Toxicity Prediction Models Research Articles

A new method for predicting the acute toxicity of carbamate pesticides based on the perspective of binding information with carrier protein

Toxicity is one of the most important factors limiting the success of new drug development. In this paper, we built a fast and convenient new method (Carrier protein binding information-toxicity relationship, CPBITR) for predicting drug acute toxicity based on the perspective of binding information with carrier protein. First, we studied the binding information between carbamate pesticides and human serum albumin (HSA) through various spectroscopic methods and molecular docking. Then a total of 16 models were established to clarify the relationship between binding information with HSA and drug toxicity. The results showed that the binding information was related to toxicity. Finally we obtained the effective toxicity prediction model for carbamate pesticides. And the “Platform for Predicting Drug Toxicity Based on the Information of Binding with Carrier Protein” was established with the Back-propagation neural network model. We proposed and proved that it was feasible to predict drug toxicity from this new perspective: binding with carrier protein. According to this new perspective, toxicity prediction model of other drugs can also be established. This new method has the advantages of convenience and fast, and can be used to screen out low-toxic drugs quickly in the early stage. It is helpful for drug research and development.

Chemometric QSAR modeling of acute oral toxicity of Polycyclic Aromatic Hydrocarbons (PAHs) to rat using simple 2D descriptors and interspecies toxicity modeling with mouse

The information of the acute oral toxicity for most polycyclic aromatic hydrocarbons (PAHs) in mammals are lacking due to limited experimental resources, leading to a need to develop reliable in silico methods to evaluate the toxicity endpoint. In this study, we developed the quantitative structure-activity relationship (QSAR) models by genetic algorithm (GA) and multiple linear regression (MLR) for the rat acute oral toxicity (LD50) of PAHs following the strict validation principles of QSAR modeling recommended by OECD. The best QSAR model comprised eight simple 2D descriptors with definite physicochemical meaning, which showed that maximum atom-type electrotopological state, van der Waals surface area, mean atomic van der Waals volume, and total number of bonds are main influencing factors for the toxicity endpoint. A true external set (554 compounds) without rat acute oral toxicity values, and 22 limit test compounds, were firstly predicted along with reliability assessment. We also compared our proposed model with the OPERA predictions and recently published literature to prove the prediction reliability. Furthermore, the interspecies toxicity (iST) models of PAHs between rat and mouse were also established, validated and employed for filling data gap. Overall, our developed models should be applicable to new or untested or not yet synthesized PAHs falling within the applicability domain (AD) of the models for rapid acute oral toxicity prediction, thus being important for environmental or personal exposure risk assessment under regulatory frameworks.

Assessment of ecotoxicity of spent fluid catalytic cracking (FCC) refinery catalysts on Raphidocelis subcapitata and predictive models for toxicity

The 17 spent fluid catalytic cracking refinery catalysts (SFCCCs) from different petroleum refineries were collected and the leachates of SFCCCs were prepared. The ecotoxicity of SFCCC leachates to Raphidocelis subcapitata was assayed. The results showed that the toxicity of the 17 SFCCCs differ greatly. Ji SFCCC was the most toxic to R. subcapitata with a 96 h EC50 value of 1.38%, while Ha SFCCC was the least toxic, with the EC50 value was >100%. The relationships between the toxicity of SFCCCs and the metal concentrations in leachates were analyzed. The concentration of Ni (p = 0.001), La (p = 0.001), Mn (p = 0.014), Ce (p = 0.017), Co (p = 0.018), and Ca (p = 0.031) in leachates showed significant correlation with EC50 values. The predictive model for the ecotoxicity of SFCCCs were established with the concentrations of Ni and La in leachates as:ln(EC50) = 0.817 + exp(1.356 – 1.736 × CNi - 0.262 × CLa) (R2 = 0.926).The main toxic ingredients of SFCCC to microalgae were identified for the first time in this work. The results and predictive model of this study are significance for toxicity determination and management of SFCCCs.

Using quantitative ion character–activity relationship (QICAR) method in evaluation of metal toxicity toward wheat

It is important to assess the toxic effects posed by soil pollutants toward plants. However, plant toxicology experiments normally involve a considerable amount of manpower, consumables and time. Therefore, the use of metal toxicity prediction models, independent of toxicity tests, is critical. In this study, we investigated the toxicity of different metal ions to wheat using hydroponic experiments. We employed the methods of soft–hard ion grouping, soft–hard ligand theory and K (conditional binding constant based on the biotic ligand model principle) in combination with hydroponic experiments to explore the application of quantitative ion character-activity relationships in predicting phytotoxicity. The results showed that the toxicity of the 19 metal ions tested varied significantly, with EC50 ranging from 0.27 μM to 4463.36 μM. The linear regression relationships between the toxicity of these metal ions and their physicochemical properties were poor (R2 = 0.237–0.331, p < 0.05). These relationships were improved after grouping the metals according to the soft–hard theory (R2 = 0.527–0.744 and p < 0.05 for soft ions; R2 = 0.445–0.743 and p < 0.05 for hard ions). The application of soft–hard ligand theory, based on the binding affinity of the metals to the ligands, showed poor prediction of the phytotoxicity of metals, with R2 = 0.413 (p = 0.024) for the softness consensus scale (σCon) and R2 = 0.348 (p = 0.218) for the normalized hard ligands scale (HLScale). However, the method of K provided the closest fit in predicting toxicity (R2 = 0.803, p < 0.001). Our results showed that the application of soft–hard ion grouping and log K can improve prediction of the phytotoxicity of metals relatively well, which can potentially be used for deriving the toxicity of elements with limited toxicity data.

Predictive Toxicity Models for Chemically Related Substances: A Case Study with Nonionic Alcohol Ethoxylate Surfactant.

Predictive toxicity models, including interspecies correlation estimation (ICE) models, are reliable alternatives to animal toxicity testing. The ICE models describe mathematical relationships facilitating toxicity prediction from one surrogate test species to a species of unknown sensitivity. The ICE models were developed from curated aquatic toxicity data for 19 nonionic alcohol ethoxylate (AE) surfactants manufactured through the same process. Comparison of AE-ICE predictions with observed values from external validation datasets indicates a reasonable predictive accuracy. Model predictions were also closer to observed values than predictions from previously published ICE models for other substance groups. Comparison of acute fifth percentile hazard concentrations (HC5s) from species sensitivity distributions enhanced with AE-ICE predictions with chronic HC5s published elsewhere produced an acute-to-chronic ratio of 5, which was used to estimate chronic HC5s. With both acute and chronic HC5s for 14 AE substances, regressions were made relative to their respective liposome-water partitioning coefficients (log K lipw ), resulting in HC5-log K lipw relationships useful in estimating HC5s for AE substances with known composition but with limited or unavailable toxicity data. The findings from this case study further demonstrate that ICE models are viable alternatives to toxicity testing and could be used as weight of evidence in hazard assessment evaluations. Environ Toxicol Chem 2021;40:2073-2082. © 2021 SETAC.

Radiomics for prediction of radiation-induced lung injury and oncologic outcome after robotic stereotactic body radiotherapy of lung cancer: results from two independent institutions

ObjectivesTo generate and validate state-of-the-art radiomics models for prediction of radiation-induced lung injury and oncologic outcome in non-small cell lung cancer (NSCLC) patients treated with robotic stereotactic body radiation therapy (SBRT).MethodsRadiomics models were generated from the planning CT images of 110 patients with primary, inoperable stage I/IIa NSCLC who were treated with robotic SBRT using a risk-adapted fractionation scheme at the University Hospital Cologne (training cohort). In total, 199 uncorrelated radiomic features fulfilling the standards of the Image Biomarker Standardization Initiative (IBSI) were extracted from the outlined gross tumor volume (GTV). Regularized models (Coxnet and Gradient Boost) for the development of local lung fibrosis (LF), local tumor control (LC), disease-free survival (DFS) and overall survival (OS) were built from either clinical/ dosimetric variables, radiomics features or a combination thereof and validated in a comparable cohort of 71 patients treated by robotic SBRT at the Radiosurgery Center in Northern Germany (test cohort).ResultsOncologic outcome did not differ significantly between the two cohorts (OS at 36 months 56% vs. 43%, p = 0.065; median DFS 25 months vs. 23 months, p = 0.43; LC at 36 months 90% vs. 93%, p = 0.197). Local lung fibrosis developed in 33% vs. 35% of the patients (p = 0.75), all events were observed within 36 months. In the training cohort, radiomics models were able to predict OS, DFS and LC (concordance index 0.77–0.99, p < 0.005), but failed to generalize to the test cohort. In opposite, models for the development of lung fibrosis could be generated from both clinical/dosimetric factors and radiomic features or combinations thereof, which were both predictive in the training set (concordance index 0.71– 0.79, p < 0.005) and in the test set (concordance index 0.59–0.66, p < 0.05). The best performing model included 4 clinical/dosimetric variables (GTV-Dmean, PTV-D95%, Lung-D1ml, age) and 7 radiomic features (concordance index 0.66, p < 0.03).ConclusionDespite the obvious difficulties in generalizing predictive models for oncologic outcome and toxicity, this analysis shows that carefully designed radiomics models for prediction of local lung fibrosis after SBRT of early stage lung cancer perform well across different institutions.

A novel toxicity prediction model for hydrazine compounds based on 1D–3D molecular descriptors

The present study analyzed the lethal dose (LD50) of 73 hydrazine compounds in rats and established a comprehensive toxicity database of hydrazine compounds. Based on the existing toxicity database, the 1D–3D molecular descriptors of hydrazine compounds were obtained by quantum chemistry calculations and using DRAGON 6.0 software, and subsequently a quantitative structure–activity relationship prediction model was constructed. Aiming at the existing defects of the traditional neural network, such as easy to fall into local optimum and limited by the number of data samples, this paper proposed an improved BP neural network to simulate the toxicity prediction of hydrazine compounds. The determination coefficient and error coefficient of the prediction data set were found to be 0.864 and 0.679, respectively. The simulation results showed that the improved neural network model offered better prediction results for the toxicity of hydrazine compounds, which met the actual accuracy requirements for the toxicity prediction for hydrazine compounds.

Comparison of seven in silico tools for evaluating of daphnia and fish acute toxicity: case study on Chinese Priority Controlled Chemicals and new chemicals

BackgroundA number of predictive models for aquatic toxicity are available, however, the accuracy and extent of easy to use of these in silico tools in risk assessment still need further studied. This study evaluated the performance of seven in silico tools to daphnia and fish: ECOSAR, T.E.S.T., Danish QSAR Database, VEGA, KATE, Read Across and Trent Analysis. 37 Priority Controlled Chemicals in China (PCCs) and 92 New Chemicals (NCs) were used as validation dataset.ResultsIn the quantitative evaluation to PCCs with the criteria of 10-fold difference between experimental value and estimated value, the accuracies of VEGA is the highest among all of the models, both in prediction of daphnia and fish acute toxicity, with accuracies of 100% and 90% after considering AD, respectively. The performance of KATE, ECOSAR and T.E.S.T. is similar, with accuracies are slightly lower than VEGA. The accuracy of Danish Q.D. is the lowest among the above tools with which QSAR is the main mechanism. The performance of Read Across and Trent Analysis is lowest among all of the tested in silico tools. The predictive ability of models to NCs was lower than that of PCCs possibly because never appeared in training set of the models, and ECOSAR perform best than other in silico tools.ConclusionQSAR based in silico tools had the greater prediction accuracy than category approach (Read Across and Trent Analysis) in predicting the acute toxicity of daphnia and fish. Category approach (Read Across and Trent Analysis) requires expert knowledge to be utilized effectively. ECOSAR performs well in both PCCs and NCs, and the application shoud be promoted in both risk assessment and priority activities. We suggest that distribution of multiple data and water solubility should be considered when developing in silico models. Both more intelligent in silico tools and testing are necessary to identify hazards of Chemicals.

Toxicogenomic Profiling of 28 Nanomaterials in Mouse Airways.

Toxicogenomics opens novel opportunities for hazard assessment by utilizing computational methods to map molecular events and biological processes. In this study, the transcriptomic and immunopathological changes associated with airway exposure to a total of 28 engineered nanomaterials (ENM) are investigated. The ENM are selected to have different core (Ag, Au, TiO2, CuO, nanodiamond, and multiwalled carbon nanotubes) and surface chemistries (COOH, NH2, or polyethylene glycosylation (PEG)). Additionally, ENM with variations in either size (Au) or shape (TiO2) are included. Mice are exposed to 10 µg of ENM by oropharyngeal aspiration for 4 consecutive days, followed by extensive histological/cytological analyses and transcriptomic characterization of lung tissue. The results demonstrate that transcriptomic alterations are correlated with the inflammatory cell infiltrate in the lungs. Surface modification has varying effects on the airways with amination rendering the strongest inflammatory response, while PEGylation suppresses toxicity. However, toxicological responses are also dependent on ENM core chemistry. In addition to ENM‐specific transcriptional changes, a subset of 50 shared differentially expressed genes is also highlighted that cluster these ENM according to their toxicity. This study provides the largest in vivo data set currently available and as such provides valuable information to be utilized in developing predictive models for ENM toxicity.

In silico prediction of chemical respiratory toxicity via machine learning

Chemical respiratory toxicity usually causes serious harms to human body, so it is necessary to identify drugs or compounds with potential respiratory toxicity in early drug discovery stage. In this study, we collected 2,529 compounds from public databases and literature, and used six machine learning methods together with nine types of molecular fingerprints to construct a series of binary classification models for prediction of chemical respiratory toxicity. The accuracy of the best performing model was 0.869 for test set, and 0.933 for external validation set. Meanwhile, we defined the applicability domain of the models based on molecular similarity. We also identified the structural alerts about chemical respiratory toxicity through information gain and substructure frequency analysis, which could be used to elucidate their mechanisms and optimize the structures with less toxicity. Our study would be very helpful for prediction of chemical respiratory toxicity in early stage of drug discovery and environmental risk assessment.

Status Quo in Data Availability and Predictive Models of Nano-Mixture Toxicity.

Co-exposure of nanomaterials and chemicals can cause mixture toxicity effects to living organisms. Predictive models might help to reduce the intensive laboratory experiments required for determining the toxicity of the mixtures. Previously, concentration addition (CA), independent action (IA), and quantitative structure–activity relationship (QSAR)-based models were successfully applied to mixtures of organic chemicals. However, there were few studies concerning predictive models for toxicity of nano-mixtures before June 2020. Previous reviews provided comprehensive knowledge of computational models and mechanisms for chemical mixture toxicity. There is a gap in the reviewing of datasets and predictive models, which might cause obstacles in the toxicity assessment of nano-mixtures by using in silico approach. In this review, we collected 183 studies of nano-mixture toxicity and curated data to investigate the current data and model availability and gap and to derive research challenges to facilitate further experimental studies for data gap filling and the development of predictive models.

Utility of the Geriatric 8 for the Prediction of Therapy-Related Toxicity in Older Adults with Diffuse Large B-Cell Lymphoma.

The management of severe adverse events (AEs) is important in safely and effectively providing chemotherapy to older adults with diffuse large B-cell lymphoma (DLBCL). However, reports on simple and DLBCL-specific predictive models for treatment-related toxicity in elderly individuals are scarce. The aim of this study was to examine the usefulness of Geriatric 8 (G8) in predicting treatment-related severe AEs, nonhematological toxicity, and febrile neutropenia in older adults with DLBCL in real-world practice. We conducted a multicenter, retrospective study on 398 consecutive patients with DLBCL (aged ≥65 years) who received standard therapy at three centers in Japan (University of Fukui Hospital, the Fukui Prefectural Hospital, and the Japanese Red Cross Fukui Hospital), between 2007 and 2017. Multivariate logistic analysis demonstrated that the G8 score was an independent predictive factor for severe AEs. Moreover, a logistic regression model with restricted cubic spline showed a nonlinear association between the incidence of severe AEs and the G8 score. According to receiver operating characteristic analysis, the most discriminative cutoff value of the G8 for the incidence of severe AEs was 11, with an area under the curve value of 0.670. AEs occurred most often in the first course of chemotherapy and decreased as the course progressed. The G8 score, an easy-to-use geriatric assessment tool, can be a useful prediction model of treatment-related severe AEs during standard therapy in older adults with DLBCL. In older patients with diffuse large B-cell lymphoma (DLBCL), to accurately predict the risk of severe adverse events (AEs) in advance is essential for safe and effective treatment. This study demonstrated that the Geriatric 8 score, a simple and established geriatric assessment tool, indicated a high predictive ability for occurrence of therapy-related severe AEs in elderly patients with DLBCL who were treated with standard treatment.

Selection of low-toxic and highly efficient ionic liquids for the separation of palladium and platinum in acidic solution, and prediction of the metal affinity of ionic liquids

Separation studies were conducted to separate palladium (Pd) or platinum (Pt) from acidic solutions using 15 hydrophobic ionic liquids (ILs) capable of forming a binary phase with acidic aqueous solution. This study considered two cases. The first was to find high-efficient ILs capable of separating metals from a single solution, and the second was to find ILs capable of selectively separating Pt from an acidic mixture solution of Pd and Pt. In addition, we also considered toxic effect of ILs for greener separation process. For this studies, we experimentally examined the separation efficiencies of the ILs, and discuss the properties with their toxicities, theoretically predicted by previously presented toxicity prediction model, were compared. The results revealed that the extraction efficiency of ILs for Pd and Pt from acidic solutions depends on the types of IL and metal concentration. In the separation study on Pd in a single metal solution, trihexyltetradecylammonium bromide [P6,6,6,14]Br and trioctylmethylammonium choride [N8,8,8,1]Cl efficiently separates Pd from acidic solutions, while the separation on Pt in its single metal solution, 1-benzyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide[IM1-1Bz][(CF3SO2)2N], 1-methyl-3-dodecylimidazolim bis(trifluoromethylsulfonyl)imide[IM12][(CF3SO2)2N], 1-hexyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide[IM16][(CF3SO2)2N], [P6,6,6,14]Br and [N8,8,8,1]Cl led to high efficiencies. Moreover, in the study to selectively separate Pt from a mixed solution of Pd and Pt, [IM1-1Bz][(CF3SO2)2N] and [IM16][(CF3SO2)2N] have significant separation efficiency. When comparing the efficiencies and toxicities of the two ILs, [IM1-1Bz][(CF3SO2)2N] is more efficient, but it is more toxic than [IM16][(CF3SO2)2N]. Thus, for sustainable development, [IM16][(CF3SO2)2N] can be considered as a better option. Moreover, in order to understand the partitioning mechanisms of ILs we developed quantitative structure–activity relationship models with R2 values greater than 0.91.

Radiomics analysis of 3D dose distributions to predict toxicity of radiotherapy for lung cancer

Purpose(Chemo)–radiotherapy (RT) is the gold standard treatment for patients with locally advanced lung cancer non accessible for surgery. However, current toxicity prediction models rely on clinical and dose volume histograms (DVHs) and remain unsufficient. The goal of this work is to investigate the added predictive value of the radiomics approach applied to dose maps regarding acute and late toxicities in both the lungs and esophagus. MethodsAcute and late toxicities scored using the CTCAE v4.0 were retrospectively collected on patients treated with RT in our institution. Radiomic features were extracted from 3D dose maps considering Gy values as grey-levels in images. DVH and usual clinical factors were also considered. Three toxicity prediction models (clinical only, clinical + DVH and combined, i.e., including clinical + DVH + radiomics) were incrementally trained using a neural network on 70% of the patients for prediction of grade ≥2 acute and late pulmonary toxicities (APT/LPT) and grade ≥2 acute esophageal toxicities (AET). After bootstrapping (n = 1000), optimal cut-off values were determined based on the Youden Index. The trained models were then evaluated in the remaining 30% of patients using balanced accuracy (BAcc). Results167 patients were treated from 2015 to 2018: 78% non small-cell lung cancers, 14% small-cell lung cancers and 8% other histology with a median age at treatment of 66 years. Respectively, 22.2%, 16.8% and 30.0% experienced APT, LPT and AET. In the training set (n = 117), the corresponding BAcc for clinical only/clinical + DVH/combined were 0.68/0.79/0.92, 0.66/0.77/0.87 and 0.68/0.73/0.84. In the testing evaluation (n = 50), these trained models obtained a corresponding BAcc of 0.69/0.69/0.92, 0.76/0.80/0.89 and 0.58/0.73/0.72. ConclusionIn patients with a lung cancer treated with RT, radiomic features extracted from 3D dose maps seem to surpass usual models based on clinical factors and DVHs for the prediction of APT and LPT.

Identification of Pharmacogenomic Single Nucleotide Polymorphism Variants As Contributors to Toxicity Phenotype in the Treatment of Acute Childhood Leukemia

Introduction: Despite more than 80% long-term survival in acute lymphoblastic leukemia (ALL), morbidity due to drug-related toxicities remains high. Treatment interruptions and omissions from these toxicities may affect survivorship outcomes and morbidities in pediatric cancer. Although multiple factors contribute to a patient's risk of toxicity pharmacogenetic factors have been shown to play critical roles. Genetic variation within genes involved in pharmacokinetic and pharmacodynamic pathways of chemotherapies can influence gene expression and/or activity resulting in inter-patient variation in drug levels and thus toxicity risk or therapeutic efficacy. Identification of SNPs of clinical relevance that are predictive of toxicity can allow clinicians to optimize therapy to manage toxicity phenotypes. Objective: The objective of this cross-sectional study was to explore pharmacogenomic biomarkers associated with clinically relevant toxicity phenotypes in children receiving ALL therapy. Methods: The protocol was approved by the University of Florida (UF) Institutional Review Board (IRB201802623). All participants provided informed consent. Patients ≤ 26 years of age with a diagnosis of de novo or secondary ALL and who had received induction and consolidation chemotherapy after May 2012 at UF were eligible for participation. Chart review was performed and CTCAE-graded toxicity data was abstracted for gastrointestinal (GI), neurological, and endocrine toxicities along with prolonged hospitalization (&amp;gt; 4 days) due to febrile neutropenia. Genomic DNA was obtained from peripheral blood. SNPs in candidate pharmacological genes in cytarabine, vincristine, methotrexate, daunorubicin/doxorubicin, mercaptopurine/thioguanine pathways were selected through literature search and PharmGKB database. Genotyping was performed using Sequenom-based based chemistry. SNPs with low call rate and minor allele frequency of &amp;lt;0.10 were filtered and 105 SNPs were tested for association with each toxicity endpoint using logistic regression models with additive, dominant, and recessive modes of inheritance. Odds ratio (OR) and 95% confidence interval were calculated for each test. SNPs with association P-value &amp;lt; 0.05 were considered significant. Results: Prevalence of toxicity and top genes with significant SNPs are summarized in Figure 1. SNP in a carbonyl reductase (CBR1) involved in anthracycline metabolism was associated with higher risk of GI toxicity (OR 2.49, p=0.016). Multiple SNPs in cytarabine pathway genes were associated with increased risk of GI toxicities (rs12067645 in CTPS1, rs11853372 in SLC28A1 and rs10916819 in CDA) whereas SNP in an influx transporter SLCO1B1, implicated in methotrexate uptake was associated with lower risk of GI toxicity (rs2291075, OR 0.224, p =0.017). Six SNPs were associated with higher incidence of hematological toxicity with most interesting SNP being rs12067645 in CTPS1, a gene involved in cytarabine metabolism (OR 5.89, p=0.026). rs1128503 in drug efflux transporter ABCB1, rs2228110 in ALDH3A1, and rs2853539 in TYMS were all associated with lower incidence of neurologic toxicities. rs4673 in CYBA (OR 3.76, p=0.015) associated with higher neurological toxicity. Of the 5 SNPs associated with endocrinopathies, 4 were associated with increased and one with reduced incidence of toxicity. Interestingly, 2 SNPs in cytarabine uptake transporter SLC29A1, rs507964 (OR 2.89, p=0.02) and rs324148 (OR 1.89, p=0.042) and one missense SNP rs1051266 in SLC19A implicated in methotrexate influx was predictive of prolonged hospitalizations from febrile neutropenia. Figure 1 highlights some of these results. Conclusion: In conclusion, we identified common SNPs in genes associated with pharmacology of most commonly used anti-leukemic agents that were predictive of interpatient variability in incidence of drug toxicity phenotypes in a pilot cohort of 51 patients with ALL. Though limited by sample size, our studies demonstrate exciting results with ongoing enrollment and analysis showing promise in developing a SNP based model for prediction of toxicity in pediatric patients. Our goal is to integrate SNPs into a toxicity score for each patient which, once validated in collaborative multi-site cohorts, will hold value when transitioned to the clinic for personalizing treatment regimens to achieve therapeutic benefit and minimize morbidities. Disclosures No relevant conflicts of interest to declare.

Targeted Alpha Therapy in mCRPC (Metastatic Castration-Resistant Prostate Cancer) Patients: Predictive Dosimetry and Toxicity Modeling of 225Ac-PSMA (Prostate-Specific Membrane Antigen).

Radioligand therapy is a type of internal radiotherapy combining a short-range radioisotope labeled to a carrier with a high affinity for a specific receptor expressed on tumor cells. Targeted alpha therapy (TAT) combines a high-linear energy transfer (LET) emitter (225Ac) with a prostate-specific membrane antigen (PSMA) carrier, specifically binding tumor cells in patients with metastatic castration-resistant prostate cancer. Although the antitumor activity of 225Ac-PSMA is well-documented, this treatment is nowadays only used as salvage therapy because the high incidence of xerostomia limits the therapeutic window. Thus, methods to reduce salivary toxicity and models able to describe xerostomia incidence are needed. We recently studied the efficacy of salivary gland protectors administered in combination with 177Lu-PSMA therapy. Starting from these data, we performed a predictive dosimetric evaluation of 225Ac-PSMA to assess the impact of salivary gland protectors in TAT. 225Ac-PSMA predictive dosimetry was performed in 13 patients treated with 177Lu-PSMA. Sequential whole-body planar images were acquired 0.5–1, 16–24, 36–48, and 120 h post-injection. 177Lu time-activity curves were corrected for 225Ac physical decay and assumed in equilibrium for all daughters. The OLINDA/EXM spherical model was used for dose estimation of the parotid and submandibular glands. The dose for each daughter was calculated and summed for the total dose estimation. The biologically effective dose formalism was extended to high-LET emitters. For the total biologically effective dose formalism extended to high-LET emitters, including the contribution of all daughter isotopes, the brachytherapy formalism for a mixture of radionuclides was implemented. Equivalent doses in 2 Gy/fraction (EQD2) were then calculated and compared with the normal tissue complication probability model derived from external beam radiotherapy for grade ≥2 xerostomia induction. Median predictive doses were 0.86 BdRBE5/MBq for parotid glands and 1.05 BdRBE5/MBq for submandibular glands, with a 53% reduction compared with previously published data. The results show that the radiobiological model implemented is conservative, as it overestimates the complication rate with respect to the clinical data. Our data shows the possibility of reducing salivary gland uptake in TAT with the coadministration of organ protectors, but these results should be confirmed for TAT with 225Ac-PSMA by carrying out prospective trials with defined toxicity endpoints and dosimetry procedures.

External Validation of a Predictive Model for Acute Skin Radiation Toxicity in the REQUITE Breast Cohort.

Background: Acute skin toxicity is a common and usually transient side-effect of breast radiotherapy although, if sufficiently severe, it can affect breast cosmesis, aftercare costs and the patient's quality-of-life. The aim of this study was to develop predictive models for acute skin toxicity using published risk factors and externally validate the models in patients recruited into the prospective multi-center REQUITE (validating pREdictive models and biomarkers of radiotherapy toxicity to reduce side-effects and improve QUalITy of lifE in cancer survivors) study.Methods: Patient and treatment-related risk factors significantly associated with acute breast radiation toxicity on multivariate analysis were identified in the literature. These predictors were used to develop risk models for acute erythema and acute desquamation (skin loss) in three Radiogenomics Consortium cohorts of patients treated by breast-conserving surgery and whole breast external beam radiotherapy (n = 2,031). The models were externally validated in the REQUITE breast cancer cohort (n = 2,057).Results: The final risk model for acute erythema included BMI, breast size, hypo-fractionation, boost, tamoxifen use and smoking status. This model was validated in REQUITE with moderate discrimination (AUC 0.65), calibration and agreement between predicted and observed toxicity (Brier score 0.17). The risk model for acute desquamation, excluding the predictor tamoxifen use, failed to validate in the REQUITE cohort.Conclusions: While most published prediction research in the field has focused on model development, this study reports successful external validation of a predictive model using clinical risk factors for acute erythema following radiotherapy after breast-conserving surgery. This model retained discriminatory power but will benefit from further re-calibration. A similar model to predict acute desquamation failed to validate in the REQUITE cohort. Future improvements and more accurate predictions are expected through the addition of genetic markers and application of other modeling and machine learning techniques.

Assessing tumor centrality in lung stereotactic ablative body radiotherapy (SABR): the effects of variations in bronchial tree delineation and potential for automated methods

Accurate delineation of the proximal bronchial tree (PBT) is crucial for appropriate assessment of lung tumor centrality and choice of Stereotactic Ablative Body Radiotherapy (SABR) dose prescription. Here, we investigate variabilities in manual PBT delineation and their potential to influence assessing lesion centrality. A fully automatic, intensity-based tool for PBT contouring and measuring distance to the target is also described. This retrospective analysis included a total of 61 patients treated with lung SABR. A subset of 41 patients was used as a training dataset, containing clinical PBT contour and additional subsequently generated manual contours. The tool was optimized and compared against manual contours in terms of volume, distance to the target and various overlap/similarity metrics. The remaining 20 patients were used as a validation dataset to investigate the dosimetric effects of variations between manual and automatic PBT contours. Considerable interobserver variability was observed, particularly in identifying the superior and inferior borders of the PBT. Automatic PBT contours were comparable to manual contours with average Dice of 0.63 to 0.79 and mean distance to agreement of 1.78 to 3.34 mm. No significant differences in dosimetric parameters were found between automatically and manually generated contours. A moderate negative correlation was found between PBT maximum dose and distance to the lesion (p < 0.05). Variability in manual PBT delineation may result in inconsistent assessment of tumor centrality. Automatic contouring can help standardize clinical practice, support investigations into the link between SABR outcomes and lesion proximity to central airways and the development of predictive toxicity models that incorporate precise measurements of tumor location in relation to high-risk organs.

2-D chemical structure image-based in silico model to predict agonist activity for androgen receptor

BackgroundAbnormal activation of human nuclear hormone receptors disrupts endocrine systems and thereby affects human health. There have been machine learning-based models to predict androgen receptor agonist activity. However, the models were constructed based on limited numerical features such as molecular descriptors and fingerprints.ResultIn this study, instead of the numerical features, 2-D chemical structure images of compounds were used to build an androgen receptor toxicity prediction model. The images may provide unknown features that were not represented by conventional numerical features. As a result, the new strategy resulted in a construction of highly accurate prediction model: Mathews correlation coefficient (MCC) of 0.688, positive predictive value (PPV) of 0.933, sensitivity of 0.519, specificity of 0.998, and overall accuracy of 0.981 in 10-fold cross-validation. Validation on a test dataset showed MCC of 0.370, sensitivity of 0.211, specificity of 0.991, PPV of 0.882, and overall accuracy of 0.801. Our chemical image-based prediction model outperforms conventional models based on numerical features.ConclusionOur constructed prediction model successfully classified molecular images into androgen receptor agonists or inactive compounds. The result indicates that 2-D molecular mimetic diagram would be used as another feature to construct molecular activity prediction models.

Artificial intelligence and machine learning for protein toxicity prediction using proteomics data.

Instead of only focusing on the targeted drug delivery system, researchers have a great interest in developing peptide-based therapies for the procurement of numerous class of diseases. The main idea behind this is to anchor the properties of the receptor to design peptide-based therapeutics. As these macromolecules have distinct physicochemical properties over small molecules, it becomes an obligatory field for the treatment of diseases. For this, various in silico models have been developed to speculate the proteins by virtue of the application of machine learning and artificial intelligence. By analysing the properties and structural alert of toxic proteins, researchers aim to dissert some of the mechanisms of protein toxicity from which therapeutic insights may be drawn. Numerous models already exist worldwide emphasizing themselves as leading paramount for toxicity prediction in protein macromolecules. Few of them comparatively compete with the other predictive protein toxicity models and convincingly give a high-performance result in terms of accuracy. But their foundation is quite ambiguous, and varying approaches are found at the level of toxicoproteomic data utilization while building a machine learning model. In this review work, we present the contribution of artificial intelligence and machine learning approaches in prediction of protein toxicity using proteomics data.