| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Letters to the Editor |
Department of Radiology, University of California, San Francisco, Box 0628, 505 Parnassus Avenue, San Francisco, CA 94143-0628, e-mail: jeanne.laberge@radiology.ucsf.edu
Editor:
I read with interest the article by Dr Nikolic and colleagues in the January 2000 issue of Radiology (1). I believe that the subject addressed by the authors is important, and I was pleased to read the results of this study in Radiology.
Because this is an important topic and because the results of the study are data driven (and not particularly intuitive), I spent some time analyzing the data presented. I wish to call to your attention two problems in the reporting of these data.
First, the table on page 123 lists the estimated ovarian dose for all 20 patients examined with 22.34 cGy ± 1.37 (mean ± SD). I believe that this number is misleading. The actual mean dose for these 20 patients is 22.34 cGy, but the SD (ie, the SD of the 20 mean values) is 17.11 cGy. The number listed, 1.37, is the mean of the within-patient SDs and does not reflect the between-patient differences.
Is this discrepancy important? I believe that it is important, and understanding this statistical point is more than a minor technical exercise. In fact, if the authors had reported a value of 22.34 cGy ± 17.11, the reader could easily observe that there is variation in the radiation dose among the patients examined. (This variation would also be evident if the authors had pointed out that the mean estimated doses ranged from 4 to 65 cGy.) A value of 22.34 cGy ± 1.37 suggests that most patients were exposed to a dose close to 22.34 cGy (ie, 95% of patients were exposed to a dose between 19.6 and 25.1 cGy).
The number, 1.37, is the mean value of the SD of readings for a given patient (mean within-patient SD). This number provides the reader with a sense of the reliability or reproducibility of the radiation dose measurements, but it does not reflect the variation in measurements between the patients studied.
The second problem to which I would like to bring to your attention is contained in the first paragraph of page 123. The authors state that Nx, the number of women of childbearing age in the country, is 58,540 million. I believe that this is typographic error and that the correct number is 58.54 million.
Once again, I applaud Dr Nikolic and colleagues on their important research and wish to point out these discrepancies only as a means of improving the accuracy of their work.
REFERENCES
Department of Radiology, Georgetown University Hospital, 3800 Reservoir Road NW, Washington DC, 20007, e-mail: nikolib@gusun.georgetown.edu
We thank Dr LaBerge for her letter regarding our article (1). We greatly appreciate the opportunity to respond to this letter.
First, Dr LaBerge refers to the mean value of the absorbed ovarian dose and the SD listed in the table on page 123. In this table, a mean ovarian dose and a corresponding within-patient SD are listed for each patient (patients 1–20, column 4). These numbers are generated from five measurements for each patient. The mean absorbed ovarian dose and the mean of the within-patient SDs were then calculated and are listed in the last row of the same column as 22.34 cGy ± 1.37. The mean absorbed ovarian dose was, therefore, obviously calculated from those numbers that were listed above it in the same column. This calculation was correctly understood by Dr LaBerge. The listed numbers are all accurate; their computation and meaning are clear. Moreover, they are explained in the text (see first paragraph of Results).
The within-patient SDs were considered important and were included, as they reflect the consistency of the measurements for each patient. The SD of 1.37 within the mean of all patients is related to the mean estimated ovarian dose of 22.34 cGy. These numbers were listed together in one column in an admittedly unconventional fashion due to editing constraints imposed by the staff of Radiology.
There is indeed "variation in the radiation dose among the patients examined," as Dr LaBerge stated. This variation can be expected, as the same interventional procedure may require variable fluoroscopic times and may therefore result in variable absorbed radiation doses. Nevertheless, the range of the absorbed radiation dose is, in fact, mentioned in the article (see Results, second paragraph, page 123), as suggested.
We regret that this range was not made clearer and included in the table. In addition, the individual numbers for the radiation doses are all listed in the table that supplements the text. We therefore did not opt to point out the variability of the absorbed radiation dose by additionally calculating the SD of the mean (17.11 cGy) although this is correct, as noted in the erratum of this issue of Radiology wherein we also address the other item noted next.
Second, there is indeed a typographic error in the first paragraph of page 123 (58,540 million instead of 58.54 million) that we regret and that should not have occurred.
Thank you very much for your critical and constructive evaluation of our work.
REFERENCES
This article has been cited by other articles:
|
|
 |
|
  C. H. Tan, K. H. Tay, K. Sheah, K. Kwek, K. Wong, H. K. Tan, and B. S. Tan Perioperative Endovascular Internal Iliac Artery Occlusion Balloon Placement in Management of Placenta Accreta Am. J. Roentgenol., November 1, 2007; 189(5): 1158 - 1163. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
