Radio-Immune Response Model for Radiotherapy Plans with Heterogeneous Dose Distribution

PURPOSE/OBJECTIVE(S)

In order to model the immune response in tumor of the patients under radiotherapy for cancer. Characteristics of the numerical model and preliminary application are presented.

MATERIALS/METHODS

Immune response was modelled by 4 set of ordinary difference equations (ODE) as a function of biomedical variables including the amount of tumor antigen naturally released by tumor, the activation of cytotoxic T-lymphocytes (CTL) by radiation, immune suppression by tumor volume and the use of immunotherapy drugs. The effect of heterogeneous radiation dose distribution is also considered by hyperbolic tangent function to account for the immunogenic response of tissue under highly heterogeneous dose distribution intentionally modulated in spatially fractionated radiotherapy. Boundary behaviors of the model were investigated for tumors with different biomedical characteristics and under different treatment conditions. The developed model was applied to the tumor volume change in a mouse with 67NR tumors after radiation of 10 Gy to full or half volume of tumor and a clinical patient treated for sarcoma three times over 4 years.

RESULTS

Tumor growth is exponential at early phase, slow down over time with increasing immune response and eventually reaches an equilibrium condition (known as terminal tumor volume) for tumors with little to no immune suppression capability (ISC) even in the lack of radiation treatment. Breaking-through the equilibrium for tumor to grow exponentially happens when ISC is larger than the bifurcation threshold, analytically calculated from the proposed model. Tumor with ISC close to the bifurcation threshold can show complex growth behavior depending on the treatment condition and it should be carefully considered for the optimal treatment. Tumor volume change over 30 days period on mouse was modeled well with this model. Full dose irradiation reduced the tumor volume faster in the first 10 days but half volume irradiation reduced the tumor volume faster at later stage due to the improved immune response which cannot be modelled with traditional linear quadratic model. Tumor volume on a patient retreated three times over 4 years was also accurately estimated.

CONCLUSION

The proposed immune response model is capable to estimate tumor volume response under full or partial volume irradiation considering the complex immune characteristics of the tissue.

An open-source deep learning model for predicting effluent concentration in capacitive deionization

To effectively evaluate the performance of capacitive deionization (CDI), an electrochemical ion separation technology, it is necessary to accurately estimate the number of ions removed (effluent concentration) according to energy consumption. Herein, we propose and evaluate a deep learning model for predicting the effluent concentration of a CDI process. The developed deep learning model exhibited excellent prediction accuracy for both constant current and constant voltage modes (R² ≥ 0.968), and the accuracy increased with the data size. This model was based on the open-source language, Python, and the code has since been distributed with proper instructions for general use. Owing to the nature of the data-oriented deep learning model, the findings of this study are not only applicable to conventional CDI but also to various types of CDI (membrane CDI, flow CDI, faradaic CDI, etc.). Therefore, by referring to the examples shown in this study, we hope that this open-source deep learning code will be widely used in CDI research.

Automation of membrane capacitive deionization process using reinforcement learning

Capacitive deionization (CDI) is an alternative desalination technology that uses electrochemical ion separation. Although several attempts have been made to maximize the energy efficiency and productivity of CDI with conventional control methods, it is difficult to optimize the CDI processes because of the complex correlation between the operational conditions and the composition of feed water. To address these challenges, we applied deep reinforcement learning (DRL) to automatically control the membrane capacitive deionization (MCDI) process, which is one of the representative CDI processes, to accomplish high energy efficiency while desalinating water. In the DRL model, the numerical model is combined as the environment that provides states according to the actions. The feed water conditions, that is, the input state of the DRL, were assumed to have a random salt concentration and constant foulant concentration. The model was constructed to minimize energy consumption and maximize desalted water volume per cycle. After training of 1,000 episodes, the DRL model achieved a 22.07% reduction in specific energy consumption (from 0.054 to 0.042 kWh m^-3) and 11.60% increase in water desalted water volume per cycle (from 1.96×10^-5 to 2.19×10^-5 m³), achieving the desired degree of desalination, compared to the first episode. This improved performance was because the trained model selected the optimized operating conditions of current, voltage, and the number and intensity of flushing. Furthermore, it was possible to train the model depending on demand by modifying the reward function of the DRL model. The fundamental principle described in this study for applying the DRL model in MCDI operations can be the cornerstone of a fully automated water desalination process.

Deep reinforcement learning in an ultrafiltration system: Optimizing operating pressure and chemical cleaning conditions

Enhancing engineering efficiency and reducing operating costs are permanent subjects that face all engineers over the world. To effectively improve the performance of filtration systems, it is necessary to determine an optimal operating condition beyond conventional methods of periodic and empirical operation. Herein, this paper proposes an effective approach to finding an optimal operating strategy using deep reinforcement learning (DRL), particularly for an ultrafiltration (UF) system. Deep learning was developed to represent the UF system utilizing a long-short term memory and provided an environment for DRL. DRL was designed to control three actions; operating pressure, cleaning time, and cleaning concentration. Ultimately, DRL proposed the UF system to actively change the operating pressure and cleaning conditions over time toward better water productivity and operating efficiency. DRL denoted ∼20.9% of specific energy consumption can be reduced by increasing average water flux (39.5-43.7 L m^-2 h^-1) and reducing operating pressure (0.617-0.540 bar). Moreover, the optimal action of DRL was reasonable to achieve better performance beyond the conventional operation. Crucially, this study demonstrated that due to the nature of DRL, the approach is tractable for engineering systems that have structurally complex relationships among operating conditions and resultants.

Pharmaceutical removal at low energy consumption using membrane capacitive deionization

The performance of the membrane capacitive deionization (MCDI) system was evaluated during the removal of three selected pharmaceuticals, neutral acetaminophen (APAP), cationic atenolol (ATN), and anionic sulfamethoxazole (SMX), in batch experiments (feed solution: 2 mM NaCl and 0.01 mM of each pharmaceutical). Upon charging, the cationic ATN showed the highest removal rate of 97.65 ± 1.71%, followed by anionic SMX (93.22 ± 1.66%) and neutral APAP (68.08 ± 5.24%) due to the difference in electrostatic charge and hydrophobicity. The performance parameters (salt adsorption capacity, specific capacity, and cycling efficiency) and energy factors (specific energy consumption and recoverable energy) were further evaluated over ten consecutive cycles depending on the pharmaceutical addition. A significant decrease in the specific adsorption capacity (from 24.6 to ∼3 mg-NaCl g^-1) and specific capacity (from 17.6 to ∼2.5 mAh g^-1) were observed mainly due to the shortened charging and discharging time by pharmaceutical adsorption onto the electrode. This shortened charging time also led to an immediate drop in specific energy consumption from 0.41 to 0.04 Wh L^-1. Collectively, these findings suggest that MCDI can efficiently remove pharmaceuticals at a low energy demand; however, its performance changes dramatically as the pharmaceuticals are present in the target water.

Influence of natural organic matter on membrane capacitive deionization performance

Membrane capacitive deionization (MCDI) is a prospective desalination technology that removes ions using an electric potential difference across charged porous carbon electrodes. Natural organic matter (NOM) in feed water could influence the electrochemical process by leading to pore-blockages or forming a cake layer on the ion-exchange membrane-coated porous carbon electrode, thereby hindering ion removal. In this study, we explored the influence of different types of NOM, namely, humic acid (HA) and tannic acid (TA), on the MCDI desalination process for feed waters with inorganic salts (NaCl and CaCl₂). HA significantly interfered with the adsorption process and reduced the salt removal rate by up to 68% in the case of NaCl-based feed water. However, the influence of HA on salt removal in the case of CaCl₂-based feed water was marginal owing to the formation of a charge-neutralized complex, which was caused by the egg-box effect between Ca²⁺ and HA. TA reduced removal rates of salts (NaCl and CaCl₂) by 37% and 60%, respectively. This is because of the lower molecular weight and smaller hydrodynamic diameter of TA relative to that of HA, owing to which TA exhibits a stronger adhesion to the electrode pore structure. Furthermore, as TA substantially reduces MCDI performance with regard to the adsorption of inorganic salts, its presence in feed water results in higher electrical resistance and energy consumption.

P99 Rapid ventricular stimulation induces augmented conduction delay in Brugada syndrome patients

Background

The exact mechanism for Brugada Syndrome (BrS) is still not clear. There are two main physiologic hypotheses that have been suggested: the repolarization and the depolarization disorder models. Right ventricular (RV) activation delay was verified by echocardiography, conduction time in an explanted heart or in computer simulation. Verification of prolonged longitudinal activation time in human RV of only 5 patients of type-1 BrS and 5 controls was reported in 2008.

Methods

Bidirectional longitudinal activation times were assessed between RV outflow tract (RVot) and RV-apex (RVa) by stimulating and mapping RV endocardium in BrS patients. Conduction velocity was calculated considering ventricle size and distance between catheters.

Results

The studies were performed in controls (n = 18) and BrS patients (n = 6). There was no statistical difference in RP interval and QRS duration (PR 146 ± 21.7 vs 167 ± 45.2 ms, p = 0.325; QRS 102 ± 28.2 vs 122 ± 32.2 ms, p = 0.163). There was no difference of longitudinal activation time on stimulation at 500 ms (RVa to RVot: 63 ± 14.3 versus 80 ± 34.2 ms, p = 0.290; RVot to RVa: 50 ± 12.2 versus 76 ± 35.1 ms, p = 0.122). The BrS patients had longer longitudinal activation time on stimulation at 400 ms (RVa to RVot: 61 ± 15.2 versus 87 ± 28.7 ms, p = 0.009; RVot to RVa: 52 ± 11.1 versus 76 ± 35.3 ms, p = 0.029). The difference was not significant when isoproterenol was infused.

Conclusions

BrS patients display bidirectional longitudinal conduction delay when rapid stimulation. These findings support that BrS might be partly attributable to depolarization abnormality.

A comparison of two mathematical models of the impact of mass drug administration on the transmission and control of schistosomiasis

•

This paper compares two mathematical models describing the transmission dynamics of schistosome infection and the impact of mass drug administration.

•

The models differ structurally in a number of ways, including the dynamics of the intermediate snail host and the treatment of adult worms within the human host.

•

The models are validated against data taken from a mass-drug administration trial in Mozambique.

•

The differences between the model predictions and the data are discussed in the context of the structural differences between the models.

The predictions of two mathematical models describing the transmission dynamics of schistosome infection and the impact of mass drug administration are compared. The models differ in their description of the dynamics of the parasites within the host population and in their representation of the stages of the parasite lifecycle outside of the host. Key parameters are estimated from data collected in northern Mozambique from 2011 to 2015. This type of data set is valuable for model validation as treatment prior to the study was minimal. Predictions from both models are compared with each other and with epidemiological observations. Both models have difficulty matching both the intensity and prevalence of disease in the datasets and are only partially successful at predicting the impact of treatment. The models also differ from each other in their predictions, both quantitatively and qualitatively, of the long-term impact of 10 years’ school-based mass drug administration. We trace the dynamical differences back to basic assumptions about worm aggregation, force of infection and the dynamics of the parasite in the snail population in the two models and suggest data which could discriminate between them. We also discuss limitations with the datasets used and ways in which data collection could be improved.