Cancer incidence after exposure to low dose and high dose ionising radiation

Abstract

Ionising radiation is a well-established carcinogen in humans. However, there still remain many unsolved questions even about such basic aspects as the dose-response relationship both in the low dose and in the high dose range. During the period 1930-1965 11,967 infants were treated with radium-226 for skin hemangioma at the Radiotherapy Department, Sahlgrenska Hospital. The median age at the first treatment was 5 months. The frequency of the treatments decreased in the beginning of the 1960s when information about the spontaneous regression was generally accepted. The individuals have been followed up in the Swedish Cancer Register (SCR) from 1958 to 1993 and absorbed doses to different risk organs have been calculated in a computed dose-planning system. The Standardised Incidence Ratio (SIR) for all malignancies was 1.2 (95% CI 1.1-1.4).47 intracranial tumours were observed, SIR=1.8 (95% CI 1.3-2.4). The mean brain dose was only 7 cGy. The cohort was pooled with a similar cohort from Stockholm and the analysis of the pooled data, including 86 intracranial tumours in 28,008 individuals, revealed a linear dose-response relationship, which was modified by age at exposure. Exposure during the first months of life resulted in higher risks. The excess relative risk (ERR) per Gy was 2.7 (95% CI 1.0-5.6).An analysis of the risk for breast cancer with pooled data from both the Gothenburg and the Stockholm hemangioma cohorts revealed 245 women with breast cancer, SIR=1.2 (95% CI 1.1-1.4). The mean breast dose was 29 cGy. A linear dose-response relationship was established, but no significant dose-effect modification of age at exposure, dose-rate or ovarian dose could be detected. The ERR/Gy was 0.35 (95% CI 0.18-0.59).In order to study dose-response relationships in the high dose range, the risk of soft tissue sarcoma (STS) was studied in breast cancer populations via the SCR between 1958 and 1992. The SIR was 1.9 (95% CI 1.5-2.2), but the absolute risk was low, 1.3/10,000 person-years. The energy imparted for the breast irradiations was estimated in a case-control study. 67 of 107 STS were clustered in the breast region. For STS of other types than angiosarcoma the risk increased linearly with dose up to an energy imparted of 150 J. At higher doses the risk stabilised, corresponding to a linear-quadratic dose-response relationship, with a negative coefficient for the quadratic term. The odds ratio was 2.4 (95% CI 1.4-4.2) for an energy imparted corresponding t dose range, where a negative quadratic term indicated a decline of the risks at higher doses.o a tangential irradiation of the breast after breast-conserving surgery. For angiosarcomas only oedema of the arm, but not integral dose, was a significant predictor.Linear dose-response relationships were thus found in the low dose range and a linear-quadratic model fitted best in the high