Assessment of Simplified Blood Dose Protocols for the Estimation of the Maximum Tolerable Activity in Thyroid Cancer Patients Undergoing Radioiodine Therapy Using ¹²⁴I

Study Design

All patients signed a written informed consent form, and the study was approved by the local medical research ethics committee. This study retrospectively analyzed ¹²⁴I dosimetry data from high-risk differentiated thyroid cancer patients, who received PET/CT lesion and blood dosimetries. The following data were extracted: age, sex, weight, height, tumor histology, stage of disease, level of thyroid-stimulating hormone (TSH), stimulation type (exogenous or endogenous), level of thyroglobulin, and the number of preceding radioiodine therapies before ¹²⁴I dosimetry. Diffuse pulmonary uptake was identified using functional images, that is, 24-h ¹²⁴I PET/CT images (18) or whole-body ¹³¹I planar scintigraphy images acquired 7–10 d after radioiodine therapy.

Patient Preparation, Blood Sampling, and WC

All patients had undergone total thyroidectomy. The patients were either under endogenous TSH stimulation (by levothyroxine withdrawal for 4 wk) or under exogenous TSH stimulation (by use of recombinant human TSH) with sufficient TSH levels (≥25 mIU/L). The patients were instructed to be on a low-iodine diet for 4 wk before the ¹²⁴I dosimetry. Iodine contamination was excluded by urine testing before individualized ¹²⁴I dosimetry; patients were excluded if there were more than 250 μg of iodine per gram of creatinine. A capsule containing 20–50 MBq of ¹²⁴I was administered after mandatory micturition. The patient endured without micturition (or defecation) until completion of the first WC measurement.

The blood ¹²⁴I dosimetry protocol involved serial BS and WC. More precisely, heparinized blood samples were collected at 2, 4, 24, 48, and ≥96 h after ¹²⁴I administration and measured in a well counter. The details of the calibrated ¹²⁴I activity measurements have been published previously (10). The absolute activity concentration of the whole blood was determined and used to calculate the (projected) ¹³¹I uptake curve of the blood compartment. In the calculation, the individual patient’s total blood volume was predicted using the body-surface method recommended by the International Council for Standardization in Haematology (19). WC occurred at time points similar to the BS, except for the early time point (1 h instead of 2 h). The patient was positioned approximately 3 m in front of an uncollimated γ-camera detector. Anterior and posterior counts were acquired for the calculation of a geometric mean. After dead time correction, all geometric mean counts were normalized to the first data point to construct the (projected) whole-body ¹³¹I uptake curve. For BS or WC, reference standard was counted to examine the consistency of the well counter system or γ-camera system.

Reference and Simplified Protocols and Their Naming Convention

The reference protocol included the complete set of BS and WC data points. Simplified protocols included a reduced number of the combined BS and respective WC data points; this type of protocol is termed protocols C. Protocols that used data points from BS (type B) or from WC (type W) alone were also included. For obvious reasons, the data points of the first BS (2 h) and the first WC (1 h) were included in all of the simplified protocols.

Twelve simplified protocols were considered: C(24), C(48), C(24, 96), C(48, 96); W(24), W(48), W(24, 96), W(48, 96); and B(24), B(48), B(24, 96), B(48, 96). The first symbol of the protocol’s naming convention denotes the protocol type. The numbers within parentheses are the time points (in h) of the measured uptake values that were included in the respective protocols (in addition to the first uptake value). For instance, C(24) is a simplified protocol that included both BS data points acquired at approximately 2 and 24 h and WC data points acquired at approximately 1 and 24 h.

Blood and Whole-Body Residence Times and Absorbed Dose to Blood

To avoid ambiguity and maintain consistency for all protocols, no curve-fitting was applied. The blood and whole-body residence times were determined by separating the blood and whole-body uptake curves into 3 phases, that is, an early, a mid, and a late phase. Figure 1 schematically illustrates the 3 phases and the piecewise parameterization of representative uptake curves. In particular, for all protocols, an instant uptake was assumed, that is, the uptake at the zero time point equaled the uptake value of the early time point (early phase). The mid phase, the region between the early and the last uptake point (≥96 h), was parameterized using multiple single monoexponential functions. This approach is similar to the trapezoid method for computing the area under the curve, except that it approximates the uptake values between 2 measurements using a piecewise exponential rather than a piecewise linear function. In detail, the uptake values between 2 adjacent measured data points were interpolated (or extrapolated to 96 h if the last measured data point is not included) using a monoexponential function with an effective ¹³¹I half-life calculated by linear regression. Thus, the number of the single monoexponential functions within the mid phase was 4 for the reference protocol C(4, 24, 48, 96) (Fig. 1A) and 2 or 1 for the protocol C(24, 96) (Fig. 1B) or protocol C(24) (Fig. 1C), respectively. In the late phase, the uptake decreased exponentially with a ¹³¹I half-life, which equaled the effective ¹³¹I half-life of the last exponential function of the mid phase (Figs. 1B and 1C), unless its value was larger than the physical ¹³¹I half-life. In such a case, physical decay was assumed after the last measured uptake at ≥96 h (Fig. 1D); this occurred only for the blood uptake curve (in only 5 instances). For comparison purposes, an additional integration approach was applied. The uptake curves were fitted using biexponential functions, an approach recommended by the SOP (5). The fitting procedure was restricted to uptake curves that included the complete BS or WC data points. This protocol is referred to as SOP protocol and can be considered an alternative to the reference protocol.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Schematic representation of uptake curves of reference protocol (A) and examples of simplified protocols (B–D). Different phases are indicated by vertical dotted lines. Numbers adjacent to lines represent monoexponential function. Dashed lines indicate extrapolated values. No numbers on uptake axes are shown.

For protocols of type C and the SOP protocol, both the residence times of the blood (τ_BS) and the residence times of the whole-body compartments (τ_WC) were used to calculate the blood dose per unit of administered (¹³¹I) activity (BDpA). The following numeric value equation was used (5):Eq. 1The symbols m_P and V_BS denote the mass and total blood volume of the patient, respectively. The first term is the contribution of the blood dose from β particles derived from BS; the second term is the effective contribution of the γ ray to the blood dose obtained from the WC. For protocols of type W and B, the concept proposed by Thomas et al. (15) was applied to calculate the blood dose per unit of administered activity. Because of the strong correlation between the activity of the blood and the whole-body compartments, either the BS data or the WC data can be disregarded. More precisely, the concept assumes that the blood and whole-body uptake curves can be linearly scaled onto each other. Historically, Thomas et al. (15) disregarded the BS data and used a scaling factor to predict the blood uptake curve for protocol of type W. Thomas’ scaling factor was fixed for all patients and was the mean ratio of the blood residence time to the whole-body residence time. For example, Thomas et al. (15) found on average a mean residence time ratio of 0.14 for 27 differentiated thyroid cancer patients bearing lesions. For the protocols of type B, a similar approach was used to scale the predicted whole-body uptake curve to the measured blood uptake curve. Because the whole-body uptake has to be 100% at the zero time point, the scaling factor for matching the predicted whole-body uptake curve to the measured blood uptake curve was calculated by dividing the 100-percentage value to the measured blood uptake at 2 h. Thus, dividing the blood residence time by the measured blood uptake at approximately 2 h revealed an approximation of the respective whole-body residence time. Therefore, inserting the respective scaling factors into Equation 1 allows estimation of the blood dose per unit of administered activity based on WC or BS alone.

MTA, Patient Groups, and Criteria for Protocol Equivalence

The MTA was calculated on the basis of a blood dose threshold of 2 Gy and an activity threshold of 3.0 or 4.4 GBq retained in the whole-body at 48 h in the presence or absence of diffuse pulmonary uptake, respectively. For each protocol, the blood dose per administered activity and the 48-h whole-body retention (R₄₈) were used to determine the respective critical therapy activity. More precisely, 2 Gy divided by BDpA yielded the critical activity for possible bone marrow toxicity and 3.0 or 4.4 GBq divided by R₄₈ yielded the critical activities for possible lung toxicities. The lowest critical therapy activity was taken as the MTA.

The patients were categorized into 2 patient groups. The first patient group comprised individuals who had not undergone any radioiodine therapies (group before treatment). The patients in the second group had undergone one or more radioiodine therapies (group after treatment). For each patient, the reference MTA and the residence time ratio (τ_BS/τ_WC) were determined using the reference protocol, and the mean residence time ratio was calculated for each patient group. Thereafter, the estimated MTAs derived from using the 12 simplified protocols and the SOP protocol were determined.

Comparisons between simplified protocols or SOP protocol and the reference protocols were assessed using the percentage MTA deviation from reference MTA and Lin’s concordance correlation coefficients (ρ_c). For each protocol, the percentage MTA deviation was illustrated using box plots, and the percentage of patients with percentage MTA deviations within the margin of ± 20%, the mean absolute percentage MTA deviation, and the absolute maximum percentage deviation were determined. A simplified protocol was considered equivalent to the reference protocol when both the estimated MTAs of 95% of the patients were within the ±20% deviation range and the maximum absolute percentage MTA deviation did not exceed the limit of 30% in a few cases. Lin’s concordance correlation coefficient along with the 2-sided lower and upper 95% confidence intervals were used to assess the strength of protocol agreement. With respect to a previous study (13), we applied Partik’s criteria, which designates ρ_c values >0.95 as excellent and >0.90 as very good.

Statistics

The descriptive statistics included the mean, median, SD, minimum, and maximum, which are provided in the following form: mean (median) ± SD (minimum to maximum). The correlation between different quantities was evaluated using the Pearson correlation coefficient; a significance level (P value) of <5% was considered statistically significant.