Biphasic and monophasic repair: comparative implications for biologically equivalent dose calculations in pulsed dose rate brachytherapy of cervical carcinoma.

OBJECTIVE: To consider the implications of the use of biphasic rather than monophasic repair in calculations of biologically-equivalent doses for pulsed-dose-rate brachytherapy of cervix carcinoma. METHODS: Calculations are presented of pulsed-dose-rate (PDR) doses equivalent to former low-dose-rate (LDR) doses, using biphasic vs monophasic repair kinetics, both for cervical carcinoma and for the organ at risk (OAR), namely the rectum. The linear-quadratic modelling calculations included effects due to varying the dose per PDR cycle, the dose reduction factor for the OAR compared with Point A, the repair kinetics and the source strength. RESULTS: When using the recommended 1 Gy per hourly PDR cycle, different LDR-equivalent PDR rectal doses were calculated depending on the choice of monophasic or biphasic repair kinetics pertaining to the rodent central nervous and skin systems. These differences virtually disappeared when the dose per hourly cycle was increased to 1.7 Gy. This made the LDR-equivalent PDR doses more robust and independent of the choice of repair kinetics and α/ß ratios as a consequence of the described concept of extended equivalence. CONCLUSION: The use of biphasic and monophasic repair kinetics for optimised modelling of the effects on the OAR in PDR brachytherapy suggests that an optimised PDR protocol with the dose per hourly cycle nearest to 1.7 Gy could be used. Hence, the durations of the new PDR treatments would be similar to those of the former LDR treatments and not longer as currently prescribed. ADVANCES IN KNOWLEDGE: Modelling calculations indicate that equivalent PDR protocols can be developed which are less dependent on the different α/ß ratios and monophasic/biphasic kinetics usually attributed to normal and tumour tissues for treatment of cervical carcinoma.

Effects of very low dose-rate (90)Sr/(90)Y exposure on the acute moist desquamation response of pig skin.

BACKGROUND AND OBJECTIVES: Previous data, predominantly involving high dose-rate fractionated irradiation with incomplete repair intervals, had indicated that the kinetics of repair of sublethal damage for acute radiation reactions in pig skin could best be defined by a biphasic repair model with half-times for repair of 0.2 and 5.4 h, partition coefficient 0.5. To further test the validity of this finding and obtain a better estimate of the repair rate of the slow component of repair, the acute response of pig skin to very low dose-rates (VLDR), originally estimated to be 0.0067-0.0244 Gy/min, was investigated as part of a 4 fraction irradiation protocol involving an overall treatment time of <9 days to avoid confounding factors such as induced repopulation and enhanced radio-sensitivity in this animal tissue. MATERIALS AND METHODS: The flank skin of female Large White pigs, 3-4 months of age, was locally irradiated (8 sites/flank) with 22.5 mm diameter (90)Sr/(90)Y plaques. Irradiation with a 4 fraction protocol included 3 equal, high dose-rate, fractions with full repair, followed by a fourth VLDR fraction. The total doses administered were originally planned to represent the dose associated with the predicted ED(20), ED(50) and ED(80) (75% of total biological dose given at high dose-rate and 25% at VLDR) calculated on the basis of the repair kinetic parameters obtained from earlier studies. However, during the analysis a revision to the physical dosimetry was identified; this had been overlooked prior to the start of the study. Following completion of irradiation the irradiated sites were examined weekly and the presence or absence of moist desquamation recorded. RESULTS: The incidence of moist desquamation was slightly higher than expected on the basis of the parameters used to calculate iso-effective doses, at least in part as a consequence of the change to the dosimetry. Using likelihood methods and the original dose estimates, the best model based estimate of the dose-rate correction factor for the LDR and VLDR plaques was 1.29. This was comparable with the physical calibration factor, median value 1.23. The VLDR fraction associated with a 50% incidence of moist desquamation, based on experimental observation, was 23.2+/-0.84, 27+/-2.6 and 30.1+/-3.2 Gy, for corrected VLDRs of 0.0247, 0.0093 and 0.0068 Gy/min, respectively. A biphasic model, which incorporated a dose-rate correction factor, provided a better fit than a monophasic repair model to the total data set, which now included the new VLDR data. Moreover, the monophasic repair model suggested a dose-rate correction factor of 1.63, well outside the range derived from the re-evaluation of the physical dosimetry. CONCLUSION: Using the total data (with model based corrected dose-rates), the analysis revealed two components of repair with half-times of 0.103 (0.0594-0.177) and 2.97 (1.96-4.50) h; partition coefficient 0.375 (0.225-0.526). These are comparable with the estimates for other tissues (the CNS in particular) and suggest that the kinetics of repair may be relatively species and tissue independent with variation observed being more related to experimental design rather than any true differences.

Clinical implications of incomplete repair parameters for rat spinal cord: the feasibility of large doses per fraction in PDR and HDR brachytherapy.

PURPOSE: To evaluate the clinical implications of the repair parameters determined experimentally in rat spinal cord and to test the feasibility of large doses per fraction or pulses in daytime high-dose-rate (HDR) or pulsed-dose-rate (PDR) brachytherapy treatment schedules as an alternative to continuous low-dose-rate (CLDR) brachytherapy. METHODS AND MATERIALS: BED calculations with the incomplete repair LQ-model were performed for a primary CLDR-brachytherapy treatment of 70 Gy in 140 h or a typical boost protocol of 25 Gy in 50 h after 46-Gy conventional external beam irradiation (ERT) at 2 Gy per fraction each day. Assuming biphasic repair kinetics and a variable dose rate for the iridium-192- (192Ir) stepping source, the LQ-model parameters for rat spinal cord as derived in three different experimental studies were used: (a) two repair processes with an alpha/beta ratio = 2.47 Gy and repair half-times of 0.2 h (12 min) and 2.2 h (Pop et. al.); (b) two repair processes with an alpha/beta ratio = 2.0 Gy and repair half-times of 0.7 h (42 min) and 3.8 h (Ang et al.); and (c) two repair processes with an alpha/beta ratio = 2.0 Gy and repair half-times of 0.25 h (15 min) and 6.4 h (Landuyt et al.). For tumor tissue, an alpha/beta ratio of 10 Gy and a monoexponential repair half time of 0.5 h was assumed. The calculated BED values were compared with the biologic effect of a clinical reference dose of conventional ERT with 2 Gy/day and complete repair between the fractions. Subsequently, assuming a two-catheter implant similar to that used in our experimental study and with the repair parameters derived in our rat model, BED calculations were performed for alternative PDR- and HDR-brachytherapy treatment schedules, in which the irradiation was delivered only during daytime. RESULTS: If the repair parameters of the study of Pop et al., Ang et al., or Landuyt et al. are used, for a CLDR-treatment of 70 Gy in 140 h, the calculated BED values were 117, 193, or 216 Gy(sc) (Gy(sc) was used to express the BED value for the spinal cord), respectively. These BED values correspond with total doses of conventional ERT of 65, 96, or 104 Gy. The latter two are unrealistic high values and illustrate the danger of a straightforward comparison of BED values if repair parameters are used in situations quite different from those in which they were derived. For a brachytherapy boost protocol, the impact of the different repair parameters is less, due to the fact that the percentage increase in total BED value by the brachytherapy boost is less than 50%. If a primary treatment with CLDR brachytherapy delivering 70 Gy in 140 h has to be replaced, high doses per fraction or pulses (> 1 Gy) during daytime can only be used if the overall treatment time is prolonged with 3-4 days. The dose rate during the fraction or pulse should not exceed 6 Gy/h. For a typical brachytherapy boost protocol after 46 Gy ERT, it seems to be safe to replace CLDR delivering a total dose of 25 Gy in 50 h by a total dose of 24 Gy in 4 days with HDR or PDR brachytherapy during daytime only. Total dose per day should be limited to 6 Gy, and the largest time interval as possible between each fraction or pulse should be used. CONCLUSION: Extrapolations based on longer repair half-times in a CLDR reference scheme may lead to the calculation of unrealistically high BED values and dangerously high doses for alternative HDR and PDR treatment schedules. Based on theoretical calculations with the IR model and using the repair parameters derived in our rat spinal cord model, it is estimated that with certain restrictions, large doses per fraction or pulses can be used during daytime schedules of HDR or PDR brachytherapy as an alternative to CLDR brachytherapy, especially for those treatment conditions in which brachytherapy is used after ERT for only less than 50% of the total dose.

Radiation tolerance of rat spinal cord to pulsed dose rate (PDR-) brachytherapy: the impact of differences in temporal dose distribution.

PURPOSE: To investigate the impact of a time-variable dose rate during a high dose rate (HDR-) or pulsed dose rate (PDR-) brachytherapy fraction with the HDR-microSelectron and to compare this with the biological effect of a constant dose rate treatment with the same average dose rate (as in the case of (192)Ir-wires). Moreover, the kinetics of repair in rat spinal cord are investigated using a wide spectrum of temporal dose distributions. MATERIALS AND METHODS: Two parallel catheters are inserted on each side of the vertebral bodies of the rat spinal column (Th(10)-L(4)) and connected to the HDR-microSelectron. Interstitial irradiation (IRT) is performed with a stepping (192)Ir-point source, varying the activity of the point source between 0.3 and 6.5 Ci. Three different groups of experiments are defined, varying the overall treatment time and average dose rates in the range of 3-8, 28-53 and 82-182 min and 312-489 Gy/h, 32-56 Gy/h and 13-15 Gy/h, respectively. Difference in temporal dose distribution (dose rate variation) during almost the same overall treatment time is obtained by varying the number of pulses per dwell position in either one or ten runs through the implant. For reasons of comparison, previously reported results of continuous irradiation at a constant dose rate obtained with two (192)Ir-wires in a fixed position are reanalyzed. Paralysis of the hindlegs after 5-6 months and histopathological examination of the spinal cord of each animal are used as experimental endpoints. RESULTS: During one run of the (192)Ir-point source, the peak dose rate is at least 25 times higher as compared with the minimum local dose rate and almost four times higher as compared with the average dose rate. For the three different groups of varying overall treatment times and average dose rates there is a significant difference in biological effect, with an ED(50)-value of 23.1-23.6 Gy (average dose rate 312-489 Gy/h), 25.4-27.9 Gy (average dose rate 312-489 Gy/h) and 29.3-33 Gy (average dose rate 13-15 Gy/h). For these range of single doses, difference in temporal dose distribution with either one or ten runs is only significant for treatment times less then 1 h. For the prolonged treatment times at lower average dose rates, the difference between one or ten run is no longer significant. However, the results with the (192)Ir-point source at an average dose rate/run of 13-15 Gy/h are significantly different from the ED(50)-value of 33 Gy using (192)Ir-wires at the same but constant dose rate. Using different types of analysis to estimate the repair parameters, the best fit of the data is obtained assuming biphasic repair kinetics and a variable dose rate (geometrically dependent) for the (192)Ir-point source. On the basis of the incomplete repair LQ model, two repair processes with an alpha/beta ratio=2.47 Gy and repair halftimes of 0.19 and 2.16 h are detected. The partition coefficient for the longer repair process is 0.98. This results in the proportion of total damage associated with the longer repair halftime being 0.495 for short sharp fractions with complete repair in between. CONCLUSIONS: Even in the range of high dose rates of 15-500 Gy/h, spinal cord radiation tolerance is significantly increased by a reduction in dose rate. For larger doses per fraction in PDR-brachytherapy dose rate variation is important, especially for tissues with very short repair half times (components). In rat spinal cord the repair of sublethal damage (SLD) is governed by a biphasic repair process with repair halftimes of 0.19 and 2.16 h.

Haemopoietic injury after irradiation: analysis of dose responses and repair using a target-cell model.

Published dose-incidence data for haemopoietic lethality in mice have been analysed using a mathematical model based on target-cell survival. The analysis of three comprehensive data sets produced an initial D(o) value of about 1.4 Gy, decreasing to about 1.1 Gy at 3 Gy, 0.9 Gy at 5 Gy, and about 0.7 Gy at a dose of 10 Gy. The alpha/beta ratio was about 14 Gy, and the repair half-time was about 0.3 h. The level of target-cell depletion at LD37 was at about 6 x 10(-4). The D(o) values are compatible with those measured directly for several stages of early haemopoietic progenitor cells in the marrow. The additional use of 13 or alternatively 24 other less-comprehensive data sets increased the overall degree of heterogeneity, so flattening dose-response curves and increasing the deduced overall D(o) values by a factor of about 2. However, when these data sets were stratified with respect to ln(N(o)) where N(o) is the number of tissue rescuing units (TRU), the results were comparable to those obtained when the three comprehensive data sets were analysed individually. Also, the repair halftime was higher at about 1 h. Further, the implied radiosensitivity of the projected target-cell population comprising the TRU was similar to the survival curves obtained for CFU-S and other closely-related haemopoietic progenitor cell types. It has been shown that the number of critical TRU at risk in the marrow is the main feature modulating heterogeneity even when it is assumed that the cellular radiosensitivity does not vary between strains. The number of stem cells comprising a TRU may vary between strains and this may also be influenced by environmental and/or immunological factors. However, it is certainly the case that the initial complement of TRU plays a major role in the incidence of whole body radiation induced mortality.

Biphasic cellular repair and implications for multiple field radiotherapy treatments.

Data describing the response of several normal tissues to fractionated irradiation, in terms of a biphasic repair of sub-lethal damage, have now been published. Typical results of such analyses have been taken and applied to a conventional radiotherapy protocol of 60 Gy in 30 daily fractions. The effect of using a four field treatment plan is shown to reduce the biological effect of the radiation schedule by increments dependent upon the time interval between each field in a treatment fraction, with a 10% reduction in the extrapolated dose response (ERD) resulting from a 5 min interfield interval. When applied to tissues having the same repair characteristics as pig skin this reduction in ERD is predicted to result in an approximately 25% reduction in the probability of acute morbidity from a protocol of 60 Gy in 30 fractions. These results imply that the basic LQ model, which is unable to correct for interfield intervals, overestimates the effect on normal tissues of radical clinical protocols, most of which use more than a single field. Increasing the interfield interval could be used to reduce the normal tissue side effects from radical radiotherapy when multiple fields are used.

The influence of the number of fractions and bi-exponential repair kinetics on biological equivalence in pulsed brachytherapy.

A linear-quadratic radiobiological model incorporating single or bi-exponential repair kinetics has been used to show the following and other features when a continuous low dose rate (CLDR) 70 Gy/140 h brachytherapy protocol is replaced by a radiobiologically equivalent pulsed dose rate (PDR) system using 140 fractions for reasons of dosage homogeneity. (1) For equivalent effects in late-reacting tissues, the PDR dose (at 5 or 0.05 Gy min-1) x 1 h intervals needs to be reduced by up to only 3%. Progressively further reductions in dose are required when fewer larger fractions are used. (2) When equivalence using pulsed doses is achieved for one normal tissue type, and extrapolated response doses (ERD) are calculated for other tissue types in the irradiated volume, values of the ERD remain within 5% of each other using the above PDR protocol and associated parameters. (3) For tumours with alpha/beta = 10 Gy and a single repair halftime of 0.1-1.0 h, there is no significant loss of therapeutic benefit using the PDR protocol equivalenced for late normal tissue reactions. The strategy of replacing an LDR boost protocol of about 24 Gy by a PDR protocol gives similar levels to the 70 Gy PDR protocol for the expected percentage increase in the biological dose to normal tissues (due to the PDR protocol alone). These calculations also highlight the importance of the values assumed for the conventional alpha/beta ratio and the repair kinetics when estimating equivalent PDR protocols. The use of an inappropriate radiobiological parameterization will lead to erroneous conclusions with the potential to advocate PDR protocols which will, in practice, lead to an increase in late complications.

Repair kinetics in pig epidermis: an analysis based on two separate rates of repair.

Pig skin was irradiated using 90Sr/90Y plaques and the dose-related incidence of induced moist desquamation was determined. The repair of radiation-induced sublethal damage (SLD) was studied by fitting these response data to the generalized LQ equation for incomplete repair using quasilikelihood methods with binomial statistics, and either a Poisson or logistic link to relate the probability of response to the covariates. A Poisson response analysis based on the assumption that SLD was governed by two repair processes gave estimated repair half-times of 0.20 [(95% confidence limits) 0.12, 0.34] and 6.6 [4.3, 10.0] h. The estimates of the short and long repair half-times were significantly different, although there was no significant difference between the results using the Poisson and logistic modes of analyses. The partition coefficient for the longer repair process was 0.5 [0.34, 0.71] indicating that about 33% of SLD-derived lethal damage is associated with the longer repair process in the case of 'complete repair' protocols. However, this proportionation is, in general, protocol dependent for incomplete repair protocols. A chi2 test on the residual deviance showed that the assumption of two repair processes for SLD gave a superior fit to the data than a single repair process at a significance level >99%. The radiation dose to the assumed target cell population depends upon their depth from the skin surface, due to the relatively short range of the electron emission from the 90Sr/90Y plaques. However, further modelling analyses have shown that the estimated repair half-times were independent of the assumed target cell distribution in the skin. This is in contrast with the alpha/beta ratio, where different (clinically significant) estimates can be obtained depending upon the assumed target cell distribution. If the target cells were at 16 micrometer depth from the surface of the skin, the estimated value for the alpha/beta ratio using the biphasic repair model would be 4.6[3.6, 5.6] Gy(Poisson analysis). However, the estimates decrease with the assumed depth (distribution) of the target cells.

Two components of repair in irradiated kidney colony forming cells.

Data describing the response of mouse kidney colony-forming cells to fractionated X-irradiation using different interfraction intervals were analysed in order to detect the presence of one or more components of repair. A two-component repair model gave a superior fit, with reference to a single-repair rate model, giving distinct repair halftimes of 0.15 (approximate 95% confidence limits: 0.0, 0.40) and 5.03 (1.23, 8.84) h. These values are the first reported for normal cells in vivo, and they are similar to values calculated for tissue responses in skin, lung and the spinal cord. The slow component of repair is important in radiotherapy, in particular regarding novel hyperfractionation regimens when interfraction intervals much less than 1 day are employed.

Derivation and application of equations describing the effects of fractionated protracted irradiation, based on multiple and incomplete repair processes. Part I. Derivation of equations.

A general model is developed where the induction and interaction rates of sublethal radiation damage with subsequent irradiation are represented as polynomial functions of the dose-rate. The effect of incomplete multiple repair processes is also included. Equations are evaluated for fractionated protracted irradiation where the dose-rate is constant during each fraction. However, both the dose-rate and the fractional dose are permitted to change from fraction to fraction. The resultant equations show that the apparent alpha/beta ratio derived from the analysis of equivalent protocols may be protocol dependent. Also, the alpha/beta ratio calculated from experimental data, assuming a single repair half-life, will appear to be protocol dependent if in reality more than one repair process is involved in the repair of sublethal damage. It is possible that the effects due to the induction of sublethal damage or its subsequent interaction may be distinguished by designing experiments in which the dose-rate is not constant throughout the whole protocol; also that the underlying processes governing the conversion of sublethal damage may be analysed by fitting experimental data to the equations.

Derivation and application of equations describing the effects of fractionated protracted irradiation, based on multiple and incomplete repair processes. Part 2. Analysis of mouse lung data.

General equations for fractionated protracted irradiation have been applied to the analysis of mouse lung data derived from isoeffect protocols. The data were accrued from a spectrum of different protocols including different radiation qualities, interfraction times and dose-rates. This analysis, based upon two component repair rate processes, suggests that there are possibly two repair-rate processes involved in the repair of sublethal damage in the lung. The two repair half-lives were 0.32 and 1.92 h with a low dose-rate partitioning of 1:0.38 between the amount of lethal damage resulting from rapid and slow sublethal damage repair processes at low dose-rates. The analysis also indicates that there may be a dose-rate amplification of the amount of resultant lethal damage, deriving from sublethal damage, associated with the longer repair process. It is also shown that the alpha/beta ratio is probably independent of the radiation quality and that the ratios of multiple to single event damage for 240kVp X-rays and 60Co are almost identical. However, the absolute values of alpha and beta for 240 kVp X-rays may be greater than for 60Co.

Application of the linear-quadratic model with incomplete repair to radionuclide directed therapy.

The linear-quadratic (LQ) model for fractionated external beam therapy has been modified by previous authors to include the effects due to an exponentially decaying dose rate. However, the LQ model has now been extended to include a general time varying dose rate profile, and the equations can be readily evaluated if an exponential radiation damage repair process is assumed. These equations are applicable to radionuclide directed therapy, including brachytherapy. Kinetic uptake data obtained during radionuclide directed therapy may therefore be used to determine the radiobiological dosimetry of the target and non-target tissues. Also, preliminary tracer studies may be used to pre-plan the radionuclide directed therapy, provided that tracer and therapeutic amounts of the radionuclide carrier are identically processed by the tissues. It is also shown that continuous radionuclide therapy will induce less damage in late-responding tissues than 2 Gy/fraction external beam therapy if the ratio of the maximum dose rate and the sublethal damage repair half-life in the tissue is less than 1.0 Gy. Similar inequalities may be derived for beta-particle radionuclide directed therapy. For example, it can be shown that radionuclide directed therapy will induce less damage to slowly repopulating tissue than 2 Gy/fraction external beam therapy for the same total dose if the maximum percentage initial uptake in tissue is less than 0.046%/g or 0.23%/g for an injected activity of 50 mCi of 90Y or 131I, respectively.

Application of radiobiological dosimetry to radionuclide directed therapy.

The standard linear quadratic model, which has been used to assess the radiobiological damage to tissue by external beam fractionated radiotherapy, has been extended to encompass a general continuous time varying dose rate protocol such as radionuclide therapy. If the radionuclide clearance from the tissue is purely exponential, the effect is readily calculated. Otherwise, the effect can be evaluated by numerical integration if the dose rate time-1 profile is known. It can be shown that if the maximum percentage initial uptake g-1 uptake in normal or tumour tissue is less than 0.046 or 0.23 for an administered activity of 50 mCi of 90Y or 131I respectively, then the radiation-induced damage will certainly be less than for 2 Gy fraction-1 external beam therapy for the same total dose. Preliminary imaging and knowledge of the radionuclide kinetics using a non-therapeutic dose may be used to calculate the predicted radiation damage to tissues for a particular therapeutic dose provided the tracer and therapy doses have identical kinetics.

Dosimetric model for antibody targeted radionuclide therapy of tumor cells in cerebrospinal fluid.

Although encouraging results have been obtained using systemic radioimmunotherapy in the treatment of cancer, it is likely that regional applications may prove more effective. One such strategy is the treatment of central nervous system leukemia in children by intrathecal instillation of targeting or nontargeting beta particle emitting radionuclide carriers. The beta particle dosimetry of the spine is assessed, assuming that the spinal cord and the cerebrospinal fluid compartment can be adequately represented by a cylindrical annulus. The radionuclides investigated were 90Y, 131I, 67Cu, and 199Au. It is shown that the radiation dose to the cord can be significantly reduced using short range beta particle emitters and that there is little advantage in using targeting carriers with these radionuclides. 199Au and 67Cu also have the advantage of having a suitable gamma emission for imaging, permitting pretherapy imaging and dosimetric calculations to be undertaken prior to therapy. If these methods prove successful, it may be possible to replace the external beam component used in the treatment of central nervous system leukemia in children by intrathecal radionuclide therapy, thus reducing or avoiding side effects such as growth and intellectual impairment.

The rapid separation and assay of previously unreported plasminogen activators in human euglobulin fractions. The effect of venous occlusion.

Euglobulins were prepared from pre and post venous occlusion plasma by acidification and dilution at pH 5.9. It is shown by isoelectrofocussing that there are several plasminogen activators with differing pI's in euglobulin fractions. Activators with pI's 8.9/9.0, and 9.4 are the major activators responsible for the increase in post occlusion plasminogen activator activity. Fibrin was necessary for the induction of the vascular occlusion response.

Comparative study of different urokinase preparations.

Three different urokinase preparations were assayed for amidolytic and plasminogen activator activity. It was found that the preparation with the least amidolytic activity displayed a significantly greater plasminogen activator activity than the others. These results suggest that this particular preparation may be more effective clinically and that the use of assay systems for the comparison of urokinase preparations must be treated with caution.

The comparison of solid phase and fibrin plate methods for the measurement of plasminogen activators.

A rapid and highly sensitive solid phase assay was compared with the fibrin plate method for the measurement of urokinase, streptokinase and the plasminogen activators in human euglobulin fractions. The solid phase assay was run using glu - or lys - plasminogen, and significant differences were observed in the activation of the plasminogens by urokinase and streptokinase. Plasminogen activator levels in euglobulin fractions were also measureable. Very good agreement was obtained between the fibrin plate and solid phase methods in all cases.

Iodine organification defect following treatment of thyrotoxicosis with antithyroid drugs.

Two patients remained clinically euthyroid following treatment for thyrotoxicosis with antithyroid drugs in spite of persistently elevated thyroid radioiodine uptakes not suppressable by exogenous triiodothyronine. Perchlorate discharge tests showed a defect in the intrathyroidal organification of iodine. Circulating levels of thyroxine were normal. From our study of 105 patients treated for thyrotoxicosis with antithyroid drugs, apparent remission of thyrotoxicosis by this mechanism might occur in up to 2 percent of patients.