TO THE EDITOR:
Sarkar et al. recently reported on the relative merits of 123I and 131I for diagnostic whole-body scanning for thyroid tumors (1). Based on a limited number of patients, they claimed that in comparison with 123I, 131I showed superior sensitivity for identifying differentiated thyroid cancer metastases (1). However, we do not agree with the interpretation of several of the alleged differences between 131I and 123I scans in the figures shown. In addition, the authors did not include the post-therapy images, which are the optimal standard against which to judge the accuracy of diagnostic imaging.
The 96-h 131I image in Figure 1 shows prominent left cervical and pulmonary metastases that the authors state are not seen in the companion 123I image. However, the 123I image also shows a focal increase in the same left cervical area in the neck, and a focal increase in the left posterior mid-lung field is also evident in both the 24-h 123I image and the 24-h 131I whole-body image shown in this figure. We agree that the target-to-background ratio for these lesions is less in the 123I image than in the 96-h 131I image.
In describing Figure 2A, the authors mention a right cervical focus and lung uptake that were seen in the 131I images but not in the 123I images. It is unclear to us how much of the cervical uptake in the 131I image may be incidental esophageal activity, which is also seen in the companion 123I image. Comparison of this area with the post-treatment 131I images would help clarify this question. In addition, we believe that the 123I images also show at least some abnormal focal uptake in the right lower lung field, even though the target-to-background activity is once again much less than that seen with 131I.
In Figure 2B, we agree that the metastatic foci in the left hip, the right knee, and left axilla are identified both by 123I and by 131I. From the authors’ arrow in the figure and description in the legend, it is not clear what is being identified in the 131I image as the right iliac bone metastasis, which the authors claim is identified by 131I but not by 123I. If the arrow in the 131I image is pointing to the focus overlying the region of the cecum in the right lower quadrant, then we would argue that a focus in the same location is evident in the companion 123I scan. Similarly, we take issue with the claimed disparity between 123I and 131I for detection of abnormal lung uptake. Although this abnormality again shows a higher target-to-background ratio in the 131I image, the soft-tissue lung activity in the 123I image is clearly higher than that of the abdomen. Of all 6 131I-positive sites shown in this patient, only the left skull focus appears to have been more convincingly missed by the 123I image, and by itself, this factor would not have had any significant impact on the treatment algorithm.
In light of the above considerations, we believe that the authors have exaggerated the differences in sensitivity between 123I and 131I for detection of distant metastases, even though 131I did show some of them better. In contrast, other authors have reported competitive or superior sensitivity for 123I, compared with 131I, for diagnostic thyroid tumor scanning, including identification of distant metastases (2,3). In a study by Siddiqi et al. (3), diagnostic scanning with 123I correctly identified thyroid metastases in 9 of 12 patients (confirmed in post-therapy scans) in whom 131I diagnostic scanning had negative findings. In a perhaps related observation, we note that the quality of the whole-body 123I images shown by Sarkar et al. (1) does not appear as good as that found by others (2,4), possibly contributing to suboptimal sensitivity in their experience.
The ability to image 131I later after dosing than is possible with 123I, afforded by the longer half-life of the former, no doubt contributes to the improved target-to-background uptake ratio and thereby the sensitivity for detecting potential lower-avidity sites of thyroid metastases. Gerard and Cavalieri recently reported that using a larger 185-MBq 123I dose in combination with a later 48-h imaging time can improve the target-to-background ratio and, thereby, the sensitivity for detecting less-iodine-avid sites of differentiated thyroid tissue (4). Use of this approach in the patients shown by Sarkar et al. (1) would likely have improved the conspicuity of the 123I foci corresponding to the thyroid metastases in question.
A final important consideration is the potential adverse influence of stunning by diagnostic doses of 131I. Given that the evidence of such potential is now compelling (4,5), it is all the more important to optimize the sensitivity of 123I diagnostic imaging to avoid the use of 131I for this purpose, which may compromise subsequent therapeutic efficacy.
REFERENCES
REPLY:
We appreciate the comments by Drs. Gerard and Mandel and are pleased to have a second opportunity to clarify the common misperception that 123I is better than 131I at detecting thyroid cancer metastases.
First, we want to reiterate the purpose of our study (1). It was not our intent to assess the efficacy of diagnostic (pretherapy) 123I imaging in comparison with post-therapy imaging, a subject addressed in other publications. Instead, we compared 123I imaging directly with diagnostic 131I imaging using comparable (74–185 MBq) amounts of radiotracer. To our knowledge, this has been the only study with a head-to-head comparison of 123I and 131I in patients with thyroid cancer, including those with distant metastases.
Drs. Gerard and Mandel cite several studies to bolster their case for 123I imaging. The study by Shankar et al. (2) compared diagnostic (pretherapy) 123I imaging with post-therapy 131I imaging; that is, there was no comparison with diagnostic 131I studies. Also, was the “medium energy” collimator used in that study optimal for 131I? Gerard and Cavalieri assessed the sensitivity of 123I in a similar fashion (3). That 123I provides “acceptable levels of sensitivity” when compared with post-therapy imaging, as claimed in their article, does not necessarily imply it is as good as, let alone better than, 131I. Another limitation of their study was that it focused on detection of cervical tissue including thyroid remnants, not extracervical metastases. The last study cited, by Siddiqui et al., had a similar theme (4). The main thrust of this study was the comparison of pretherapy 123I imaging to post-therapy 131I scans. The authors also appear to suggest that 123I is superior to 131I, but the data provided are far from convincing. The diagnostic 131I and 123I studies were not done (sequentially) at the same time. Although details are lacking, it appears that they were done up to 5 mo apart, rendering any comparison moot. Furthermore, neither the amounts of 131I used for diagnostic imaging nor the imaging times or counts were included, and the only figure in the entire article has a very count-poor 131I image. Thus, none of the 3 studies cited as showing the superiority of 123I directly compared the 2 agents at the same sitting using comparable amounts of tracer.
Needless to say, we do not agree with Drs. Gerard and Mandel’s interpretation of the images. However, we do applaud their painstaking attempts to find abnormalities on the 123I images corresponding to obvious lesions on the 131I studies because it proves our point that metastases are better visualized with 131I. In our view, many of the lesions that were seen on the 123I images would not have been appreciated without the benefit of the accompanying 131I scans. We also concur with their statement that “target-to-background activity [for 123I] is… much less than that seen with 131I.”
Our study did not address such other issues as stunning or the need for routine pretherapy whole-body imaging in the first place (5). But we do emphasize that development of an appropriate diagnostic algorithm must take into account the relative insensitivity of 123I for thyroid cancer metastases in the 74- to 185-MBq range.
In conclusion, it is misleading to claim that 123I is superior to 131I for the detection of thyroid cancer metastases before therapy without actually comparing the 2 tracers. Having made a direct comparison, we have found just the opposite—that 131I is the better imaging agent. Although editorial constraints limited the number of figures in our article, the images that we provided adequately proved this point. We realize that a direct comparison of diagnostic 123I and 131I images is difficult and that patients with distant metastases are few, but we hope that studies similar to ours will be done by others.