RU2417736C1 - Method for individual dosage of radioiodine-131 in treating patients with toxic goiter - Google Patents

Abstract

FIELD: medicine.

SUBSTANCE: invention refers to medicine, specifically to endocrinology, radiology, therapy. It involves physical examination of a patient. A percentage of iodine-131 test dose capture in 4 hours, a percent of iodine-131 test dose capture in 24 hours, ultrasound thyroid size, blood free thyroxine (pmol/l), duration of the disease, sex of the patient, duration of thyrostatic withdrawal before a therapeutic dose of radioiodine-131, age of the patient at a first manifestation of said disease, presence of thyroid resection in anamnesis, mercasolyl (thyrozol) dose just before drug withdrawal. Further, upper and lower limits of recommended individual radioiodine (I) dosage range for the certain patient. The lower dosage limit is derived from solving a problem of predicting a dose for negative therapeutic effect events, and the upper limit - for positive effect events. The reliability of positive or negative therapeutic effect is estimated by the prescribed dosage on the basis of solving the classification problem wherein a clinical data base of the prescribed doses supplements input data composition. The therapeutic effect is considered to be an output parametre, while the reliability estimation is represented in the form of a number range 0 to 1; a zero value of the range is interpreted as the negative effect, while the unit value shows the positive effect. The calculated range of the recommended doses is specified by examining the doses in vicinities of the borders with a preset dose look providing therapy confidence as 0.5 for the left border and 0.9 for the right border. The recommended dose is that one that corresponds to the right border of the specified range of the recommended doses.

EFFECT: method provides higher accuracy of individual dosage ensured by considering a great amount of patient's parametres having an action on the therapeutic effect.

3 ex

Description

The proposed method relates to medicine, namely to endocrinology, radiology, therapy, and is intended to determine an individual dose of radioactive iodine-131 ( ¹³¹ I) in the treatment of patients with diffuse toxic goiter (DTZ).

One of the most effective methods of treating patients with diffuse toxic goiter is radioiodine therapy, so special attention should be paid to its dosimetric planning.

Currently, when radioiodine therapy of patients with DTZ has found application, a method based on the introduction of a standard (identical for all patients) dose of ¹³¹ I.

This method is economical in terms of material and time costs, however, the question of the optimal dose is still open: doses in the range of 185-555 MBq are usually used. In some cases, a large volume of the thyroid gland (thyroid gland) or low drug uptake after 24 hours, the dose is usually heuristically increased.

The disadvantages of the method include ignoring all individual clinical and laboratory parameters of the patient, which have a significant impact on the outcome of therapy. The appointment of low fixed doses (185 MBq) leads to a high incidence of relapse (up to 40%), while increasing the dose to 555 MBq reduces the number of relapses on average to 19% [1].

At the same time, along with the introduction of a standard dose of ¹³¹ I, methods are known for determining the individual dose of radioactive iodine-131 ( ¹³¹ I) in the treatment of patients with diffuse toxic goiter, taking into account the mass of the thyroid gland (specific dose) and / or capture of radioactive iodine of the thyroid gland after 24 hours .

One of these methods is a method for determining an individual dose of ¹³¹ I for the treatment of patients with DTZ by calculating it according to the formula:

individual dose value ¹³¹ I = 296 MBq · 100 / capture of ¹³¹ I by the thyroid gland after 24 hours.

Moreover, this method takes into account only one indicator of the patient’s body, namely: the capture of ¹³¹ I by the thyroid gland after 24 hours.

The number of relapses in patients who received an individual dose determined by this method in the annual period is up to 14% [2].

The closest in combination of essential features selected by the authors as a prototype is the known method for determining the individual dose of radioactive ¹³¹ I in the treatment of patients with diffuse toxic goiter, which includes taking into account two clinical parameters of the patient: the mass of his thyroid gland and thyroid uptake ¹³¹ I after 24 hours, with the subsequent determination of the individual dose by the formula:

individual dose value ¹³¹ I = Z · M _SHCHZH · 100 / capture ¹³¹ I after 24 hours (%),

Where

Z is the value of the specific dose, usually selected in the range of 3.7-7.4 MBq;

M _thyroid - the mass of the thyroid gland [1].

Despite the fact that when determining the dose, clinical parameters such as thyroid mass and thyroid uptake ¹³¹ I after 24 hours are taken into account, which presumably have the most significant effect on the outcome of radioiodine therapy in patients with DTZ, dose determination by this method leads to relapse in patients with an average of not less than in 10-15% of cases [1].

The authors of the invention in the clinic reproduced this method for determining the dose of ¹³¹ I for 52 patients with DTZ. The results showed that the frequency of hyperthyroidism in this group of patients after treatment remained high throughout the year and amounted to 30-45% of cases, which is 3-4 times higher than the relapse data indicated in the source.

The present invention is to develop a method for determining the individual dose of radioactive iodine ( ¹³¹ I) in the treatment of patients with DTZ with high accuracy.

The proposed problem is solved by the proposed method for determining the individual dose of radioactive iodine-131 in the treatment of patients with diffuse toxic goiter, including conducting a clinical examination of the patient with the determination of its indicators, selected on the basis of the available database of clinical data that have a significant impact on the outcome of treatment (positive or negative) and provide the greatest accuracy in predicting the outcome of treatment, such as: the percentage of capture of the test dose of iodine-131 after 4 hours, the percentage of capture of the test dose iodine-131 after 24 hours, the volume of the thyroid gland according to ultrasound, free blood thyroxine (pmol / l), the duration of the disease, the sex of the patient, the duration of cancellation of thyreostatics before taking a therapeutic dose of radioactive iodine-131, the patient's age at the first manifestation of this disease , the presence of thyroid resection history dose merkazolila (tyrosol) the day before drug withdrawal, then determining upper and lower limit of the recommended individual dose of radioactive iodine ⁽¹³¹ I) for the Paci nt based on certain values of these indicators of the patient, while the determination of the lower limit of the range is carried out as a result of solving the problem of predicting the dose for cases of negative treatment results, the upper - for cases of positive results, assessing the reliability of a positive or negative outcome of treatment depending on the dose assigned to the patient solving the classification problem, in which the clinical database of prescribed doses is added to the input data, the result of treatment used as an output parameter, and the reliability estimate is presented as a range of numbers from 0 to 1, while the zero value of the range is interpreted as a negative outcome, a single value is interpreted as positive, then the calculated range of recommended doses is refined by enumerating doses in the vicinity of the boundaries with a given step for search for a dose that ensures confidence in the success of therapy 0.5 for the left border and 0.9 for the right border, as a result, the dose corresponding to the right border of the specified dose is selected as the recommended dose range of recommended doses.

A new technical result of the proposed method is to increase the accuracy of determining the individual dose in the treatment of patients with DTZ, which reduces the number of relapses in treatment to 9-10%.

This result is due to the examination of a larger number of patients and, based on this study, obtaining additional clinical and laboratory parameters of patients that affect the outcome of therapy, determining the most significant of all the obtained indicators and calculating the individual patient dose using the exact mathematical apparatus.

The proposed method involves examining a larger number of patients and, on the basis of this study, obtaining clinical and laboratory parameters of the patient that affect the outcome of therapy, determining the most significant of them and applying the values of clinical and laboratory parameters of the patient to calculate an individual dose using an exact mathematical apparatus, which based on the use of the obtained clinical database, including 373 patients with DTZ disease (Graves disease) . The average age is 48 years (39; 56), the duration of the disease is 5 years (3; 9), the thyroid gland is 27 ml (19.1; 41.0), the minimum thyroid gland is 3.6 ml, and the maximum is 185 5 ml. The therapeutic dose of iodine-131 was prescribed empirically and amounted to 350 MBq (250; 480), the minimum - 72 MBq, the maximum - 1180 MBq. 1 month after the start of treatment, hypothyroidism was observed in 64 patients, euthyroidism in 194, hyperthyroidism in 115. Hypothyroidism and euthyroidism were assessed as a positive result of treatment, hyperthyroidism as the absence of the expected result.

The mathematical apparatus used to calculate the individual dose in the proposed method includes solving a combination of classification and forecasting problems:

- the forecasting task is posed as follows: on the basis of the set of known values of the patient indicators set, presented in the form of a table, and the known outcome of therapy, construct a function (algorithm) for estimating the dose recommended for the patient based on the values of the patient indicators set and the outcome of therapy;

- the classification task is posed in a similar way with the difference that the outcome of the therapy is divided into a given set of classes of two elements (positive and negative), and consists in constructing a function (algorithm) for assessing the outcome of therapy for the patient and its reliability.

One of the most effective methods for solving the problems of forecasting and classification is the construction of hybrid systems of artificial intelligence (AI). The construction of such systems includes two main steps:

1) Identification of the structure and parameters of the system.

2) Parametric system optimization.

The identification of the system structure is carried out by clustering the available input data and includes: linguistic variables are selected (by the number of input parameters of the problem), their values (terms) are determined, and fuzzy production rules of the Takagi-Sugeno-Kang 0-order logical derivation system (TSK0) of the form :

- значения входных лингвистических переменных x_i, b_k - значения выходного параметра у. Нечеткие продукционные правила формируются с помощью алгоритма сферической кластеризации входных данных на основе модифицированной схемы конкурентного обучения. При этом каждому сформированному кластеру соответствует ровно одно правило.where K is the number of fuzzy rules,

are the values of the input linguistic variables x _i , b _k are the values of the output parameter y. Fuzzy production rules are formed using an algorithm for spherical clustering of input data based on a modified competitive training scheme. Moreover, each formed cluster corresponds to exactly one rule.

The parametric identification of the system consists in determining the parameters c _ik and a _{ik of the} Gaussian membership functions of the values of linguistic variables:

and parameters b _k :

в условиях продукционных правил, a x(t) и y(t) - значения входных и выходного параметра задачи.Membership functions determine the degree of validity of a statement

under the conditions of production rules, ax (t) and y (t) are the values of the input and output parameters of the problem.

The structure of a structured and parameterized system consists of a fuzzification step and a fuzzy production output step.

At the fuzzification step, the degree of confidence in the condition of each system inference rule is calculated. To determine the logical connective AND in the rules, the product operator (Larsen operator) is used:

At the output step, in accordance with the general output scheme of systems of type TSK0 for given values of the input vector x (0), the value of the output variable y (0) is calculated by the formula:

In the case of solving the classification problem, a set of such systems is formed, the output variable of each of which is identified with a separate class of the problem, and the values of b _k determine the degree of confidence in the element belonging to this class.

The parametric optimization of the system consists in clarifying the values of the parameters of membership functions (2) and the parameters b _k that make up the vector of parameters w in the parameter space W by solving the optimization problem, the essence of which is to minimize the empirical risk functional

, определенная на пространстве входных переменных X и пространстве параметров W, определяет выход описанной выше системы TSK0 при поступлении на ее вход конкретного элемента x(t) обучающей выборки D_N.where L (y (t), ϕ (x (t), w)) = (y (t) -ϕ (x (t), w)) ² is the quadratic loss function for a particular element (x (t), y (t)) of the training set D _N , the function ϕ: X × W → Y,

, defined on the space of input variables X and the parameter space W, determines the output of the TSK0 system described above when a specific element x (t) of the training sample D _N arrives at its input.

To solve this problem, the optimization algorithm for fuzzy systems TSK0 is used, based on a combination of the immune algorithm and the Particle Swarm Optimization method.

The implementation of the proposed method is as follows.

Initially, a larger number of patients are examined and, on the basis of this study, a set of the most significant clinical and laboratory parameters of the patient is obtained that affect the outcome of therapy (positive or negative). Then they solve the problem of predicting the dose value separately for each indicator, evaluate the accuracy of the forecast obtained and select indicators that provide the highest forecast accuracy. Based on the obtained results of solving the forecasting problem, a clinical data base is formed for the selected indicators, which is used as input to solve further forecasting and classification problems (see table). Next, a rough assessment of the upper and lower boundaries of the recommended dose range of ¹³¹ I is performed depending on the values of the patient’s clinical and laboratory parameters selected at the first stage obtained as a result of his examination. For this, the input data table formed at the first stage is divided into two parts, the first of which corresponds to positive treatment outcomes, and the second to negative ones. Dose prediction is carried out for both parts, with the result of dose prediction for negative outcomes being a lower estimate of the dose range, and for positive outcomes, the upper. After that, an assessment of the reliability of a positive or negative outcome of treatment is formed depending on the dose prescribed to the patient. To do this, solve the classification problem, in which a clinical database of prescribed doses is added to the input data, and the outcome of the therapy is used as an output parameter. The result of the stage is a classification system, which for each patient and any given dose provides an assessment of the outcome of therapy and a confidence value in it in the form of a number lying in the range from 0 to 1. At the same time, a zero value is interpreted as an exactly negative outcome, and a single value as definitely a positive outcome, a value of 0.5 - as "50 to 50".

To clarify the range of recommended doses calculated at the second stage by enumerating doses in the vicinity of coarse borders with a given step of 10 MBq to find the minimum dose that ensures confidence in the success of the therapy is greater than 0.5 for the left border and 0.9 for the right border, confidence values in the outcome of therapy for the tested dose determined using a classification system built in the third stage. As the recommended individual dose of ¹³¹ I, choose the dose corresponding to the right border of the specified range of recommended doses.

In addition, the constructed classification system for assessing the reliability of a positive or negative outcome of treatment depending on the dose prescribed to the patient allows, by enumerating doses of the entire allowable range, to build the efficiency curve of the assigned individual doses for the patient, i.e. dependence of the outcome of therapy in the form of confidence in its positive outcome in the range (0, 1) of the prescribed dose of ¹³¹ I.

As an example of the application of the calculation of individual doses of ¹³¹ I, a three-month forecasting period is considered. In this case, the available statistics database contained 294 records of treated patients. The following table shows the results of automatic determination of conditional coefficients of the relative importance of factors (input variables) on the outcome of therapy and selection of their optimal set before the first stage of calculating individual doses.

Clinical and laboratory parameters of the patient The coefficient of influence of the indicator A subset of indicators for constructing a classification hybrid system Classification accuracy
her hybrid system Therapeutic activity of ¹³¹ I, in MBq (dose) 1.4189 dose 0.82 Percentage of capture of a test dose of iodine-131 after 24 hours (3) 0.9183 dose, 3 0.75 The volume of the thyroid gland according to ultrasound, in ml (V _thyroid ) 0.7124 dose, 3, V _thyroid 0.79 Free blood thyroxin pmol / l (svT4) 0.7049 dose, 3, V _SHCHZH , svT4 0.83 Duration of DTZ disease, in years (dLDT) 0.6649 dose, 3, V _thyroid gland, svT4, for DDTZ 0.81 Patient gender (gender) 0.5467 dose, 3, V _thyroid gland, svT4, for DDTZ, gender 0.81 The duration of the cancellation of thyreostatics before taking a therapeutic dose of iodine-131, in days (marks) 0.5296 dose, 3, V _thyroid gland, svT4, dlDTZ, gender 0.86 Patient's age at the first manifestation of Graves disease, in years (debt DTZ) 0.5265 dose, 3, V _thyroid gland, svT4, dlDTZ, sex, otm, debDTZ 0.84 The percentage of capture test dose of iodine-131 after 4 hours (2) 0.5079 dose, 3, V _thyroid gland, svT4, dlDTZ, sex, otm, debDTZ, 2 0.88 A history of thyroid resection (O) 0.4054 dose, 3, V _thyroid gland, svT4, dlDTZ, sex, otm, debDTZ, 2, O 0.91 The dose of merkazolil (tyrosol) on the eve of drug withdrawal (CT) 0.4007 dose, 3, V _thyroid gland, svT4, dlDTZ, sex, otm, debDTZ, 2, O, ST 0.91

The values of the calculated influence coefficients confirm that the greatest influence on the outcome of therapy is exerted by such indicators as the value of the prescribed dose of radioactive iodine-131, as well as the capture of radioactive iodine-131 by the thyroid gland after 24 hours and the volume of the thyroid gland. At the same time, the results show that additional consideration of the influence of such indicators of the patient as free blood thyroxin, the duration of DTZ disease, the sex of the patient, the duration of the thyrostatic withdrawal before taking a therapeutic dose of iodine-131, the patient's age at the first manifestation of DTZ disease (in years), the percentage of capture of the test dose of iodine-131 after 4 hours, the presence of a thyroid resection in the anamnesis, allows to increase the accuracy of the classifying hybrid system from 79 to 91% and thereby determines the recommended for use in races ete individual dose of a set of indicators. In addition, in the process of selecting the optimal set of indicators, such indicators as the age and percentage of capture of the test dose of iodine-131 after 2 hours were automatically excluded from the analysis, since the calculated Pearson correlation coefficients for them with the values of debDTZ and 2, respectively, were greater than 0.8 and, with on the other hand, the latter indicators had a greater influence coefficient on the outcome of therapy.

To evaluate the diagnostic information parameters of the proposed method, a cross-validation method was used, at each iteration of which disjoint training and test samples of the initial statistics database were used. The training sample was used to build hybrid systems in accordance with the proposed four-stage method of calculating individual doses, and the test sample was used to calculate individual doses to patients of this sample. At each iteration, the accuracy parameters of the classifying hybrid system were calculated, since it is she who makes the conclusion about the outcome of the therapy based on the given values of the input parameters and dose. The parameters of sensitivity, specificity and effectiveness were also calculated. All of the indicated parameters of diagnostic information content were averaged over iterations of the cross-check and issued as the result for the proposed method for calculating individual doses. The following average values were obtained: accuracy - 91%, sensitivity - 90%, specificity - 81%, efficiency - 85%. The high accuracy value of the diagnostic method allows calculating the percentage of cases of relapses of the disease not more than 9-10%, which, in turn, allows to increase the effectiveness (reduce the number of relapses) of treating patients with DTZ with radioactive iodine compared with the currently used methods for calculating individual doses.

The specific application of the proposed method

The proposed method was determined by the individual dose of radioactive iodine-131 for the treatment of 25 patients with DTZ. The following treatment results were obtained within a three-month period: the number of positive treatment outcomes was 24, the number of negative treatment outcomes was 1. Thus, the total percentage of relapses in the three-month period was 4%, and the accuracy was 96%.

Example 1. Patient B., age 52 years. The indicators of the patient were as follows:

- the percentage of capture of the test dose of iodine-131 after 4 hours (2): 89%;

- the percentage of capture of the test dose of iodine-131 after 24 hours (3): 92%;

- thyroid volume (V _TG): 23.5 ml;

- free blood thyroxine (svT4): 18.03 pmol / l;

- the duration of the disease DTZ (dLTZ): 12 years;

- gender of the patient (gender): female;

- the duration of the cancellation of thyrostatics before taking a dose of iodine-131 (otm): 12 days;

- age of the patient at the first manifestation of DTZ (debtDTZ): 40 years;

- the presence of thyroid resection in the anamnesis (O): presence;

- the dose of merkazolil on the eve of drug withdrawal (CT): 15.

The dose of ¹³¹ I for the treatment of the patient, determined by the proposed method, was 288.2 MBq. The patient had a positive result of therapy (hypothyroidism) by the three-month period after treatment.

Example 2. Patient C., age 32 years. The indicators of the patient were as follows:

- the percentage of capture of the test dose of iodine-131 after 4 hours (2): 58%;

- the percentage of capture of the test dose of iodine-131 after 24 hours (3): 74%;

- thyroid volume (V _TG) 28 ml;

- free blood thyroxine (svT4): 5.15 pmol / l;

- the duration of the disease DTZ (dLTZ): 9 years;

- gender of the patient (gender): female;

- the duration of the cancellation of thyrostatics before taking a dose of iodine-131 (otm): 9 days;

- age of the patient at the first manifestation of DTZ (debtDTZ): 23 years;

- the presence of thyroid resection in the anamnesis (O): absence;

- the dose of merkazolil on the eve of drug withdrawal (CT): 15.

The dose of ¹³¹ I for the treatment of the patient, determined by the proposed method, was 387.2 MBq. The patient had a positive result of therapy (hypothyroidism) by the three-month period after treatment.

Example 3. Patient G., age 54 years. Patient performance was as follows:

- the percentage of capture of the test dose of iodine-131 after 4 hours (2): 49%;

- the percentage of capture of the test dose of iodine-131 after 24 hours (3): 76%;

- thyroid volume (V _TG) 26 ml;

- free blood thyroxine (svT4): 29.58 pmol / l;

- the duration of the disease DTZ (dLTZ): 2 years;

- gender of the patient (gender): male;

- the duration of the cancellation of thyrostatics before taking a dose of iodine-131 (otm): 6 days;

- age of the patient at the first manifestation of DTZ (debtDTZ): 52 years;

- the presence of thyroid resection in the anamnesis (O): absence;

- dose of merkazolil on the eve of drug withdrawal (CT): 20.

The dose of ¹³¹ I for the treatment of the patient, determined by the proposed method, amounted to 553.0 MBq. The patient showed a positive result of therapy (hypothyroidism) by a period of three months after treatment.

As can be seen from the obtained results, the proposed method for determining the individual dose of radioactive iodine-131 in the treatment of patients with DTZ has high accuracy, which allows to reduce the number of relapses during treatment to 9-10%.

INFORMATION SOURCES

1. Andrei Iagaru and I. Ross McDougall. Treatment of Thyrotoxicosis. J. Nucl. Med. 2007; 48: 379-389.

2. Alexander E., Larsen P.R. High dose of (131) I therapy for the treatment of hyperthyroidism caused by Graves ′ disease. J. Clin Endocrinol Metab. 2002; 87: 1073-1077.

Claims (1)

A method for determining an individual dose of radioactive iodine-131 in the treatment of patients with diffuse toxic goiter, including conducting a clinical examination of the patient with determining the patient’s indicators, selected on the basis of the available clinical database, which have a significant impact on the outcome of treatment (positive or negative) and ensure the most accurate outcome forecast treatment, such as: the percentage of capture of the test dose of iodine-131 after 4 hours, the percentage of capture of the test dose of iodine-131 after 24 hours, the volume of the thyroid gland according to ultrasound data, free blood thyroxin (pmol / L), duration of the disease, gender of the patient, duration of thyreostatic withdrawal before taking a therapeutic dose of radioactive iodine-131, patient age at the first manifestation of this disease, history of thyroid resection, dose of mercazolil ( tyrosol) on the eve of drug withdrawal, the subsequent determination of the upper and lower limits of the recommended individual dose of radioactive iodine (I ¹³¹ ) for this patient based on certain values of the indicated patient indicators, while determining the lower limit of the range is carried out as a result of solving the problem of predicting the dose for cases of negative treatment results, the upper one for cases of positive results, assessing the reliability of a positive or negative outcome of treatment depending on the dose assigned to the patient based on the solution of the classification problem, in which a database of clinical data on the prescribed doses is added to the input data, the treatment result is used as an output parameter, and the assessment is up to the characteristics are represented as a range of numbers from 0 to 1, while the zero value of the range is interpreted as a negative outcome, a single value as positive, then the calculated range of recommended doses is specified by sorting the doses in the vicinity of the borders with a given step to find the dose that ensures confidence in success 0.5 for the left border and 0.9 for the right border, as a result, the dose corresponding to the right border of the specified range of recommended doses is chosen as the recommended dose.