TO THE EDITOR:

Dorn et al. have described a dosimetry-guided approach for calculating the largest “safe” dose of ¹³¹I for treatment of differentiated thyroid cancer (1). With this method, a maximal dose was determined assuming an upper limit of <3 Gy for the absorbed dose to the bone marrow or <30 Gy to the lungs. Their method used an initial tracer dose of 150–400 MBq (4–10.8 mCi) of ¹³¹I with serial dosimetric imaging for 4–5 d, followed by administration of the maximal treatment dose of ¹³¹I. The authors stated that this approach was intended to achieve “the maximum therapeutic benefit.” However, the outcome data presented in Table 4 do not appear to support this claim. From this table, thyroglobulin was >100 in 29 therapy instances. After ¹³¹I treatment, 20 of these 29 instances (69%) still had a thyroglobulin level of >100. The average survival of 29 patients who received “curative-intent” therapy was 4.1 y. Even after excluding patients older than 59 y (representing a group with poorer prognosis), average survival of the 10 younger patients was still only 5.2 y. We would be eager to have the authors explain why these outcome considerations support the contention of maximum therapeutic benefit by using this dosimetric approach, compared with that achieved using standardized ¹³¹I treatment algorithms.

The authors state in the Discussion, “In our patients, no stunning effect could be objectively demonstrated.” They go on to comment that “no decrease in thyroid uptake was seen after dosimetry with a 370-MBq test activity (with estimated absorbed doses as high as 50–100 Gy to the thyroid remnants).” However, the authors have also cited in the Discussion that dose estimates of 300 Gy should be sufficient to fully ablate normal thyroid tissue (their reference 18). It does not seem plausible that doses as large as one third of that estimated to be sufficient to ablate the thyroid would not have been responsible for a substantial stunning effect.

Methodology to assess for stunning was not described by the authors. Therefore, one might presume that such assessment was attempted by subjective comparison of the dosimetric and the posttreatment scans, assuming the latter were obtained (not described in the Materials and Methods). As has been summarized elsewhere by both of us (2,3), we believe that the potential liability of stunning by diagnostic doses of ¹³¹I before administration of the therapeutic dose is real and potentially of concern and that this liability was likely of consequence in the authors’ dosimetric procedure.

As has been noted by McMenemin et al. (4), qualitative comparison of diagnostic and posttreatment ¹³¹I scans may be highly insensitive for the detection of stunning. They described 3 cases in which significant reductions in uptake of the ¹³¹I treatment dose, compared with the prior ¹³¹I diagnostic dose (14%, 41%, and 60% reduced), were qualitatively inapparent by comparison of the pre- and posttreatment scans. The unreliability of such qualitative comparisons is potentially related to several factors, including the challenge of comparing target-to-background ratios at sites of differentiated thyroid tissue across a large dose range, differences between diagnostic and therapeutic postdose scanning times, and variable-differential washout.

Conversely, the data presented in Table 4 by Dorn et al. permit estimation of potentially significant stunning to thyroid tumor from the diagnostic dose in a substantial proportion of 26 of the therapeutic administrations for which the therapeutic estimates of tumor dosimetry were given. Although the exact diagnostic doses for these patients were not provided, one could assume for the sake of estimation an average of the diagnostic dose range given by the authors (150–400 MBq), equal to 275 MBq (7.5 mCi). By applying a factor equal to the ratio of the average diagnostic dose to the therapeutic dose for each of these 26 patients, the estimated dosimetry from the tracer dose of ¹³¹I was 6.4 ± 9 Gy (mean ± SD). Those doses estimated for patients 3, 13, 23, 25, and 28 were 16.6, 18.8, 12.3, 25.1, and 39.6 Gy, respectively. A recent report by Postgård et al. (5) described an in vitro porcine thyroid-cell bicameral tissue-culture model that showed a convincing stunning effect by exposure to ¹³¹I in the culture medium. A range of ¹³¹I exposures from 3 to 80 Gy significantly reduced the subsequent ¹²⁵I transport from the basal to the apical chamber. The lowest ¹³¹I dose tested, 3 Gy, caused a nearly 50% reduction, with a precipitous dose–response falloff for higher-level exposures. With respect to the 26 administrations that could be evaluated from the authors’ Table 4, we note that 13 (50%) could be estimated to have received >3 Gy to the thyroid tumor from the average diagnostic dose. This is a conservative quantitative estimate, since 12 of the therapeutic doses listed in Table 4 as “>” a particular Gy value were assumed to equal that value for the purposes of these calculations. Another recent report, by Lees et al. (6), described a statistically significant 3-fold increase in additional ¹³¹I treatments and a 50% increase in the total cumulative ¹³¹I treatment dose required for ablation of the remnant when a 185-MBq dose of ¹³¹I was used for the diagnostic scan before radioablation with ¹³¹I. This ¹³¹I diagnostic dose is approximately one third less than the average diagnostic dose used by the authors. These quantitative considerations would argue that at least some, if not many, of the cases described by the authors could be expected to have had a stunning effect, with consequent reduced fractional uptake of the therapeutic administration caused by the diagnostic dose.

The authors have noted that the initial ¹³¹I treatment, “the first strike,” has the highest therapeutic effect and that risk–benefit assessment before giving high-dose ¹³¹I is essential. We fully agree with these assertions, both of which allude to the importance of assessing the potential liabilities of stunning by the pretreatment diagnostic ¹³¹I dose. We believe that the methodology for estimating a maximal safe therapeutic ¹³¹I dose incorporating an initial diagnostic ¹³¹I dose in the range of that used by the authors most likely would contribute a significant risk of stunning. It is ironic that this effect may in fact offset a significant component of the incremental therapeutic benefit intended by the method. We wonder whether comparable, if not even better, outcomes might have been achieved with substantially lower therapeutic ¹³¹I doses had pretreatment exposure to the relatively larger diagnostic ¹³¹I doses been avoided. If used, dosimetric approaches for calculating a maximal ¹³¹I treatment dose should apply the lowest possible ¹³¹I dose to minimize stunning.