Patients undergoing total thyroidectomy require long-term thyroid-stimulating hormone (TSH) suppression therapy post-surgery. Among these, oral levothyroxine sodium tablets (L-T4) are one of the important maintenance treatment methods. To enhance treatment efficacy, the research team at the First Affiliated Hospital of Guangdong Pharmaceutical University utilized artificial intelligence (AI) technology to establish a digital model for L-T4 dosage and designed a precise medication plan based on this model, achieving a “customized” dosage. Coupled with 3D printing technology for precise dosing, this approach allows for a tailored administration of L-T4. Additionally, clinical pharmacists and the healthcare team conduct long-term follow-ups to continuously optimize the medication dosage, implementing pharmaceutical care and forming a “precision treatment ecological chain” to address the challenges of precise treatment for post-thyroid cancer patients.
AI Models Facilitate Customized L-T4 Dosage
By employing AI technology to construct a digital model for L-T4 dosage, a precise medication plan for post-thyroid cancer patients can be designed. To identify suitable AI technologies, the research team conducted a literature review, screening eight L-T4 dosage prediction models from a vast database, and validated these models using data from over 100 previous thyroid cancer patients undergoing L-T4 replacement therapy.
The results indicated that traditional multiple linear regression models have limitations such as inconsistent influencing factors and insufficient data correlation mining, leading clinical doctors to still rely on experience to determine L-T4 dosages for post-thyroid cancer patients.
To overcome these challenges, the research team innovatively introduced quantitative pharmacology theory and computer simulation technology, employing machine learning methods to reanalyze key indicators such as patient age, body mass index (BMI), pulse pressure, basal metabolic rate, heart rate, and TSH levels, ultimately determining that the adaptive boosting (Adaboost) regression algorithm was the optimal modeling method.
Compared to traditional models, the L-T4 dosage digital model constructed using the Adaboost regression algorithm can dynamically integrate multidimensional physiological indicators, facilitating a “customized” medication dosage for post-thyroid cancer patients.
3D Printing Technology for Tailored L-T4 Dosage
Currently, the common specifications for L-T4 in clinical settings are 50μg or 100μg per tablet, necessitating a fractional dosing approach to achieve precise administration and individualized dosage adjustments. However, traditional fractional dosing methods such as powdering and manual splitting are not only lacking in accuracy, safety, and compliance but also face issues such as drug contamination, poor adherence, and labor intensity, which can adversely affect treatment outcomes.
To address the challenges posed by traditional L-T4 fractional dosing methods, the research team, based on the previously established L-T4 dosage digital model, innovatively employed semi-solid extrusion (SSE) 3D printing technology to tailor L-T4 dosages.
3D printing technology is a process based on three-dimensional digital models, which combines materials using a layer-by-layer manufacturing approach according to discrete and stacking principles. This 3D printing technology can utilize excipients such as binders to prepare tablets with precise dosages, featuring characteristics such as accurate dosing, arbitrary dosages, good compliance, high mechanical integrity, and high recognizability.
The specific production steps are as follows: using commercially available L-T4 as raw material, optimizing the ratios of excipients such as hydroxypropyl methylcellulose (HPMC) and ethanol to ensure the drug’s formability, uniformity, and stability, thereby printing tablets that meet the preset dosage and shape. Test results show that the quality inspection items of the 3D printed tablets, including content uniformity, weight variation, dissolution, stability, and microbial limits, all meet the standards of the “Pharmacopoeia of the People’s Republic of China,” with content uniformity and weight variation significantly outperforming traditional fractional dosing methods, and the tablets being complete and aesthetically pleasing.
Building a “Precision Treatment Ecological Chain”
Clinical pharmacists, utilizing the L-T4 dosage digital model and 3D printing technology, collaborate with the healthcare team to conduct precise treatment for post-thyroid cancer patients.
The specific implementation pathway is as follows: first, the physician assesses and determines the necessity and treatment goals for L-T4 therapy; then, the clinical pharmacist obtains parameters affecting L-T4 dosage through consultations and uses the AI-constructed dosage digital model to design individualized medication dosages; subsequently, pharmacists at the precision medication research center produce the corresponding dosages of 3D printed tablets, which are then dispensed by the pharmacy to nurses or patients.
During the treatment period, clinical pharmacists provide medication monitoring and guidance to patients. Patients are required to undergo outpatient follow-ups 4 to 6 weeks post-surgery, and those whose thyroid function indicators do not meet standards will have their dosages optimized through model iteration by the clinical pharmacist until treatment goals are achieved, forming a “precision treatment ecological chain” that realizes a comprehensive and closed-loop precision treatment process.
The research team collected data from 37 post-thyroid cancer patients receiving precision treatment, showing that 56.76% of patients achieved standard TSH and other thyroid function indicators during the first follow-up at 4 to 6 weeks, a rate 2.2 times higher than that of previous experience-based dosing (27.84%); 86.49% of patients achieved standards after two dosage adjustments, with the average daily dosage of L-T4 (μg)/body weight (kg) ratio being 1.60, lower than the experience-based dosing ratio of 1.84.
Validated by real-world studies, the treatment model combining the L-T4 dosage digital model with 3D printing technology effectively helps patients shorten the time to reach standard levels, reduce the daily dosage required to achieve standards, and improve the achievement rate. Related research has obtained one utility model patent and one special project from the Guangdong Pharmaceutical Association for surgical pharmacy.
Written by: Pharmacy Department, First Affiliated Hospital of Guangdong Pharmaceutical University, General Surgery Department One
Edited by: Wu Fenggang
Reviewed by: Sun Meng
Let us know if you “like” this!
