| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Mammography Specialists Medical Group, Inc., 14651 S Bascom Ave., Suite 210,
Los Gatos, CA 95032.
2 Departments of Radiology and Biomedical Engineering, University of California
Davis Medical Center, Sacramento, CA.
3 Lynn Sage Breast Center, Northwestern University Feinberg School of Medicine,
Chicago, IL.
4 Queen of the Valley Hospital, Napa, CA.
5 Radiological Associates Medical Group of Santa Clara Valley, San Jose,
CA.
Received July 30, 2006;
accepted after revision November 8, 2006.
Address correspondence to R. L Smathers
(rs{at}mammo.net).
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. Mean glandular dose to the breast during screening mammography was measured for 206 women who had undergone breast augmentation. For 13 of these women, mean glandular dose from preoperative mammography also was measured. Effective tube current, peak kilovoltage, and breast thickness were measured, and mean glandular dose was calculated for 1,632 images. Two screen-film combinations and three target-filter combinations were studied.
RESULTS. For four-view augmentation mammography with a molybdenum-molybdenum (Mo-Mo) target-filter combination, mean glandular dose was reduced 35%, from 10.7 to 7.0 mGy, by changing the screen-film combination from 100 to 190 speed. For four-view augmentation mammography, mean glandular dose was reduced 24% by changing the target-filter combination from Mo-Mo to rhodium-rhodium (Rh-Rh) for full views of breasts containing implants. For four-view augmentation mammography, mean glandular dose was reduced 50% by changing the screen-film combination from 100 to 190 speed and changing the target-filter combination from Mo-Mo to Rh-Rh for implant-full views.
CONCLUSION. Mean glandular dose per breast from four-view augmentation mammography with the 100-speed screen-film and Mo-Mo target-filter combinations averaged 10.7 mGy, which is 3.1 times higher than the 3.4 mGy for conventional two-view mammography of breasts without implants. In 40 years of screening, this number represents a more than tripled lifetime attributable risk of radiation-induced breast cancer—an unacceptable level. Use of faster screen-film combinations, use of Rh-Rh target-filter combinations, and acquisition of three rather than four views are dose-reduction methods that together result in a 66% dose reduction, from 10.7 to 3.6 mGy. Mean glandular dose should be kept less than 7.0 mGy per breast for screening mammography of patients with breast implants.
Keywords: breast cancer • implantable devices • mammography • physics • radiation dose • screening
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
We measured mean glandular dose from screening mammography performed with the routine four-view augmentation technique in examinations of 206 women with cosmetic breast implants. Preoperative and postoperative mammograms of 13 women who had undergone breast augmentation were compared. The mammographic exposure information included peak kilovoltage, effective tube current, and compressed breast thickness. Mean glandular dose for each image and total mean glandular dose per breast for nonimplant and implant examinations were calculated. The influence of screen-film combination speed, target-filter combination, and number of views obtained was evaluated.
|
|
|
|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Two hundred six patients, referred to as the single group, had data sheets completed in a 30-month period. Before augmentation surgery, 13 of the 206 single-group patients had undergone screening mammography in which exposure information was recorded on the images. These 13 patients were referred to as the paired group. Each of the 13 paired-group patients had exposure information recorded for both the preoperative and the postoperative mammograms. Implant status was designated non-implant when breast implants were absent. All data sheets were entered into an Excel (Microsoft) spreadsheet for analysis.
"ID" is the standard American College of Radiology designation for "implant-displaced" views [5]. "IF" stands for "implant-full" and is defined here as a mammographic view that intentionally includes the implant in the image and does not attempt to displace the implant out of the image. Most mammographic screening centers routinely use additional views for breast implant patients. Eklund et al. [2] recommend four views: craniocaudal and mediolateral oblique (MLO) without and with implant displacement. Implant mammograms in this study were obtained with two of the Eklund views and two similar but slightly different views. The craniocaudal implant-full view was replaced with a slightly laterally exaggerated craniocaudal view. The MLO implant-displaced view was replaced with a lateromedial implant-displaced view. These slight changes from the standard Eklund views avoided duplication of views when capsular contraction did not allow implant displacement.
Routine implant screening mammography was performed in the following four views: craniocaudal implant-displaced, craniocaudal implant-full, MLO implant-full, and lateromedial implant-displaced. Figure 1A, 1B, 1C, 1D shows an example of these four views. Eight images were evaluated for each patient in the single group: four views per breast. Twelve images were evaluated for each patient in the paired group: two preoperative views per breast and four postoperative views per breast.
The following information was recorded for each patient: date of examination, X-ray unit used, technologist, and radiologist. The following information was recorded for each breast: percentage of breast volume occupied by dense tissue; implant type—silicone, saline, or double lumen; implant location—subglandular or submuscular. The following information was recorded for each mammo-graphic image: side—right or left; view—craniocaudal implant-displaced, craniocaudal implant-full, lateromedial implant-displaced, MLO implant-full; breast compression force in newtons, tube angle in degrees, compressed breast thickness in centimeters, peak kilovoltage, and effective tube current in milliampere-seconds.
Percentage of breast volume occupied by dense breast tissue was visually assessed from the area of dense breast tissue seen on each image. Combining the area seen on two or more views allowed estimation of the percentage of breast volume occupied by dense breast tissue for each breast. This subjective visual method is similar to counting fatty (gray) pixels and dense (white) pixels on a digital mammographic histogram. Each breast was assigned a percentage of breast volume occupied by dense breast tissue of 10-90% in increments of 10% and assigned to one of three groups used to calculate glandular dose: g0, g.5, and g1 [6]: g0, fatty tissue replaced (dense tissue, 30% or less); g0.5, approximately one half dense and one half fatty tissue (dense tissue, 40-60%); and g1, dense tissue with little fat (dense tissue, 70% or more). In most cases the percentage of breast volume occupied by dense breast tissue was indicated in the original mammographic report. Example mammograms showing 10-90% of breast volume occupied by dense breast tissue are available [7]. The use of 10% categories for breast density allowed easy correlation with the three density groups used in the Sobol equation [6] and the four density groups used in the BI-RADS system.
Eight mammographic X-ray units were used. The source-image distance was 66 cm for five units and 75 cm for three units. All units were manufactured between 1990 and 1998 and had received annual physicist inspections. The eight units used were one GE DMR unit, one GE DMR+ unit, and two GE 800T units (GE Healthcare); one Lorad MIII unit (Lorad Medical Systems); and one Bennett Profile unit and two Bennett Contour units (Trex Medical). The 100-speed screen-film combination was a Kodak Min-R screen, single emulsion film (Microvision, Sterling/DuPont). and Kodak M-35 and X-Omat Multiloader 300 processors in standard cycle with Kodak RP X-Omat chemistry at 95°F (35°C) (Eastman Kodak Company). The 190-speed screen-film combination was a Kodak Min-R 2190 screen, Kodak Min-R 2000 film, and a Kodak Min-R processor in standard cycle with Kodak RP X-Omat chemistry at 95°F (35°C).
A physicist measured the following characteristics of each of the eight mammographic X-ray units: half-value layer (HVL), peak kilovoltage calibration, and radiation exposure in air. Equipment used included an ion chamber (model 303 3 cm3 parallel plate with Keithley 35080 electrometer) (CNMC Company) and a model 35080A peak kilo-voltage meter with Cd k-Edge Mammo Pack and Linear Mammo Pack (Keithley Instruments). Aluminum, the electrometer, and the ion chamber were used to measure the HVL for each peak kilovoltage setting for each mammographic X-ray unit. The electrometer and chamber were calibrated, and the calibration was traceable to the National Institute of Standards and Technology. The range of peak kilo-voltage tested matched the range used for the patient exposures for all mammographic X-ray units.
The peak kilovoltage of each mammographic X-ray unit was measured with the electrometer and the noninvasive peak kilovoltage meter. The values recorded with the electrometer were used to correct nominal peak kilovoltage values on each machine. The peak kilovoltage meter was calibrated, and the calibration was traceable to the National Institute of Standards and Technology. The radiation exposure was measured in air at 4.5 cm from the input surface of the screen-film holder with a fixed technique of 50 mAs and for each peak kilovoltage. The compression device was in the beam in contact with the ion chamber.
The following method was used for calculating mean glandular dose and normalized glandular absorbed dose for each mammographic image [3, 8-10]. For each image, the HVL was derived from the physicist's calibration tables for the specific X-ray unit and peak kilovoltage. Given the peak kilovoltage and HVL value, the ratio of exposure (in milliroentgens) to effective tube current (in milliampere-seconds) was derived from the physicist's calibration tables. This value was multiplied by the recorded effective tube current for each image. Skin entrance exposure in air was corrected with the inverse square law for the difference between actual compressed breast thickness and a 4.5-cm calibration phantom at the source-image distance of the specific X-ray unit. The normalized glandular absorbed dose was calculated in milligrays per roentgen from the equations published by Sobol and Wu [6] for each image. This calculation required the following for each image: target-filter combination, HVL, peak kilovoltage, compressed breast thickness, and breast density group (g0, g0.5, g1). The normalized glandular absorbed dose was multiplied by the corrected skin entrance exposure in air to yield the mean glandular dose.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
The 100-speed screen-film combination was similar to many of the standard mammographic screen-film combination systems currently used in the United States. The 190-speed combination is similar to the newer Kodak EV 190-screen and EV film. For the 100-speed Mo-Mo group, mean glandular dose per breast was 1.6 mGy for implant-displaced views and 3.7 mGy for implant-full views (mean glandular dose for implant-full view was 2.3 times higher than for implant-displaced view). The total mean glandular dose per breast for a four-view 100-speed Mo-Mo implant mammogram was 10.7 mGy. This dose was 3.1 times greater than that for a conventional two-view nonimplant 100-speed Mo-Mo mammogram. The mean glandular dose per breast for conventional two-view nonimplant mammography was 3.4 mGy for 100-speed Mo-Mo images and 2.6 mGy for 190-speed Mo-Mo images.
The total number of images in Table 1 includes extra, or repeated, images (obtained mostly for technical reasons). In this study, an extra view was any image obtained in addition to the standard four implant views for any reason, including positioning, technical factors, and repeated imaging. One extra view was obtained for 21 patients and two extra views for four patients. Therefore, at least one extra view was obtained for 12% of patients. Extra views accounted for 1.65% of all images. The average number of images was 4.02 per breast. For 100-speed Mo-Mo, the average total mean glandular dose for a four-view implant screening mammogram with extra views was 10.9 mGy, 3% higher than the 10.7-mGy dose without extra views. These results are similar to those of Burch and Goodman [11], who found the extra views in the two-view group contributed 3.2% of the total dose.
Table 2 shows techniques and mean glandular doses for nonimplant and implant mammograms. Before augmentation, the images are all nonimplant (conventional). After augmentation, they are implant-displaced and implantfull. Table 2 shows average values for implantdisplaced and implant-full view parameters, including breast compression force, compressed breast thickness, peak kilovoltage, effective tube current, and mean glandular dose. The average tube angle was 5.5° for craniocaudal implant-full views and 50.5° for MLO implant-full views. Average breast compression force was 51 N for all implant views. Average compressed breast thickness was 7.7 cm for implant-full views and 3.6 cm for implant-displaced views. For the 100-speed Mo-Mo images, the average effective tube current was 236 mAs for implant-full views and 80.2 mAs for implant-displaced views. The average HVL values were 0.285-0.508 mm aluminum. In this study, peak kilovoltage averaged 24.9 kVp for nonimplant, 25.5 kVp for implant-displaced, and 27.7 kVp for implant-full views. The following ratios were calculated for the 100-speed Mo-Mo group individually for each breast and then were averaged. The average compressed breast thickness ratio was 2.1 for implant-full over implant-displaced views. The average peak kilovoltage ratio was 1.1 for implant-full over implant-displaced views. The average effective tube current ratio was 2.9 for implant-full over implant-displaced views.
|
The 13 patients in the paired group under-went mammography before and after augmentation surgery. Table 2 shows the average values for the paired group for the two conventional preoperative nonimplant mammographic views (craniocaudal and MLO). The average MLO view was obtained with 51° of tube angulation. The average breast compression force for nonimplant views was 85 N. This force was 67% higher than the average breast compression force of 51 N for implant views.
Compressed breast thickness was similar for nonimplant views and implant-displaced views. Peak kilovoltage was slightly higher for implant-full views compared with implant-displaced and nonimplant views. The effective tube current was slightly lower for implant-displaced views compared with nonimplant views. Implant-full views had more effective tube current than implant-displaced and non-implant views. The mean glandular dose was averaged per view and per breast for each technique (nonimplant, implant-displaced, and implant-full). The mean glandular dose for the implant-displaced views was slightly lower than for the nonimplant views. All implant-displaced views were obtained with a Mo-Mo target-filter combination. Implant-full views were obtained with four techniques: 100-speed Mo-Mo, 190-speed Mo-Mo, 190-speed Mo-Rh, and 190-speed Rh-Rh. For the 100-speed Mo-Mo group, the mean glandular dose was 1.7 mGy for nonimplant, 1.6 mGy for implant-displaced, and 3.7 mGy for implant-full views. For both the 100-speed Mo-Mo and 190-speed Mo-Mo groups, implant-full views had more than twice the dose of the nonimplant views and implant-displaced views. Mean glandular dose for the 190-speed Rh-Rh group was almost the same as that for the 100-speed Mo-Mo nonimplant group (1.7 mGy).
Table 3 shows mean glandular dose before and after augmentation. These values are shown for the paired group before and after augmentation and for all images, both paired and unpaired, after augmentation. Twenty-six breasts were imaged in the paired before and after groups compared with 411 in the after (both) group. The table shows the averages for the three types of views (nonimplant, implant-displaced, and implant-full). It also shows total mean glandular dose for a complete examination per breast with the conventional two views, implant three views, and implant four views.
|
A two-view nonimplant mammogram had a total mean glandular dose of 3.4 mGy. A four-view implant mammogram had a total mean glandular dose of 10.7 mGy and a dose of 9.4 mGy for the paired group. The ratio of mean glandular dose for a four-view implant mammogram to that for a two-view nonimplant mammogram was 2.8 according to the single-group data and 3.1 according to the paired-group data. The mean glandular dose values in Tables 2 and 3 are close to those found by Klein [12], who reported an average mean glandular dose of 1.86 mGy per view for the Mo-Mo target-filter combination, 25 kVp, and compressed breast thickness less than 4 cm. The mean glandular doses in Tables 2 and 3 also are similar to those found by Jamal et al. [13], who reported 1.54 mGy for the craniocaudal view, 1.82 mGy for the MLO view, and 3.37 mGy mean per woman. The mean glandular dose per woman in our study was 3.4 mGy for the nonimplant two-view examination.
A four-view 100-speed Mo-Mo implant mammogram had a total per breast mean glan-dular dose of 10.7 mGy. The mean glandular dose in the before group decreased from 3.4 to 2.6 mGy with the change from 100 to 190 speed film. For all implant groups, a considerable mean glandular dose reduction was achieved when three rather than four views were obtained. A three-view mammogram obtained with the 190 speed Rh-Rh system had a mean glandular dose of 3.6 mGy. This value was only slightly (6%) more than the 3.4 mGy for a conventional two-view 100-speed Mo-Mo mammogram. Figure 2 shows the relation between mean glandular dose and compressed breast thickness for four-view augmentation mammograms in six groups. All implant views (implant-full and implant-displaced) were plotted. These curves were derived from the data on 1,128 individual images, including implant-displaced and implant-full views. The curves were calculated from the original data by use of exponential curve fitting.
|
The two curves for implant-displaced views had a more exponential upsweep than those of the four implant-full views. Comparison of our Figure 2 with Figure 4 in the article by Kruger and Schueler [3] shows the exponential trend is comparable. The Mo-Mo target-filter combination had the highest doses, and the Rh-Rh combination the lowest. The implant-displaced views were obtained predominantly at a breast thickness of 1-4 cm. The implant-full views were obtained predominantly at a breast thickness of 3-9 cm. There was a decrease in mean glandular dose for both Mo-Mo groups with the change from 100- to 190-speed film. Considerable dose reduction also occurred for the four implant-full views with the change from the 190-speed Mo-Mo combination to 190-speed Mo-Rh or 190-speed Rh-Rh.
In this study, percentage of breast volume occupied by dense breast tissue varied considerably, the average being 42% with an SD of 18%. Percentage of breast volume occupied by dense breast tissue and what Klein [12] calls fraction of glandular tissue are similar measures. Klein found fraction of glandular tissue values of 43% and 35% in two groups of patients. In Figure 8 in the article by Beckett and Kotre [4], median mean glandular dose is similar for augmented and unaugmented breasts at lesser breast thickness. With greater breast thickness (compressed breast thickness > 6.5 cm), Beckett and Kotre found the median mean glandular dose for augmented breasts was considerably lower than that of unaugmented breasts. Our Figure 2 shows mean glandular dose continued to increase for 6-11 cm of compressed breast thickness. Compared with Beckett and Kotre, we examined more patients who had undergone augmentation, calculated mean glandular dose for more images, used the actual recorded effective tube current for each film, and directly measured the percentage of breast volume occupied by dense breast tissue for each breast. Percentage of breast volume occupied by dense breast tissue is an important factor in the Sobol equation for calculating normalized glandular absorbed dose and may be an important reason for the different results at high breast thickness.
Table 4 shows estimates of
the risk of radiation-induced breast cancer based on mean glandular dose for
the augmentation examinations shown. The lifetime attributable risk values
were calculated from Table 12 D-1 of the seventh report on the biologic
effects of ionizing radiation
[14]. For the lifetime
attributable risk figures, it is assumed that women with breast implants
undergo 40 augmentation mammograms, starting at age 40 years and continuing
annually up to and including age 79. The life-time attributable risk was
calculated annually for each mammogram from ages 40 to 79, and the values were
summed. The far right column of Table
4 shows that the lifetime attributable risk for four-view
mammograms obtained with a 100-speed Mo-Mo system is 206. This finding
indicates that the risk of radiation-induced breast cancer is 206 among
100,000 for women (
1 in 500) undergoing annual mammograms with that
technique from age 40 to 79. This risk is three times greater than the
lifetime attributable risk of 69 for mammograms with a three-view 190-speed
Rh-Rh system, which translates to a risk of radiation-induced breast cancer of
approximately 1 in 1,500. This lifetime attributable risk of three-view
190-speed Rh-Rh mammograms is only slightly higher than the lifetime
attributable risk of 65 for nonaugmentation conventional two-view 100-speed
Mo-Mo mammography.
|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Mean glandular dose and compressed breast thickness are closely related [15]. The average compressed breast thickness for implant-full views (7.7 cm) was more than double that for implant-displaced (3.5 cm) and nonimplant (3.5 cm) views. The implant added more than 4 cm of breast thickness in the implant-full views. Mean glandular dose increases as compressed breast thickness increases, as shown by higher mean glandular dose values for implant-full views compared with implant-displaced and nonimplant views (Table 2 and Fig. 2). Increased compression (decreased compressed breast thickness) improves mammographic image quality by reducing motion blur, lowering magnification lack of sharpness, and separating overlapping structures. The result is better image uniformity and reduced dose [16]. Scatter of the primary beam is increased when compressed breast thickness is increased. For long image exposure times, film reciprocity failure can occur. Exposure times of 1.5 seconds or longer for implant-full views are not unusual for 100-speed Mo-Mo images.
To compensate for thick and dense breasts, peak kilovoltage is typically raised to increase beam penetration, especially for implant-full views. When peak kilovoltage is increased, tissue contrast decreases because the fraction of interactions by photoelectric effect is decreased. This loss of tissue contrast can decrease the conspicuity of microcalcifications, making detection and analysis difficult. After testing 60 clinical mammographic units, LaVoy et al. [17] found that increasing the X-ray tube potential from 25 to 28 kVp reduced average glandular dose 26%. Manual or automated selection of a higher peak kilovoltage for implantfull views is a simple way to shorten exposure time, decrease motion blur, and reduce dose.
Judicious selection of target-filter combinations can be important in breast dose reduction. Mean glandular dose can be reduced by target-filter combinations other than Mo-Mo [16]. With Mo-Rh or Rh-Rh target-filter combinations, dose reduction for dense or thick breasts (> 6 cm) has been reported [18]. In thinner breasts, image quality may be diminished with the Mo-Rh or Rh-Rh combinations compared with the Mo-Mo target-filter combination.
Technologists should be well trained in the performance of implant mammography to minimize repeated imaging or acquisition of extra images, especially repeated implant-full views, because these images are associated with the highest mean glandular dose. Our results suggest that implant-full views be obtained with rhodium (target-filter combination Mo-Rh or Rh-Rh). The Mo-Mo target-filter combination should be avoided for implantfull views because of the relatively high doses involved. Thilander-Klang et al. [19] found that Mo-Mo can be used for high contrast at compressed breast thicknesses less than 3 cm but that Mo-Rh at 27 kVp is recommended for compressed breast thicknesses up to 6 cm. Dance et al. [20] found that in imaging of thicker breasts, in which the peak kilovoltage can be increased, 20% improvement in contrast can be achieved without dose penalty when Mo-Rh or Rh-Rh is used.
The use of faster screen-film combinations can reduce dose at the cost of image quality. Tolerable limits for image quality are still achievable with a 190-speed system with the benefit of 35% reduction in dose. We found clinically that both the 190-speed system used in the study and the 190-speed system we currently use (Kodak EV190 and EV film) provide acceptable image quality [21]. Rhodium in the target-filter combination is necessary only for implant-full views. Mo-Mo technique should still be used for most implantdisplaced views.
There are compromises in decreasing the number of views in a mammogram. The issue is reduced dose versus cancer detection. As shown in Table 3, considerable reduction in radiation dose is achieved with three-view as opposed to four-view augmentation mammography. This reduction is accomplished by obtaining both implant-displaced views but only one implant-full view. On the basis of feedback from practicing radiologists about the relative merits of the views, in selected routine screening cases, we currently obtain the MLO implant-full view and eliminate the craniocaudal implant-full view. Abnormalities detected only on implant-full views usually are in the posterior aspect of the breast or on the borders of the implant (such as in the axillary tail and near the inframammary fold). Coverage of these regions is usually better on MLO implant-full views than on craniocaudal implant-full views. This three-view augmentation mammogram consists of craniocaudal implant-displaced, lateromedial implant-displaced, and MLO implant-full views.
Elimination of one of two implant-full views may decrease the rate of cancer detection [22]. The amount of the decrease is unknown and must be weighed against the documented additional radiation risk of the extra implant-full view. The significant radiation dose reduction achieved by eliminating one of the implant-full views translates over a lifetime into a reduction in the rate of mammography-induced breast cancer, and this fact cannot be easily dismissed.
When Eklund et al. [2] introduced implant-displaced views, acquisition of implant-full views probably continued for a standard mammogram, and implant-displaced views were considered the new extras. Eklund et al. did not assess dose or the risk of radiation-induced breast cancer associated with the recommended increase in the number of views. Most of the value of augmentation mammography comes from implant-displaced views, and a smaller contribution comes from implant-full views. Implant-full views have limited image quality and poorer parenchymal detail than implant-displaced views. The implant, being a firm spheroid of silicone or saline solution or both, is less compressible than native breast tissue and limits the amount of compression used during acquisition of an implant-full view. For implant-full views, breast tissue directly between the compression surfaces and the implant can be properly compressed, but this tissue is usually obscured by the implant. The visualized breast tissue is suboptimally compressed. Sufficient breast compression force is needed to immobilize the breast and implant to minimize motion blur. Additional force does not improve the image quality of the surrounding tissue.
The average breast compression force for the nonimplant (conventional) views in this study was 85 N. This value is similar to that in other reports—that is, 89.5 N [12] and 89 N (craniocaudal and MLO views) [8]. The average breast compression force was 51 N for all implant views in our study. All technologists at the two facilities at which mammograms were obtained used lower compression force on implant patients than on nonimplant patients. The 67% higher breast compression force for nonimplant (conventional) than implant views raises the question why technologists compress these patients' breasts differently. Technologists state they use lower breast compression force on implant-full views because they learned it in training and fear rupturing the implant. They may use just enough force so that the breast does not move during the X-ray exposure. Technologists may use less force for implant-displaced views because some patients report pain during compression of the thinner anterior breast tissue, including nipple and areola. Many technologists are afraid of causing implant rupture during mammography. In our experience and that of other authors [2], mammography-induced implant rupture either has not happened or has been so rare as to be medically insignificant compared with the considerable rate of rupture not related to mammography.
Factors in the calculation of the rate of breast cancer induced by versus that detected with mammography include age at first mammogram, frequency of screening, number of images obtained per mammogram, and radiation dose per mammogram [23, 24]. In the case of an augmentation patient who undergoes 40 annual mammograms from age 40 to age 79, the lifetime attributable risk of radiation-induced breast cancer from these 40 mammograms obtained in four views with a 100-speed Mo-Mo system is three times greater than that from acquisition of three views with a Mo-Rh or Rh-Rh system (Table 4). We believe the mean glandular dose per breast for a four-view augmentation mammogram obtained with 100-speed Mo-Mo technique is too high at 10.7 mGy. Diagnostic radiation levels should always be as low as reasonably achievable, balancing the unavoidable compromise between dose reduction and cancer detection. We believe dose-reduction methods should be used to keep the total mean glandular dose less than 7.0 mGy for a routine screening implant mammogram.
Concerning the use of three-view augmentation mammography, we believe reduction in the lifetime attributable risk of radiation-induced breast cancer may outweigh the unestablished and uncertain loss of lesion detection for selected patients. Future studies should be designed to quantitatively evaluate the risk-to-benefit ratio of radiation dose and induced-cancer risk versus image quality and cancer detection for implant-full views. Such studies should determine whether more than one implant-full view is warranted and if they are, how many. The relative value of each additional implant-full view should be determined on the basis of its strengths and weaknesses in visualization of tissue in the axillary, posterior, and medial locations.
For implant-full views, we use higher peak kilovoltage and rhodium in the target-filter combination (Mo-Rh or Rh-Rh), and we recommend this practice for other facilities. We obtain routine screening implant mammograms with a 190-speed screen-film system (Kodak EV). We perform conventional four-view implant studies for all patients undergoing diagnostic mammography, as for palpable lesions, high risk, and a strong family history of breast cancer. For routine screening, we use an intermediate approach that balances the increased cancer risk to older patients with the increased risk of radiation-induced cancer among younger patients. This age-balanced approach entails three views for patients 40-59 years old and four views for patients 60 years and older. This practice reduces lifetime attributable risk 27-32% compared with acquisition of four views for all age groups. Given the sizable and ever growing fraction of screening patients with implants, mammography facilities need to consider which dose-reduction techniques are best for their practices. A multifactorial approach combining selected dose-reduction techniques (target-filter selection, screen-film combination, and images acquired) and patient age considerations will have the greatest benefit.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
