Second primary malignancies in females with primary fallopian tube cancer
Article first published online: 31 JAN 2007
Copyright © 2007 Wiley-Liss, Inc.
International Journal of Cancer
Volume 120, Issue 9, pages 2047–2051, 1 May 2007
How to Cite
Riska, A., Pukkala, E., Scélo, G., Mellemkjaer, L., Hemminki, K., Weiderpass, E., McBride, M. L., Pompe-Kirn, V., Tracey, E., Brewster, D. H., Kliewer, E. V., Tonita, J. M., Kee-Seng, C., Jonasson, J. G., Martos, C., Boffetta, P. and Brennan, P. (2007), Second primary malignancies in females with primary fallopian tube cancer. Int. J. Cancer, 120: 2047–2051. doi: 10.1002/ijc.22562
- Issue published online: 28 FEB 2007
- Article first published online: 31 JAN 2007
- Manuscript Accepted: 24 NOV 2006
- Manuscript Received: 14 NOV 2006
- Finnish Cultural Foundation
- The Finnish Cancer Foundation
- The Finnish Medical Foundation
- Lilly Foundation
- The National Cancer Institute. Grant Number: R03 CA101442-02
- primary fallopian tube cancer;
- second primary cancer;
- multi-centre cohort study
Primary fallopian tube cancer (PFTC) is a rare disease, and its aetiological factors are poorly understood. Studies on PFTC in the setting of 2nd primary malignant neoplasms can provide clues on aetiology and also define the possible side effects of different treatment modalities for PFTC. A cohort of 2,084 cases with first PFTC was extracted from the data from 13 cancer registries from Europe, Canada, Australia and Singapore and followed for second primary cancers within the period 1943–2000. Standardized incidence ratios (SIRs) were calculated and Poisson regression analyses were done to find out the RRs related to age at, period of and time since the PFTC diagnosis. There were 118 cancer cases observed after first PFTC (SIR 1.4, 95%CI 1.1–1.6). Elevated SIRs were seen for colorectal cancer (1.7, 95%CI 1.0–2.6), for breast cancer (1.5, 95%CI 1.1–2.2), for bladder cancer (2.8, 95%CI 1.0–6.0), for lung cancer (1.8, 95% CI 0.9–3.2) and for nonlymphoid leukaemia (3.7, 95%CI 1.0–9.4). Significant risk increases were detected for colorectal cancer during the 2nd to 5th year after the first PFTC diagnosis, for breast cancer in follow-up 10+ and for nonlymphoid leukaemia during the 2nd to 10th year. The clustering of cancers of the lung and bladder in PFTC patients may suggest shared smoking aetiology. The excess of colorectal and breast cancers after PFTC may indicate a genetic aetiology. © 2007 Wiley-Liss, Inc.
Primary fallopian tube cancer (PFTC) is a rare disease, contributing 0.3%–1.6% of female genital tract malignancies in Western countries.1, 2, 3 Studies from the United States indicated an incidence rate of 0.4 per 100,000 in the 1980s; with the rate being higher in Caucasian women including Hispanics than in African Americans.4 In Finland, the age-adjusted incidence rate increased from 0.1/100,000 in 1953–1957 to 0.5/100,000 in 1993–1997.3 In Denmark, the incidence rate during a 5-year period around 1980 was 0.3/100,000.5 The 5-year survival rate has varied between 22 and 57% over times and regions.5, 6, 7, 8
Little is known about aetiology of PFTC. More than 90% of the cases are serous adenocarcinomas and the histological appearance resembles that of ovarian serous carcinomas.1, 4, 9 Fallopian tube is a hormone sensitive organ. The epithelial cells of fallopian tube undergo changes during the menstrual cycle in response to changes in hormonal levels, particularly oestrogen. PFTC has been generally treated as ovarian cancer since both tumors have similar histologies and biological behavior.10 Studies have reported clustering of breast cancers and PFTC in same patients.2, 11, 12 There are also several findings on higher BRCA1 and BRCA2 mutation prevalence among PFTC patients than in average healthy population.13, 14, 15, 16
The present investigation is a unique multi-centre study including cancer data from 13 population-based cancer registries in Europe, Australia, Canada and Singapore.
Material and methods
This study is part of an international multi-centre study of second primary cancers coordinated by the International Agency for Research on Cancer (IARC) including data from 13 cancer registries in Europe, Australia, Canada and Singapore that have been in operation for at least 25 years (Table I). Details of data handling and standardization between the 13 participating registries have been described elsewhere.17, 18 Coding of multiple primaries in the cancer registries has followed a common set of rules proposed by the International Association of Cancer Registries (IACR) and the IARC.19 This was possible as all participating cancer registries currently use the IARC/IACR rules or a local set of more extensive or detailed rules. According to the rules, a primary cancer is one which originates in a primary site or tissue and is thus neither an extension, nor a recurrence nor a metastasis. Only one tumor shall be recognized as arising in an organ or pair of organs or tissue unless the histology is different. Second tumors occurring in the fallopian tube were however not analyzed, because rules were not compatible between registries or over time. Cancers of the brain and nervous system, bladder cancer and nonmelanoma skin cancer have been registered differently in the various registries. In the present study, we followed the same rules as applied in the series of volumes of Cancer Incidence in 5 Continents.20
|Characteristics||PFTC||Second primary cancer|
|n||%||n||Cumulative incidence of second primary cancers (%)|
|Age at PFTC diagnosis|
|Calendar period of PFTC diagnosis|
|Time since PFTC diagnosis (full years)|
|Australia, New South Wales||72||3.4||7||9.7|
|Canada, British Columbia||112||5.4||12||10.7|
The data from the 13 cancer registries were checked carefully for inconsistencies and missing information, with verification provided by the local registry when required. A small proportion of subjects (0.06%) were excluded because of remaining missing or inconsistent information. Furthermore, patients for whom the first primary cancer diagnosis and death were recorded at the same time or who had 2 first primary cancers recorded simultaneously (same dates of diagnosis) were excluded (8%). After these exclusions, there were 2,084 women with a first PFTC. No treatment data for PFTC were available.
All cases of PFTC were followed up for second primary cancer from date of first diagnosis (1943–2000) to date of second primary cancer (1943–2000), date of death, date of migration or end of follow-up (1992–2000). The number of second primary cancers observed was compared with the expected number of cancers calculated from accumulated person-years and rates among females of first PFTC specific for each registry and 5-year age and calendar-periods. The standardized incidence ratios (SIRs) were stratified for time since PFTC diagnosis (<1year, 1–4, 5–9 and 10+ years after the diagnosis first PFTC), for calendar-period of PFTC (<1975, 1975+) and for age at PFTC diagnosis (21–49 years, 50–59 years, 60–92 years). A Poisson regression analysis was done for selected cancer sites to quantify the independent risk ratios (RR) related to each variable.
The study population of 2,084 women with a first PFTC contributed 11,047 person-years of follow-up (median 2.5, mean 5.3, maximum 49.6 years). At the time of the PFTC diagnosis, 52% of women were 60+ years-old, 24% were diagnosed before 1975 (Table I). The majority of cases were from Europe (87%). The proportion of patients having a second primary cancer after PFTC during the follow-up was 5.7% (Table I).
The SIR of second primary cancer in all sites combined after a PFTC was 1.4 (95% CI = 1.1–1.6) (Table II). Elevated risks were seen for colorectal cancer (SIR 1.7, 95% CI 1.0–2.6), breast cancer (SIR 1.5, 95% CI 1.1–2.2), bladder cancer (SIR 2.8, 95% CI 1.0–6.0), nonlymphoid leukaemia (SIR 3.7, 95% 1.0–9.4) and lung cancer (SIR 1.8, 95% CI 0.9–3.2). There were no second cancers with significantly decreased risk.
|Cancer site (ICD-9)||Observed||SIR||95% CI|
|All malignant (140–208)||118||1.4||1.1–1.6|
|Oral cavity, pharynx (140–149)||2||1.5||0.2–5.6|
|Small intestine (152)||1||3.5||0.1–20|
|Colorectal (153, 154)||20||1.7||1.0–2.6|
|Liver (155.0, 155.1)||0||–||0–5.0|
|Gallbladder, bile ducts (156)||3||1.9||0.4–5.5|
|Soft tissue sarcoma (171)||1||2.5||0.1–14|
|Melanoma of skin (172)||1||0.4||0.0–2.5|
|Other neoplasm of skin (173)||9||1.5||0.7–2.9|
|Cervix uteri (180)||2||0.7||0.1–2.4|
|Corpus uteri (182)||2||0.3||0.0–1.3|
|Other female genital (179, 184)||2||1.8||0.2–6.7|
|Bladder (188, 189.3–4)||6||2.8||1.0–6.0|
|Kidney (189.0–2, 189.5–9)||2||0.9||0.1–3.1|
|Brain, nervous system (191–192)||0||–||0–3.3|
|Thyroid gland (193)||2||2.3||0.3–8.3|
|Hodgkin's disease (201)||0||–||0–15|
|Non-Hodgkin's lymphoma (200, 202)||2||0.9||0.1–3.1|
|Multiple myeloma (203)||1||0.8||0.0–4.4|
The SIR for second cancer (all sites combined) was highest if the time since PFTC diagnosis was 10+ years, age at PFTC diagnosis <60 years, or the year of PFTC diagnosis <1984 (Table III).
|Factor||Cancer site (ICD-9)|
|All malignant (140–208)||Female breast (174)||Colorectal (153,154)|
|Obs||SIR||95% CI||Obs||SIR||95% CI||Obs||SIR||95% CI|
|Age at PFTC diagnosis|
|Calendar period of PFTC diagnosis|
|Time since PFTC diagnosis (years)|
For breast cancer, a significantly elevated SIR of 2.3 (95% CI 1.2–3.8) was seen over 10 years after the diagnosis of PFTC (Table III). Also in the multivariate analysis there was an indication of an increase of breast cancer cases 10+ years after the PFTC as compared with shorter follow-up times. The age or calendar period of the diagnosis had no effect.
For colorectal cancer, a significant increase (SIR 2.1, 95% CI 1.3–4.9) was detected during the 2nd to 5th year after PFTC diagnosis (Table III). If the PFTC was diagnosed before the age of 60 years, the relative risk of colorectal cancer tended to be high (combined SIR of age categories <50 and 50–59 years 2.2, 95% CI 1.1–4.0). Multivariate analysis gave essentially the same result: the RR 1–4 years after the PFTC was 3.3 as compared with follow-up 10+ years (95% CI 1.0–11) and the RR related to PFTC diagnosed in ages 60+ years or in 1991+ as compared with younger ages or earlier calendar periods, respectively.
In multivariate analysis for the remaining cancers combined (after exclusion of breast and colorectal cancers) there was some indication of a higher risk related to PFTC diagnoses before 1984. Elevated risks after first PFTC were also detected for nonlymphoid leukaemia during the 2nd to 10th year after PFTC diagnosis (4 cases, SIR 6.9, 95% CI 1.9–17.8) and for bladder cancer in follow-up 5+ years (7 cases, SIR 4.2, 95% CI 1.4–9.7). There were 2 cases of endometrial cancer and 2 cases of cervical cancer all of them diagnosed within 1 year after the PFTC.
The number of fallopian tube cancer survivors has increased because of advances in earlier detection, treatment and supportive care. Therefore the risk of second primary cancers has become clinically more relevant. Our study––the largest one on second primary cancers after the first PFTC reported so far––showed that women with first PFTC had a 40% increased risk of developing a new primary cancer as compared with the general population. There were excesses in breast cancer, colorectal cancer, leukaemia, bladder cancer and possibly lung cancer.
Little is known about the aetiology and epidemiology of PFTC. Its aetiology is largely unknown but may be similar to that of ovarian carcinoma, since they both arise from the Müllerian duct. Previous studies have suggested that nulliparous women have an increased risk of PFTC, as for ovarian and breast cancer.1, 21, 22 There are earlier studies reporting a concomitant presentation of breast cancer, other gynaecological cancers and colon cancer among patients with PFTC,2, 12, 23, 24 suggested to reflect their similar hormone responsiveness.24 First degree relatives of 44 PFTC cases had a suggestion of an increase in risk of ovarian cancer (RR = 2.2, 95% CI 0.4–6.3) and of early-onset breast cancer (RR = 2.4, 95% CI 0.6–6.1)14; germline BRCA1 mutations were found in 5 (11%) patients with PFTC and BRCA2 mutations in 2 (5%) patients. In comparison, out of unselected ovarian patients 3.5% were found to have BRCA1 mutations13 and the prevalence of BRCA1 or BRCA2 mutations in the general population is 0.1%–0.2%.15 Brose et al. reported a cumulative risk of 3.0% (95% CI 1.3–4.7%) for PFTC among BRCA1 carriers as compared with 0.025% in the general population.16 Other reports have confirmed an increased risk of PFTC in BRCA1 or BRCA2 mutation carriers.25, 26, 27, 28 A population-based study from Sweden confirmed a very high risk (SIR 93, 95% CI 55–146) of PFTC in Swedish families with both breast and ovarian cancer.29 In families with a case of breast cancer diagnosed before the age of 35 the SIR of PFTC was 5.5 (95% CI 1.0–16). In our study the excess risk of breast cancer became apparent more than 10 years after the PFTC diagnosis and the SIR was not highest in the youngest age groups, findings that appear unlikely to be explained by genetic susceptibility. The latency would fit better with a hypothesis of an effect of ionising radiation.30
Studies have advocated either pelvic or whole abdominal radiotherapy for treatment of PFTC, but because of low efficacy and a high rate of serious complications this should no longer be used except for palliation of specific symptoms.2, 8, 31, 32, 33 The total doses used in the pelvic or total abdominal radiotherapy have ranged from 25 to 70 Gy (median 50.0 Gy).34 There are no studies on the effect of radiation therapy, used as a treatment for PFTC, on breast or other second cancer risk. On the basis of studies concerning second malignancies among survivors of ovarian cancer35 the risk effect of radiation therapy on PFTC should be low. The dose delivered to the breast during whole abdominal radiation therapy for PFTC should be similar to the dose delivered to the contralateral breast from tangential field treatments of the breast cancer. In our data the excess of breast cancer, bladder cancer, colorectal cancer and leukaemia could be in part late effects of pelvic or abdominal radiation.
Studies on late effects of chemotherapy treatments for cancer have revealed an increased risk of second primary leukaemia, mainly acute myeloid leukaemia.36 The relative risk (RR) of leukaemia after platinum-based chemotherapy for ovarian cancer was 4.0 (95% CI 1.4–11).37 For carboplatin the relative risk was 6.5 (95% CI 1.2–37) and for cisplatin 3.3 (95% CI 1.1–9.4). The effect was dose related, with the RR reaching 7.6 (95% CI 2.3–25) at doses of 1,000 mg or more of platinum. The platinum-based chemotherapy has been the main adjuvant therapy for PFTC and could therefore explain the high rates of nonlymphoid leukaemia after PFTC. Since the 1990s taxanes have been combined with platinum-based drugs in the treatment of PFTC, especially in advanced disease with a residual tumor after the operation. There are case reports of secondary acute myeloid leukaemias after the treatment with taxanes.38, 39
All cases of nonlymphoid leukaemia and the majority of excess cases of breast cancer were diagnosed among PFTC patients treated after 1975, i.e., the hypothesis that the elevated SIRs are due to earlier cancer treatment protocols seems not to hold. There were no nonlymphoid leukaemia cases immediately after the diagnosis of the PFTC, but only during the 2nd to 10th year after the PFTC, which may indicate an effect of platinum-based chemotherapy.37 The overall excess of breast cancer, colorectal cancer and bladder cancer may reflect an effect of treatments with radiotherapy and/or chemotherapy.
There were 2 cases of cervical cancer and 2 endometrial cancers diagnosed within 1 year after the diagnosis of PFTC. These may be related to the additional diagnostic and treatment activities due to the PFTC.
Tobacco-smoking is a strong established risk factor for lung cancer and bladder cancer; it has been estimated that 79% of lung cancers and 32% bladder cancers around the year 2000 among women in the Nordic countries would be avoided if cigarette smoking were eliminated.40 The raised SIRs for those cancers as second primary cancers after PFTC suggest that smoking may also have a role in the aetiology of PFTC.
Our results are based on records collected from 13 cancer registries with different registration practises in multiple cancer coding. Misclassification of metastases as new primary cancers does not seem to have been a problem, since we observed reduced risk estimates for sites frequently exposed to metastasis, such as liver, bone and brain. This may reflect a conservative coding principle: a cancer is not accepted as new primary cancer if there is any doubt that it could be a metastasis.
The conclusion of this study is that history of a PFTC has an effect on the subsequent risk of breast, colorectal, bladder and lung cancers and nonlymphoid leukaemia. Although with the largest PFTC patient cohort ever collected for this kind of study, due to the rarity of this cancer and the lack of individual data on risk factors and treatment it is difficult to determine how much of the excess risks might be associated with shared genetic or life habit factors and how much is related to treatment of the PFTC.
- 19Classification and coding for neoplasms. In: JensenOM, ParkinDM, MacLennanR, MuirCS, SkeetRG, eds. Cancer registration: principles and methods. Lyon: IARC Scientific Publication, 1991. 64–81. IARC scientific publication no. 95., .
- 20Classification and coding. In: ParkinDM, WhelanSL, FerlayJ, TeppoL, ThomasDL, eds. Cancer incidence in five continents. Lyon: IARC, 2003. 35–41..