A blood biomarker for monitoring response to anti-EGFR therapy.

BACKGROUND AND OBJECTIVE: To monitor therapies targeted to epidermal growth factor receptors (EGFR) in non-small cell lung cancer (NSCLC), we investigated Peroxiredoxin 6 (PRDX6) as a biomarker of response to anti-EGFR agents. METHODS: We studied cells that are sensitive (H3255, HCC827) or resistant (H1975, H460) to gefitinib. PRDX6 was examined with either gefitinib or vehicle treatment using enzyme-linked immunosorbent assays. We created xenograft models from one sensitive (HCC827) and one resistant cell line (H1975) and monitored serum PRDX6 levels during treatment. RESULTS: PRDX6 levels in cell media from sensitive cell lines increased significantly after gefitinib treatment vs. vehicle, whereas there was no significant difference for resistant lines. PRDX6 accumulation over time correlated positively with gefitinib sensitivity. Serum PRDX6 levels in gefitinib-sensitive xenograft models increased markedly during the first 24 hours of treatment and then decreased dramatically during the following 48 hours. Differences in serum PRDX6 levels between vehicle and gefitinib-treated animals could not be explained by differences in tumor burden. CONCLUSIONS: Our results show that changes in serum PRDX6 during the course of gefitinib treatment of xenograft models provide insight into tumor response and such an approach offers several advantages over imaging-based strategies for monitoring response to anti-EGFR agents.

Patterns of Vasculature in Mouse Models of Lung Cancer Are Dependent on Location.

PURPOSE: Preclinical studies of hypoxia are generally done using ectopic xenograft tumors, which behave differently from human tumors. Our previous findings have shown that subcutaneously implanted lung tumors exhibit more hypoxia than their orthotopic implanted or spontaneous K-ras-induced counterparts. We hypothesize that differences in hypoxia are due to site-specific differences in vascularity and perfusion. PROCEDURES: To compare the presence and functionality of vessels in these tumor models, we studied vascular perfusion in vivo in real time. RESULTS: Orthotopically implanted and spontaneous K-ras-induced lung tumors showed elevated perfusion, demonstrating vasculature functionality. Little contrast agent uptake was observed within the subcutaneously implanted tumors, indicating vascular dysfunction. These findings were corroborated at the microscopic level with Hoechst 33342 and Meca-32 staining. CONCLUSIONS: From these observations, we concluded that differences in hypoxia in experimental models is related to vessel perfusion. Thus, appropriate selection of preclinical lung tumor models is essential for the study of hypoxia, angiogenesis and therapies targeting these phenomena.

Assessing early therapeutic response to bevacizumab in primary breast cancer using magnetic resonance imaging and gene expression profiles.

Antiangiogenic therapy is a promising approach for the treatment of breast cancer. In practice, however, only a subset of patients who receive antiangiogenic drugs demonstrate a significant response. A key challenge, therefore, is to discover biomarkers that are predictive of response to antiangiogenic therapy. To address this issue, we have designed a window-of-opportunity study in which bevacizumab is administered as a short-term first-line treatment to primary breast cancer patients. Central to our approach is the use of a detailed pharmacodynamic assessment, consisting of pre- and post-bevacizumab multi-parametric magnetic resonance imaging scans and core biopsies for exon array gene expression analysis. Here, we illustrate three intrinsic patterns of response to bevacizumab and discuss the molecular mechanisms that may underpin each. Our results illustrate how the combination of dynamic imaging data and gene expression profiles can guide the development of biomarkers for predicting response to antiangiogenic therapy.

Survival and death signals can predict tumor response to therapy after oncogene inactivation.

Cancers can exhibit marked tumor regression after oncogene inhibition through a phenomenon called "oncogene addiction." The ability to predict when a tumor will exhibit oncogene addiction would be useful in the development of targeted therapeutics. Oncogene addiction is likely the consequence of many cellular programs. However, we reasoned that many of these inputs may converge on aggregate survival and death signals. To test this, we examined conditional transgenic models of K-ras(G12D)--or MYC-induced lung tumors and lymphoma combined with quantitative imaging and an in situ analysis of biomarkers of proliferation and apoptotic signaling. We then used computational modeling based on ordinary differential equations (ODEs) to show that oncogene addiction could be modeled as differential changes in survival and death intracellular signals. Our mathematical model could be generalized to different imaging methods (computed tomography and bioluminescence imaging), different oncogenes (K-ras(G12D) and MYC), and several tumor types (lung and lymphoma). Our ODE model could predict the differential dynamics of several putative prosurvival and prodeath signaling factors [phosphorylated extracellular signal-regulated kinase 1 and 2, Akt1, Stat3/5 (signal transducer and activator of transcription 3/5), and p38] that contribute to the aggregate survival and death signals after oncogene inactivation. Furthermore, we could predict the influence of specific genetic lesions (p53⁻/⁻, Stat3-d358L, and myr-Akt1) on tumor regression after oncogene inactivation. Then, using machine learning based on support vector machine, we applied quantitative imaging methods to human patients to predict both their EGFR genotype and their progression-free survival after treatment with the targeted therapeutic erlotinib. Hence, the consequences of oncogene inactivation can be accurately modeled on the basis of a relatively small number of parameters that may predict when targeted therapeutics will elicit oncogene addiction after oncogene inactivation and hence tumor regression.

Use of DNA microarray and small animal positron emission tomography in preclinical drug evaluation of RAF265, a novel B-Raf/VEGFR-2 inhibitor.

Positron emission tomography (PET) imaging has become a useful tool for assessing early biologic response to cancer therapy and may be particularly useful in the development of new cancer therapeutics. RAF265, a novel B-Raf/vascular endothelial growth factor receptor-2 inhibitor, was evaluated in the preclinical setting for its ability to inhibit the uptake of PET tracers in the A375M(B-Raf(V600E)) human melanoma cell line. RAF265 inhibited 2-deoxy-2-[(18)F]fluoro-d-glucose (FDG) accumulation in cell culture at 28 hours in a dose-dependent manner. RAF265 also inhibited FDG accumulation in tumor xenografts after 1 day of drug treatment. This decrease persisted for the remaining 2 weeks of treatment. DNA microarray analysis of treated tumor xenografts revealed significantly decreased expression of genes regulating glucose and thymidine metabolism and revealed changes in apoptotic genes, suggesting that the imaging tracers FDG, 3-deoxy-3-[(18)F]fluorothymidine, and annexin V could serve as potential imaging biomarkers for RAF265 therapy monitoring. We concluded that RAF265 is highly efficacious in this xenograft model of human melanoma and decreases glucose metabolism as measured by DNA microarray analysis, cell culture assays, and small animal FDG PET scans as early as 1 day after treatment. Our results support the use of FDG PET in clinical trials with RAF265 to assess early tumor response. DNA microarray analysis and small animal PET studies may be used as complementary technologies in drug development. DNA microarray analysis allows for analysis of drug effects on multiple pathways linked to cancer and can suggest corresponding imaging tracers for further analysis as biomarkers of tumor response.

Wall motion classification of stress echocardiography based on combined rest-and-stress data.

In this paper, we represent a new framework that performs automated local wall motion analysis based on the combined information derived from a rest and stress sequence (a full stress echocardiography study). Since cardiac data inherits time-varying and sequential properties, we introduce a Hidden Markov Model (HMM) approach to classify stress echocardiography. A wall segment model is developed for a normal and an abnormal heart and experiments are performed on rest, stress and rest-and-stress sequences. In an assessment using n = 44 datasets, combined rest-and-stress analysis shows an improvement in classification (84.17%) over individual rest (73.33%) and stress (68.33%).

Semi-supervised learning of probabilistic models for ECG segmentation.

We present a novel semi-supervised learning algorithm, based upon the EM algorithm for maximum likelihood estimation, which can be used to learn probabilistic models from subjectively labelled data. We demonstrate the method on the task of automated ECG segmentation, with a particular emphasis on the accurate measurement of the QT interval. In addition we discuss the use of wavelet methods for the representation of the ECG, and advanced duration modelling techniques for hidden Markov models applied to ECG segmentation.