who developed a second neoplasm within the first 3 years after the diagnosis of primary cancer. Five cases of secondary acute myeloblastic leukemia and one case of Hodgkins lymphoma would therefore be excluded, if the period at risk were defined as starting 3 years after the diagnosis of first malignancy. Treatment with alkylating agents or/and epipodophylotoxins was a part of therapy in five of six patients with early occurrence of SN. Both agents may be involved in induction of secondary leukemia after short latency period [14,15]. In general, the estimation of risk varies between hospital based and population based studies [2], probably due to more complete follow-up in the former registries. Varying cure rates in different time periods also have impact on estimated risk. In the study of Neglia et al. [16] 91% of the patients in his cohort were alive. Only 5 year survivors were included and the period of recrutation of the cohort overlaps with the time when treatment of childhood cancer started to be more successful. Our study included all childhood cancer patients from 1960, counting the period at risk from the time of diagnosis,

therefore overall survival in our cohort is 64%. The cumulative risk of second neoplasm in our cohort was 7.4% at 20 years and 12.6% at 25 years of follow-up. In previous studies, authors reported cumulative risk of second neoplasm to be from 3.7 to 12.1% at 25 years [17– 19]. A population based Nordic study reported cumulative risk of 3.5% at 25 years after the diagnosis of first cancer [20] which is similar to the results of population study in United Kingdom [19]. Taking into consideration all cases of second neoplasms, including meningiomas and basal cell carcinoma as well as the differences in definition of the time at risk and the time period analyzed, might contribute to higher cumulative risk calculated in our study. Cumulative risk for developing second neoplasm increases almost exponentially in time without apparent trend to form a plateau [21] and our result is consistent with this observation. In patients with acute leukemia cumulative risk for second neoplasm at 20 years after diagnosis exceeds the risk for entire cohort (13% vs. 7.4%). Lo¨ning et al. [22] reported cumulative incidence of second neoplasm in patients after acute lymphoblastic leukemia 0.5% after 5 years, 1.3% after 10 years, and 3.3% after 15 years, with significantly lower risk in nonirradiated patients (3.5% vs. 1.2%, P ¼ 0.048). In recently published Polish study, in which 3,252 ALL patients were included, even lower cumulative risk was reported (0.05% at 5 years, 0.5% at 10, and 0.95% at 15 years) [23]. The authors relate the differences to different treatment protocols although incomplete registration of second malignancies could not be excluded. Kimball-Dalton

TABLE VI. Standardized Incidence Risk for SN After Treatment of Childhood Cancer in Slovenia Between Years 1961 and 2000 at Varying Times After the Diagnosis of First Malignancy Years from diagnosis

Observed (O) Expected (E) O/E P 95% CI

0–4

5–9

10–14

15–19

20–24

25–29

30–98

All

9.00 0.77 11.68 0.000 2.81–39.9

11.00 0.55 20.18 0.000 9.98–35.8

4.00 0.58 6.93 0.003 1.88–17.7

5.00 0.63 8.00 0.000 2.58–18.5

8.00 0.63 12.70 0.000 5.48–25

0.00 0.66 0.00 1.000 0–5.59

3.00 0.79 3.81 0.046 0.78–11.1

40.00 4.59 8.71 0.000 6.23–11.9

580

Jazbec et al.

Fig. 3. Overall survival after second neoplasm after treatment of childhood cancer (dotted lines indicate 95% CI) in Slovenia.

et al. [24] found 10-fold excess of other events (induction failure, relapse, or remission death) in patients with acute leukemia, when compared to second neoplasm incidence (2.7% at 18 years) and stresses the importance of focusing on reducing leukemia relapse and on enhancement of life quality in future therapeutic protocols, as well as on preventing second neoplasms. Patients with Hodgkins disease had the highest risk for SN exceeding 24% after 25 years of observation in our study. These results are comparable to studies of Green et al. [25] who found 26.27% (6.75%) risk at 30 years after diagnosis. The study of Dores on 32,591 patients with Hodgkins disease (1.111 survived 25 years after diagnosis) found cumulative risk at 25 years 21.9% and decreasing with the age at diagnosis [26]. Beaty et al. [27] compared the cumulative risk for second neoplasm for adolescent and preadolescent group of patients with Hodgkins disease and found significantly higher risk for SN in the adolescent group, where the incidence at 15 years after diagnosis was 10% versus 1.6% in preadolescent group. His data also indicate that female sex is a risk factor for second malignancy following treatment of Hodgkins disease, even when cases of secondary breast cancer were excluded. In our study, thyroid carcinoma was predominated as second malignancy after Hodgkins disease. In the study by Bhatia et al. [28], however, breast cancer seem to be the highest concern after long follow-up period, as the probability reaches 35% at 40 years of age. Therefore, for this group of patients clinical and mamographic screening are recommended at a younger age and performed more frequently than in general population and patient follow-up and education regarding the risk and self-examination is of major importance [29]. Although the high overall survival of patients with SN after treatment of cancer in childhood in our group is influenced by inclusion of patients with secondary meningiomas, even after the exclusion of these patients

the survival rate is still over 60% at 5 years after the diagnosis of SN. The high survival rate of patients with secondary carcinomas may be due to early detection in the outpatient clinic for late effects after treatment of cancer in childhood. Early detection may allow successful treatment in early stages of disease. However, in the cases of secondary high grade gliomas and secondary myeloblastic leukemias the prognosis is poor in our series. Although all pediatric cancer survivors are at risk for SMN, identification of subgroups at especially high risk is important. In these high risk patients, treatment modifications to minimize the risk for second neoplasm may be appropriate, provided that treatment efficacy is not compromised. In patients with acute leukemia, for instance, modern treatment protocols reduce or eliminate the use of prophylactic cranial irradiation, as it is clearly related to the incidence of SN [20]. In addition to treatment modifications, preventive measures, such as education of survivors about high risk life habits and environmental hazards, may lower the rate of secondary malignancies. Accurate and timely information about the risk of second cancers and carefully designed public preventive programs, are necessary for early detection of second malignancies which is essential for successful treatment. CONCLUSIONS

Patients who were treated for cancer in childhood are at constantly increasing risk for second neoplasm, which reaches 12% at 25 years after the diagnosis in our study and can even be much higher for particular types of cancer like Hodgkins disease where it is almost 25% at 25 years. We found a significant difference in the latency period from the time of diagnosis of first malignancy to SN occurrence between patients with acute leukemia and patients with CNS tumors. The prognosis of SN occurring after treatment for cancer in childhood depends on the type of SN. In many cases long-term cure can be achieved. Long-term follow-up of patents cured for childhood cancer is mandatory among other reasons also because of the increasing risk for second neoplasms. ACKNOWLEDGMENT

The authors thank Richard Aplenc, MD, MSCE from Children’s Hospital of Philadelphia, for editorial assistance. REFERENCES 1. Kodym R. Secondary malignancies after multimodality treatment regimens. In: Do¨rr W, Engenhart-Cabille R, Zimmermann JS, editors. Normal tissue reactions in radiotherapy and oncology, vol. 37. Bassel, Krager: Front Radiat Ther Oncol, 2002. pp 84–91. 2. Rosso P, Terracini B, Fears TR, et al. Second malignant tumors after the elective end of therapy for a first cancer in childhood: A multicenter study in Italy. Int J Cancer 1994;59:451–456.

Second Neoplasms After Treatment of Childhood Cancer 3. Jereb B. Model for long-term follow-up of survivors of childhood cancer. Med Pediatr Oncol 2000;34:256–258. 4. Pompe-Kirn V, Zakotnik B, Volk N, et al. Cancer patients survival in Slovenia 1963–1990. Ljubljana: Institute of Oncology, 1995. 5. Parkin DM, Whelan SL, Ferlay J, et al., editors. Cancer incidence in five continents, vol. VIII. IARC Scientific Publications No. 155. Lyon, France: International Agency for Research on Cancer, 2002. 6. Jereb B, Anzˇicˇ J. Pediatric oncology in Slovenia. Pediatr Hematol Oncol 1996;13(5):401–404. 7. Parkin DM, Chen VW, Ferlay J, Gacelan J, Storm HH, Whelan SL. Cancer registration: Principles and methods. Lyon: IACRSci Publ, 1991. No. 95. 8. Pui CH, Cheng C, Leung W, et al. Extended follow-up of long-term survivors of childhood acute lymphoblastic leukemia. N Engl J Med 2003;349(7):640–649. 9. Anzˇicˇ J, Jereb B. Secondary malignancies or late recurencies. Ann NY Acad Sci 1997;824:229–231. 10. Strojan P, Popovic´ M, Petricˇ-Grabnar G, et al. Two unusual recurrences of intracranial germinoma. Pediatr Hematol Oncol 1996;13(4):381–385. 11. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53:457–481. 12. Abramson DH, Ronner HJ, Ellsworth RM. Second tumors in nonirradiated bilateral retinoblastoma. Am J Ophthalmol 1979; 87(5):624–627. 13. De Vathaire F, Schweisguth O, Rodary C, et al. Long-term risk of second malignant neoplasm after a cancer in childhood. Cancer 1989;59:448–452. 14. Hawkins MM, Kinner Wilson LM, Stovall MA, et al. Epipodophylotoxins, alkylating agents, and radiation and risk of secondary leukemia after childhood cancer. Br Med J 1992;304:951–958. 15. Ching-Hon P, Ribeiro RC, Hancock ML, et al. Acute myeloid leukemia in children treated with epipodophylotoxins for acute lymphoblastic leukemia. N Engl J Med 1991;325(24):1682–1687. 16. Neglia JP, Friedman DL, Yasui Y, et al. Second malignant neoplasms in five-year survivors of childhood cancer: Childhood cancer survivor study. J Natl Cancer Inst 2001;93(8):618–629. 17. Li FP. Second malignant tumors after cancer in childhood. Cancer 1977;40:1899–1902.

581

18. Hawkins MM, Draper GJ, Kingston JE. Incidence of second primary tumors among childhood cancer survivors. Br J Cancer 1987;56:339–343. 19. Greene DM, Zevon MA, Reese PA, et al. Second malignant tumors following treatment during childhood and adolescence for cancer. Med Pediatr Oncol 1994;22:1–10. 20. Olsen JH, Garwicz S, Hertz H, et al. Second malignant neoplasms after cancer in childhood or adolescence. Br Med J 1993;307: 1030–1036. 21. de Vathaire F, Hawkins M, Campbell S, et al. Second malignant neoplasms after a first cancer in childhood: Temporal pattern of risk according to type of treatment. Br J Cancer 1999;79(11–12):1884– 1893. 22. Lo¨ning L, Zimmerman M, Reiter A, et al. Secondary neoplasms subsequent to Berlin–Frankfurt–Muenster therapy of acute lymphoblastic leukemia in childhood: Significantly lower risk without cranial radiotherapy. Blood 2000;95(9):2770–2775. 23. Kowalczyk JR, Nuzinska J, Armata J, et al. Second malignant neoplasms in children: A multicenter study of the Polish Pediatric Leukemia/Lymphoma Group. Med Pediatr Oncol 2002;38:421– 423. 24. Kimball-Dalton VM, Gelber RD, Li F, et al. Second malignancies in patients treated for childhood acute lymphoblastic leukemia. J Clin Oncol 1998;16(8):2848–2853. 25. Green DM, Hyland A, Barcos MP, et al. Second malignant neoplasm after treatment for Hodkins disease in childhood and adolescence. J Clin Oncol 2000;18(7):1492–1499. 26. Dores GM, Metayer C, Curtis RE, et al. Second malignant neoplasms among long-term survivors of Hodgkins disease: A population based evaluation over 25 years. J Clin Oncol 2002; 20(16):3484–3494. 27. Beaty O, Hudson MM, Greenwald C, et al. Subsequent malignancies in children and adolescents after treatment for Hodgkins disease. J Clin Oncol 1995;13(3):603–609. 28. Bhatia S, Robinson LL, Oberlin O, et al. Breast cancer and other second neoplasms after childhood Hodgkins disease. N Engl J Med 1996;334(12):745–751. 29. Kaste SC, Hudson MM, Fryear F, et al. Breast masses in women treated for childhood cancer. Cancer 1998;82(4):784–792.