Correction of Italian under-reporting in the first COVID-19 wave via age-specific deconvolution of hospital admissions

When the COVID-19 pandemic first emerged in early 2020, healthcare and bureaucratic systems worldwide were caught off guard and largely unprepared to deal with the scale and severity of the outbreak. In Italy, this led to a severe underreporting of infections during the first wave of the spread. The lack of accurate data is critical as it hampers the retrospective assessment of nonpharmacological interventions, the comparison with the following waves, and the estimation and validation of epidemiological models. In particular, during the first wave, reported cases of new infections were strikingly low if compared with their effects in terms of deaths, hospitalizations and intensive care admissions. In this paper, we observe that the hospital admissions during the second wave were very well explained by the convolution of the reported daily infections with an exponential kernel. By formulating the estimation of the actual infections during the first wave as an inverse problem, its solution by a regularization approach is proposed and validated. In this way, it was possible to compute corrected time series of daily infections for each age class. The new estimates are consistent with the serological survey published in June 2020 by the National Institute of Statistics (ISTAT) and can be used to speculate on the total number of infections occurring in Italy during 2020, which appears to be about double the number officially recorded.

Early detection of variants of concern via funnel plots of regional reproduction numbers

Early detection of the emergence of a new variant of concern (VoC) is essential to develop strategies that contain epidemic outbreaks. For example, knowing in which region a VoC starts spreading enables prompt actions to circumscribe the geographical area where the new variant can spread, by containing it locally. This paper presents 'funnel plots' as a statistical process control method that, unlike tools whose purpose is to identify rises of the reproduction number ([Formula: see text]), detects when a regional [Formula: see text] departs from the national average and thus represents an anomaly. The name of the method refers to the funnel-like shape of the scatter plot that the data take on. Control limits with prescribed false alarm rate are derived from the observation that regional [Formula: see text]'s are normally distributed with variance inversely proportional to the number of infectious cases. The method is validated on public COVID-19 data demonstrating its efficacy in the early detection of SARS-CoV-2 variants in India, South Africa, England, and Italy, as well as of a malfunctioning episode of the diagnostic infrastructure in England, during which the Immensa lab in Wolverhampton gave 43,000 incorrect negative tests relative to South West and West Midlands territories.

Just an artifact? The concordance between peer review and bibliometrics in economics and statistics in the Italian research assessment exercise

During the Italian research assessment exercise (2004–2010), the governmental agency (ANVUR) in charge of its realization performed an experiment on the concordance between peer review and bibliometrics at an individual article level. The computed concordances were at most weak for science, technology, engineering, and mathematics. The only exception was the moderate concordance found for the area of economics and statistics. In this paper, the disclosed raw data of the experiment are used to shed light on the anomalous results obtained for economics and statistics. In particular, the data permit us to document that the protocol of the experiment adopted for economics and statistics was different from the one used in the other areas. Indeed, in economics and statistics the same group of scholars developed the bibliometric ranking of journals for evaluating articles, managing peer reviews and forming the consensus groups for deciding the final scores of articles after having received the referee’s reports. This paper shows that the highest level of concordance in economics and statistics was an artifact mainly due to the role played by consensus groups in boosting the agreement between bibliometrics and peer review.

Modeling vaccination rollouts, SARS-CoV-2 variants and the requirement for non-pharmaceutical interventions in Italy

Despite progress in clinical care for patients with coronavirus disease 2019 (COVID-19)1, population-wide interventions are still crucial to manage the pandemic, which has been aggravated by the emergence of new, highly transmissible variants. In this study, we combined the SIDARTHE model2, which predicts the spread of SARS-CoV-2 infections, with a new data-based model that projects new cases onto casualties and healthcare system costs. Based on the Italian case study, we outline several scenarios: mass vaccination campaigns with different paces, different transmission rates due to new variants and different enforced countermeasures, including the alternation of opening and closure phases. Our results demonstrate that non-pharmaceutical interventions (NPIs) have a higher effect on the epidemic evolution than vaccination alone, advocating for the need to keep NPIs in place during the first phase of the vaccination campaign. Our model predicts that, from April 2021 to January 2022, in a scenario with no vaccine rollout and weak NPIs ([Formula: see text] = 1.27), as many as 298,000 deaths associated with COVID-19 could occur. However, fast vaccination rollouts could reduce mortality to as few as 51,000 deaths. Implementation of restrictive NPIs ([Formula: see text] = 0.9) could reduce COVID-19 deaths to 30,000 without vaccinating the population and to 18,000 with a fast rollout of vaccines. We also show that, if intermittent open-close strategies are adopted, implementing a closing phase first could reduce deaths (from 47,000 to 27,000 with slow vaccine rollout) and healthcare system costs, without substantive aggravation of socioeconomic losses.

Insulin Pulse Characteristics and Insulin Action in Non-diabetic Humans

OBJECTIVE

Pulsatile insulin secretion is impaired in diseases such as type 2 diabetes that are characterized by insulin resistance. This has led to the suggestion that changes in insulin pulsatility directly impair insulin signaling. We sought to examine the effects of pulse characteristics on insulin action in humans, hypothesizing that a decrease in pulse amplitude or frequency is associated with impaired hepatic insulin action.

METHODS

We studied 29 nondiabetic subjects on two occasions. On 1 occasion, hepatic and peripheral insulin action was measured using a euglycemic clamp. The deuterated water method was used to estimate the contribution of gluconeogenesis to endogenous glucose production. On a separate study day, we utilized nonparametric stochastic deconvolution of frequently sampled peripheral C-peptide concentrations during fasting to reconstruct portal insulin secretion. In addition to measuring basal and pulsatile insulin secretion, we used approximate entropy to measure orderliness and Fourier transform to measure the average, and the dispersion of, insulin pulse frequencies.

RESULTS

In univariate analysis, basal insulin secretion (R2 = 0.16) and insulin pulse amplitude (R2 = 0.09) correlated weakly with insulin-induced suppression of gluconeogenesis. However, after adjustment for age, sex, and weight, these associations were no longer significant. The other pulse characteristics also did not correlate with the ability of insulin to suppress endogenous glucose production (and gluconeogenesis) or to stimulate glucose disappearance.

CONCLUSIONS

Overall, our data demonstrate that insulin pulse characteristics, considered independently of other factors, do not correlate with measures of hepatic and peripheral insulin sensitivity in nondiabetic humans.

Diabetes-associated genetic variation in TCF7L2 alters pulsatile insulin secretion in humans

BACKGROUNDMetabolic disorders such as type 2 diabetes have been associated with a decrease in insulin pulse frequency and amplitude. We hypothesized that the T allele at rs7903146 in TCF7L2, previously associated with β cell dysfunction, would be associated with changes in these insulin pulse characteristics.METHODSTwenty-nine nondiabetic subjects (age 46 ± 2, BMI 28 ± 1 kg/m2) participated in this study. Of these, 16 were homozygous for the C allele at rs7903146 and 13 were homozygous for the T allele. Deconvolution of peripheral C-peptide concentrations allowed the reconstruction of portal insulin secretion over time. These data were used for subsequent analyses. Pulse orderliness was assessed by approximate entropy (ApEn), and the dispersion of insulin pulses was measured by a frequency dispersion index (FDI) after a Fast Fourier Transform (FFT) of individual insulin secretion rates.RESULTSDuring fasting conditions, the CC genotype group exhibited decreased pulse disorderliness compared with the TT genotype group (1.10 ± 0.03 vs. 1.19 ± 0.04, P = 0.03). FDI decreased in response to hyperglycemia in the CC genotype group, perhaps reflecting less entrainment of insulin secretion during fasting.CONCLUSIONDiabetes-associated variation in TCF7L2 is associated with decreased orderliness and pulse dispersion, unchanged by hyperglycemia. Quantification of ApEn and FDI could represent novel markers of β cell health.FUNDINGThis work was funded by US NIH (DK78646, DK116231), University of Padova research grant CPDA145405, and Mayo Clinic General Clinical Research Center (UL1 TR000135).

Citation gaming induced by bibliometric evaluation: A country-level comparative analysis

It is several years since national research evaluation systems around the globe started making use of quantitative indicators to measure the performance of researchers. Nevertheless, the effects on these systems on the behavior of the evaluated researchers are still largely unknown. For investigating this topic, we propose a new inwardness indicator able to gauge the degree of scientific self-referentiality of a country. Inwardness is defined as the proportion of citations coming from the country over the total number of citations gathered by the country. A comparative analysis of the trends for the G10 countries in the years 2000-2016 reveals a net increase of the Italian inwardness. Italy became, both globally and for a large majority of the research fields, the country with the highest inwardness and the lowest rate of international collaborations. The change in the Italian trend occurs in the years following the introduction in 2011 of national regulations in which key passages of professional careers are governed by bibliometric indicators. A most likely explanation of the peculiar Italian trend is a generalized strategic use of citations in the Italian scientific community, both in the form of strategic author self-citations and of citation clubs. We argue that the Italian case offers crucial insights on the constitutive effects of evaluation systems. As such, it could become a paradigmatic case in the debate about the use of indicators in science-policy contexts.

Systemic Exposure of Rituximab Increased by Ibrutinib: Pharmacokinetic Results and Modeling Based on the HELIOS Trial

INTRODUCTION

In the HELIOS trial, bendamustine/rituximab (BR) plus ibrutinib (BR-I) improved disease outcomes versus BR plus placebo in previously treated chronic lymphocytic leukemia/small lymphocytic lymphoma. Here, we describe the pharmacokinetic (PK) observations, along with modeling to further explore the interaction between ibrutinib and rituximab.

METHODS

578 subjects were randomized to ibrutinib or placebo with BR (6 cycles). Ibrutinib PK samples and tumor measurements were obtained from all subjects; a subset was evaluated for bendamustine and rituximab PK. Population rituximab PK was assessed using nonlinear mixed-effects modeling.

RESULTS

Dose-normalized plasma concentration-time bendamustine data were comparable between the arms. Systemic rituximab exposure was higher with BR-I versus BR; mean trough serum concentrations were 2- to 3-fold higher in the first three cycles and 1.2- to 1.7-fold higher subsequently. No relevant safety differences were observed. In the modeling, including treatment arm as a categorical covariate and tumor burden as a continuous time-varying covariate on overall rituximab clearance significantly improved fitting of the data.

CONCLUSIONS

BR-I led to higher dose-normalized systemic rituximab exposure versus BR and more rapid steady-state achievement. The modeling data suggest that rituximab disposition is, at least in part, target mediated. Determining the clinical significance of these findings requires further assessments.

TRIAL REGISTRATION

This study is registered at https://clinicaltrials.gov/ct2/show/NCT01611090 .

Assessment of pulsatile insulin secretion derived from peripheral plasma C-peptide concentrations by nonparametric stochastic deconvolution

The characteristics of pulsatile insulin secretion are important determinants of type 2 diabetes pathophysiology, but they are understudied due to the difficulties in measuring pulsatile insulin secretion noninvasively. Deconvolution of either peripheral C-peptide or insulin concentrations offers an appealing alternative to hepatic vein catheterization. However, to do so, there are a series of methodological challenges to overcome. C-peptide has a relatively long half-life and accumulates in the circulation. On the other hand, peripheral insulin concentrations reflect relatively fast clearance and hepatic extraction as it leaves the portal circulation to enter the systemic circulation. We propose a method based on nonparametric stochastic deconvolution of C-peptide concentrations, using individually determined C-peptide kinetics, to overcome these limitations. The use of C-peptide (instead of insulin) concentrations allows estimation of portal (and not post-hepatic) insulin pulses, whereas nonparametric stochastic deconvolution allows evaluation of pulsatile signals without any a priori assumptions of pulse shape and occurrence. The only assumption required is the degree of smoothness of the (unknown) secretion rate. We tested this method first on simulated data and then on 29 nondiabetic subjects studied during euglycemia and hyperglycemia and compared our estimates with the profiles obtained from hepatic vein insulin concentrations. This method produced satisfactory results both in the ability to fit the data and in providing reliable estimates of pulsatile secretion, in agreement with hepatic vein measurements. In conclusion, the proposed method enables reliable and noninvasive measurement of pulsatile insulin secretion. Future studies will be needed to validate this method in people with type 2 diabetes.

Model individualization for artificial pancreas

BACKGROUND AND OBJECTIVE

The inter-subject variability characterizing the patients affected by type 1 diabetes mellitus makes automatic blood glucose control very challenging. Different patients have different insulin responses, and a control law based on a non-individualized model could be ineffective. The definition of an individualized control law in the context of artificial pancreas is currently an open research topic. In this work we consider two novel identification approaches that can be used for individualizing linear glucose-insulin models to a specific patient.

METHODS

The first approach belongs to the class of black-box identification and is based on a novel kernel-based nonparametric approach, whereas the second is a gray-box identification technique which relies on a constrained optimization and requires to postulate a model structure as prior knowledge. The latter is derived from the linearization of the average nonlinear adult virtual patient of the UVA/Padova simulator. Model identification and validation are based on in silico data collected during simulations of clinical protocols designed to produce a sufficient signal excitation without compromising patient safety. The identified models are evaluated in terms of prediction performance by means of the coefficient of determination, fit, positive and negative max errors, and root mean square error.

RESULTS

Both identification approaches were used to identify a linear individualized glucose-insulin model for each adult virtual patient of the UVA/Padova simulator. The resulting model simulation performance is significantly improved with respect to the performance achieved by a linear average model.

CONCLUSIONS

The approaches proposed in this work have shown a good potential to identify glucose-insulin models for designing individualized control laws for artificial pancreas.

MPBPK-TMDD models for mAbs: alternative models, comparison, and identifiability issues

The aim of the present study was to evaluate model identifiability when minimal physiologically-based pharmacokinetic (mPBPK) models are integrated with target mediated drug disposition (TMDD) models in the tissue compartment. Three quasi-steady-state (QSS) approximations of TMDD dynamics were explored: on (a) antibody-target complex, (b) free target, and (c) free antibody concentrations in tissue. The effects of the QSS approximations were assessed via simulations, taking as reference the mPBPK-TMDD model with no simplifications. Approximation (a) did not affect model-derived concentrations, while with the inclusion of approximation (b) or (c), target concentration profiles alone, or both drug and target concentration profiles respectively deviated from the reference model profiles. A local sensitivity analysis was performed, highlighting the potential importance of sampling in the terminal pharmacokinetic phase and of collecting target concentration data. The a priori and a posteriori identifiability of the mPBPK-TMDD models were investigated under different experimental scenarios and designs. The reference model and QSS approximation (a) on antibody-target complex were both found to be a priori identifiable in all scenarios, while under the further inclusion of QSS approximation (b) target concentration data were needed for a priori identifiability to be preserved. The property could not be assessed for the model including all three QSS approximations. A posteriori identifiability issues were detected for all models, although improvement was observed when appropriate sampling and dose range were selected. In conclusion, this work provides a theoretical framework for the assessment of key properties of mathematical models before their experimental application. Attention should be paid when applying integrated mPBPK-TMDD models, as identifiability issues do exist, especially when rich study designs are not feasible.

Current mathematical models for cancer drug discovery

INTRODUCTION

Pharmacometric models represent the most comprehensive approaches for extracting, summarizing and integrating information obtained in the often sparse, limited, and less-than-optimally designed experiments performed in the early phases of oncology drug discovery. Whilst empirical methodologies may be enough for screening and ranking candidate drugs, modeling approaches are needed for optimizing and making economically viable the learn-confirm cycles within an oncology research program and anticipating the dose regimens to be investigated in the subsequent clinical development. Areas covered: Papers appearing in the literature of approximately the last decade reporting modeling approaches applicable to anticancer drug discovery have been listed and commented. Papers were selected based on the interest in the proposed methodology or in its application. Expert opinion: The number of modeling approaches used in the discovery of anticancer drugs is consistently increasing and new models are developed based on the current directions of research of new candidate drugs. These approaches have contributed to a better understanding of new oncological targets and have allowed for the exploitation of the relatively sparse information generated by preclinical experiments. In addition, they are used in translational approaches for guiding and supporting the choice of dosing regimens in early clinical development.

Temporal pharmacokinetic/pharmacodynamic interaction between human CD3ε antigen-targeted monoclonal antibody otelixizumab and CD3ε binding and expression in human peripheral blood mononuclear cell static culture

Otelixizumab is a monoclonal antibody (mAb) directed to human CD3ε, a protein forming part of the CD3/T-cell receptor (TCR) complex on T lymphocytes. This study investigated the temporal interaction between varying concentrations of otelixizumab, binding to human CD3 antigen, and expression of CD3/TCR complexes on lymphocytes in vitro, free from the confounding influence of changing lymphocyte frequencies observed in vivo. A static in vitro culture system was established in which primary human peripheral blood mononuclear cells (PBMCs) were incubated over an extended time course with titrated concentrations of otelixizumab. At each time point, free, bound, and total CD3/TCR expression on both CD4+ and CD8+ T cells and the amount of free otelixizumab antibody in the supernatant were measured. The pharmacokinetics of free otelixizumab in the culture supernatants was saturable, with a shorter apparent half-life at low concentration. Correspondingly, a rapid, otelixizumab concentration-, and time-dependent reduction in CD3/TCR expression was observed. These combined observations were consistent with the phenomenon known as target-mediated drug disposition (TMDD). A mechanistic, mathematical pharmacokinetic/pharmacodynamic (PK/PD) model was then used to characterize the free otelixizumab-CD3 expression-time relationship. CD3/TCR modulation induced by otelixizumab was found to be relatively fast compared with the re-expression rate of CD3/TCR complexes following otelixizumab removal from supernatants. In summary, the CD3/TCR receptor has been shown to have a major role in determining otelixizumab disposition. A mechanistic PK/PD model successfully captured the PK and PD in vitro data, confirming TMDD by otelixizumab.

Abstract B19: Population pharmacokinetic-pharmacodynamic (PKPD) modeling of ibrutinib in patients with B-cell malignancies

Ibrutinib (IBRU) is an oral Bruton's tyrosine kinase (BTK) inhibitor, approved by US FDA for the treatment of chronic lymphocytic leukemia (CLL/SLL) and mantle cell lymphoma (MCL) patients having received at least one prior therapy. A nonlinear mixed-effects population model was developed to describe the PK of IBRU in patients with B-Cell malignancies and to establish the effect of pathophysiological covariates on its PK behavior. The relationship between PK and BTK engagement in peripheral blood mononuclear cells (PBMC) was also explored. IBRU PK data (3477 observations in 245 patients) were available in patients with MCL, CLL/SLL and recurrent B-cell malignancies at dose levels from 1.25 to 12.5 mg/kg and at fixed doses from 420 to 840 mg once daily. An additional phase 2 study in 119 patients with MCL (772 observations) treated at 560 mg once daily was used to validate the PK model. BTK occupancy was assessed (694 observations in 127 patients) in PBMCs using a fluorescent affinity probe. Various models were tested on the data using the first-order conditional estimation method as implemented in NONMEM version 7.1.

A 2-compartment linear model with sequential zero-first order absorption and first order elimination was able to accommodate available PK data, including those of the validation dataset (prediction errors <15%). PK was dose- and time- independent. IBRU was rapidly absorbed, extensively distributed (volume of distribution at steady-state ~ 10,000 L) and cleared (apparent oral clearance ~1000 L/h). Relative bioavailability in the fasting state was about one third lower compared to the fed condition used in the clinical trials. No significant effect of other pathophysiological covariates on the PK was found (including sex, age or indication) except for body weight and coadministration of antacids, which had a marginal effect on the volume of distribution and duration of absorption, respectively. Analysis of PK-BTK engagement suggested that IBRU is a potent inhibitor of the BTK activity and that its interaction with BTK is rapid and durable.

Citation Format: Italo Poggesi, Maria Luisa Sardu, Eleonora Marostica, Juthamas Sukbuntherng, Betty Y. Chang, Jan de Jong, Xavier Woot de Trixhe, An Vermeulen, Giuseppe De Nicolao, Susan Mary O'Brien, John C Byrd, Ranjana H Advani, Danelle Frances James, William Deraedt, Darrin Beaupre, Michael Wang. Population pharmacokinetic-pharmacodynamic (PKPD) modeling of ibrutinib in patients with B-cell malignancies. [abstract]. In: Proceedings of the AACR Special Conference on Hematologic Malignancies: Translating Discoveries to Novel Therapies; Sep 20-23, 2014; Philadelphia, PA. Philadelphia (PA): AACR; Clin Cancer Res 2015;21(17 Suppl):Abstract nr B19.

Biomarker- versus drug-driven tumor growth inhibition models: an equivalence analysis

The mathematical modeling of tumor xenograft experiments following the dosing of antitumor drugs has received much attention in the last decade. Biomarker data can further provide useful insights on the pathological processes and be used for translational purposes in the early clinical development. Therefore, it is of particular interest the development of integrated pharmacokinetic-pharmacodynamic (PK-PD) models encompassing drug, biomarker and tumor-size data. This paper investigates the reciprocal consistency of three types of models: drug-to-tumor, such as established drug-driven tumor growth inhibition (TGI) models, drug-to-biomarker, e.g. indirect response models, and biomarker-to-tumor, e.g. the more recent biomarker-driven TGI models. In particular, this paper derives a mathematical relationship that guarantees the steady-state equivalence of the cascade of drug-to-biomarker and biomarker-to-tumor models with a drug-to-tumor TGI model. Using the Simeoni TGI model as a reference, conditions for steady-state equivalence are worked out and used to derive a new biomarker-driven model. Simulated and real data are used to show that in realistic cases the steady-state equivalence extends also to transient responses. The possibility of predicting the drug-to-tumor potency of a new candidate drug based only on biomarker response is discussed.

Population modelling of patient responses in antidepressant studies: a stochastic approach

This paper addresses the problem of modelling longitudinal data describing patients' responses in clinical trials. In particular, a systematic approach relying on a system theoretic paradigm is proposed to deal with contexts where limited physiopathological knowledge is available on disease, drug response, and patients' characteristics. The model relies on the notion of patient's health state which summarizes the patient's condition. In order to cope with the limited number of clinical data usually available, the paper considers a very parsimonious realization where the two state variables are the clinical endpoint and its derivative. Within a population framework, the individual response is modelled as the sum of an individual shift and the average response of subjects belonging to the same study, both described as Markovian processes and identified by empirical Bayes techniques. The proposed approach is validated with experimental data from a Phase II, flexible-dose, depression trial. The dose changes due to the flexible-dose scheme are handled as perturbations on the state. The connection between inter-individual variability and model stability is evaluated showing that the introduction of stable poles helps to describe populations whose range of individual responses does not diverge with time. In this way, good individual fittings and visual predictive checks were obtained for the clinical data. The proposed analysis provides a systematic approach to semi-mechanistic modelling when a precise knowledge of the physiological mechanisms of the disease is incomplete or missing.

Continuous-time Markov modelling of flexible-dose depression trials

The aim of this paper is to provide a systematic methodology for modelling longitudinal data to be used in contexts of limited or even absent knowledge of the physiological mechanism underlying the disease time course. Adopting a system-theoretic paradigm, a population response model is developed where the clinical endpoint is described as a function of the patient's health state. In particular, a continuous-time stochastic approach is proposed where the clinical score and its time-derivative summarize the patient's health state affected by a random term accounting for exogenous unpredictable factors. The proposed approach is validated on experimental data from the placebo and drug arms of a Phase II depression trial. Since some subjects in the trial may undergo changes in their treatment dose due to the flexible dosing scheme, dose escalations are modelled as instantaneous perturbations on the state. In its simplest form--an integrated Wiener process--was able to correctly capture the individual responses in both treatment arms. However, a better description of inter-individual variability was obtained by means of a stable Markovian model. Parameter estimation has been carried out according to the empirical Bayes method.

Multicenter closed-loop insulin delivery study points to challenges for keeping blood glucose in a safe range by a control algorithm in adults and adolescents with type 1 diabetes from various sites

BACKGROUND

The Control to Range Study was a multinational artificial pancreas study designed to assess the time spent in the hypo- and hyperglycemic ranges in adults and adolescents with type 1 diabetes while under closed-loop control. The controller attempted to keep the glucose ranges between 70 and 180 mg/dL. A set of prespecified metrics was used to measure safety.

RESEARCH DESIGN AND METHODS

We studied 53 individuals for approximately 22 h each during clinical research center admissions. Plasma glucose level was measured every 15-30 min (YSI clinical laboratory analyzer instrument [YSI, Inc., Yellow Springs, OH]). During the admission, subjects received three mixed meals (1 g of carbohydrate/kg of body weight; 100 g maximum) with meal announcement and automated insulin dosing by the controller.

RESULTS

For adults, the mean of subjects' mean glucose levels was 159 mg/dL, and mean percentage of values 71-180 mg/dL was 66% overall (59% daytime and 82% overnight). For adolescents, the mean of subjects' mean glucose levels was 166 mg/dL, and mean percentage of values in range was 62% overall (53% daytime and 82% overnight). Whereas prespecified criteria for safety were satisfied by both groups, they were met at the individual level in adults only for combined daytime/nighttime and for isolated nighttime. Two adults and six adolescents failed to meet the daytime criterion, largely because of postmeal hyperglycemia, and another adolescent failed to meet the nighttime criterion.

CONCLUSIONS

The control-to-range system performed as expected: faring better overnight than during the day and performing with variability between patients even after individualization based on patients' prior settings. The system had difficulty preventing postmeal excursions above target range.

Population model of longitudinal FEV1 data in asthmatics: meta-analysis and predictability of placebo response

Asthma is an obstructive lung disease where the mechanism of disease progression is not fully understood hence motivating the use of empirical models to describe the evolution of the patient's health state. With reference to placebo response, measured in terms of FEV1 (Forced Expiratory Volume in 1 s), a range of empirical models taken from the literature were compared at a single trial level. In particular, eleven GSK trials lasting 12 weeks in mild-to-moderate asthma were used for the modelling of longitudinal placebo responses. Then, the chosen exponential model was used to carry out an individual participant data meta-analysis on eleven trials. A covariate analysis was also performed to find relevant covariates in asthma to be accounted for in the meta-analysis model. Age, gender, and height were found statistically significant (e.g. the taller the patients the higher the FEV1, the older the patients the lower the FEV1, and females have lower FEV1). By truncating each trial at week 4, the predictive properties of the meta-analysis model were also investigated, showing its ability to predict long-term FEV1 response from truncated trials. Summarizing, the study suggests that: (i) the exponential model effectively describes the placebo response; (ii) the meta-analysis approach may prove helpful to simulate new trials as well as to reduce trial duration in view of its predictive properties; (iii) the inclusion of available covariates within the meta-analysis model provides a reduction of the inter-individual variability.

Modeling of human tumor xenografts and dose rationale in oncology

Xenograft models are commonly used in oncology drug development. Although there are discussions about their ability to generate meaningful data for the translation from animal to humans, it appears that better data quality and better design of the preclinical experiments, together with appropriate data analysis approaches could make these data more informative for clinical development. An approach based on mathematical modeling is necessary to derive experiment-independent parameters which can be linked with clinically relevant endpoints. Moreover, the inclusion of biomarkers as predictors of efficacy is a key step towards a more general mechanism-based strategy.

Day and night closed-loop control in adults with type 1 diabetes: a comparison of two closed-loop algorithms driving continuous subcutaneous insulin infusion versus patient self-management

OBJECTIVE

To compare two validated closed-loop (CL) algorithms versus patient self-control with CSII in terms of glycemic control.

RESEARCH DESIGN AND METHODS

This study was a multicenter, randomized, three-way crossover, open-label trial in 48 patients with type 1 diabetes mellitus for at least 6 months, treated with continuous subcutaneous insulin infusion. Blood glucose was controlled for 23 h by the algorithm of the Universities of Pavia and Padova with a Safety Supervision Module developed at the Universities of Virginia and California at Santa Barbara (international artificial pancreas [iAP]), by the algorithm of University of Cambridge (CAM), or by patients themselves in open loop (OL) during three hospital admissions including meals and exercise. The main analysis was on an intention-to-treat basis. Main outcome measures included time spent in target (glucose levels between 3.9 and 8.0 mmol/L or between 3.9 and 10.0 mmol/L after meals).

RESULTS

Time spent in the target range was similar in CL and OL: 62.6% for OL, 59.2% for iAP, and 58.3% for CAM. While mean glucose level was significantly lower in OL (7.19, 8.15, and 8.26 mmol/L, respectively) (overall P = 0.001), percentage of time spent in hypoglycemia (<3.9 mmol/L) was almost threefold reduced during CL (6.4%, 2.1%, and 2.0%) (overall P = 0.001) with less time ≤2.8 mmol/L (overall P = 0.038). There were no significant differences in outcomes between algorithms.

CONCLUSIONS

Both CAM and iAP algorithms provide safe glycemic control.

Artificial pancreas: model predictive control design from clinical experience

BACKGROUND

The objective of this research is to develop a new artificial pancreas that takes into account the experience accumulated during more than 5000 h of closed-loop control in several clinical research centers. The main objective is to reduce the mean glucose value without exacerbating hypo phenomena. Controller design and in silico testing were performed on a new virtual population of the University of Virginia/Padova simulator.

METHODS

A new sensor model was developed based on the Comparison of Two Artificial Pancreas Systems for Closed-Loop Blood Glucose Control versus Open-Loop Control in Patients with Type 1 Diabetes trial AP@home data. The Kalman filter incorporated in the controller has been tuned using plasma and pump insulin as well as plasma and continuous glucose monitoring measures collected in clinical research centers. New constraints describing clinical knowledge not incorporated in the simulator but very critical in real patients (e.g., pump shutoff) have been introduced. The proposed model predictive control (MPC) is characterized by a low computational burden and memory requirements, and it is ready for an embedded implementation.

RESULTS

The new MPC was tested with an intensive simulation study on the University of Virginia/Padova simulator equipped with a new virtual population. It was also used in some preliminary outpatient pilot trials. The obtained results are very promising in terms of mean glucose and number of patients in the critical zone of the control variability grid analysis.

CONCLUSIONS

The proposed MPC improves on the performance of a previous controller already tested in several experiments in the AP@home and JDRF projects. This algorithm complemented with a safety supervision module is a significant step toward deploying artificial pancreases into outpatient environments for extended periods of time.

Deconvolution-based assessment of pituitary GH secretion stimulated with GHRH+arginine in Prader-Willi adults and obese controls

OBJECTIVE

The assessment of GH deficiency in adult patients with Prader-Willi syndrome (PWS) has been previously assessed through the evaluation of quantitative parameters, such as the peak value of GH response to exogenous stimuli. A comprehensive description of the pattern of secretory response obtainable by deconvolution analysis is still lacking. The aim of our study was to characterize the time evolution of responses of PWS subjects compared with obese controls.

DESIGN AND SUBJECTS

GH responsiveness was measured following the combined administration of GHRH+arginine to 65 PWS adults (24 males, 41 females) aged 18-41·2 years, and 17 age-, gender- and body mass index-matched obese controls. PWS subjects were analysed considering the stratification on different genotypes.

MEASUREMENTS

GH response to GHRH+arginine was analysed in terms of peak values, standard area under the curves (AUCs), AUCs due to the stimulus, AUCs of the Instantaneous Secretion Rate signal and Secretion Response Analysis.

RESULTS

In terms of both peak values and AUC, GH responses were statistically different between PWS UPD15 and PWS DEL15 subjects as well as between PWS UPD15 and obese controls. PWS subjects showed a lower and a more delayed GH response compared with obese controls. Moreover, PWS UPD15 subjects had the most delayed GH response.

CONCLUSIONS

Our findings demonstrate that impaired GH secretion in PWS subjects compared with obese controls regards not only amplitude parameters such as peak value and AUC, but also the shape of the secretory response, which is more delayed, especially for UPD15 subjects.

A multi-task learning approach for the extraction of single-trial evoked potentials

Evoked potentials (EPs) are of great interest in neuroscience, but their measurement is difficult as they are embedded in background spontaneous electroencephalographic (EEG) activity which has a much larger amplitude. The widely used averaging technique requires the delivery of a large number of identical stimuli and yields only an "average" EP which does not allow the investigation of the possible variability of single-trial EPs. In the present paper, we propose the use of a multi-task learning method (MTL) for the simultaneous extraction of both the average and the N single-trial EPs from N recorded sweeps. The technique is developed within a Bayesian estimation framework and uses flexible stochastic models to describe the average response and the N shifts between the single-trial EPs and this average. Differently from other single-trial estimation approaches proposed in the literature, MTL can provide estimates of both the average and the N single-trial EPs in a single stage. In the present paper, MTL is successfully assessed on both synthetic (100 simulated recording sessions with N=20 sweeps) and real data (11 subjects with N=20 sweeps) relative to a cognitive task carried out for the investigation of the P300 component of the EP.

A PCA approach to population analysis: with application to a Phase II depression trial

For psychiatric diseases, established mechanistic models are lacking and alternative empirical mathematical structures are usually explored by a trial-and-error procedure. To address this problem, one of the most promising approaches is an automated model-free technique that extracts the model structure directly from the statistical properties of the data. In this paper, a linear-in-parameter modelling approach is developed based on principal component analysis (PCA). The model complexity, i.e. the number of components entering the PCA-based model, is selected by either cross-validation or Mallows' Cp criterion. This new approach has been validated on both simulated and clinical data taken from a Phase II depression trial. Simulated datasets are generated through three parametric models: Weibull, Inverse Bateman and Weibull-and-Linear. In particular, concerning simulated datasets, it is found that the PCA approach compares very favourably with some of the popular parametric models used for analyzing data collected during psychiatric trials. Furthermore, the proposed method performs well on the experimental data. This approach can be useful whenever a mechanistic modelling procedure cannot be pursued. Moreover, it could support subsequent semi-mechanistic model building.

Fully integrated artificial pancreas in type 1 diabetes: modular closed-loop glucose control maintains near normoglycemia

Integrated closed-loop control (CLC), combining continuous glucose monitoring (CGM) with insulin pump (continuous subcutaneous insulin infusion [CSII]), known as artificial pancreas, can help optimize glycemic control in diabetes. We present a fundamental modular concept for CLC design, illustrated by clinical studies involving 11 adolescents and 27 adults at the Universities of Virginia, Padova, and Montpellier. We tested two modular CLC constructs: standard control to range (sCTR), designed to augment pump plus CGM by preventing extreme glucose excursions; and enhanced control to range (eCTR), designed to truly optimize control within near normoglycemia of 3.9-10 mmol/L. The CLC system was fully integrated using automated data transfer CGM→algorithm→CSII. All studies used randomized crossover design comparing CSII versus CLC during identical 22-h hospitalizations including meals, overnight rest, and 30-min exercise. sCTR increased significantly the time in near normoglycemia from 61 to 74%, simultaneously reducing hypoglycemia 2.7-fold. eCTR improved mean blood glucose from 7.73 to 6.68 mmol/L without increasing hypoglycemia, achieved 97% in near normoglycemia and 77% in tight glycemic control, and reduced variability overnight. In conclusion, sCTR and eCTR represent sequential steps toward automated CLC, preventing extremes (sCTR) and further optimizing control (eCTR). This approach inspires compelling new concepts: modular assembly, sequential deployment, testing, and clinical acceptance of custom-built CLC systems tailored to individual patient needs.

Bayesian population approaches to the analysis of dose escalation studies

In dose escalation studies cohorts of subjects are given increasing doses of a candidate drug to assess safety and tolerability, pharmacokinetics and pharmacological response. The escalation is carried on until a predefined stopping limit is achieved, often identified by a pharmacokinetic endpoint such as peak plasma concentration or area under the plasma concentration-time profile. In the present work, the application of Bayesian methodologies to Phase I dose escalation studies is explored. A Bayesian population model is devised, which provides predictions of dose-response and dose-risk curves, both for individuals already enrolled in the trial and for a new, previously untested subject. Empirical and fully Bayesian estimation algorithms are worked out. Such methods provide equivalent performances on both experimental and simulated datasets. With respect to previous work, it is quantitatively proven not only that a more general and flexible model is identifiable, but also that such flexibility is needed in real scenarios.

Enhancing the accuracy of subcutaneous glucose sensors: a real-time deconvolution-based approach

Minimally invasive continuous glucose monitoring (CGM) sensors can greatly help diabetes management. Most of these sensors consist of a needle electrode, placed in the subcutaneous tissue, which measures an electrical current exploiting the glucose-oxidase principle. This current is then transformed to glucose levels after calibrating the sensor on the basis of one, or more, self-monitoring blood glucose (SMBG) samples. In this study, we design and test a real-time signal-enhancement module that, cascaded to the CGM device, improves the quality of its output by a proper postprocessing of the CGM signal. In fact, CGM sensors measure glucose in the interstitium rather than in the blood compartment. We show that this distortion can be compensated by means of a regularized deconvolution procedure relying on a linear regression model that can be updated whenever a pair of suitably sampled SMBG references is collected. Tests performed both on simulated and real data demonstrate a significant accuracy improvement of the CGM signal. Simulation studies also demonstrate the robustness of the method against departures from nominal conditions, such as temporal misplacement of the SMBG samples and uncertainty in the blood-to-interstitium glucose kinetic model. Thanks to its online capabilities, the proposed signal-enhancement algorithm can be used to improve the performance of CGM-based real-time systems such as the hypo/hyper glycemic alert generators or the artificial pancreas.

Neural network incorporating meal information improves accuracy of short-time prediction of glucose concentration

Diabetes mellitus is one of the most common chronic diseases, and a clinically important task in its management is the prevention of hypo/hyperglycemic events. This can be achieved by exploiting continuous glucose monitoring (CGM) devices and suitable short-term prediction algorithms able to infer future glycemia in real time. In the literature, several methods for short-time glucose prediction have been proposed, most of which do not exploit information on meals, and use past CGM readings only. In this paper, we propose an algorithm for short-time glucose prediction using past CGM sensor readings and information on carbohydrate intake. The predictor combines a neural network (NN) model and a first-order polynomial extrapolation algorithm, used in parallel to describe, respectively, the nonlinear and the linear components of glucose dynamics. Information on the glucose rate of appearance after a meal is described by a previously published physiological model. The method is assessed on 20 simulated datasets and on 9 real Abbott FreeStyle Navigator datasets, and its performance is successfully compared with that of a recently proposed NN glucose predictor. Results suggest that exploiting meal information improves the accuracy of short-time glucose prediction.

Modelling of the outcome of non-inferiority trials by integration of historical data

The approval and differentiation of new compounds in clinical development often demands non-inferiority trials, in which the test drug is compared against a reference treatment. However, non-inferiority trials impose major operational burden with serious ethical and scientific implications for the development of new medicines. Traditional approaches make limited use of historical information on placebo and neglect inter-trial variability, relying on the constancy assumption that the control-to-placebo effect size is maintained across trials. We propose a model-based approach that overcomes such limitations and may be used as a tool to explore differentiation during clinical development. Parameter distributions are introduced which reflect the heterogeneity of trials. The method is illustrated using data from impetigo trials. Based on simulation scenarios, this Bayesian technique yields a definitive, consistent increase in the statistical power over two accepted statistical methods, allowing lower sample size requirements for the assessment of non-inferiority.

Multinomial logistic functions in markov chain models of sleep architecture: internal and external validation and covariate analysis

Mixed-effect Markov chain models have been recently proposed to characterize the time course of transition probabilities between sleep stages in insomniac patients. The most recent one, based on multinomial logistic functions, was used as a base to develop a final model combining the strengths of the existing ones. This final model was validated on placebo data applying also new diagnostic methods and then used for the inclusion of potential age, gender, and BMI effects. Internal validation was performed through simplified posterior predictive check (sPPC), visual predictive check (VPC) for categorical data, and new visual methods based on stochastic simulation and estimation and called visual estimation check (VEC). External validation mainly relied on the evaluation of the objective function value and sPPC. Covariate effects were identified through stepwise covariate modeling within NONMEM VI. New model features were introduced in the model, providing significant sPPC improvements. Outcomes from VPC, VEC, and external validation were generally very good. Age, gender, and BMI were found to be statistically significant covariates, but their inclusion did not improve substantially the model's predictive performance. In summary, an improved model for sleep internal architecture has been developed and suitably validated in insomniac patients treated with placebo. Thereafter, covariate effects have been included into the final model.

Bayesian population modeling of phase I dose escalation studies: Gaussian process versus parametric approaches

The early stages of the drug development process are often characterized by a limited number of subjects participating the study and a limited number of measurements per individual that can be collected, mainly due to technical, ethical, and cost reasons. The so-called dose escalation studies, performed during phase I, usually involve about 40 subjects or less, and feature observations at no more than three (rarely four or five) dose levels-per-subject. Depending on the complexity of the underlying pharmacokinetics, simple linear models or nonlinear ones (e.g., power, E(max) models) may be appropriate to describe the relationship between the metrics of systemic exposure to the drug (C(max), AUC) and the administered dose. However, in such data-poor scenarios, formulating models based on parametric descriptions is generally hard, and may easily result in model misspecification. Hence, nonparametric or "model-free" solutions, borrowed from the machine learning field, are deemed appealing. We resort to Gaussian process theory to work out Bayesian posterior expectations of a population (a.k.a mixed-effects) regression problem, namely Population Smoothing Splines (PSS). We show that in seven experimental dose escalation studies, Population Smoothing Splines improve on three widely used parametric population methods. Superiority of the model-free technique is confirmed by a simulated benchmark: Population Smoothing Splines compare very favorably even with the true parametric model structure underlying the simulated data.

Multinational study of subcutaneous model-predictive closed-loop control in type 1 diabetes mellitus: summary of the results

BACKGROUND

In 2008-2009, the first multinational study was completed comparing closed-loop control (artificial pancreas) to state-of-the-art open-loop therapy in adults with type 1 diabetes mellitus (T1DM).

METHODS

The design of the control algorithm was done entirely in silico, i.e., using computer simulation experiments with N=300 synthetic "subjects" with T1DM instead of traditional animal trials. The clinical experiments recruited 20 adults with T1DM at the Universities of Virginia (11); Padova, Italy (6); and Montpellier, France (3). Open-loop and closed-loop admission was scheduled 3-4 weeks apart, continued for 22 h (14.5 h of which were in closed loop), and used a continuous glucose monitor and an insulin pump. The only difference between the two sessions was that insulin dosing was performed by the patient under a physician's supervision during open loop, whereas insulin dosing was performed by a control algorithm during closed loop.

RESULTS

In silico design resulted in rapid (less than 6 months compared to years of animal trials) and cost-effective system development, testing, and regulatory approvals in the United States, Italy, and France. In the clinic, compared to open-loop, closed-loop control reduced nocturnal hypoglycemia (blood glucose below 3.9 mmol/liter) from 23 to 5 episodes (p<.01) and increased the amount of time spent overnight within the target range (3.9 to 7.8 mmol/liter) from 64% to 78% (p=.03).

CONCLUSIONS

In silico experiments can be used as viable alternatives to animal trials for the preclinical testing of insulin treatment strategies. Compared to open-loop treatment under identical conditions, closed-loop control improves the overnight regulation of diabetes.

Modelling the genesis and treatment of cancer: the potential role of physiologically based pharmacodynamics

Physiologically based modelling of pharmacodynamics/toxicodynamics requires an a priori knowledge on the underlying mechanisms causing toxicity or causing the disease. In the context of cancer, the objective of the expert meeting was to discuss the molecular understanding of the disease, modelling approaches used so far to describe the process, preclinical models of cancer treatment and to evaluate modelling approaches developed based on improved knowledge. Molecular events in cancerogenesis can be detected using 'omics' technology, a tool applied in experimental carcinogenesis, but also for diagnostics and prognosis. The molecular understanding forms the basis for new drugs, for example targeting protein kinases specifically expressed in cancer. At present, empirical preclinical models of tumour growth are in great use as the development of physiological models is cost and resource intensive. Although a major challenge in PKPD modelling in oncology patients is the complexity of the system, based in part on preclinical models, successful models have been constructed describing the mechanism of action and providing a tool to establish levels of biomarker associated with efficacy and assisting in defining biologically effective dose range selection for first dose in man. To follow the concentration in the tumour compartment enables to link kinetics and dynamics. In order to obtain a reliable model of tumour growth dynamics and drug effects, specific aspects of the modelling of the concentration-effect relationship in cancer treatment that need to be accounted for include: the physiological/circadian rhythms of the cell cycle; the treatment with combinations and the need to optimally choose appropriate combinations of the multiple agents to study; and the schedule dependence of the response in the clinical situation.

Bayesian online multitask learning of Gaussian processes

Standard single-task kernel methods have recently been extended to the case of multitask learning in the context of regularization theory. There are experimental results, especially in biomedicine, showing the benefit of the multitask approach compared to the single-task one. However, a possible drawback is computational complexity. For instance, when regularization networks are used, complexity scales as the cube of the overall number of training data, which may be large when several tasks are involved. The aim of this paper is to derive an efficient computational scheme for an important class of multitask kernels. More precisely, a quadratic loss is assumed and each task consists of the sum of a common term and a task-specific one. Within a Bayesian setting, a recursive online algorithm is obtained, which updates both estimates and confidence intervals as new data become available. The algorithm is tested on two simulated problems and a real data set relative to xenobiotics administration in human patients.

Control to range for diabetes: functionality and modular architecture

BACKGROUND

Closed-loop control of type 1 diabetes is receiving increasing attention due to advancement in glucose sensor and insulin pump technology. Here the function and structure of a class of control algorithms designed to exert control to range, defined as insulin treatment optimizing glycemia within a predefined target range by preventing extreme glucose fluctuations, are studied.

METHODS

The main contribution of the article is definition of a modular architecture for control to range. Emphasis is on system specifications rather than algorithmic realization. The key system architecture elements are two interacting modules: range correction module, which assesses the risk for incipient hyper- or hypoglycemia and adjusts insulin rate accordingly, and safety supervision module, which assesses the risk for hypoglycemia and attenuates or discontinues insulin delivery when necessary. The novel engineering concept of range correction module is that algorithm action is relative to a nominal open-loop strategy-a predefined combination of basal rate and boluses believed to be optimal under nominal conditions.

RESULTS

A proof of concept of the feasibility of our control-to-range strategy is illustrated by using a prototypal implementation tested in silico on patient use cases. These functional and architectural distinctions provide several advantages, including (i) significant insulin delivery corrections are only made if relevant risks are detected; (ii) drawbacks of integral action are avoided, e.g., undershoots with consequent hypoglycemic risks; (iii) a simple linear model is sufficient and complex algorithmic constraints are replaced by safety supervision; and (iv) the nominal profile provides straightforward individualization for each patient.

CONCLUSIONS

We believe that the modular control-to-range system is the best approach to incremental development, regulatory approval, industrial deployment, and clinical acceptance of closed-loop control for diabetes.

Closed-loop artificial pancreas using subcutaneous glucose sensing and insulin delivery and a model predictive control algorithm: preliminary studies in Padova and Montpellier

New effort has been made to develop closed-loop glucose control, using subcutaneous (SC) glucose sensing and continuous subcutaneous insulin infusion (CSII) from a pump, and a control algorithm. An approach based on a model predictive control (MPC) algorithm has been utilized during closed-loop control in type 1 diabetes patients. Here we describe the preliminary clinical experience with this approach. Six type 1 diabetes patients (three in each of two clinical investigation centers in Padova and Montpellier), using CSII, aged 36 +/- 8 and 48 +/- 6 years, duration of diabetes 12 +/- 8 and 29 +/- 4 years, hemoglobin A1c 7.4% +/- 0.1% and 7.3% +/- 0.3%, body mass index 23.2 +/- 0.3 and 28.4 +/- 2.2 kg/m(2), respectively, were studied on two occasions during 22 h overnight hospital admissions 2-4 weeks apart. A Freestyle Navigator(R) continuous glucose monitor and an OmniPod insulin pump were applied in each trial. Admission 1 used open-loop control, while admission 2 employed closed-loop control using our MPC algorithm. In Padova, two out of three subjects showed better performance with the closed-loop system compared to open loop. Altogether, mean overnight plasma glucose (PG) levels were 134 versus 111 mg/dl during open loop versus closed loop, respectively. The percentage of time spent at PG > 140 mg/dl was 45% versus 12%, while postbreakfast mean PG was 165 versus 156 mg/dl during open loop versus closed loop, respectively. Also, in Montpellier, two patients out of three showed a better glucose control during closed-loop trials. Avoidance of nocturnal hypoglycemic excursions was a clear benefit during algorithm-guided insulin delivery in all cases. This preliminary set of studies demonstrates that closed-loop control based entirely on SC glucose sensing and insulin delivery is feasible and can be applied to improve glucose control in patients with type 1 diabetes, although the algorithm needs to be further improved to achieve better glycemic control.

Run-to-run tuning of model predictive control for type 1 diabetes subjects: in silico trial

BACKGROUND

The technological advancements in subcutaneous continuous glucose monitoring and insulin pump delivery systems have paved the way to clinical testing of artificial pancreas devices. The experience derived by clinical trials poses technological challenges to the automatic control expert, the most notable being the large interpatient and intrapatient variability and the inherent uncertainty of patient information.

METHODS

A new model predictive control (MPC) glucose control system is proposed. The starting point is an MPC algorithm applied in 20 type 1 diabetes mellitus (T1DM) subjects. Three main changes are introduced: individualization of the ARX model used for prediction; synthesis of the MPC law on top of the open-loop basal/bolus therapy; and a run-to-run approach for implementing day-by-day tuning of the algorithm. In order to individualize the ARX model, a sufficiently exciting insulin profile is imposed by splitting the premeal bolus into two smaller boluses (40% and 60%) injected 30 min before and 30 min after the meal.

RESULTS

The proposed algorithm was tested on 100 virtual subjects extracted from an in silico T1DM population. The trial simulates 44 consecutive days, during which the patient receives breakfast, lunch, and dinner each day. For 10 days, meals are multiplied by a random variable uniformly distributed in [0.5, 1.5], while insulin delivery is based on nominal meals. Moreover, for 10 days, either a linear increase or decrease of insulin sensitivity (+/-25% of nominal value) is introduced.

CONCLUSIONS

The ARX model identification procedure offers an automatic tool for patient model individualization. The run-to-run approach is an effective way to auto-tune the aggressiveness of the closed-loop control law, is robust to meal variation, and is also capable of adapting the regulator to slow parameter variations, e.g., on insulin sensitivity.

A minimal model of tumor growth inhibition

The preclinical development of antitumor drugs greatly benefits from the availability of models capable of predicting tumor growth as a function of the drug administration schedule. For being of practical use, such models should be simple enough to be identifiable from standard experiments conducted on animals. In the present paper, a stochastic model is derived from a set of minimal assumptions formulated at cellular level. Tumor cells are divided in two groups: proliferating and nonproliferating. The probability that a proliferating cell generates a new cell is a function of the tumor weight. The probability that a proliferating cell becomes nonproliferating is a function of the plasma drug concentration. The time-to-death of a nonproliferating cell is a random variable whose distribution reflects the nondeterministic delay between drug action and cell death. The evolution of the expected value of tumor weight obeys two differential equations (an ordinary and a partial differential one), whereas the variance is negligible. Therefore, the tumor growth dynamics can be well approximated by the deterministic evolution of its expected value. The tumor growth inhibition model, which is a lumped parameter model that in the last few years has been successfully applied to several antitumor drugs, is shown to be a special case of the minimal model presented here.

Diabetes: Models, Signals, and Control

The control of diabetes is an interdisciplinary endeavor, which includes a significant biomedical engineering component, with traditions of success beginning in the early 1960s. It began with modeling of the insulin-glucose system, and progressed to large-scale in silico experiments, and automated closed-loop control (artificial pancreas). Here, we follow these engineering efforts through the last, almost 50 years. We begin with the now classic minimal modeling approach and discuss a number of subsequent models, which have recently resulted in the first in silico simulation model accepted as substitute to animal trials in the quest for optimal diabetes control. We then review metabolic monitoring, with a particular emphasis on the new continuous glucose sensors, on the analyses of their time-series signals, and on the opportunities that they present for automation of diabetes control. Finally, we review control strategies that have been successfully employed in vivo or in silico, presenting a promise for the development of a future artificial pancreas and, in particular, discuss a modular architecture for building closed-loop control systems, including insulin delivery and patient safety supervision layers. We conclude with a brief discussion of the unique interactions between human physiology, behavioral events, engineering modeling and control relevant to diabetes.

A model-based approach to the in vitro evaluation of anticancer activity

PURPOSE

The use of in vitro screening tests for characterizing the activity of anticancer agents is a standard practice in oncology research and development. In these studies, human A2780 ovarian carcinoma cells cultured in plates are exposed to different concentrations of the compounds for different periods of time. Their anticancer activity is then quantified in terms of EC(50) comparing the number of metabolically active cells present in the treated and the control arms at specified time points. The major concern of this methodology is the observed dependency of the EC(50) on the experimental design in terms of duration of exposure. This dependency could affect the efficacy ranking of the compounds, causing possible biases especially in the screening phase, when compound selection is the primary purpose of the in vitro analysis. To overcome this problem, the applicability of a modeling approach to these in vitro studies was evaluated.

METHODS

The model, consisting of a system of ordinary differential equations, represents the growth of tumor cells using a few identifiable and biologically relevant parameters related to cell proliferation dynamics and drug action. In particular, the potency of the compounds can be measured by a unique and drug-specific parameter that is essentially independent of drug concentration and exposure time. Parameter values were estimated using weighted nonlinear least squares.

RESULTS

The model was able to adequately describe the growth of tumor cells at different experimental conditions. The approach was validated both on commercial drugs and discovery candidate compounds. In addition, from this model the relationship between EC(50) and the exposure time was derived in an analytic form.

CONCLUSIONS

The proposed approach provides a new tool for predicting and/or simulating cell responses to different treatments with useful indications for optimizing in vitro experimental designs. The estimated potency parameter values obtained from different compounds can be used for an immediate ranking of anticancer activity.

Evaluating the efficacy of closed-loop glucose regulation via control-variability grid analysis

BACKGROUND

Advancements in subcutaneous continuous glucose monitoring and subcutaneous insulin delivery are stimulating the development of a minimally invasive artificial pancreas that facilitates optimal glycemic regulation in diabetes. The key component of such a system is the blood glucose controller for which different design strategies have been investigated in the literature. In order to evaluate and compare the efficacy of the various algorithms, several performance indices have been proposed.

METHODS

A new tool-control-variability grid analysis (CVGA)-for measuring the quality of closed-loop glucose control on a group of subjects is introduced. It is a method for visualization of the extreme glucose excursions caused by a control algorithm in a group of subjects, with each subject presented by one data point for any given observation period. A numeric assessment of the overall level of glucose regulation in the population is given by the summary outcome of the CVGA.

RESULTS

It has been shown that CVGA has multiple uses: comparison of different patients over a given time period, of the same patient over different time periods, of different control laws, and of different tuning of the same controller on the same population.

CONCLUSIONS

Control-variability grid analysis provides a summary of the quality of glycemic regulation for a population of subjects and is complementary to measures such as area under the curve or low/high blood glucose indices, which characterize a single glucose trajectory for a single subject.

Model predictive control of type 1 diabetes: an in silico trial

BACKGROUND

The development of artificial pancreas has received a new impulse from recent technological advancements in subcutaneous continuous glucose monitoring and subcutaneous insulin pump delivery systems. However, the availability of innovative sensors and actuators, although essential, does not guarantee optimal glycemic regulation. Closed-loop control of blood glucose levels still poses technological challenges to the automatic control expert, most notable of which are the inevitable time delays between glucose sensing and insulin actuation.

METHODS

A new in silico model is exploited for both design and validation of a linear model predictive control (MPC) glucose control system. The starting point is a recently developed meal glucose-insulin model in health, which is modified to describe the metabolic dynamics of a person with type 1 diabetes mellitus. The population distribution of the model parameters originally obtained in healthy 204 patients is modified to describe diabetic patients. Individual models of virtual patients are extracted from this distribution. A discrete-time MPC is designed for all the virtual patients from a unique input-output-linearized approximation of the full model based on the average population values of the parameters. The in silico trial simulates 4 consecutive days, during which the patient receives breakfast, lunch, and dinner each day.

RESULTS

Provided that the regulator undergoes some individual tuning, satisfactory results are obtained even if the control design relies solely on the average patient model. Only the weight on the glucose concentration error needs to be tuned in a quite straightforward and intuitive way. The ability of the MPC to take advantage of meal announcement information is demonstrated. Imperfect knowledge of the amount of ingested glucose causes only marginal deterioration of performance. In general, MPC results in better regulation than proportional integral derivative, limiting significantly the oscillation of glucose levels.

CONCLUSIONS

The proposed in silico trial shows the potential of MPC for artificial pancreas design. The main features are a capability to consider meal announcement information, delay compensation, and simplicity of tuning and implementation.