Predicting time to castration resistance in hormone sensitive prostate cancer by a personalization algorithm based on a mechanistic model integrating patient data.

BACKGROUND: Prostate cancer (PCa) is a leading cause of cancer death of men worldwide. In hormone-sensitive prostate cancer (HSPC), androgen deprivation therapy (ADT) is widely used, but an eventual failure on ADT heralds the passage to the castration-resistant prostate cancer (CRPC) stage. Because predicting time to failure on ADT would allow improved planning of personal treatment strategy, we aimed to develop a predictive personalization algorithm for ADT efficacy in HSPC patients. METHODS: A mathematical mechanistic model for HSPC progression and treatment was developed based on the underlying disease dynamics (represented by prostate-specific antigen; PSA) as affected by ADT. Following fine-tuning by a dataset of ADT-treated HSPC patients, the model was embedded in an algorithm, which predicts the patient's time to biochemical failure (BF) based on clinical metrics obtained before or early in-treatment. RESULTS: The mechanistic model, including a tumor growth law with a dynamic power and an elaborate ADT-resistance mechanism, successfully retrieved individual time-courses of PSA (R(2) = 0.783). Using the personal Gleason score (GS) and PSA at diagnosis, as well as PSA dynamics from 6 months after ADT onset, and given the full ADT regimen, the personalization algorithm accurately predicted the individual time to BF of ADT in 90% of patients in the retrospective cohort (R(2) = 0.98). CONCLUSIONS: The algorithm we have developed, predicting biochemical failure based on routine clinical tests, could be especially useful for patients destined for short-lived ADT responses and quick progression to CRPC. Prospective studies must validate the utility of the algorithm for clinical decision-making.

Population pharmacokinetic meta-analysis of five beta-lactams antibiotics to support dosing regimens in dogs for surgical antimicrobial prophylaxis.

The Pharmacokinetic/Pharmacodynamic (PK/PD) relationship of antimicrobial drugs (AMD) for surgical prophylaxis has been poorly studied, hampering evidence-based decision making around AMD dosing and timing. Our objective is to use PK/PD principles to inform (1) the timing of administration and (2) the interval for re-administration of AMD used peri-operatively in dogs. Raw plasma concentrations of cefazolin, cefuroxime, cefalexin, amoxicillin and ampicillin were retrieved from original intravenous studies performed in dogs. E. coli and methicillin-susceptible staphylococci were identified as possible intraoperative contaminants and their epidemiological cut-offs (ECOFF) were retrieved from the EUCAST database. Individual PK data were refitted with non-linear mixed effect models (Phoenix®). We performed Monte Carlo simulation to compute i) the 95th percentile of time of peak concentration in the peripheral compartment (informing timing between administration and first incision) and ii) the duration for which at least 90% of dogs maintain a free plasma concentration above ECOFF (informing timing of re-administration: 1.5-4 h). Cefazolin (22-25 mg/kg), cefuroxime (20 mg/kg), cefalexin (15 mg/kg) and amoxicillin (16.7 mg/kg) reached peak peripheral concentrations within 30 min, but ampicillin (20 mg/kg) required 82 min, respectively. For methicillin-susceptible staphylococci, cefazolin and cefuroxime require re-administration every 2 h, whereas cefalexin and both amoxicillin and ampicillin can be readministered every 3 and 4 h, respectively. For E. coli, only cefazolin provided adequate perioperative coverage with 2-hourly administration, where cefuroxime and cefalexin failed uniformly. Alternatively, ampicillin and amoxicillin (critically ill dogs) may cover E. coli contaminations, but only if readministered every 1.5 h. These PK-derived conclusions provide a rationale for perioperative AMD administration timing.

Predicting [177Lu]Lu-HA-DOTATATE kidney and tumor accumulation based on [68Ga]Ga-HA-DOTATATE diagnostic imaging using semi-physiological population pharmacokinetic modeling.

BACKGROUND: Prediction of [177Lu]Lu-HA-DOTATATE kidney and tumor uptake based on diagnostic [68Ga]Ga-HA-DOTATATE imaging would be a crucial step for precision dosing of [177Lu]Lu-HA-DOTATATE. In this study, the population pharmacokinetic (PK) differences between [177Lu]Lu-HA-DOTATATE and [68Ga]Ga-HA-DOTATATE were assessed and subsequently [177Lu]Lu-HA-DOTATATE was predicted based on [68Ga]Ga-HA-DOTATATE imaging. METHODS: A semi-physiological nonlinear mixed-effects model was developed for [68Ga]Ga-HA-DOTATATE and [177Lu]Lu-HA-DOTATATE, including six compartments (representing blood, spleen, kidney, tumor lesions, other somatostatin receptor expressing organs and a lumped rest compartment). Model parameters were fixed based on a previously developed physiologically based pharmacokinetic model for [68Ga]Ga-HA-DOTATATE. For [177Lu]Lu-HA-DOTATATE, PK parameters were based on literature values or estimated based on scan data (four time points post-injection) from nine patients. Finally, individual [177Lu]Lu-HA-DOTATATE uptake into tumors and kidneys was predicted based on individual [68Ga]Ga-HA-DOTATATE scan data using Bayesian estimates. Predictions were evaluated compared to observed data using a relative prediction error (RPE) for both area under the curve (AUC) and absorbed dose. Lastly, to assess the predictive value of diagnostic imaging to predict therapeutic exposure, individual prediction RPEs (using Bayesian estimation) were compared to those from population predictions (using the population model). RESULTS: Population uptake rate parameters for spleen, kidney and tumors differed by a 0.29-fold (15% relative standard error (RSE)), 0.49-fold (15% RSE) and 1.43-fold (14% RSE), respectively, for [177Lu]Lu-HA-DOTATATE compared to [68Ga]Ga-HA-DOTATATE. Model predictions adequately described observed data in kidney and tumors for both peptides (based on visual inspection of goodness-of-fit plots). Individual predictions of tumor uptake were better (RPE AUC -40 to 28%) compared to kidney predictions (RPE AUC -53 to 41%). Absorbed dose predictions were less predictive for both tumor and kidneys (RPE tumor and kidney -51 to 44% and -58 to 82%, respectively). For most patients, [177Lu]Lu-HA-DOTATATE tumor accumulation predictions based on individual PK parameters estimated from diagnostic imaging outperformed predictions based on population parameters. CONCLUSION: Our semi-physiological PK model indicated clear differences in PK parameters for [68Ga]Ga-HA-DOTATATE and [177Lu]Lu-HA-DOTATATE. Diagnostic images provided additional information to individually predict [177Lu]Lu-HA-DOTATATE tumor uptake compared to using a population approach. In addition, individual predictions indicated that many aspects, apart from PK differences, play a part in predicting [177Lu]Lu-HA-DOTATATE distribution.

Developing Tools to Evaluate Non-linear Mixed Effect Models: 20 Years on the npde Adventure.

This article revisits 20 years of our work in developing evaluation tools adapted to non-linear mixed effect models. These hierarchical models involve a large number of assumptions concerning the structural evolution of the outcomes, the link between different outcomes, the variabilities in the parameters and model evaluation aims at assessing these various components, both to help guide the model building and to communicate on model adequacy for a given purpose. During our career, we have developed and extended simulation-based evaluation tools called normalised prediction discrepancies (npd) and normalised prediction distribution errors (npde), providing informative diagnostics through graphs and tests.

Population pharmacokinetic analysis of nanoparticle-bound and free camptothecin after administration of NLG207 in adults with advanced solid tumors.

PURPOSE: NLG207 (formerly CRLX101) is a nanoparticle-drug conjugate (NDC) of the potent topoisomerase I inhibitor, camptothecin (CPT). The present study sought to characterize the complex pharmacokinetics (PK) of NLG207 and better describe CPT release from nanoparticles using a population PK (popPK) model. METHODS: From 27 patients enrolled on two phase II clinical trials (NCT02769962 and NCT03531827), dense sampling was performed up to 48 h post-administration of NLG207 during cycle one and six of treatment; samples were also collected at ~ 360 h post-dose. Conjugated and free CPT concentrations were quantified from each sample, resulting in 477 observations to build a popPK model using non-linear mixed-effects modeling. RESULTS: The PK of NLG207 was characterized by combining two linear two-compartment models with first-order kinetics each to describe nanoparticle-bound (conjugated) and free CPT. Allometric scaling based on body weight provided the best body-size descriptor for all PK parameters. The typical volumes of distribution of the conjugated CPT central and free CPT central compartments were 3.16 L (BSV CV%; 18.1%) and 21.1 L (CV%; 79.8%), respectively. CPT release from the nanoparticle formulation was characterized via an initial rapid clearance of 5.71 L/h (CV%; 62.6%), which decreased via first-order decay (estimated half-life of 0.307 h) to the steady-state value of 0.0988 L/h (CV%; 33.5%) by ~ 4 h after end of infusion. Renal clearance of free CPT was 0.874 L/h (CV%; 42.2%). CONCLUSION: The popPK model confirmed nanoparticle behavior of conjugated CPT and mechanistically characterized CPT release from NLG207. The current analysis provides a strong foundation for future study as a potential predictive tool in ongoing NLG207 clinical trials.

Application of Item Response Theory to Model Disease Progression and Agomelatine Effect in Patients with Major Depressive Disorder.

INTRODUCTION: In this paper, we studied the effect over time of agomelatine, an antidepressant drug administered in patient with major depressive disorder, through item response theory (IRT), taking into account a strong placebo effect and missing not at random. We also assessed the informativeness of the HAMD-17 scale's item. MATERIALS AND METHODS: The data includes five phase III clinical trials sponsored by Servier Institute, totalling 1549 patients followed during a maximum of 1 year. At each observation, individual scores for the 17 items of the HAMD scale were recorded. The probability for each score was modelled with IRT. A non-linear mixed effects model was used to describe the evolution of the disease and was coupled with a time to event model to predict dropout. Clinical trial simulations were then used to compare placebo and active treatment. Informativeness of each item was evaluated using the Fisher information theory. RESULTS: The best model combined an IRT model, a longitudinal model for underlying depression which describes the remission and then a possible relapse, and a hazard model for dropout depending on the evolution from baseline. The drug effect was best modelled as an effect on the remission and the relapse phases. The median predicted drop in HAMD between baseline and 6 weeks was 8.8 (90% PI, 8.3-9.2) when on placebo and 13.1 (90% PI, 12.8-13.4) when treated. Nine items were found to be the most informative. CONCLUSION: The IRT framework allowed to characterise the evolution of depression with time and estimate the effect of agomelatine, as well as the link between symptoms and disease.