Page 420 - 2014 Printable Abstract Book
P. 420
(PS7-72) Confounding of the association between radiation exposure from CT scans and cancer risk by
2
1
2
cancer susceptibility syndromes. Johanna M. Meulepas ; Cecile M. Ronckers ; Johannes H.M. Merks ; Jay
1
1
3
H. Lubin ; and Michael Hauptmann , Netherlands Cancer Institute, Amsterdam, Netherlands ; Academic
3
Medical Center Amsterdam, Amsterdam, Netherlands ; and National Cancer Institute, Bethesda, MD
2
Background: Given the enormous number of diagnostic Computed Tomography (CT) scans
performed each year, many in children, it is of great public health importance to understand and quantify
the cancer risks associated with these scans. Three large cohort studies conducted in the UK, Australia
and Taiwan have shown elevated cancer risks following pediatric radiation exposure from CT scans, and
more studies are underway. Commentators have suggested these results may be partially due to
confounding by indication. Cancer susceptibility syndromes (CSS) represent one potential confounder
since CSS increase the risk of cancer and may be associated with an increased likelihood of having a CT
scan, thereby confounding the association between radiation exposure from CT scans and cancer risk.
Methods: We determined the strength of the association between CSS and specific cancers and estimated
the magnitude of potential confounding of CT-related cancer risk for leukemia and brain tumors, based
on assumptions about the frequency of CT scans among CSS patients compared with the general
population. We separately evaluated scenarios for external comparisons (standardized incidence ratios,
SIR) and internal comparisons (relative risks, RR). Results: Based on current knowledge and assumptions,
most CSS are too rare to cause a more than 10% overestimation of SIRs or RRs. The most common
syndromes can substantially bias CT-related cancer risk upwards if CT scans are used much more often
among affected patients than among others. Available data on the use of CT scans in diagnosis of and care
for patients with CSS do not support these scenarios. Unfortunately, reliable data are scarce; conversely,
biases related to CSS cannot be ruled out entirely. Conclusions: Single CSS are unlikely to substantially
confound radiation-related risks of leukemia and brain tumors in studies of CT scans. Nevertheless,
empirical data on the likelihood of CT scanning among CSS patients vs. the general population are urgently
needed in order to inform the interpretation of previous and future studies on CT scanning and cancer
risk.
(PS7-73) Organ dose calculations for patients undergoing computed tomography examinations with
automatic exposure control. Dayton D. McMillan; Michael Choi; Vana Derderian; and Choonsik Lee
National Cancer Institute, Rockville, MD
Background: In many countries radiological imaging has been increasingly utilized in medicine,
making these procedures significant sources of radiation exposure to individual patients and the public.
Of these procedures, computed tomography (CT) scans are of importance because of the relatively high
dose patients receive from them compared to other radiation diagnostic procedures. To date, several
methods have been introduced to estimate average organ dose a patient is expected to receive from
conventional CT scan. However, methods to estimate accurate organ doses for newer CT scanning
techniques such as automatic exposure control (AEC), a technique to reduce doses, are not available. In
this work we developed a method to estimate organ doses for patients undergoing CT scans with AEC
technique. Methods: We extended a CT dosimetry method that we previously developed for conventional
CT scans (not for AEC techniques), where we modeled a reference CT scanner (Siemens Sensation 16) and
experimentally validated using anthropomorphic physical phantoms. We selected 10 adult patients from
National Institute of Health Clinical Center, who underwent AEC-implemented chest-abdomen-pelvis CT
2
1
2
cancer susceptibility syndromes. Johanna M. Meulepas ; Cecile M. Ronckers ; Johannes H.M. Merks ; Jay
1
1
3
H. Lubin ; and Michael Hauptmann , Netherlands Cancer Institute, Amsterdam, Netherlands ; Academic
3
Medical Center Amsterdam, Amsterdam, Netherlands ; and National Cancer Institute, Bethesda, MD
2
Background: Given the enormous number of diagnostic Computed Tomography (CT) scans
performed each year, many in children, it is of great public health importance to understand and quantify
the cancer risks associated with these scans. Three large cohort studies conducted in the UK, Australia
and Taiwan have shown elevated cancer risks following pediatric radiation exposure from CT scans, and
more studies are underway. Commentators have suggested these results may be partially due to
confounding by indication. Cancer susceptibility syndromes (CSS) represent one potential confounder
since CSS increase the risk of cancer and may be associated with an increased likelihood of having a CT
scan, thereby confounding the association between radiation exposure from CT scans and cancer risk.
Methods: We determined the strength of the association between CSS and specific cancers and estimated
the magnitude of potential confounding of CT-related cancer risk for leukemia and brain tumors, based
on assumptions about the frequency of CT scans among CSS patients compared with the general
population. We separately evaluated scenarios for external comparisons (standardized incidence ratios,
SIR) and internal comparisons (relative risks, RR). Results: Based on current knowledge and assumptions,
most CSS are too rare to cause a more than 10% overestimation of SIRs or RRs. The most common
syndromes can substantially bias CT-related cancer risk upwards if CT scans are used much more often
among affected patients than among others. Available data on the use of CT scans in diagnosis of and care
for patients with CSS do not support these scenarios. Unfortunately, reliable data are scarce; conversely,
biases related to CSS cannot be ruled out entirely. Conclusions: Single CSS are unlikely to substantially
confound radiation-related risks of leukemia and brain tumors in studies of CT scans. Nevertheless,
empirical data on the likelihood of CT scanning among CSS patients vs. the general population are urgently
needed in order to inform the interpretation of previous and future studies on CT scanning and cancer
risk.
(PS7-73) Organ dose calculations for patients undergoing computed tomography examinations with
automatic exposure control. Dayton D. McMillan; Michael Choi; Vana Derderian; and Choonsik Lee
National Cancer Institute, Rockville, MD
Background: In many countries radiological imaging has been increasingly utilized in medicine,
making these procedures significant sources of radiation exposure to individual patients and the public.
Of these procedures, computed tomography (CT) scans are of importance because of the relatively high
dose patients receive from them compared to other radiation diagnostic procedures. To date, several
methods have been introduced to estimate average organ dose a patient is expected to receive from
conventional CT scan. However, methods to estimate accurate organ doses for newer CT scanning
techniques such as automatic exposure control (AEC), a technique to reduce doses, are not available. In
this work we developed a method to estimate organ doses for patients undergoing CT scans with AEC
technique. Methods: We extended a CT dosimetry method that we previously developed for conventional
CT scans (not for AEC techniques), where we modeled a reference CT scanner (Siemens Sensation 16) and
experimentally validated using anthropomorphic physical phantoms. We selected 10 adult patients from
National Institute of Health Clinical Center, who underwent AEC-implemented chest-abdomen-pelvis CT
