FIBER-ML, an Open-Source Supervised Machine Learning Tool for Quantification of Fibrosis in Tissue Sections.

Pathologic fibrosis is a major hallmark of tissue insult in many chronic diseases. Although the amount of fibrosis is recognized as a direct indicator of the extent of disease, there is no consentaneous method for its quantification in tissue sections. This study tested FIBER-ML, a semi-automated, open-source freeware that uses a machine-learning approach to quantify fibrosis automatically after a short user-controlled learning phase. Fibrosis was quantified in sirius red-stained tissue sections from two fibrogenic animal models: acute stress-induced cardiomyopathy in rats (Takotsubo syndrome-like) and HIV-induced nephropathy in mice (chronic kidney disease). The quantitative results of FIBER-ML software version 1.0 were compared with those of ImageJ in Takotsubo syndrome, and with those of inForm in chronic kidney disease. Intra- and inter-operator and inter-software correlation and agreement were assessed. All correlations were excellent (>0.95) in both data sets. The values of discriminatory power between the pathologic and healthy groups were <10-3 for data on Takotsubo syndrome and <10-4 for data on chronic kidney disease. Intra-operator agreement, assessed by intra-class coefficient correlation, was good (>0.8), while inter-operator and inter-software agreement ranged from moderate to good (>0.7). FIBER-ML performed in a fast and user-friendly manner, with reproducible and consistent quantification of fibrosis in tissue sections. It offers an open-source alternative to currently used software, including quality control and file management.

Machine Learning of Multi-Modal Tumor Imaging Reveals Trajectories of Response to Precision Treatment.

The standard assessment of response to cancer treatments is based on gross tumor characteristics, such as tumor size or glycolysis, which provide very indirect information about the effect of precision treatments on the pharmacological targets of tumors. Several advanced imaging modalities allow for the visualization of targeted tumor hallmarks. Descriptors extracted from these images can help establishing new classifications of precision treatment response. We propose a machine learning (ML) framework to analyze metabolic-anatomical-vascular imaging features from positron emission tomography, ultrafast Doppler, and computed tomography in a mouse model of paraganglioma undergoing anti-angiogenic treatment with sunitinib. Imaging features from the follow-up of sunitinib-treated (n = 8, imaged once-per-week/6-weeks) and sham-treated (n = 8, imaged once-per-week/3-weeks) mice groups were dimensionally reduced and analyzed with hierarchical clustering Analysis (HCA). The classes extracted from HCA were used with 10 ML classifiers to find a generalized tumor stage prediction model, which was validated with an independent dataset of sunitinib-treated mice. HCA provided three stages of treatment response that were validated using the best-performing ML classifier. The Gaussian naive Bayes classifier showed the best performance, with a training accuracy of 98.7 and an average area under curve of 100. Our results show that metabolic-anatomical-vascular markers allow defining treatment response trajectories that reflect the efficacy of an anti-angiogenic drug on the tumor target hallmark.

Molecular Analysis of the Superior Efficacy of a Dual Epidermal Growth Factor Receptor (EGFR)-DNA-Targeting Combi-Molecule in Comparison with Its Putative Prodrugs 6-Mono-Alkylamino- and 6,6-Dialkylaminoquinazoline in a Human Osteosarcoma Xenograft Model.

Background: ZR2002 is a dual EGFR-DNA-targeting combi-molecule that carries a chloroethyl group at the six-position of the quinazoline ring designed to alkylate DNA. Despite its good pharmacokinetics, ZR2002 is metabolized in vivo into dechlorinated metabolites, losing the DNA-alkylating function required to damage DNA. To increase the DNA damage activity in tumor cells in vivo, we compared ZR2002 with two of its 6-N,N-disubstituted analogs: "JS61", with a nitrogen mustard function at the six-position of the quinazoline ring, and "JS84", with an N-methyl group. Methods: Tumor xenografts were performed with the human Saos-2 osteosarcoma cell line expressing EGFR. Mice were treated with ZR2002, JS84 or JS61, and the tumor burden was measured with a caliper and CT/PET imaging. Drug metabolism was analyzed with LC-MS. EGFR and ɣ-H2AX phosphorylation were quantified via Western blot analysis and immunohistochemistry. Results: In vivo analysis showed that significant tumor growth inhibition was only achieved when ZR2002 was administered in its naked form. The metabolic dealkylation of JS61 and JS84 did not release sufficient concentrations of ZR2002 for the intratumoral inhibition of P-EGFR or enhanced levels of P-H2AX. Conclusions: The results in toto suggest that intratumoral concentrations of intact ZR2002 are correlated with the highest inhibition of P-EGFR and induction of DNA damage in vivo. ZR2002 may well represent a good drug candidate for the treatment of EGFR-expressing osteosarcoma.

The out-of-field dose in radiation therapy induces delayed tumorigenesis by senescence evasion.

A rare but severe complication of curative-intent radiation therapy is the induction of second primary cancers. These cancers preferentially develop not inside the planning target volume (PTV) but around, over several centimeters, after a latency period of 1-40 years. We show here that normal human or mouse dermal fibroblasts submitted to the out-of-field dose scattering at the margin of a PTV receiving a mimicked patient's treatment do not die but enter in a long-lived senescent state resulting from the accumulation of unrepaired DNA single-strand breaks, in the almost absence of double-strand breaks. Importantly, a few of these senescent cells systematically and spontaneously escape from the cell cycle arrest after a while to generate daughter cells harboring mutations and invasive capacities. These findings highlight single-strand break-induced senescence as the mechanism of second primary cancer initiation, with clinically relevant spatiotemporal specificities. Senescence being pharmacologically targetable, they open the avenue for second primary cancer prevention.

Sunitinib-induced cardiac hypertrophy and the endothelin axis.

Anti-angiogenics drugs in clinical use for cancer treatment induce cardiotoxic side effects. The endothelin axis is involved in hypertension and cardiac remodelling, and addition of an endothelin receptor antagonist to the anti-angiogenic sunitinib was shown to reduce cardiotoxicity of sunitinib in mice. Here, we explored further the antidote effect of the endothelin receptor antagonist macitentan in sunitinib-treated animals on cardiac remodeling. Methods: Tumor-bearing mice treated per os daily by sunitinib or vehicle were imaged before and after 1, 3 and 6 weeks of treatment by positron emission tomography using [18F]fluorodeoxyglucose and by echocardiography. Non-tumor-bearing animals were randomly assigned to be treated per os daily by vehicle or sunitinib or macitentan or sunitinib+macitentan, and imaged by echocardiography after 5 weeks. Hearts were harvested for histology and molecular analysis at the end of in vivo exploration. Results: Sunitinib treatment increases left ventricular mass and ejection fraction and induces cardiac fibrosis. Sunitinib also induces an early increase in cardiac uptake of [18F]fluorodeoxyglucose, which is significantly correlated with increased left ventricular mass at the end of treatment. Co-administration of macitentan prevents sunitinib-induced hypertension, increase in ejection fraction and cardiac fibrosis, but fails to prevent increase of the left ventricular mass. Conclusion: Early metabolic changes predict sunitinib-induced cardiac remodeling. Endothelin blockade can prevent some but not all cardiotoxic side-effects of sunitinib, in particular left ventricle hypertrophy that appears to be induced by sunitinib through an endothelin-independent mechanism.

Concurrent imaging of vascularization and metabolism in a mouse model of paraganglioma under anti-angiogenic treatment.

Rationale: Deregulation of metabolism and induction of vascularization are major hallmarks of cancer. Using a new multimodal preclinical imaging instrument, we explored a sequence of events leading to sunitinib-induced resistance in a murine model of paraganglioma (PGL) invalidated for the expression of succinate dehydrogenase subunit B (Sdhb-/-). Methods: Two groups of Sdhb-/- tumors bearing mice were treated with sunitinib (6 weeks) or vehicle (3 weeks). Concurrent Positron Emission Tomography (PET) with 2' -deoxy-2'-[18F]fluoro-D-glucose (FDG), Computed Tomography (CT) and Ultrafast Ultrasound Imaging (UUI) imaging sessions were performed once a week and ex vivo samples were analyzed by western blots and histology. Results: PET-CT-UUI enabled to detect a rapid growth of Sdhb-/- tumors with increased glycolysis and vascular development. Sunitinib treatment prevented tumor growth, vessel development and reduced FDG uptake at week 1 and 2 (W1-2). Thereafter, imaging revealed tumor escape from sunitinib treatment: FDG uptake in tumors increased at W3, followed by tumor growth and vessel development at W4-5. Perfused vessels were preferentially distributed in the hypermetabolic regions of the tumors and the perfused volume increased during escape from sunitinib treatment. Finally, initial changes in total lesion glycolysis and maximum vessel length at W1 were predictive of resistance to sunitinib. Conclusion: These results demonstrate an adaptive resistance of Sdhb-/- tumors to six weeks of sunitinib treatment. Early metabolic changes and delayed vessel architecture changes were detectable and predictable in vivo early during anti-angiogenic treatment. Simultaneous metabolic, anatomical and functional imaging can monitor precisely the effects of anti-angiogenic treatment of tumors.

Performance evaluation of the PET component of a hybrid PET/CT-ultrafast ultrasound imaging instrument.

We recently introduced a hybrid imaging instrument, PETRUS, based on a combination of positron emission tomography (PET) for molecular imaging, x-ray computed tomography (CT) for anatomical imaging, co-registration and attenuation correction, and ultrafast ultrasound imaging (UUI) for motion-correction, hemodynamic and biomechanical imaging. In order to ensure a precise co-registration of simultaneous PET-UUI acquisitions, ultrasound probes attached to an ultrafast ultrasound scanner are operated in the field of view (FOV) of a small animal PET/CT scanner using a remote-controlled micro-positioner. Here we explore the effect of the presence of ultrasound probes on PET image quality. We compare the performance of PET and image quality with and without the presence of probes in the PET field of view, both in vitro following the NEMA-NU-4-2008 standard protocol, and in vivo in small animals. Overall, deviations in the quality of images acquired with and without the ultrasound probes were under 10% and under 7% for the NEMA protocol and in vivo tests, respectively. Our results demonstrate the capability of the PETRUS device to acquire multimodal images in vivo without significant degradation of image quality.