Mortality from Second Tumour Among Long-Term Survivors of Retinoblastoma: a Retrospective Analysis of the Italian Retinoblastoma Registry

Home , Esthesioneuroblastoma, Retinoblastoma

Oncogene (2006) 25, 5350–5357 & 2006 Nature Publishing Group All rights reserved 0950-9232/06 $30.00 www.nature.com/onc REVIEW Mortality from second tumour among long-term survivors of retinoblastoma: a retrospective analysis of the Italian retinoblastoma registry

A Acquaviva1, L Ciccolallo2, R Rondelli3, A Balistreri4, R Ancarola1, R Cozza5, D Hadjistilianou6, S De Francesco6, P Toti7, G Pastore8, R Haupt9, M Carli10, N Santoro11, A Di Cataldo12, A Fiorillo13, P Indolﬁ14, P Nucci15, A Sandri16, FPorta 17, AB Porcaro18, P Tamaro19 and G Morgese1

1Italian Retinoblastoma Registry, Department of Pediatrics, Obstetrics and Reproductive Medicine, University of Siena, Siena, Italy; 2Epidemiology Unit, Department of Preventive and Predictive Medicine, Istituto Nazionale per lo Studio e la Cura dei Tumori, Milano, Italy; 3Department of Pediatrics, University of Bologna, Bologna, Italy; 4Department of Surgery and Bioengineering, University of Siena, Siena, Italy; 5Department of Pediatric Oncology, Bambino Gesu Pediatric Hospital, Roma, Italy; 6Department of Ophthalmological Science, University of Siena, Siena, Italy; 7Department of Human Pathology and Oncology, University of Siena, Siena, Italy; 8Childhood Cancer Registry of Piedmont, Cancer Epidemiology Unit, University of Torino, Torino, Italy; 9Department of Hematology-Oncology Istituto Gaslini, Genoa, Italy; 10Division of Hematology/Oncology, Department of Pediatrics, University- Hospital of Padova, Padova, Italy; 11Department of Biomedicine in Childhood, University of Bari, Bari, Italy; 12Centre of Paediatric Haematology and Oncology, University of Catania, Catania, Italy; 13Department of Pediatrics, University of Federico II of Napoli, Napoli, Italy; 14Pediatric Oncology Service, Second University of Napoli, Napoli, Italy; 15Pediatric Ophthalmology & Strabismus Unit, San Paolo Hospital, University of Milano, Milano, Italy; 16Pediatric Oncology Department, University of Torino, Torino, Italy; 17Department of Pediatrics, University of Brescia, Brescia, Italy; 18Department of Urology, Civil Major Hospital, Verona, Italy and 19Department of Pediatrics, Ospedale Infantile Burlo Garofolo, University of Trieste, Trieste, Italy

Survivors of retinoblastoma (Rb) are at high riskof dying Introduction from second malignant tumour. The occurrence of second malignant neoplasm (SMN) and related mortality in a Retinoblastoma (Rb), although rare, is the most cohort of 1111 cases from the Italian Retinoblastoma frequent malignant ocular neoplasm in children. At Registry was analysed, considering the possible role of diagnosis, it arises as unilateral (about 70%) or bilateral both genetic and iatrogenic causes. Rb patients had a (about 30%) and appears as unifocal or multifocal. greater than 10-fold excess in overall mortality compared Rb (OMIM no. 180200) is a model of hereditary with the general population (standardized mortality ratio tumour in man. It can occur as familial or sporadic. (SMR) 10.73, 95% CI 9.00–12.80). Their excess risk The role of inactivating mutation of Rb1 gene in the attributable to cancers other than Rb was 14.93 95% CI development of these tumours (Friend et al., 1986) is 10.38–21.49). Survivors of hereditary Rb had an SMR for well known. all causes of 16.25 (95% CI 13.20–20.00), whereas their Bilateral Rb arises from a germline mutation in one SMR for all cancers was 25.72 (95% CI 17.38–38.07). copy of Rb1 combined with a somatic loss or mutation Survivors of unilateral sporadic Rb had an SMR of of the corresponding wild-type copy. 4.12 from all cancers (95% CI 1.55–10.98) and a much For unilateral sporadic Rb is proved both the higher excess for overall mortality (SMR 13.34, 95% CI possibility to acquire a somatic mutation in both copies 10.74–16.56). As expected, survivors of hereditary Rb had and to carry a new germline mutation in one copy higher mortality from cancers of the bone (SMR 391.90, developing a somatic mutation in the other copy. 95% CI 203.90–753.20) and soft tissue (SMR 453.00, 95% Further studies demonstrated the important func- CI 203.50–1008.40), small intestine (SMR 1375.50, 95% tions in different cellular process played by p105Rb, the CI 344.00–5499.70), nasal cavity (SMR 13.71, 95% CI protein encoded by Rb1 (proliferation, DNA replica- 1.93–97.35) and cancers of the brain and central nervous tion, DNA repair and apoptosis). system (SMR 41.14, 95% CI 13.2–127.55) The role of p105Rb and of the protein that interacts Oncogene (2006) 25, 5350–5357. doi:10.1038/sj.onc.1209786 with it appears more and more decisive in the occurrence of other cancers. In fact, in many sarcomas, small-cell Keywords: retinoblastoma; second primary neoplasm; lung and bladder cancers, primary breast tumours, chemotherapy; radiotherapy; mortality glioblastomas and other cancers, a mutation of Rb1 or modiﬁed expression of p105 Rb have been signalled. The risk of additional cancer in long-term survivors Correspondence: Dr A Acquaviva, Italian Retinoblastoma Registry, Department of Pediatrics, Obstetrics and Reproductive Medicine, of hereditary Rb (bilateral and unilateral carriers of University of Siena, Viale Bracci, 16, Siena 53100, Italy. germline Rb1 mutation) is reported by many studies E-mail: [email protected] (Abramson et al., 1979; Lueder et al., 1986; DerKinderen Analysis of the mortality of the Italian Rb patients A Acquaviva et al 5351 et al., 1987; Roarty et al., 1988; Eng et al., 1993; Moll Table 1 Characteristic of 1111 Rb patients sporadic unilateral cases et al., 1997). vs hereditary In particular, hereditary Rb survivors, compared with Characteristic Sporadic Hereditary Total the general population, have an increased risk of developing numerous cancers as soft-tissue sarcoma, No. % No. % No. % osteosarcoma, melanoma, brain tumour and lung cancer No. of subjectsa 703 63 408 37 1111 100 (Kaye and Harbour, 2004). They also experience an increased risk of dying from these second malignant Study centre neoplasms (SMNs) (Fletcher et al., 2004). Siena OC 167 24 103 25 270 24 Until now, about 35 different types of second malig- Other 536 76 305 75 841 76 nant tumour have been described. Sex Many authors have investigated the role of radio- Male 374 53 221 54 595 54 therapy combined or not with chemotherapy in the Female 329 47 187 46 516 46 growth of second malignant tumours in Rb survivors. Age at Rb diagnosis (years) The aim of this study is to describe, for the ﬁrst time, a o1 171 24 235 58 406 37 cohort of 1111 Italian Rb patients, analysing their 1 216 31 96 24 312 28 mortality from Rb and from SMNs in Rb survivors. 2 150 21 49 12 199 18 3–15 166 24 28 7 194 17

Results Year of Rb diagnosis 1923–1964 26 4 36 9 62 6 1965–1974 120 17 64 16 184 17 Our cohort consists of 1111 patients diagnosed with Rb 1975–1984 172 24 107 26 279 25 in Italy between 1923 and 2003. About a quarter of our 1985–1994 181 26 83 20 264 24 cases were observed at the Ophthalmology Clinic of the 1995–2003 204 29 118 29 322 29 University of Siena. Age at last observation (years) 0–9 248 35 179 44 427 38 Characteristics of the cohort 10–19 171 24 85 21 256 23 The characteristics of the cohort are summarized in 20–29 157 22 76 19 233 21 Table 1. 30–39 90 13 40 10 130 12 40–76 37 5 28 7 65 6 Seven hundred and three patients (63%) were sporadic and 408 (37%) hereditary. Hereditary cases Laterality were mostly bilateral (87%). No important difference in Unilateral 703 100 55 13 758 68 distribution between males and females was observed. Bilateral 0 0 353 87 353 32 The majority of cases were diagnosed within 36 months Surgery of life (83%) and after 1974 (78%). The median age of Performed 552 79 301 74 853 77 all patients was 18 months, and 48% of cases were Not performed 35 5 32 8 67 6 diagnosed before 1985. Unknown 116 17 75 18 191 17 At the last patient’s status control, 913 patients Treatments reported were alive: 86% of the 603 sporadic cases and 76% of Radiotherapy only 51 7 41 10 92 8 the 310 hereditary cases. Of the 171 patients deceased, Chemotherapy only 42 6 33 8 75 7 132 (12%) died of progression of disease (76 sporadic Both of them 79 11 88 22 167 15 and 56 hereditary), 29 (3%) of an SMN (four sporadic None of them 121 17 21 5 142 13 and 25 hereditary) and 10 (1%) of other causes. Only Unknown 410 58 225 55 635 57 27 patients (2%) were lost to follow-up. The median Vital status follow-up time since Rb diagnosis was 12 years for Alive 603 86 310 76 913 82 hereditary and 13 years for sporadic cases. At the last Dead – Rb 76 11 56 14 132 12 observation, 18% of hereditary and 17% of sporadic Dead – SMN 4 1 25 6 29 3 Dead – other causes 2 0 8 2 10 1 patients were older than 30 years. Lost to follow-up 18 3 9 2 27 2 Surgery was performed for 77% of cases (79% sporadic and 74% hereditary). Chemotherapy alone Abbreviations: Rb, retinoblastoma; SMN, second malignant neo- or in combination with radiotherapy was employed to plasm. aRow proportions (sporadic vs hereditary). treat 17% of sporadic and 30% of hereditary cases. Radiotherapy alone or combined with chemotherapy was employed to cure 18% of sporadic patients and At 76 years of age, the overall cumulative mortality was 32% of hereditary patients. Most of the 635 patients 43% (95% conﬁdence interval (CI): 27–64). Cumulative with unknown chemotherapy or radiotherapy treatment risk of death from Rb was 18% (95% CI: 15–22) and underwent surgical treatment at diagnosis. from SMNs was 18% (95% CI: 9–34). Figure 2 shows cumulative survival by period of Cumulative mortality diagnosis. Curves are graphed for the ﬁrst 20 years Figure 1 describes the cumulative mortality from all of follow-up and patients are grouped depending on causes (a), from 132 Rbs (b) and from 29 SMNs (c). whether they were diagnosed before or after 1985, which

Oncogene Analysis of the mortality of the Italian Rb patients A Acquaviva et al 5352

Figure 1 Cumulative mortality (time scale age in years) – all causes, Rb and SMNs.

is the year assumed as a turning point in the application of new therapeutic protocols (Acquaviva et al., 1985). Before 1985, survival rate was 76% (95% CI 72–79) and after 1985 it was 93% (95% CI 90–95). At 49 years of follow-up, the overall cumulative mortality for the hereditary group was twofold that for the sporadic group (Figure 3).

Rb patients with SMNs Forty-one second primary tumours were notified as occurring in the cohort in the follow-up period, but only 38 could be confirmed and had complete information; their characteristics in relation to hereditary status and treatment for Rb are described in Table 2. In Table 3, Figure 2 Cumulative survival (time-scale years since diagnosis) – year of occurrence, cancer site, morphology and vital cases diagnosed before and after 1985 (follow-up scale truncated at status are reported. In both tables, SMNs are grouped 20 years). according to occurrence inside or outside (45%) the irradiation field. Of the 38 patients who developed an SMN 21 (55%) were males, 26 (68%) had bilateral Rb and 31 (82%) were hereditary Rb cases. The range of latency time between the Rb diagnosis and the SMN diagnosis was 1–48 years. By the time of last control, 29 (83%) patients were deceased. Twenty-one SMNs (55%) occurred in a site inside the irradiation field. The most frequent site of the second tumour was bone (nine). Sarcomas were the prevalent morphology (twenty), followed by carcinomas (six), brain tumours (three) and LANL (three). Three patients were treated with surgery only and died, respectively, of melanoma (after 2003), bowel carcinoma and prostate carcinoma. Of 30 patients treated with radiotherapy, 18 developed a malignant tumour inside the irradiation field and 12 outside. Of the 19 patients treated with radiotherapy and Figure 3 Cumulative mortality (time-scale years since diagnosis) chemotherapy, 12 developed a malignant tumour inside by hereditary vs sporadic Rbs. the irradiation field and six outside. Seventeen of them

Oncogene Analysis of the mortality of the Italian Rb patients A Acquaviva et al 5353 Table 2 SMNs in Rb patients by irradiation ﬁeld – Rb treatment Patient Year of Sex Hereditary Laterality Year of Rb EBRT Interstitial Chemotherapy birth diagnosis RT

Inside ﬁeld 1 1976 FYes Bilateral 1977 No Yes 25 Gy VCR-END 2 1975 M Yes Bilateral 1976 RE 40 Gy/LE 40 Gy No VCR, ACT-D, CPM BCNU 3 1973 M Yes Bilateral 1975 LE 50 Gy/RE 45 Gy No VCR-CPM 4 1991 M No Unilateral 1992 LE 40 Gy No C E 5 1976 M Yes Bilateral 1977 RE 55 Gy No VCR-ADR- CPM 6 1955 M Yes Bilateral 1957 RE 55 Gy Yes ? No 7 1972 M Yes Bilateral 1973 RE 50Gy No VCR-ADR- CPM 8 1958 M Yes Bilateral 1959 RE 40 Gy/LE 30 Gy No No 9 1967 M Yes Bilateral 1968 LE 40 Gy Yes No 10 1966 FYes Bilateral 1966 RE 30 Gy/LE 40 Gy SKULL 30 Gy No VCR-ADR- RACHIS 39 Gy CPM-ACT-D 11 1984 M Yes Bilateral 1984 LE 43 Gy/RE 43 Gy SKULL 36 Gy No VCR-CPM 12 1983 FYes Bilateral 1984 LE 40 Gy/RE 40 Gy No No 13 1982 FYes Bilateral 1983 RE ?/LE ? No No 14 1966 M Yes Unilateral 1966 No No No 15 1984 FYes Bilateral 1984 LE 30 Gy No VCR-CPM 16 1979 FYes Bilateral 1980 RE ? No No 17 1974 FYes Bilateral 1976 LE 40 Gy/RE 40 Gy SKULL 24 Gy No VCR-CPM- BCNU MTX 18 1947 FNo Unilateral 1955 RE 30 Gy No No 19 1962 M No Unilateral 1965 RE 300 Gy No TEM 20 1964 FNo Unilateral 1965 RE ? No VCR-CPM-PBZ 21 1981 FYes Bilateral 1982 LE 40 Gy No VCR-CPM

Outside ﬁeld 1 1964 M Yes Bilateral 1966 ? ? ? 2 1950 FYes Unilateral 1952 No No No 3 1970 M Yes Bilateral 1971 ? ? ? 4 1976 M Yes Bilateral 1977 LE 35 Gy Yes ? No 5 1987 M Yes Bilateral 1989 LE 42 Gy/RE 40 Gy Yes 22 Gy No 6 1980 FYes Bilateral 1981 RE 30 Gy/LE 30 Gy Yes ? VCR-ADR- CPM 7 1981 M Yes Bilateral 1984 LE 40 Gy/RE 40 Gy No VCR-CPM- BCNU MTX 8 1973 M Yes Bilateral 1974 RE ?/LE ? ? ? 9 1972 FYes Bilateral 1972 RE 55 Gy No VCR-CPM 10 1961 FYes Bilateral 1962 ? No No 11 1964 FNo Unilateral 1965 RE 32 Gy No VCR-CPM-PBZ 12 1926 M Yes Unilateral 1933 No No No 13 1943 M Yes Unilateral 1951 RE 40 Gy No No 14 1974 FNo Unilateral 1975 No Yes ? CPM 15 1968 FNo Unilateral 1970 LE 45 Gy No VCR-CPM 16 1958 M Yes Unilateral 1959 LE 40 Gy No No 17 1979 M Yes Bilateral 1980 LE 40 Gy No VCR-CPM- BCNU-ACT-D

Abbreviation: ACT-D, actinomycin-D; ADR, doxorubicine; CPM, cyclophosphamide; C, carboplatin; E, etoposide; EBRT, external beam radiation therapy; F, female; LE, left eye; M, male; PBZ, procarbazine; Rb, retinoblastoma; RE ¼ right eye, SMN, second malignant neoplasm; VCR, vincristine.

(89.4%) were treated with cyclophosphamide included Survivors of hereditary Rbs had an SMR for all in their chemotherapy protocol. causes of 16.25 (95% CI 13.20–20.00), whereas their SMR for all cancers was 25.72 (95% CI 17.38–38.07). Survivors of unilateral sporadic Rbs had an SMR Standardized mortality ratios of 4.12 from all cancers (95% CI 1.55–10.98) and a Mortality of Rb patients from cancer other than Rb and much higher excess for overall mortality (SMR 13.34, other causes is shown in Table 4. 95% CI 10.74–16.56). Rb patients had a greater than 10-fold excess in Survivors of hereditary Rb had an SMR for all causes overall mortality compared with that in the general of 16.25 (95% CI 13.20–20.00), whereas their SMR for population (standardized mortality ratio (SMR) 10.73, all cancers was 25.72 (95% CI 17.38–38.07). Survivors 95% CI 9.00–12.80). Their excess risk attributable to of unilateral sporadic Rb had an SMR of 4.12 from all cancers other than Rbs was 14.93, 95% CI 10.38–21.49). cancers (95% CI 1.55–10.98) and a much higher excess

Oncogene Analysis of the mortality of the Italian Rb patients A Acquaviva et al 5354 Table 3 SNMs in Rb patients by irradiation ﬁeld – SMN characteristics Patient SMN diagnosis year SMN site – ICD9 SMN site – description SMN morphology Vital status

Inside field 1 1983 1719 Soft tissue Rhabdomyosarcoma Dead 2 1985 1719 Soft tissue Rhabdomyosarcoma Dead 3 1980 1600 Nasal cavities Glioma Dead 4 2004 1609 Nasal cavities, accessory sinus Carcinoma Alive 5 1987 1700 Bone, skull and face Osteosarcoma Dead 6 1994 1700 Bone, skull and face Esthesioneuroblastoma Dead 7 1987 1719 Soft tissue STS Dead 8 1963 1719 Soft tissue STS Dead 9 1985 1719 Soft tissue Rhabdomyosarcoma Alive 10 1977 1719 Soft tissue STS Dead 11 1989 1709 Bone, unspecified Osteosarcoma Dead 12 1996 1709 Bone, unspecified Osteosarcoma Dead 13 1996 1710 Soft tissue, head and neck Leiomyosarcoma Dead 14 2003 1729 Melanoma Melanoma Alive 15 2002 1902 Eye, lacrimal gland Adenoma Alive 16 1981 1911 Brain, frontal lobe Brain tumor Dead 17 1983 1911 Brain, frontal lobe Brain tumor Dead 18 1991 1921 CNS, cerebral meninges Meningioma Alive 19 1966 2059 Myeloid leukaemia LANL Dead 20 1991 2059 Myeloid leukaemia LANL Dead 21 1988 2059 Myeloid leukaemia LANL Dead

Outside ﬁeld 1 1999 1521 Small intestine, jejunum Carcinoma Dead 2 2000 1521 Small intestine, jejunum Carcinoma Dead 3 1986 1704 Bone, long bones of upper limb Osteosarcoma Dead 4 1980 1704 Bone, long bones of upper limb Osteosarcoma Dead 5 2003 1707 Bone, long bones of lower limb Osteosarcoma Alive 6 1997 1707 Bone, long bones of lower limb Osteosarcoma Dead 7 1997 1707 Bone, long bones of lower limb Osteosarcoma Alive 8 1990 1707 Bone, long bones of lower limb Osteosarcoma Dead 9 1987 1707 Bone, long bones of lower limb Osteosarcoma Dead 10 1993 1744 Breast Carcinoma Dead 11 1979 1830 Ovary Sarcoma Dead 12 1969 1855 Prostate Carcinoma Dead 13 1979 1869 Testis Seminoma Dead 14 1996 1719 Soft tissue Leiomyosarcoma Alive 15 2000 1890 Kidney Carcinoma Dead 16 1997 1921 CNS, cerebral meninges Meningioma Dead 17 2005 1719 Soft tissue Leiomyosarcoma Alive

Abbreviations: LANL, leukaemia acute non-lymphoblastic; Rb, retinoblastoma; SNM, second malignant neoplasm; STS, soft tissue sarcoma.

Table 4 Numbers of Obs. and Exp. deaths and SMRs with 95% CIs – SMN mortality by cancer site, overall and in sporadic and hereditary cases Cause of death All cases Sporadic Hereditary

No. of 95% CI No. of 95% CI No. of 95% CI deaths deaths deaths

Obs. Exp. SMR Upper Lower Obs. Exp. SMR Upper Lower Obs. Exp. SMR Upper Lower

Malignant tumours 29 1.94 14.93 10.38 21.49 4 0.97 4.12 1.55 10.98 25 0.97 25.72 17.38 38.07 other than Rb Small intestine 2 o0.01 775.10 193.80 3099.10 0 o0.01 — — 2 0.00 1375.50 344.00 5499.70 Nasal cavity 1 0.18 5.51 0.78 39.15 0 0.11 — — 1 0.07 13.71 1.93 97.35 Bone 9 0.06 148.72 77.38 285.83 0 0.04 — — 9 0.02 391.90 203.90 753.20 Soft tissue 6 0.04 165.13 74.18 367.55 0 0.02 — — 6 0.01 453.00 203.50 1008.40 Brain, CNS 3 0.18 16.54 5.34 51.29 0 0.11 — — 3 0.07 41.14 13.27 127.55 Myeloid leukaemia 3 0.36 8.28 2.67 25.69 2 0.22 9.03 2.26 36.12 1 0.14 7.11 1.00 50.45 Other cancersa 5 1.08 4.65 1.94 11.17 2 0.45 4.41 1.10 17.63 3 0.62 4.82 1.56 14.96 Other causes of deathb 10 9.68 1.03 0.56 1.92 2 5.18 0.39 0.10 1.54 8 4.50 1.78 0.89 3.55 Retinoblastoma 132 76 56 All causes 171 11.55 10.73 9.00 12.80 82 6.15 13.34 10.74 16.56 89 5.48 16.25 13.20 20.00

Abbreviations: CI, conﬁdence interval; CNS, central nervous system; Exp., expected; Obs., observed; Rb, retinoblastoma; SMN, second malignant neoplasm; SMR, standardized mortality ratio. aFemale breast, ovary, prostate, testis, kidney. bOne suicide, ﬁve iatrogenic toxicity cases, four unknown causes.

Oncogene Analysis of the mortality of the Italian Rb patients A Acquaviva et al 5355 for overall mortality (SMR 13.34, 95% CI 10.74–16.56). diagnosis. Besides the osteosarcoma, many second As expected, survivors of hereditary Rb had higher malignancies were found in Rb patients: rhabdomyo- mortality from cancers of the bone (SMR 391.90, 95% sarcoma and soft-tissue sarcoma, leukaemia, breast CI 203.90–753.20) and soft tissue (SMR 453.00, 95% CI cancer, bladder cancer, lung cancer, brain tumours, 203.50–1008.40), small intestine (SMR 1375.50, 95% CI melanoma, Hodgkin’s disease and others. Abramson 344.00–5499.70), nasal cavity (SMR 13.71, 95% CI et al. (2001) described many cases of Rb survivors who 1.93–97.35) and cancers of the brain and central nervous developed third, fourth and fifth as multiple nonocular system (SMR 41.14, 95% CI 13.2–127.55). tumours. Compared with the general population, survivors of Subsequently, the observation of SMNs occurred in unilateral sporadic Rb had statistically significantly Rb patients not cured with radiotherapy and SMNs in higher mortality from myeloid leukaemia (SMR 9.03, patients who did not receive both radiotherapy and 95% CI 2.26–36.12) and other cancers (ovary and chemotherapy stimulated many studies finalized to kidney) (SMR 4.41, 95% CI 1.10–17.63). explain this interesting biological phenomenon. Draper et al. (1986) signalled a high risk for second primary cancers in Rb patients treated with alkylating Discussion agents. The high rates of SMN in hereditary Rb than The aim of this study is to describe a large cohort of sporadic Rb suggested an underlying genetic suscepti- Italian Rb patients and to evaluate overall and cause- bility to a large variety of cancers. Hereditary Rb specific mortality in comparison with the general carrying RB1 mutations have a lifetime predisposition population. to SMN. Blanquette et al. (1995) have reported that Rb Cumulative mortality from all causes at 76 years patients with mutations in exon 19 tend to develop of age was 43%, from Rb was 18% and from SMNs was SMN. On the contrary, Lohman et al. (1996) did not 18%. Cases diagnosed before 1985 had a survival rate of confirm the correlation between RB1 gene mutations 76%, whereas for those diagnosed subsequently, it was and second cancers. Matsumoto and Kobayashi (2003) significantly higher (93%). Forty-one second primary described a patient with bilateral Rb who developed tumours were notified as occurring in the cohort; 38 of multifocal bone cancer 10 years after undergoing them had complete information. Thirty-one SMNs treatment for primary tumour and carrying a novel occurred in hereditary Rb cases. The range of latency germline mutation in the RB1 gene. time between the Rb diagnosis and the SMN diagnosis Although the occurrence of SMNs is actually well was 1–48 years. Sarcomas were the prevalent morpho- recognized, there is some discussion about their logy. Rb patients had a greater than 10-fold excess in incidence. Actually the incidence of second malignancies overall mortality compared with that in the general at 10 years from initial diagnosis has been reported to population (SMR 10.73). In particular, survivors of range from 4 to 20% (Abramson et al., 1984; Draper hereditary Rb had higher mortality from cancers of the et al., 1986; Lueder et al., 1986). This can be explained bone and soft tissue, whereas unilateral sporadic Rb had by multiple reasons: different radiotherapy techniques statistically significantly higher mortality from myeloid and dosages, different proportion of patients exposed, leukaemia and other cancers (ovary and kidney). different chemotherapy protocols, the length of obser- The Rb was among the first cancer for which an vation time, the size of cohort, the proportion between unusual frequency of second primary tumours was hereditary and sporadic Rb, follow-up, selection bias observed. It is well known that in developed countries, and choice of method for statistical analysis. (Aerts SMN is the primary cause of death for hereditary Rb, et al., 2004). but also sporadic Rbs are at risk for SMN. The next step of the study will be to perform the At first, the occurrence of SMN in Rb was attributed incidence follow-up for SMNs and to retrieve complete to radiotherapy alone. In fact, the majority of patients information on radiation treatment in order to clarify were treated with high dose of radiotherapy. Sagerman the effect of age at radiation on the subsequent risk et al. (1969) showed that the higher occurrence of of second nonocular tumours. secondary cancers in these patients is dose-dependent. Wong et al. (1997) reported that in a large US cohort, almost 90% of patients were cured with radiotherapy Patients and methods and many of those received between 1937 and 1965 very high dose of radiotherapy (111 Gy was the mean orbital We started in 1985 a retrospective and prospective clinical– dose among patients who developed osteosarcoma). The epidemiological study on Rb in Italy. second tumour may develop either inside or outside the We sent a form for collecting clinical, genetical, pathologi- field of radiation. The high incidence of cancers outside cal, therapeutic and follow-up data to ophthalmologic and the irradiation field in survivors of hereditary Rb paediatric clinics, radiotherapy centres, pathology and genetic institutes and cancer registries, requiring information on suggested in those a high risk of tumours in adult- patients with a diagnosis of Rb. Many patients were recruited hood. Chauveinc et al. (2001) have estimated that in Rb inquiring the database of the Italian Association Paediatric patients, osteosarcoma occurred 1.2 years earlier inside Hematology-Oncology (AIEOP). than outside the radiation field with a latency period of We also contacted paediatric and ophthalmic referral 1.3 years between radiotherapy and osteosarcoma centres in London, Paris, Essen, Barcelona, Lausanne, Zurich

Oncogene Analysis of the mortality of the Italian Rb patients A Acquaviva et al 5356 and Utrecht in order to identify Italian patients who went Reference mortality rates from the general Italian popu- abroad for diagnosis and treatment. lation to calculate expected deaths in our cohort were available As our study is both retrospective and multicentric, data since 1955, so all patients enter the follow-up after 1 January collection is incomplete. 1955; we used mortality rates specific for calendar year A total of 1209 patients with Rb diagnosed between 1923 (annual), age (quinquennial classes) and sex published by the and 2005 were collected in the Italian Retinoblastoma Statistics Office of ISS (http://www.iss.it). Registry. Stata statistical software was used to perform all the We checked each patient’s vital status at municipal registry analyses (StataCorp, 2006). offices every 2 years, requiring a copy of the death certificate. Patients were classified on the basis of their clinical records Acknowledgements as either unilateral or bilateral and as either hereditary (positive family history for Rb) or apparently sporadic (unilateral with no family history for Rb) and as genetic We thank the following colleagues for their kind collaboration (all bilateral and multifocal unilateral cases) or non-genetic in the data collection: C Costagliola, Department of Oph- (unifocal unilateral cases). thalmology, University of Napoli; P Fiorani, Department Exclusion criteria for the mortality analysis were as follows: of Pediatrics, University of Ancona; ML Pinello and G Perilongo, Department of Pediatrics, University of Padova; (a) residence outside Italy, PL Marradi, Department of Pediatrics, University of Verona, (b) diagnosis after 2003, A Garaventa and B De Bernardi, Istituto G Gaslini, Genova; (c) age at diagnosis >15 years and L Notarangelo, Department of Pediatric, University of (d) death before 1955. Brescia; G Bernini, Department of Pediatrics, University of Firenze; C Ferrari and P Picci, Istituto Ortopedico Rizzoli; We considered eligible for our statistical evaluation 1111 V Ninfo and R Alaggio, Department of Pathology, University patients, diagnosed between 1 January 1923 and 31 December of Padova; A Schiavetti, Department of Pediatrics, University 2003 (last vital status control). ‘La Sapienza’, Roma; FFossati-Bellaniand R Luksch, We classified our cohort by age at Rb diagnosis (o1, 1, 2, Department of Pediatrics, Istituto Nazionale per lo studio e 3–15 years), year of diagnosis (1923–1964, 1965–1974, 1975– la cura dei tumori, Milano; B Ricci, Department of 1984, 1985–1994, 1995–2003), age at last observation (1–9, Ophthalmology, University Cattolica del Sacro Cuore, Roma; 10–19, 20–29, 30–39, 40,76), family history and treatment M Arico` , Unit of Pediatric, Hematology-Oncology, Ospedale reported (surgery only, radiotherapy only, chemotherapy only, ‘G. Di Cristina’, Palermo; A Zanasso, Department of chemoradiotherapy, uncertain therapy). Pediatric, University of Trieste; M Comis, Divisione di We defined SMNs according to the Warren and Gates Ematologia, Azienda Ospedaliera di Reggio Calabria, Reggio criteria: (a) each of the tumours must present a definite picture Calabria; G Murgia, Department of Pediatrics, University of malignancy; (b) each tumour must be distinct and (c) the of Cagliari; S Calabro` , Divisione di Oculistica, Ospedale probability that the SMN is a metastatic lesion from the Santobono, Napoli; N Delle Noci, Department of Oph- primary tumour must be excluded. As a pinealoblastoma is thalmology, University of Bari; P Perri, Department of histologically identical to Rb, in agreement to some authors, Ophthalmology, University of Ferrara; Magnani C, Registro we excluded pinealoblastomas from SMNs counting them Tumori Infantili del Piemonte; S Ferretti, Registro dei as Rbs. Tumori della Provincia of Ferrara, Dipartimento di Medicina No systematic follow-up has been attempted until now in sperimentale e diagnostica, University of Ferrara; P Zambon, order to ascertain incidence of SMNs in our cohort. We could Registro Tumori del Veneto, Azienda Ospedaliera of study mortality for all causes included SMN, collecting this Padova, Padova; FBellu’, Registro Tumori Alto-Adige; JE information from death certificates and the characteristics Kingston, St Bartholomew’s Hospital, London; Munier F of patients for whom we could register the occurrence of an and Balmer A, Clinique Ophthalmic, Universite´ de Lausanne; SMN (no benign tumours were notified). T Lehnert, Klinische Forschergruppe Ophthalmologische We applied the Kaplan–Meier method (Kaplan and Meier, Onkologie un Genetik, Universitatsklinikum Essen , Germany; 1958) for estimating the cumulative mortality for all causes, J Schipper, Arnhems Radiotherapeutisch Instituut, Arnhem, Rb and SMNs; all statistical significance tests were two-sided. The Netherlands. This study was supported partly by a grant We estimated the SMRs and their exact Poisson 95% CIs from the Italian Association for Research on Retinoblastoma for all causes, SMNs and other causes. (AIRR).

References

Abramson DH, Ellsworth RM, Kitchin FD, Tung G. (1984). Chauveinc L, Mosseri E, Quintana E, Desjardin L, Schlienger Ophthalmology 91: 1351–1355. P, Doz F et al. (2001). Ophthalmol Genet 2: 77–88. Abramson DH, Melson MR, Dunkel IJ, Frank CM. (2001). DerKinderen DJ, Koten JW, Wolterbeck R, Beemer FA, Tan Ophthalmology 180: 1868–1876. KE, Den Otter W. (1987). Ophthalmol Pediatr Genet 8: 23–25. Abramson DH, Romnner HJ, Ellsworth RM. (1979). Am J Draper GJ, Sanders BM, Kingston JE. (1986). Br J Cancer 53: Ophthalmol 87: 624–627. 661–671. Acquaviva A, Bardelli AM, Bernardini C, D’Ambrosio A, Eng C, Li FP, Abramson DH, Ellsworth RM, Wong FL, Fois A, Frezzotti R et al. (1985). Boll Ocul 64: 141–159. Goldman MB et al. (1993). J Natl Cancer Inst 85: 1121–1128. Aerts I, Pacquement H, Doz F, Mosseri V, Desjardin L, Sastre Fletcher O, Easton D, Anderson K, Gilham C, Jay M, Peto J. X. (2004). Eur J Cancer 40: 1522–1529. (2004). J Natl Cancer Inst 96: 357–363. Blanquette V, Turleau C, Gross-Morand MS, Senamaud- Friend SH, Bernards R, Rogelj S, Weinberg RA, Rapaport Beaufort C, Doz F, Besmond C. (1995). Hum Mol Genet 4: JM, Albert DM et al. (1986). Nature 323: 643–646. 383–388. Kaplan EL, Meier P. (1958). JAMA 53: 457–481.

Oncogene Analysis of the mortality of the Italian Rb patients A Acquaviva et al 5357 Kaye Fj, Harbour JW. (2004). J Natl Cancer Inst 96: 342–343. Roarty JD, McLean IW, Zimmerman LE. (1988). Ophthal- Lohman DR, Brandt B, Hopping W, Passarge E, Horsthemke mology 95: 1583–1587. B. (1996). Am J Hum Genet 58: 940–949. Sagerman RH, Cassidy JR, Tretter P, Ellsworth RM. (1969). Lueder GT, Judish F, O’Gorman TW. (1986). Arch Ophthal- Am J Roentgend Ther Nucl Med 105: 529–535. mol 104: 372–373. StataCorp (2006). Stata Statistical Software: Release 9.1. Matsumoto H, Kobayashi O, Tamura K, Semine I, Stata: College Station, TX. Yamamoto M. (2003). Pediatr Int 45: 728–730. Wong FL, Boice Jr JD, Abramson DH, Tarone RE, Moll AC, Imhof SM, Kuik DJ, Bouter LM, Den Otter W, Kleinerman RA, Stovall M et al. (1997). JAMA 278: Bezemer PD et al. (1997). Hum Genet 98: 109–112. 1262–1267.

Oncogene