TO THE EDITOR:
Recently, 2 articles (1,2) on thyroid stunning have been published in The Journal of Nuclear Medicine. Yeung et al. (1) quantitatively confirmed significantly lower uptake of therapeutic 131I in the neck after diagnostic imaging with low activity of 131I (3) and presented evidence of thyroid stunning even after the diagnostic administration of ≤40 MBq 131I. Cholewinski et al. (2) found no visually apparent post-therapeutic thyroid stunning after diagnostic scintigraphy with 185 MBq 131I performed 3 d before radioiodine therapy.
Thyroid stunning is usually defined as decreased ability of normal thyroid or metastatic tissue to trap or retain therapeutic radioiodine, as a result of a previously administered diagnostic amount of 131I. Such a definition implies that the effect is quantitative and that measurement of fractional lesion uptake must be taken. Visual assessment of scintigraphic changes in thyroid remnants or cancer kinetics of radioiodine cannot be objective in every respect. Visualizing the same number of foci and the same relative intensity of uptake (2) does not prove the absence of stunning, because the ratio of observed uptake to predicted uptake can still be lower than the ratio of administered therapeutic activity to diagnostic activity. As a rule, quantitative studies on thyroid stunning have found decreased efficacy per unit of therapeutic activity of 131I after the diagnostic administration of 131I (1, 3–5), whereas visual assessment has often failed to recognize it (2). Furthermore, radiation absorbed dose (Gy) and dose rate (Gy/h) are the parameters that most clearly explain the impact of ionizing radiation on living organisms, though consideration of these 2 parameters is usually neglected.
It is to be regretted that both recent papers (1,2) failed to refer to the work of Jeevanram et al. (4), which to our knowledge is the only article in the literature that considers the influence of the diagnostic absorbed dose of 131I on the subsequent therapeutic uptake of 131I. In their article, Jeevanram et al. (4) showed that the higher the diagnostic absorbed dose, the higher the reduction of therapeutic uptake of 131I will be. In their study, the mean therapeutic uptake at 72 h was only about one half of the diagnostically expected uptake, whereas the diagnostic absorbed dose of up to 17.5 Gy appeared as a dose level of acceptably negative influence with, on average, 24% reduced therapeutic uptake (4).
After surgery on our patients for well-differentiated thyroid cancer, we calculated both the diagnostic and the therapeutic absorbed doses on the basis of determined functional mass of the target organ and the integration of the time–activity curves, which were constructed from serial uptake measurements (5). The mean therapeutic absorbed dose was only about one half of the value predicted by a work-up study with 75 MBq 131I (5). The correlation between the diagnostic absorbed dose and the reduction of therapeutic absorbed dose was clear, whereas the diagnostic absorbed dose above which thyroid stunning becomes significant was found to be 10 Gy (5).
We believe that any future quantitative investigation of diagnostic and therapeutic kinetics of radioiodine should take into consideration 2 points. First, different diagnostic and therapeutic fractional lesion uptake at a specific time point may result from different shapes of the retention curves (different maximums, different half-lives). Therefore, single measurements of the uptake at the same time after administration of diagnostic and therapeutic radioiodine are indicative (1,3,4) but not conclusive. For a more reliable comparison of activity retention, as many measurements should be taken as possible. Second, different diagnostic and therapeutic fractional lesion uptake may result from changes in lesion mass as a function of time. Potential loss or gain of the lesion mass during the study (radiation damage, thyroptin stimulation) may change the absolute amount of the activity in the lesion, even if the uptake per gram of the target tissue remains constant.
We made a few other noteworthy observations regarding the 2 recent papers. In Yeung et al. (1), Figure 2 shows the correlation (r = 0.75; slope = 0.42) between percentage uptake of diagnostic and therapeutic activity (1); it is obvious that the uptake of patient 2 (listed in Table 1) is not present on the graph, whereas the slope of the regression line is much greater than 0.42. Recalculation in both lesion-to-lesion and patient-to-patient cases gives very good correlation between diagnostic and therapeutic uptake (r = 0.95; P < 0.01; slope ≥ 0.83). Furthermore, excluding the most outlying point (patient 3) from the plot on Figure 3 makes the correlation between the ratio of therapeutic to diagnostic uptake (%T/D) and diagnostic activity significant (r = −0.72; P < 0.02). Thus, it seems that the stated unpredictability of the degree of stunning (1) does not hold, as has been reported previously (3,4,5).
In Cholewinski et al. (2), the camera tracking speed was 10–12 cm/min for diagnostic whole-body scans and 20 cm/min for post-therapeutic whole-body scans, and the administered diagnostic and therapeutic activities were 185 and 5,550 MBq 131I, respectively. Assuming similar diagnostic and therapeutic retention of radioiodine, visual comparison of 2 scintigrams for stunning with counting statistics different for an order of magnitude would be difficult. Nevertheless, on Figure 1, which should show no evidence of stunning by the diagnostic dose (2), the lesions in the thyroid bed appear scintigraphically less prominent on the therapeutic than on the diagnostic scan, and 131I concentration in the left shoulder was not mentioned at all.
REFERENCES
REPLY:
At Memorial Sloan-Kettering Cancer Center (New York, New York), we annually perform about 200 thyroid dosimetry studies. In these studies, patients are usually administered a tracer dose of 37–185 MBq 131I, primarily to determine the maximum tolerated dose, which is defined as that activity that would result in a radiation absorbed dose of 2 Gy to blood (1,2). These patients receive serial whole-body gamma camera scanning on days 1, 2, and 4 and daily neck uptake measurements using a neck probe. Radiation absorbed dose to the tumor is estimated in these patients. However, the accuracy of these estimates depends on the mass of the lesion, which is rarely known, particularly because the tumors for many of our patients are either at or below the limits of the resolution of the gamma camera. In such instances, even CT is not a reliable estimator of tumor mass. For this purpose, we are investigating the potential of PET using 124I with the hope of improving tumor dose quantitation (3). The initial intention of the study was to correlate the magnitude of stunning with the absorbed dose from the dosimetry study before therapy, but we discovered that this was not possible in these patients.
We have recalculated the slope and correlation coefficient for the plot of percentage uptake from the therapy scan versus percentage uptake from the dosimetry scan (Fig. 2 (4)). The original Figure 2 was incorrect and the recalculated slope was 0.78 with an R2 of 0.89, which suggests a much better correlation than previously presented. However, this is a trend line with a large scatter of points around the line of best fit.
When the percentage uptake of the therapeutic dose relative to the diagnostic dose (%T/D) was plotted against the administered activity, no correlation was found (R2 = 0.001). With no stunning, all data points should lie along a horizontal line at 100%. If stunning is correlated with the amount of radioiodine administered, a progressive reduction in %T/D would be expected. One point at 388% can definitely be considered an outlier by the Thompson criterion (5). A re-estimation of the correlation coefficient with this data point omitted resulted in little improvement in the correlation coefficient (R2 = 0.0643).
These data reveal nothing about the magnitude of stunning versus radiation absorbed dose. We agree with Dr. Medvedec that a dependence of the degree of stunning on the radiation dose is to be expected. We expected it too but were unable to show it in our work. However, other factors that will also impact the magnitude of stunning should not be neglected. First, the amount of radioiodine accumulation will depend on the tumor histology, with stunning conjectured to be greater in a well-differentiated tumor with a more intact follicular histology. Second, in 1 patient with 2 neck lesions (patient 10 in our study (4)) of possibly different histologic type, we observed that the uptake in one lesion was greatly reduced during the therapy, with a concomitant increase in the second lesion of lower percentage uptake on the dosimetry scan. This was presumably a consequence of competition; that is, the first lesion acted as an iodine sink for the second lesion during the dosimetry scan but exhibited greater uptake once the radiation effects had damaged the trapping of the first lesion.
In conclusion, we generally agree with the remarks made by Dr. Medvedec. Although we did not observe a strong correlation between administered activity and the magnitude of stunning, we do anticipate an improved correlation when correlated with lesion dose, with the reminder that there are several other factors such as histologic type that will also impact the magnitude of the effect.
REPLY:
We read with attention Dr. Medvedec’s Letter to the Editor, which raises several interesting points.
Thyroid stunning implies lower uptake of the therapeutic dose than that expected from the diagnostic dose; its importance lies in its potential effect in reducing the efficacy of 131I therapy. The purpose of our study was to try to evaluate whether decreased uptake, should it exist and be sufficient to suggest reduced therapeutic efficacy, was evident in our studies. Our data suggest that this is not the case, as evidenced by our patient population’s long-term survival rates, which compare favorably with those of other centers.
We intentionally chose to analyze the data visually rather than by semiquantitative approaches used by other colleagues for several reasons. These semiquantitative approaches fail to take into account several important factors that affect the results. For example, the thyroid remnant is not the only organ to which 131I is distributed. Therefore, a therapeutic dose of 131I may not be distributed to all the organs in the same proportion as a diagnostic dose, because of nonlinear variation in rates of transfer between different compartments. Furthermore, we feel it would be overly simplistic to conclude that the fractional uptake of the therapeutic dose was reduced by the diagnostic dose merely because a region-of-interest ratio or a similar index was reduced. Hence, a conclusion of “decreased efficacy per unit therapeutic activity” would be vague and should be regarded as a possibility rather than as proven fact unless it is supported by other data, by visual or long-term follow-up, or by a similar parameter that actually demonstrates the detrimental consequences. If the numbers show a significantly large reduction that is neither visible in the images to a trained professional nor shown to affect the long-term efficacy of treatment and prognosis of a patient, then we respectfully suggest that the scientific process demands that we re-examine and refine the process generating those numbers, that is, modifying the theory to fit the facts rather than the reverse.
In view of this statement, we do not believe that a difference in tracking speed between the diagnostic and therapeutic scans or the range of 131I doses affects our conclusions. We saw the same number of lesions, with similar relative intensities in all our patients, in both sets of images. If the severity of stunning is as high as 40%–50% in certain cases, as suggested by some quantitative parameters, one would not expect to see a residual focus or metastatic lesion in a therapeutic scan that had been seen on a diagnostic study, which would suggest that it may not have been adequately treated. Of course, in this scenario, one must then also show that the lesion subsequently grew bigger to the detriment of the patient, requiring active management.
We would like to point out that the apparent difference in the thyroid bed lesion in our Figure 1 (1) is caused by slightly different thresholding of the exposed film rather than by any significant difference in uptake, as is borne out by the relative intensity of the gut between Figures 1A and 1B. However, we agree that the caption omits the metastatic lesions in the left shoulder and the right lung; our intention was to point out a metastatic lesion in addition to the thyroid bed as an example in this patient’s case. Indeed, both of the lesions not mentioned in the caption are seen at a similar relative intensity in both images, further supporting our conclusions about the lack of visually apparent stunning.
We regret that we did not refer to Jeevanram et al. (2) in our article. We derived our list of references from a computerized Medline search, which we assumed to be complete for our search keywords. However, our previous comments also apply to the methodology of this article as well.
We thank the authors for their interest in the subject and in our article.
