Anzeige
Erschienen in:
01.03.2012 | Letter to the Editor
Reply to Greenberg: indeed, what has changed!
Erschienen in: Pediatric Radiology | Ausgabe 3/2012
Einloggen, um Zugang zu erhaltenExcerpt
I would like to thank Dr. Bruce Greenberg for his interesting comments [1]. I will address his major points in their order of presentation.
1.
“The article boldly implies that this is proven and does not rely on assumptions.”
As we know, other than the deterministic effects of radiation, we are relying on statistical analysis. That is, the stochastic effects of radiation causing cancer are predicted by the most sophisticated statistical modeling. These projections are updated as follow-up is extended, dose estimates are updated and analytical methods are improved. The 2004 work of Preston et al. [2] noted by Dr. Greenberg is based on updated doses and 3 years more follow-up than were used in the Pierce and Preston 2000 paper [3].
I am afraid that Dr. Greenberg has misinterpreted this complicated paper, because he has isolated a few numbers from the big picture. The authors conclude the 2004 paper with “In summary, the introduction of DS02 (editor’s note: this paper) does not lead to any major changes in cancer risk estimates … despite real and sometimes marked improvements in individual dose estimates. …and these changes appear to have little impact on summary risk estimates.”
2.
“I do not know of any radiologist who would consider 100 mSv low dose.”
As Dr. Greenberg points out, we are discussing a range of 0–100 mSv (0–10 rads). We all know that doses for CT examinations were much higher at children’s hospitals in 2000 than they are today. In fact, a two-phase head study could reach 10 rads in older children and a three-phase abdominal study might well have exceeded 5 rads. Even today, in adult institutions, head scans in adolescents are in the 3- to 5-rad range and abdominal scans are more than 2 rads. The “low dose” relates to radiation from the Preston et al. atomic bomb patients [2], not to what we consider low doses.
3.
Interpretation of the details of reference 2.
When one views the table (Table 3, reprinted here as Table 1) that Dr. Greenberg is referring to, the fitted model predicts 4,282 background cases (i.e. cases that have nothing to do with radiation exposure). There still is a prediction of two excess cancers at the lowest doses. I asked Dr. Dale Preston about this, and he said that the mean dose to these patients was about 1 mGy (100 mrads). Due to the statistical modeling used, there is no constraint that forces the total number of fitted cases in each category to be equal to the observed number of cases. The fitted cases are only limited by the total population. Constraining the within-dose category total fitted cases to be equal to the observed number of cases reduces the fitted number of background cases in the lowest dose category by 14 (to 4,268) and has no noticeable effect on the number of fitted excess cases (same two cases). It is important that the Japanese data suggest that there are radiation-associated cases at doses in our diagnostic range.
Table 1
Observed and expected solid cancer deaths from 1950 to 2000 by dose group (Table 3 reprinted with permission from [2])
|
Weighted colon dose category (Sv)
|
Subjects
|
Person years
|
Solid cancer deaths
|
Expected background
|
Fitted excess
|
|---|---|---|---|---|---|
|
< 0.005
|
38,507
|
1,415,830
|
4,270
|
4,282
|
2
|
|
0.005–0.1
|
29,960
|
1,105,215
|
3,387
|
3,313
|
44
|
|
0.1–0.2
|
5,949
|
218,670
|
732
|
691
|
41
|
|
0.2–0.5
|
6,380
|
232,407
|
815
|
736
|
99
|
|
0.5–1
|
3,426
|
125,243
|
483
|
378
|
116
|
|
1–2
|
1,764
|
64,689
|
326
|
191
|
113
|
|
2+
|
625
|
22,302
|
114
|
56
|
64
|
|
Total
|
86,611
|
3,184,356
|
10,127
|
9,647
|
479
|
4.
“We should be able to tell parents that to the best of our knowledge there is no evidence linking cancers to computed tomographic examinations that result in patient exposures up to 5 mSv.”
This is an incorrect approach. There are, at present, no precise estimates of excess cancer risks at doses associated with typical pediatric CT scans. However, analyses of the atomic bomb survivor data, such as those described by Preston and Pierce [2, 3], demonstrate significant increases in risk with dose over the low-dose range, and the data strongly suggest (but do not prove) that doses from typical CT scans may result in a subsequent lifelong increase in cancer risks.