| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Letters to the Editor |
Department of Radiology, University Hospitals, Catholic University of Leuven, Herestraat 49, B-3000 Leuven, Belgium. e-mail: yicheng.ni@med.kuleuven.ac.be
Editor:
We read with interest the two associated articles by Dr Goldberg (1) and Dr Kim and colleagues (2) in the February 2005 issue of Radiology.
In rabbits with intramuscularly inoculated VX2 carcinoma that was further treated incompletely with radiofrequency (RF) ablation, Dr Kim and colleagues compared two magnetic resonance (MR) imaging contrast agents, namely, gadopentetate dimeglumine and SH L 643A, for their potentials in differentiation between residual tumor and benign periablational reactive tissues (2). Although both agents are tissue-nonspecific and are eliminated by glomerular filtration, they differ in behaviors after intravascular administration because of their distinctive molecular weights. Gadopentetate dimeglumine is a small molecular agent with a molecular weight of 546 Da and is distributed in extracellular fluid space, whereas SH L 643A is a macromolecular blood pool agent with a molecular weight of 17 453 Da and is exclusively distributed in the intravascular space of intact capillary endothelium (2). By analyzing MR imaging contrast enhancement patterns as a function of time and verifying the respective microvascular densities, the authors concluded that SH L 643A was superior over gadopentetate dimeglumine owing to its confined presence in the tumoral vasculature, which is more abundant than that of benign tissues (2).
In his commentary Science to Practice essay, Dr Goldberg commented that this study represents a first step in addressing a clinical issue of great importance (1).
Both articles (1,2) should be complimented for timely drawing of public attentions to this crucial topic on imaging assessment after interstitial ablation therapies. However, we have been puzzled by some of the messages in these two articles and would like to discuss them openly.
First, we do not agree with the expression of "differentiating benign periablational enhancement from residual tumor," which can be found in the title and throughout the text of the article by Dr Kim and colleagues (2) and in the essay by Dr Goldberg (1). By definition, "enhancement" is a common term used in the imaging field to describe incremental change in signal intensity or contrast after using certain contrast materials, whereas "tumor" is a pathology term denoting an abnormal lump or mass of tissue resulting from excessive cell division. Although understandable in the context, being one is an abstract noun (enhancement) and the other a concrete noun (tumor), these two terms are not directly comparable, and such an improper expression should not be officially documented in a journal like Radiology. Nevertheless, Dr Kim and colleagues could hardly be criticized for making this error because they had obviously been misinstructed by a recently published proposal for standardization of terms and reporting criteria (3) with respect to imaging-guided tumor ablation (see near the top of the first column on page 424 in reference 2 and near the top of the third column on page 339 in reference 3). The expression used in the title of this letter to the editor may be considered appropriate.
Second, one may frequently encounter erroneous or inconsistent contents when reading the article by Dr Kim and colleagues (2). (a) It is described on page 424, "In all cases, the RF application time of less than 300 seconds was chosen intentionally to ensure incomplete tumor ablation"; however, the actual RF application time counted at least 450–600 seconds per session with the given protocol. (b) Across pages 424 and 425, it is first mentioned that SH L 634A and gadopentetate dimeglumine were administered for the first and second set of images with a 24-hour interval, respectively. Then, this is contradictorily followed by the sentence, "For a randomized order of the two contrast agents, half of the animals were imaged in a crossover fashion with one agent and then 24 hours later, with the other agent." (c) On page 426, the reported imaging findings do not match with the corresponding curves of enhancement ratios. For instance, it is described, "One week after RF ablation, the rim enhancement along the ablated tumor became more prominent compared with that immediately after RF ablation on contrast-enhanced MR images," but this is not true for both SH L 634A and gadopentetate dimeglumine (see figure 2a vs 2b and figure 3a vs 3b). (d) The contrasts between periablational tissue and residual tumor as shown on figure 1c and 1d for SH L 634A and figure 1e and 1f for gadopentetate dimeglumine cannot be correctly reflected by curves in figure 2b and 3b, respectively. (e) It is said on page 426, "Four weeks after RF ablation, . . . were noted on contrast-enhanced MR images in 11 cases, and . . . was seen in one case." However, according to the study design on page 424, a total of 12 rabbits were divided into three groups of four animals each and were sacrificed at 1 day, 1 week, and 4 weeks, respectively. Then, how were 12 rabbits still remaining 4 weeks after RF ablation? (f) Also, the article contents are not properly referenced; for example, it is mentioned on page 424, "The experimental VX2 rabbit carcinoma was prepared in a manner reported previously," which is supported by reference 19 of the article. However, the manners were quite divergent between this study and that in the reference, that is, percutaneous versus transabdominal, cell suspension versus tissue fragment, and intramuscular versus intrahepatic, respectively. (g) References 8 and 29 of the article, on page 430, are irrelevant to "RF ablation of solid tumor." (h) References 19 and 28 of the article, on that same page, have nothing to do with "a more complicated pharmacokinetic model." (i) Reference 8 is cited on page 424 and particularly on page 430 to show its relevance to this study on tumor RF ablation; however, it is found that the original title of reference 8, "Creation of Radiofrequency Lesions in a Porcine Model: Correlation with Sonography, CT, and Histopathology," which has nothing to do with tumors, has been altered into "Creation of Radiofrequency Residual Tumor and Post-Ablative Changes in a Porcine Model: Correlation with Sonography, CT, and Histopathology." The latter has been very uncommon in the practice of scientific publications.
Regarding the rational aspects, Dr Kim and colleagues believed that the distinct contrasts between the residual tumor and periablational tissue were due to the larger blood volume and/or higher vascular density in the viable tumor than in the peripheral granulation tissues, features exploitable only with the use of SH L 643A (2). However, the MR images in figure 1 and the curve patterns in figure 2 rather favor a closer likelihood for the combined effects: (a) more abundant, but leaky, vasculature in the tumor than in the early scar tissues, and (b) subsequent more prominent interstitial (instead of intravascular) enhancement in malignant than in benign tissues. These are most likely the real underlying mechanisms. The authors are encouraged to prove this proposition by performing further experiments.
In regard to the clinical applicability, one should be cautious to extrapolate the study of Dr Kim and colleagues to the clinical scenarios. As recognized in both articles (1,2), this study was limited by the use of a single tumor model in rabbits and especially by the fact that the tumor was inoculated in the muscle, which is a very uncommon tumor location in patients, unlike what was stated: "Hence, the animal model provided very similar conditions to the clinical situation" (first column, page 430 in reference 2). Compared with that in the muscle, contrast enhancement patterns after tumor ablation therapies could largely differ in visceral organs such as the liver, lung, and kidney with abundant vasculature, distinct hemodynamics, and/or tissue microenvironment. Besides, both SH L 643A and gadopentetate dimeglumine are tissue-nonspecific and cleared up efficiently from both malignant and benign tissues. Their different functions rest on molecular weight–dependent compartmental distributions, which in turn rely on the normality or abnormality of the endothelial lining of capillary microcirculation. However, vascular integrity is almost always altered during both malignant and inflammatory processes, and there is unlikely any absolute qualitative or quantitative distinction between tumoral and benign neovasculatures. In reality, blood supply is unnecessarily richer in viable tumors than in granulation tissues, especially of sinusoid organs such as the liver. Therefore, it would not be surprising to see the absence of the described "distinct" contrast enhancement patterns (or even to see opposite ones) between residual tumor and periablational tissues, no matter whether large or small molecular contrast agents are used, as long as they are tissue-nonspecific. A lack of invariable contrast enhancement patterns can also be found in an article from a previous study performed with use of identical contrast agents and similar animal models and MR imaging protocols (4). For instance, with the same doses of SH L 634A and gadopentetate dimeglumine in the same intramuscularly inoculated VX2 tumor, not only the absolute values of the peak enhancement but also the shape of the curve, peak time, and slope values could vary drastically (compare figure 2 and the table in reference 2 with figure 3 and table 1 in reference 4).
As for whether there exist better solutions to the current problem, Dr Goldberg indicated that "other strategies for differentiation with other contrast agents and other imaging modalities are also being explored," without specifying details (1). One promising strategy, which was not mentioned in either article (1,2) yet has already been introduced elsewhere, is to use necrosis-avid contrast agents (NACAs) (5–8). In addition to their primary utility for imaging assessment of myocardial viability (9–14), the secondary but equally important application of NACAs is for early evaluation of therapeutic adequacy after interstitial tumor ablation (5–8,15). With their specific and nonspecific multiple functions (16,17), the resultant imaging outcomes would become unconditional, unambiguous, and indisputable for solving the problems posed here (1,2,15–17). Although, like SH L 634A, NACAs are not immediately available for clinical use, further preclinical and clinical development may prompt their eventual applications as virtual biopsy techniques (15–17).
In summary, (a) one may have reasons to doubt the validity of the conclusions and scientific values of articles analyzed here, and (b) the fact that such an article (2) with so many obvious defects that had not been detected and corrected during the generally considered strict reviewing, revision, and proofreading procedures got officially published in Radiology, and even with an accompanying Science to Practice essay (1) to stress its importance, would be serious enough to alert us to the quality of our research and submitted manuscripts.
References
Department of Radiology, Seoul National University Hospital, 28 Yongon-dong, Chongno-gu, Seoul 110–744, Korea. e-mail: moonwk@radcom.snu.ac.kr
We read with interest the letter of Dr Ni and colleagues concerning our article (1). We would like to respond to their comments as follows:
First, they inquire as to the appropriateness of using the term "benign periablational enhancement." We were also aware and partly agree with Dr Ni and colleagues that "enhancement" and "tumor" may not be directly comparable, especially in a general sense. However, as stated by the International Working Group on Image-Guided Tumor Ablation (2), the term "benign peripheral enhancement" was specifically selected by the expert committee given the fact that this finding is often caused not only by changes in tissue composition but by physiologic tissue responses to heat. We defined the term in the introduction section so as not to mislead the readers of the article.
Second, with regard to the inconsistency issue of the article, we admit the errors in reference 8 and several vague expressions in the text, although most of them were minor and were introduced by us during our responses to questions from the manuscript editor. We thank Dr Ni and colleagues for indicating the mistakes and want to correct the title of reference 8 in an erratum. However, mistakes in citing references and vague expressions in the text have nothing to do with the accuracy of animal research, and indeed we stand by the validity of our initial scientific conclusions.
Another point raised by Dr Ni and colleagues is the contrast mechanism of the residual tumor and benign periablational enhancement with the blood pool contrast agent, SH L 634A, and the nonusefulness of SH L 634A for distinguishing the residual tumor from benign periablational enhancement. As Dr Ni and colleagues commented, both the intra- and extravascular fractions with both SH L 643A and gadopentetate dimeglumine might be to some extent responsible for the observed contrast enhancement in both the periablational enhancement and the residual tumor. However, the extent should be quite different between gadopentetate dimeglumine and SH L 643A. SH L 643A revealed the obvious enhancement differences between the periablational enhancement and the residual tumor up to 30 minutes; on the contrary, gadopentetate dimeglumine did not. Furthermore, in the blinded review, all three radiologists could better differentiate residual tumor and benign periablational enhancement with SH L643A than with gadopentetate dimeglumine, and the difference was statistically significant for two of the radiologists. So we can conclude that the blood pool contrast agent SH L 643A more precisely reflects microvessel density and is less influenced by the extravascular fractions due to the leakiness of vessel compared with the conventional extracellular contrast agent, gadopentetate dimeglumine.
In regard to the issue of the discrepancies in the parameters such as peak enhancement ratio (not the absolute values of the peak enhancement as Dr Ni and colleague stated), slope values, peak time, and the shape of curves between our current and previous studies, they missed the fact that the equation for enhancement ratio in this study is different from those used in our previous studies (3,4). In this study, the enhancement ratio (ER) determined by comparison of the precontrast signal intensity (SIpre) with postcontrast signal intensity (SIpost) relative to precontrast signal intensity of the psoas muscle (SIprepsoas), as a reference, was calculated by using the following equation: ER = (SIpost – SIpre)/SIprepsoas, whereas ER = SIpost/SIpre was used in other studies (3,4). The reason why we used a new enhancement ratio relative to a common reference is as follows: According to the previous equation (ER = SIpost/SIpre), the enhancement ratio values are entirely dependent on native signal intensities of the residual tumors and periablational enhancement, which are usually relatively hypo- and hyperintense on T1-weighted images, respectively. Therefore, in theory, even if the visual contrast enhancement in the periablational enhancement is stronger than that in residual tumor, the enhancement ratio for the tumor can still be higher than the enhancement ratio for the periablational enhancement. Hence, to compare the contrast effect on two adjacent tissues, we calculated enhancement ratio relative to a common reference (precontrast signal intensity of psoas muscle) and provided the results in figures 2 and 3 and in the table. The shape of the curve and peak time of VX2 tumor for both contrast agents in this study, however, are almost identical to those in our previous studies (see figures 2 and 3 in reference 1, figure 3 in reference 3, and figures 3 and 4 in reference 4) considering tumor heterogeneity.
The last point raised by Dr Ni and colleagues is the application of NACAs for early evaluation of therapeutic adequacy after tumor ablation. They cited their data and concluded that "with their specific and nonspecific multiple functions, the resultant imaging outcomes would become unconditional, unambiguous, and indisputable for solving the problems posed here." While they stressed tissue nonspecificity and nonusefulness of blood pool contrast agents, they seem to be too optimistic about the use of NACAs for the differentiation of residual tumor and benign periablational enhancement. Our group had also performed two animal experiments with an NACA, bis-gadolinium mesoporphyrin, by using a VX2 tumor and abscess model (5,6). Unfortunately, it was impossible to specifically depict necrosis with bis-gadolinium mesoporphyrin because it also enhanced other parts of lesions, including viable tumor, inflammatory granulation tissue, hemorrhage, and fibrosis.
References
Department of Radiology, Beth Israel Deaconess Medical Center, 330 Brookline Avenue, Boston, MA 02215. e-mail: sgoldber@caregroup.harvard.edu
I have reviewed the issues raised by Dr Ni and colleagues concerning the article of Dr Kim and colleagues (1) that was recently featured in my Science to Practice piece (2), as well as the response by Drs Kim and Moon. As readily pointed out by Drs Kim and Moon, most of the issues raised by Dr Ni and colleagues are minor or textual in nature, and do not substantially impact on the underlying science, experimental design, or potential interpretation of the study. Hence, the potential impact of Dr Kim and colleague's study as outlined in my Science to Practice article remains unchanged. To reiterate, "Kim et al suggest that the MR blood pool contrast agent SHL 643A may enable more accurate differentiation of benign periablational enhancement from residual viable tumor following radiofrequency ablation. This animal study represents a first step in addressing a clinical issue of great importance" (2). I also wish to remind the readership of the highlighted caveat, namely that "additional animal and clinical studies are necessary to prove the potential utility of this agent" (2).
One issue raised by Dr Ni and colleagues that is deserving of response regards their concern with the term "benign periablational enhancement" as taken from the proposal for standardization of terms and reporting criteria (3). I wish to alert the readership that the term "benign periablational enhancement" was specifically selected by a highly representative committee and ratified by scores of additional radiologists (as well as recently by the Technology Assessment Committee of the Society for Interventional Radiology), given the fact that the finding of benign peripheral enhancement is often caused not only by changes in tissue composition but also in many situations by dynamic physiologic responses to heat (4). Within this context, the title and phrasing of the article by Dr Kim and colleagues is most appropriate as the purpose of their study was to determine whether or not one can differentiate a potentially obfuscating imaging finding (ie, "benign periablational enhancement") from the more concerning pathologic entity of residual tumor. It must also be pointed out that the Editor of Radiology requests authors to use the recommended terminology of the standardization of terms and reporting criteria (3) whenever possible, a request that Dr Kim and colleagues followed.
Another issue raised by Dr Ni and colleagues is that I did not specify an exhaustive list of "other strategies to differentiate benign periablational enhancement from tumor" within the Science to Practice article and specifically that I did not describe the use of NACAs. I wish to reiterate that the purpose of the Science to Practice piece is not to serve as a definitive review of the selected topic but rather to "both continue and expand the theme of the practical application section in the discussion of experimental study manuscripts" (5). Hence, based on the intent and structure of the Science to Practice piece, additional details of other potentially useful strategies were not included. Regardless, I am puzzled as to why Dr Ni and colleagues, a group that places so much weight on proper referencing, has used the format of a letter to the editor to forward a strategy that has yet to undergo formal peer review in the literature. Indeed, all of the cited references regarding RF ablation and this NACA appear only as abstracts, reviews, or supplemental material dating way back to 1997.
Last, Dr Ni and colleagues make many assertions about tumor and tissue vascular physiology whose relevance to the issues at hand are either unclear or debatable, especially when considering the heterogeneous nature of tumor blood supply and the complexity of how this may be altered by tissue heating. Indeed, one could characterize some of the very arguments and terminology selected by Dr Ni and colleagues (eg, the claim that viable tumor blood supply is "unnecessarily richer . . . [sic]") as nonorthodox overgeneralizations. Nevertheless, regardless of the exact mechanisms behind the findings observed by Dr Kim and colleagues, let us not lose sight of the fact that SHL 643A did permit better differentiation of residual tumor from benign periablational enhancement compared with gadolinium dimeglumine. If verified for any clinical scenario, the use of this agent would get us closer to a solution for an important, but as of yet unsolved, clinical problem—differentiating benign periablational enhancement from tumor.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
