| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Letters to the Editor |
Department of Radiology, Montefiore Medical Center, 111 East 210th Street, Bronx, NY 10467. e-mail: mlaks2000@yahoo.com
Editor:
In the article in the November issue of Radiology (1), Dr Avila and colleagues demonstrate the value of having excellent referral research centers that can focus their attention on a particular rare disease, such as lymphangioleiomyomatosis, and collect enough cases to be able to describe the spectrum of findings in this entity. We are grateful to the authors for sharing with us their experience, expertise, and images, and we very much enjoyed their review of the findings on this rare entity. Our comments are relative to the issue of diurnal variation.
First, the authors report that in 12 of the 13 patients who underwent computed tomographic (CT) examinations in both the morning and the afternoon, the volume of the lymphangioleiomyoma increased between the morning and the afternoon CT examinations, while none got smaller. Unfortunately, the morning and afternoon CT examinations were not comparable. Oral and intravenous contrast material was administered only for the morning examinations, while the afternoon examinations used only residual oral contrast material. Thus, in figure 3 (the only image demonstrating diurnal change in the article), it is clear that the increased lymphangioleiomyoma volume corresponds to a relatively diminished volume occupied by the adjacent small bowel. There is now less oral contrast material within that small bowel. What percentage of the authors’ diurnal variation is due only to the relative emptying of the oral contrast material from the patient’s bowel?
Second, in the Material and Methods section, the authors state that "[t]he volume of each lymphangioleiomyoma was estimated from CT images by measuring its maximum anterior-posterior (ap), transverse (tr), and craniocaudal (cc) diameters and applying the formula: volume = ap x tr x cc." However, since the shapes of these masses are not cubic, these calculated "volume" values are not a measure of the true volume of the lesion; at most, they might be called an index. Moreover, the true volume is not a linear function of this index number, as the lesion changes shape between imaging studies. To help the reader appreciate the limitation of the authors’ measure of "volume," consider the following thought experiment. Consider a truly isovolumic shape deformation of a cube of 100-cm3 volume to a sphere of equal volume. The "volume," as measured by Dr Avila and colleagues, would go from 100 to 191.1 cm3, yielding a 91% percent increase in "volume" (ie, greater than that of 46% [six of 13] of the lesions described by Dr Avila and colleagues), while the lesion would actually undergo a change only in shape, not in volume. The authors should use a more rigorous measurement scheme (either with software or section-by-section planimetric evaluation) for their volume measurement if they want to adequately assess the spectrum of volume differences that they are describing. What percentage of the diurnal variation in the study is due only to a shape change and not to a true volume change?
Third, the authors do not report any repetition of the imaging studies in any of their patients. It would have been prudent for the authors to have considered repeating the imaging studies and measurements on consecutive days, for a subset of the patients, to verify that they were actually observing a truly repetitive daily variation phenomenon before labeling it as diurnal variation and before entertaining speculative hypotheses as to its cause. Identical technique for each examination with equal volumes of oral bowel contrast material might then have been possible. This reimaging might have been done with magnetic resonance (MR) imaging to avoid exposing the patients to unnecessary ionizing radiation.
Additionally, in the context of entitling their article "Lymphangioleiomyomatosis: CT of Diurnal Variation of Lymphangioleiomyomas," the authors’ figure 2 is unfortunate. What at first glance appears to represent a demonstration of a substantial "diurnal variation" in volume of a pelvic mass is actually a pair of images of two different-sized left-sided pelvic masses in two different individuals. The juxtaposition to the much less impressive "diurnal variation" in size of the lesion in figure 3 (again, the only image showing variation in the article) was confusing to me. As a radiologist, I look at the pictures first!
REFERENCES
Department of Diagnostic Radiology, Warren Grant Magnuson Clinical Center, National Institutes of Health, Building 10, Room 1C-660, 10 Center Drive MSC 1182, Bethesda, MD 20892-1182. e-mail: navila@nih.gov
We thank Dr Laks for his careful reading of our article and his comments.
In the single figure that demonstrates an increase in the size of a lymphangioleiomyoma between the morning and the afternoon CT examinations, Dr Laks observes concomitant decrease in bowel distention with oral contrast material and concludes the former to be the result of the latter. In essence, he proposes that a bowel filled with contrast material develops sufficient intraluminal pressure to drive fluid from the lymphangioleiomyoma. It is unfortunate that this image was selected for presentation, as we have examples of diurnal enlargement of lymphangioleiomyomas without change in adjacent bowel distention. Nevertheless, the proposal is implausible. Normal bowel filling (in the absence of obstruction) may distort, compress, and displace other loops of the bowel. However, the distended patent bowel does not develop enough intraluminal pressure to compress retroperitoneal masses or cysts. Evidence of the direction of the deforming forces acting between the small bowel and a mass is manifested on the images in the bowel-lymphangioleiomyoma interfaces; the external surfaces of the lymphangioleiomyomas are convex and indent the bowel rather than being indented by the bowel.
In the analysis, we used the product of the three perpendicular diameters (ap x tr x cc) to quantify the percentage changes in lymphangioleiomyoma volumes over time. Dr Laks questions the validity of concluding from the results that the lymphangioleiomyomas actually increase in size; he suggests that the observed increase in the product of the diameters is due to a change in the shape of the masses rather than a true change in their volume. He supports this concern by noting that a 91% percent increase in the product of the diameters can result from an isovolumic deformation of a cube into a sphere. We agree with Dr Laks on two points: first, our method is better described as a volume index; and second, a more direct assessment of volume such as summing cross-sectional areas would have provided more accurate volume estimate. However, he overstates the potential inaccuracy and the inappropriateness of our use of the index.
The index meets the purpose of our study—support of visually apparent changes in volume with an objective measure that indicates the degree of change. The product-of-maximal-diameters index provides a precise estimate of the percentage changes in volume when the masses are ellipsoidal (or spherical) or, in general, when the masses are of any shape and maintain similar shapes as they change size. As shown on the images, the lymphangioleiomyomas approximate ellipsoids and retain similar shape as they change size—hence, the assumptions about the product-of-diameters method are reasonable. The change in shape of an isovolumic cube to a sphere thought experiment used by Dr Laks to impugn the index is an unrealistic worst-case scenario. Moreover, in the experiment, the product of diameters is calculated in a fashion that exaggerates the problem; specifically, if the diagonals of the faces of the cube are used in the calculation instead of the sides for the transaxial (ap and tr) diameters, then an isovolumic change from cube to sphere produces only a 4% change in the product of the diameters.
Dr Laks points out the absence of data from "repeating the imaging studies and measurements on consecutive days." From this he questions the validity of our conclusion that the observed size changes are a truly daily phenomenon and warrant the label "diurnal variation." It is true that we have presented no direct demonstration in individual patients that this occurs day after day. However, there is strong inferential statistical evidence for the phenomenon being a daily occurrence, as indicated by its high prevalence (12 of 13 patients). If the event occurs infrequently, then the probability of coincidentally examining 12 of 13 patients on days it occurred would be low. For example, were the lymphangioleiomyomas to change size every other day or less often, then the probability of observing the change in at least 12 of 13 cases examined on random days would be less than .002. The data also provide strong evidence that the duration of the cycles of the lymphangioleiomyoma volume changes is confined to 1 day; if the cycles were of longer duration—for example, 2 days, enlarging one day and shrinking the next—then the probability of encountering all 12 patients who showed change during the enlargement phase would be low (P < .001).
Finally, Dr Laks complains about a figure that "at first glance appears to represent a demonstration of a substantial diurnal variation in volume of a pelvic mass is actually a pair of images of two different-sized left-sided pelvic masses in two different individuals". Since this is clearly explained in the caption, this illustrates that, although images provide the most efficient means of communicating radiographic information, text remains an important additional source occasionally vital to their interpretation.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
