Dominique Barbolosi

Background/Objectives: Radioactive iodine (RAI) therapy is widely used to treat metastatic differentiated thyroid cancer. To investigate physiological determinants of treatment response, a mechanistic model was developed, formulated as a system of coupled ordinary differential equations. Methods: The model captures the interactions between tumor burden, thyroglobulin (T_g) production and clearance, and radioactive iodine activity within a pharmacokinetic–pharmacodynamic framework. Model parameters were estimated using the Monte Carlo Stochastic Approximation Expectation–Maximization (MCMCSAEM) algorithm, based on clinical data from a cohort of 50 patients. Results: Tumor radiosensitivity (ρ) and initial tumor burden (N₀) consistently emerged as the most influential factors in both responder and non-responder groups classified by disease doubling time under RAI (T_d). A reduced model using only these two parameters preserved the principal response patterns of the full model. Other parameters influenced transient dynamics but had limited effect on overall T_g variance. Conclusions: These results support the use of a reduced calibration approach focused on ρ, N₀, and the effective doubling time T_d. The findings establish a theoretical foundation for developing tractable dynamic surrogates that reproduce the main treatment kinetics and support model-based clinical decision-making in RAI therapy.

Background: Immune checkpoint inhibitors (ICIs) are now a therapeutic standard in advanced non-small cell lung cancer (NSCLC), but strong predictive markers for ICIs efficacy are still lacking. We evaluated machine learning models built on simple clinical and biological data to individually predict response to ICIs. Methods: Patients with metastatic NSCLC who received ICI in second line or later were included. We collected clinical and hematological data and studied the association of this data with disease control rate (DCR), progression free survival (PFS) and overall survival (OS). Multiple machine learning (ML) algorithms were assessed for their ability to predict response. Results: Overall, 298 patients were enrolled. The overall response rate and DCR were 15.3% and 53%, respectively. Median PFS and OS were 3.3 and 11.4 months, respectively. In multivariable analysis, DCR was significantly associated with performance status (PS) and hemoglobin level (OR 0.58, p < 0.0001; OR 1.8, p < 0.001). These variables were also associated with PFS and OS and ranked top in random forest-based feature importance. Neutrophil-to-lymphocyte ratio was also associated with DCR, PFS and OS. The best ML algorithm was a random forest. It could predict DCR with satisfactory efficacy based on these three variables. Ten-fold cross-validated performances were: accuracy 0.68 ± 0.04, sensitivity 0.58 ± 0.08; specificity 0.78 ± 0.06; positive predictive value 0.70 ± 0.08; negative predictive value 0.68 ± 0.06; AUC 0.74 ± 0.03. Conclusion: Combination of simple clinical and biological data could accurately predict disease control rate at the individual level.

Introduction: Based on a mathematical model of trabectedin-induced neutropenia, we assessed the predictive value of absolute neutrophil count (ANC) on progression-free survival (PFS) in an independent validation cohort of patients treated with trabectedin. Methods: We collected data from 87 patients in two expert centers who received at least two cycles of trabectedin for soft tissue sarcomas (STS) treatment. Correlations between ANC, patients’ characteristics, and survival were assessed, and a multivariate model including tumor grade, performance status, ANC, and hemoglobin level was developed. Results: Therapeutic ANC ≥ 7.5 G/L level was associated with shorter PFS: 3.22 months (95% confidence interval (CI), 1.57–4.87) in patients with ANC ≥ 7.5 G/L vs. 5.78 months (95% CI, 3.95–7.61) in patients with ANC < 7.5 G/L (p = 0.009). Age, primary localization, lung metastases, dose reduction, hemoglobin, and albumin rates were also associated with PFS. In multivariate analysis, ANC ≥ 7.5 G/L was independently associated with poor PFS and overall survival. Conclusion: We validated increased pre-therapeutic ANC as a predictive factor of short PFS in patients starting trabectedin for STS. ANC appears to have an impact on survival rates and may be used as a decision-making tool for personalizing second-line strategies in patients with metastatic STS.

In FDG-PET, SUV images are hampered by large potential biases. Our aim was to develop an alternative method (ParaPET) to generate 3D kinetic parametric FDG-PET images easy to perform in clinical oncology. The key points of our method are the use of a new error model of PET measurement extracted from a late dynamic PET acquisition of 15 min, centered over the lesion and an image-derived input function (IDIF). The 15-min acquisition is reconstructed to obtain five images of FDG mean activity concentration and images of its variance to model errors of PET measurement. Our approach is carried out on each voxel to derive 3D kinetic parameter images. ParaPET was evaluated and compared to Patlak analysis as a reference. Hunter and Barbolosi methods (Barbolosi-Bl: with blood samples or Barbolosi-Im: with IDIF) were also investigated and compared to Patlak. Our evaluation was carried on Ki index, the net influx rate and its maximum value in the lesion (Ki,max). This parameter was obtained from 41 non-small cell lung cancer lesions associated with 4 to 5 blood samples per patient, required for the Patlak analysis. Compare to Patlak, the median relative difference and associated range (median; [min;max]) in Ki,max estimates were not statistically significant (Wilcoxon test) for ParaPET (− 3.0%; [− 31.9%; 47.3%]; p = 0.08) but statistically significant for Barbolosi-Bl (− 8.0%; [− 30.8%; 53.7%]; p = 0.001), Barbolosi-Im (− 7.9%; [− 38.4%; 30.6%]; p = 0.007) or Hunter (32.8%; [− 14.6%; 132.2%]; p < 10− 5). In the Bland-Altman plots, the ratios between the four methods and Patlak are not dependent of the Ki magnitude, except for Hunter. The 95% limits of agreement are comparable for ParaPET (34.7%), Barbolosi-Bl (30.1%) and Barbolosi-Im (30.8%), lower to Hunter (81.1%). In the 25 lesions imaged before and during the radio-chemotherapy, the decrease in the FDG uptake (ΔSUVmax or ΔKi,max) is statistically more important (p < 0.02, Wilcoxon one-tailed test) when estimated from the Ki images than from the SUV images (additional median variation of − 2.3% [− 52.6%; + 19.1%] for ΔKi,max compared to ΔSUVmax). None of the four methodologies is yet ready to replace the Patlak approach, and further improvements are still required. Nevertheless, ParaPET remains a promising approach, offering a non-invasive alternative to methods based on multiple blood samples and only requiring a late PET acquisition. It allows deriving Ki values, highly correlated and presenting the lowest relative bias with Patlak estimates, in comparison to the other methods we evaluated. Moreover, ParaPET gives access to quantitative information at the pixel level, which needs to be evaluated in the perspective of radiomic and tumour response. NCT 02821936 ; May 2016.

Brain metastases (BMs) are associated with poor prognosis in non-small cell lung cancer (NSCLC), but are only visible when large enough. Therapeutic decisions such as whole brain radiation therapy would benefit from patient-specific predictions of radiologically undetectable BMs. Here, we propose a mathematical modeling approach and use it to analyze clinical data of BM from NSCLC. Primary tumor growth was best described by a gompertzian model for the pre-diagnosis history, followed by a tumor growth inhibition model during treatment. Growth parameters were estimated only from the size at diagnosis and histology, but predicted plausible individual estimates of the tumor age (2.1-5.3 years). Multiple metastatic models were assessed from fitting either literature data of BM probability (n = 183 patients) or longitudinal measurements of visible BMs in two patients. Among the tested models, the one featuring dormancy was best able to describe the data. It predicted latency phases of 4.4 - 5.7 months and onset of BMs 14 - 19 months before diagnosis. This quantitative model paves the way for a computational tool of potential help during therapeutic management.

Succinate dehydrogenase subunit B and D (SDHB and SDHD) mutations represent the most frequent cause of hereditary pheochromocytoma and paraganglioma (PPGL). Although truncation of the succinate dehydrogenase complex is thought to be the disease causing mechanism in both disorders, SDHB and SDHD patients exihibit different phenotypes. These phenotypic differences are currently unexplained by molecular genetics. The aim of this study is to compare disease dynamics in these two conditions via a Markov chain model based on 4 clinically-defined steady states. Our model corroborates at the population level phenotypic observations in SDHB and SDHD carriers and suggests potential explanations associated with the probabilities of disease maintenance and regression. In SDHB-related syndrome, PPGL maintenance seems to be reduced compared to SDHD (p = 0.04 vs 0.95) due to higher probability of tumor cell regression in SDHB vs SDHD (p = 0.87 vs 0.00). However, when SDHB-tumors give rise to metastases, metastatic cells are able to thrive with decreased probability of regression compared with SDHD counterparts (p = 0.17 vs 0.89). By constrast, almost all SDHD patients develop PGL (mainly head and neck) that persist throughout their lifetime. However, compared to SDHB, maintenance of metastatic lesions seems to be less effective for SDHD (p = 0.83 vs 0.11). These findings align with data suggesting that SDHD-related PPGL require less genetic events for tumor initiation and maintenance compared to those related to SDHB, but fail to initiate biology that promotes metastatic spread and metastatic cell survival in host tissues. By contrast, the higher number of genetic abnormalities required for tumor initiation and maintenance in SDHB PPGL result in a lower penetrance of PGL, but when cells give rise to metastases they are assumed to be better adapted to sustain survival. These proposed differences in disease progression dynamics between SDHB and SDHD diseases provide new cues for future exploration of SDHx PPGL behavior, offering considerations for future specific therapeutic and prevention strategies.

Modeling therapeutic response to radioiodine in metastatic thyroid cancer: a proof-of-concept study for individualized medicine

Combining radiotherapy with immune checkpoint blockade may offer considerable therapeutic impact if the immunosuppressive nature of the tumor microenvironment (TME) can be relieved. In this study, we used mathematical models, which can illustrate the potential synergism between immune checkpoint inhibitors and radiotherapy. A discrete-time pharmacodynamic model of the combination of radiotherapy with inhibitors of the PD1–PDL1 axis and/or the CTLA4 pathway is described. This mathematical framework describes how a growing tumor first elicits and then inhibits an antitumor immune response. This antitumor immune response is described by a primary and a secondary (or memory) response. The primary immune response appears first and is inhibited by the PD1–PDL1 axis, whereas the secondary immune response happens next and is inhibited by the CTLA4 pathway. The effects of irradiation are described by a modified version of the linear-quadratic model. This modeling offers an explanation for the reported biphasic relationship between the size of a tumor and its immunogenicity, as measured by the abscopal effect (an off-target immune response). Furthermore, it explains why discontinuing immunotherapy may result in either tumor recurrence or a durably sustained response. Finally, it describes how synchronizing immunotherapy and radiotherapy can produce synergies. The ability of the model to forecast pharmacodynamic endpoints was validated retrospectively by checking that it could describe data from experimental studies, which investigated the combination of radiotherapy with immune checkpoint inhibitors. In summary, a model such as this could be further used as a simulation tool to facilitate decision making about optimal scheduling of immunotherapy with radiotherapy and perhaps other types of anticancer therapies. Cancer Res; 76(17); 4931–40. ©2016 AACR.

Positron emission tomography with 2-deoxy-2-[fluorine-18]fluoro- D-glucose integrated with computed tomography (18F-FDG PET/CT) is a recommended imaging procedure in the evaluation of non-small-cell lung cancers (NSCLC). Radiochemotherapy is a mainstay for treatment of locally advanced NSCLC, for which overall survival still remains poor. Early evaluation of treatment response may help in decision-making to complete radiotherapy or to switch to other treatment modalities. The present study aimed to evaluate the performance of a new metabolic parameters based on a SKA-S-derived mathematical method, as compared with standard SUV measurement during radiotherapy. Four patients treated with radiotherapy or radiochemotherapy for NSCLC were evaluated using 18F-FDG PET/CT during radiotherapy and after treatment completion. Whole body 18F-FDG PET/CT was performed followed by four additional list-mode acquisitions centered over the target lesion. Response was evaluated at four times (i.e., PET1 to PET4) by calculating standard SUV values and T80%, the time taken to reach 80% of 18F-FDG metabolized fraction using a SKA-S-derived mathematical method. Data from SUV and T80% calculations were found to be controversial. T80%, was found to be more predictive of clinical outcome. Although results from this pilot study should be further confirmed in a large prospective study, the data suggest that T80% is a promising metabolic biomarker for assessing early response to radiotherapy. In this proof of concept study we show that T80%, defined from a mathematic model taking into account the net influx rate constant and vascular volume could be consider as a promising biomarker as compared with SUVmax.

Metronomic oral vinorelbine is effective in metastatic NSCLC and malignant pleural mesothelioma, but all the studies published thus far were based upon a variety of empirical and possibly suboptimal schedules, with inconsistent results. Mathematical modelling showed by simulation that a new metronomic protocol could lead to a better safety and efficacy profile. This phase Ia/Ib trial was designed to confirm safety (phase Ia) and evaluate efficacy (phase Ib) of a new metronomic oral vinorelbine schedule. Patients with metastatic NSCLC or malignant pleural mesothelioma in whom standard treatments failed and who exhibited ECOG performance status 0–2 and adequate organ function will be eligible. Our mathematical PK-PD model suggested an alternative weekly D1, D2 and D4 schedule (named Vinorelbine Theoretical Protocol) with a respective dose of 60, 30 and 60 mg. Trial recruitment will be two-staged, as 12 patients are planned to participate in phase Ia to confirm safety and consolidate the calibration of the model parameters. Depending on the phase Ia results and after a favourable decision from a consultative committee, the extension phase (phase Ib) will be an efficacy study including 20 patients who will receive the Optimal Vinorelbine Theoretical Protocol. The primary endpoint is the tolerance (assessed by CTC v4.0) for the phase Ia and the objective response according to RECIST 1.1 for phase Ib. An ancillary study on circulating angiogenesis biomarkers will be a subproject of the trial. This ongoing trial is the first to prospectively test a mathematically optimized schedule in metronomic chemotherapy. As such, this trial can be considered as a proof-of-concept study demonstrating the feasibility to run a computational-driven protocol to ensure an optimal efficacy/toxicity balance in patients with cancer. EudraCT N°: 2015-000138-31