Search

Iterative decoding algorithms may be viewed as high-dimensional nonlinear dynamical systems, depending on a large number of parameters. In this work, we introduce a simplified description of several iterative decoding algorithms in terms of the a posteriori average entropy, and study them as a function of a single parameter that closely approximates the signal-to-noise ratio (SNR). Using this approach, we show that virtually all the iterative decoding schemes in use today exhibit similar qualitative dynamics. In particular, a whole range of phenomena known to occur in nonlinear systems, such as existence of multiple fixed points, oscillatory behavior, bifurcations, chaos, and transient chaos are found in iterative decoding algorithms. As an application, we develop an adaptive technique to control transient chaos in the turbo-decoding algorithm, leading to a substantial improvement in performance. We also propose a new stopping criterion for turbo codes that achieves the same performance with considerably fewer iterations.

In this correspondence, we introduce a simple one-dimensional (1-D) nonlinear map to describe the iterates of the bit-error rate (BER) of parallel-concatenated convolutional codes (PCCC) on the binary-input Gaussian channel. A lower bound on this map is derived based upon the weight enumerator of the constituent codes, thus enabling the characterization of the dynamics of the decoder in terms of fixed points, along with the associated stability analysis.

We assessed the tolerability, safety, pharmacokinetics and dose-limiting toxicity (DLT) of the recombinant humanized IgG4 anti-vascular endothelial growth factor (VEGF) monoclonal antibody, HuMV833, in patients with advanced cancer. Cohorts of patients with progressive solid tumours received escalating doses of HuMV833 as a 1-h intravenous (I.V.) infusion on days 1, 15, 22, and 29. Twenty patients (median Eastern Cooperative Oncology Group (ECOG) score 1) were accrued. HuMV833 infusions were well tolerated and there were no grade III or IV toxicities definitely related to the antibody. Grade I or II toxicities probably related to the antibody included fatigue, dyspnoea and rash. There were two episodes of asymptomatic hypocalcaemia, one at grade III and one grade IV, which were recorded in early follow-up. There were eight grade I episodes of asymptomatic elevation of activated partial thromboplastin time (APTT) and two grade III events; one in a patient receiving 1 mg/kg and the other receiving extended doses of 10 mg/kg. Pharmacokinetic analysis revealed a non-linear kinetic and an elimination half-life of between 8.2 (0.3 mg/kg) and 18.7 (10 mg/kg) days. One patient with ovarian cancer experienced a partial response (PR) of 9 months duration and eight had disease stabilisation (SD) including one patient with colorectal carcinoma whose disease was stable for 14 months. In 13 of the 14 samples taken from 12 patients, the plasma concentration of hepatocyte growth factor (HGF) was reduced 24 h after drug administration. HuMV833 is safe and lacked DLT at doses up to 10 mg/kg on this schedule. Multiple doses were well tolerated, despite occasional asymptomatic elevations in APTT. By combining pharmacokinetic, pharmacodynamic and toxicity data, we can identify doses of 1 and 3mg/kg for further investigation. HuMV833 appears to possess some clinical activity.

We propose a novel density evolution approach to analyze the iterative decoding algorithms of low-density parity-check (LDPC) codes and product codes, based on Gaussian densities. Namely, for these classes of codes we derive a one-dimensional (1D) map whose iterates directly represent the error probability both for the additive white Gaussian noise (AWGN) and the Rayleigh-fading channel. These simple models allow a qualitative analysis of the nonlinear dynamics of the decoding algorithm. As an application, we compute the decoding thresholds and show that they are consistent with the simulation results available in the literature.

Pseudochaotic time hopping (PCTH) is a previously proposed encoding/modulation scheme for ultra-wideband (UWB) impulse radio. PCTH exploits concepts from symbolic dynamics to generate aperiodic spreading sequences, resulting in a noise-like spectrum. We present a multiple-access technique suitable for the PCTH scheme. In particular, we provide an analytical expression of the bit-error rate performance as a function of the number of users and validate it by simulation.

The EORTC Melanoma Group (MG) was founded in 1969 by both clinicians and scientists from various disciplines and fields of research with a common interest in malignant melanoma. This collaborative approach has always been the foundation of the groups strength. With an interest in tumour biology and especially the immunological aspects of the disease, the group has always pursued a scientific approach to treatment development in malignant melanoma. Over the years, the group has performed many clinical trials, epidemiological studies, histopathological studies defining standards and guidelines, translational research regarding prognostic factors and various metastatic and immunological aspects of melanoma, and developed quality assurance programmes for immunological and molecular biological assays in laboratory networks. At present, the EORTC MG runs the worldwide largest clinical trial programme in stages II, III and IV melanoma involving some 140 cancer centres in and outside Europe. Each trial is associated with the appropriate translational research programmes.

Early drug development at EORTC has always been subject to structural changes to adapt to the rapid changes that occur in oncological drug development. The expertise of early drug developers has always been cross-fertilised with disease-/tumour-oriented groups and also backwards to the laboratory research groups. This results in the establishment of a solid and dedicated network of medical oncologists with focused expertise in cancer drug development. The EORTC Data Center is fully equipped with all expertise to support clinical research activities and includes regulatory, safety, and quality assurance desks. The EORTC New Drug development Programme (NDDP) provides methodological expertise to early clinical trials and coordinates phase I and phase II studies addressing various approaches. Through NDDP, the early clinical groups and the disease-/tumour-oriented groups have created specific networks to address early drug development in specific tumour types. This results in very efficient networks which have the resources and the patients to address and conduct challenging clinical trials in a standardised fashion ensuring the highest standards in cancer treatment.

TPS3093 Background: Chimeric Antigen Receptor (CAR)-T therapy has potentially serious limitations related to target antigen loss, toxicity due to pre-conditioning regimen, and lack of activity in many tumor types. To overcome these limitations, we have developed a novel CAR-T, called NKR-2, incorporating the full-length human natural killer receptor NKG2D fused with the human CD3 zeta signaling domain. When expressed in T-cells, the naturally-expressed DAP10 provides co-stimulatory signals to NKR-2 to produce cytokines and selectively target tumor cells upon recognition of up to 8 different stress-induced NKG2D ligands expressed in many solid and hematologic malignancies. In preclinical studies, NKR-2 demonstrated long-term anti-tumor activity towards a breadth of tumor indications, in the absence of pre-conditioning, whilst simultaneously targeting tumor cells and cells from the local tumor neo-vasculature and suppressive immune environment. In our recently completed First-in-Human Phase 1 study (NCT02203825) in hematologic cancers, a single administration of autologous NKR-2 was safe with initial signs of clinical benefit. Methods: Exploiting the multiple ligand targeting capability and unique mode of action of NKR-2, the THINK trial (THerapeutic Immunotherapy with NKR-2) is an open-label Phase I study that will assess the safety and clinical activity of multiple infusion NKR-2 treatment (every 2 weeks x 3 infusions) in relapse/refractory patients with metastatic or locally advanced CRC, urothelial carcinoma, TNBC, pancreatic cancer, recurrent epithelial ovarian and fallopian tube carcinoma, AML/MDS and MM, post standard treatment. The study contains two consecutive segments. The dose escalation segment will enroll 18 patients in two separate hematologic and solid malignancy arms, and will evaluate 3 dose levels of NKR-2 (3x108, 1x109 and 3x109 per injection) following a 3+3 design. The expansion segment will then enroll 96 additional patients in 7 separate cohorts for each indication with 3 steps of statistical analysis (overall futility, cohort futility and final evaluation). The study is open for recruitment in both EU and US. Clinical trial information: NCT03018405.

International audience; In this paper we address the problem of multiple target tracking using passive radars. The fusion of data from several bistatic radars will be used to facilitate the detection of multiple targets. We propose a strategy for performing target localization and robustness against ghost target rejection in multistatic passive radars. We report the effect of our proposed technique on tracker performance in terms of probability of tracks loss, probability of false tracks and mean squared error

Assessment of T-cell activation is pivotal for evaluation of cancerimmunotherapy. We initiated a clinical trial in patients with MAGE-A1and/or -A3 tumors using autologous DC pulsed with MAGE peptides aimedat analyzing T-cell-derived, IFN-γ secretion by cytokine flowcytometry and ELISPOT. We also tested whether further KLH additioncould influence this response favorably. Monocyte-derived DC weregenerated from leukapheresis products. They were pulsed with therelevant MAGE peptide(s) alone in group A (n=10 pts) andadditionally with KLH in group B (n=16 pts). A specific buttransient increase in the number of peripheral blood T lymphocytessecreting IFN-γ in response to the vaccine peptide(s) was observed in6/8 patients of group A and in 6/16 patients of group B. We concludethat anti-tumor vaccination using DC pulsed with MAGE peptides inducesa potent but transient anti-MAGE, IFN-γ secretion that is notinfluenced by the additional delivery of a nonspecific, T-cellhelp.

Corrections In our original manuscript [1], the corresponding author, Michael A Morse, was missed from the authors’ list. Therefore the correct author list should be: Erika Hamilton, Kimberly Blackwell, Amy C Hobeika, Timothy M Clay, Gloria Broadwater, Xiu-Rong Ren, Wei Chen, Henry Castro, Frederic Lehmann, Neil Spector, Junping Wei, Takuya Osada, H Kim Lyerly and Michael A Morse. Also the title of our original manuscript ‘Phase 1 clinical trial of HER2-specific immunotherapy with concomitant HER2 kinase inhibtion’ is also incorrect. The title should read:‘Phase 1 clinical trial of HER2-specific immunotherapy with concomitant HER2 kinase inhibition’. Journal of Translational Medicine regret any inconvenience that this inaccuracy might have caused.

Purpose To detect a pretreatment gene expression signature (GS) predictive of response to MAGE-A3 immunotherapeutic in patients with metastatic melanoma and to investigate its applicability in a different cancer setting (adjuvant therapy of resected early-stage non–small-cell lung cancer [NSCLC]). Patients and Methods Patients were participants in two phase II studies of the recombinant MAGE-A3 antigen combined with an immunostimulant (AS15 or AS02B). mRNA from melanoma biopsies was analyzed by microarray analysis and quantitative polymerase chain reaction. These results were used to identify and cross-validate the GS, which was then applied to the NSCLC data. Results In the patients with melanoma, 84 genes were identified whose expression was potentially associated with clinical benefit. This effect was strongest when the immunostimulant AS15 was included in the immunotherapy (hazard ratio [HR] for overall survival, 0.37; 95% CI, 0.13 to 1.05; P = .06) and was less strong with the other immunostimulant AS02B (HR, 0.84; 95% CI, 0.36 to 1.97; P = .70). The same GS was then used to predict the outcome for patients with resected NSCLC treated with MAGE-A3 plus AS02B; actively treated GS-positive patients showed a favorable disease-free interval compared with placebo-treated GS-positive patients (HR, 0.42; 95% CI, 0.17 to 1.03; P = .06), whereas among GS-negative patients, no such difference was found (HR, 1.17; 95% CI, 0.59 to 2.31; P = .65). The genes identified were mainly immune related, involving interferon gamma pathways and specific chemokines, suggesting that their pretreatment expression influences the tumor's immune microenvironment and the patient's clinical response. Conclusion An 84-gene GS associated with clinical response for MAGE-A3 immunotherapeutic was identified in metastatic melanoma and confirmed in resected NSCLC.

Introduction: Conventional CAR-T cells express a single chain antibody variable fragment that restricts recognition to one tumor antigen and a limited set of cancers. This study employs a novel CAR fusing full-length human NKG2D with the CD3z signaling domain. In autologous transduced CM-CS1 T cells, NKG2D CAR receives endogenous costimulation via DAP10 to target multiple NKG2D-ligands that are upregulated in solid and hematologic malignancies but absent or poorly expressed on healthy tissues. Methods: A phase 1 dose-escalation study to establish safety and feasibility of a single infusion of CM-CS1 T cells without lymphodepleting conditioning enrolled subjects with AML/MDS-RAEB or relapsed/refractory progressive multiple myeloma (MM) without standard therapy options (NCT02203825). Eligibility criteria included suitable organ function, no CNS disease, no prior allogeneic SCT or adoptive T-cell therapy, no therapy within 3 weeks prior to infusion, no immune suppression, and no uncontrolled infection. Dose-escalation spanned 4 cohorts [half-log increments from 1x106 to 3x107 CM-CS1 T cells] according to a 3+3 design. DLTs included ≥ Grade 3 non-hematologic toxicity or ≥ Grade 2 autoimmune toxicity related to CAR T cells. Initial assessment was at 28 days. At least 1 AML/MDS and 1 MM subject were mandated in each dose level. Manufacturing included PBMC stimulation with OKT3 and IL-2 followed by 2 rounds of retroviral transduction at DFCI's Cell Manipulation Core Facility. Vector copy number (VCN) and replication-competent retrovirus (RCR) testing were performed on whole blood and PBMCs, respectively, using quantitative PCR. Results: From April 2015 to July 2016, 11 subjects were infused, and 10 completed the DLT period. Eight of 11 were male, 6 had AML/MDS, and median age was 70 (range 44 to 79) (Panel A). Median WBC was 2.3 (range 0.7 to 7.2 K/uL); median ALC was 0.74 (range 0.09-2.37 K/uL). Five had cells manufactured from peripheral blood; 6 underwent apheresis. Median percentage of blasts in bone marrow for AML/MDS patients was 50% (range 4-68%). All myeloma patients had undergone ≥ 5 therapies including ≥1 autologous SCT. Four of the 6 AML/MDS patients had secondary disease, 3 had complex cytogenetics, 3 had p53 mutations, and 1 had a FLT3-ITD mutation. Dose-escalation proceeded from 1x106 to 3x107 CM-CS1 T cells. All 11 products passed release criteria, and there were no infusion reactions. Products consisted of median 97.2% CD3+ cells and 31.0% CD8+ cells, with vector-specific NKG2D expression on median 74.6% of CD3+ and 66.3% of CD8+ cells (Panel B). The first 10 subjects completed their 28 day evaluation period without DLTs. There were no cases of cytokine release syndrome, cell-related neurotoxicity, auto-immunity, or CAR T-related death. SAEs included a Grade 4 intracochlear bleed and an episode each of grade 4 neutropenia and thrombocytopenia deemed related to disease progression. Forty percent of patients experienced some Grade 3 toxicity, all related to underlying disease or a complication thereof (Panel C). At these initial cell doses, no patient to date has had objective tumor response at the 28 day evaluation mark. Nine initiated subsequent therapies; there have been 4 deaths secondary to disease or complications of subsequent therapies. However, cases of unexpected survival without further therapy and responses to subsequent treatments were noted. For example a patient with p53-mutated AML survived 4 months despite 50% blasts at infusion, and another entered PR at 6+months after cells on an IDH-1 inhibitor with Conclusion: In the first 3+ dose-escalation cohorts of patients with AML/MDS and myeloma, a single dose of CM-CS1 T cells without lymphodepletion was feasible and well-tolerated, with no DLTs. CAR T cells generally have not persisted beyond 1 week, consistent with pre-clinical models. Correlative analyses including post-infusion immunophenotyping are in process. Future studies of multiple infusions of NKG2D CAR T cells in both hematologic malignancies and solid tumors at the higher cell doses associated with efficacy in pre-clinical models are in planning. Table Table. Disclosures Murad: Celdara Medical, LLC: Employment. Reder:Celdara Medical, LLC: Employment. Sentman:Celdara Medical, LLC: Membership on an entity's Board of Directors or advisory committees, Other: Holds patents on this technology. Wade:Celdara Medical, LLC: Employment. Schmucker:Celdara Medical, LLC: Employment. Lehmann:Celyad, SA: Employment. Snykers:Celyad, SA: Employment. Allen:Celyad, SA: Employment. Stone:Celator: Consultancy; Jansen: Consultancy; Novartis: Consultancy; Merck: Consultancy; ONO: Consultancy; Sunesis Pharmaceuticals: Consultancy; Roche: Consultancy; Celgene: Consultancy, Membership on an entity's Board of Directors or advisory committees; Juno Therapeutics: Consultancy; Xenetic Biosciences: Consultancy; Agios: Consultancy; Abbvie: Consultancy, Membership on an entity's Board of Directors or advisory committees; Seattle Genetics: Consultancy; Pfizer: Consultancy; Karyopharm: Consultancy; Amgen: Consultancy. Soiffer:Kiadis: Membership on an entity's Board of Directors or advisory committees; Juno: Consultancy. Dranoff:Novartis: Employment. Ritz:Kiadis: Membership on an entity's Board of Directors or advisory committees.

TPS3102Background: Canonical CAR-T cell constructs encode a single chain antibody variable fragment, costimulatory domain, and signaling domain of CD3 zeta. This restricts recognition to 1 tumor an...

International audience; The mitigation of 2-D intersymbol interference (ISI) has recently emerged as an important problem in the field of storage technology and wireless communications. Several solutions based on message passing have been proposed to perform equalization on 2-D channels with ISI. We present an algorithm based on Gaussian belief propagation (GaBP). We show that the proposed algorithm has interesting features compared with the corresponding message passing solutions available in the literature. First, the complexity of the proposed equalizer is independent of the size of the constellation used for modulation. Second, since the complexity is not exponential in the size of the state space, the computational burden is reduced for channels with long memory. As an application, we consider decentralized intercell interference mitigation in uplink cellular networks, where base stations communicate with their neighbors to recover their own users' signals.

International audience; We consider physical network coding on a two-way relay channel. Two source nodes send their packets simultaneously over time-varying MIMO channels to a relay node. Assuming that the relay employs the joint decode-and-forward strategy, the relay must be able to decode the individual source packets. This task requires a robust estimation of the source-to-relay channels. We design a message passing algorithm based on factor-graphs to solve this problem efficiently.

Background Patients with HER2-overexpressing metastatic breast cancer, despite initially benefiting from the monoclonal antibody trastuzumab and the EGFR/HER2 tyrosine kinase inhibitor lapatinib, will eventually have progressive disease. HER2-based vaccines induce polyclonal antibody responses against HER2 that demonstrate enhanced anti-tumor activity when combined with lapatinib in murine models. We wished to test the clinical safety, immunogenicity, and activity of a HER2-based cancer vaccine, when combined with lapatinib. Methods We immunized women (n = 12) with metastatic, trastuzumab-refractory, HER2-overexpressing breast cancer with dHER2, a recombinant protein consisting of extracellular domain (ECD) and a portion of the intracellular domain (ICD) of HER2 combined with the adjuvant AS15, containing MPL, QS21, CpG and liposome. Lapatinib (1250 mg/day) was administered concurrently. Peripheral blood antibody and T cell responses were measured. Results This regimen was well tolerated, with no cardiotoxicity. Anti-HER2-specific antibody was induced in all patients whereas HER2-specific T cells were detected in one patient. Preliminary analyses of patient serum demonstrated downstream signaling inhibition in HER2 expressing tumor cells. The median time to progression was 55 days, with the majority of patients progressing prior to induction of peak anti-HER2 immune responses; however, 300-day overall survival was 92% (95% CI: 77-100%). Conclusions dHER2 combined with lapatinib was safe and immunogenic with promising long term survival in those with HER2-overexpressing breast cancers refractory to trastuzumab. Further studies to define the anticancer activity of the antibodies induced by HER2 vaccines along with lapatinib are underway. Trial registry ClinicalTrials.gov NCT00952692

International audience; We consider blurring of binary images and corruption by ambient noise occuring on two-dimensional storage channels. Since coding is generally used in such systems, the deconvolution problem can be treated jointly with decoding. Several methods have been proposed in the literature under the name of turbo equalization to mitigate the degradation introduced by such channels. However, the problem of blur identification has rarely been addressed previously. In this paper, we propose a technique for estimating the 2D channel coefficients, along with the variance of the ambient noise. The proposed estimation algorithm is adaptive and performed jointly with turbo equalization, so as to limit the number of known pilot symbols needed to bootstrap the channel estimator. Interestingly, we found that the computational complexity of the proposed joint channel estimation and turbo equalization method depends heavily on the sensitivity of existing turbo equalization methods to 2D channel parameter mismatch

Progress in understanding the molecular basis of melanoma has made possible the identification of molecular targets with important implications in clinical practice. In fact, new therapeutic approaches are emerging from basic science and it will be important to implement their rapid translation into clinical practice by active clinical investigation. The first meeting of Melanoma Research: a bridge Naples-USA, organized by Paolo A. Ascierto (INT, Naples, Italy) and Francesco Marincola (NIH, Bethesda, USA) took place in Naples, on 6-7 December 2010. This international congress gathered more than 30 international and Italian faculty members and was focused on recent advances in melanoma molecular biology, immunology and therapy, and created an interactive discussion across Institutions belonging to Government, Academy and Pharmaceutical Industry, in order to stimulate new approaches in basic, translational and clinical research. Four topics of discussion were identified: New pathways in Melanoma, Biomarkers, Clinical Trials and New Molecules and Strategies.

In order to define the antigens recognized by cytolytic T lymphocytes (CTLs) on autologous tumors, we derived tumor-specific CTL clones from autologous mixed lymphocyte tumor cell cultures. The gene coding for a tumor rejection antigen expressed on a melanoma was isolated by transfecting genomic DNA of the tumor into an antigen-loss variant of the melanoma. Transfectants were identified on the basis of their ability to stimulate tumor necrosis factor release by the CTL clone. The gene that transferred the expression of the antigen was named MAGE-1. It is a new gene, silent in normal tissues with the exception of testis, but expressed in several types of tumors. The antigen recognized by the CTL clone is a nonapeptide derived from the protein encoded by gene MAGE-1, and presented by the HLA class I molecule HLA-A1. Using two other antimelanoma CTL clones, we identified the tyrosinase gene as coding for an antigen presented by HLA-A2 on this type of tumors. The identification of these tumor rejection antigens open new possibilities for the specific immunotherapy of cancer.

International audience; We consider the problem of semi-supervised segmentation of textured images. Existing model-based approaches model the intensity field of textured images as a Gauss-Markov random field to take into account the local spatial dependencies between the pixels. Classical Bayesian segmentation consists of also modeling the label field as a Markov random field to ensure that neighboring pixels correspond to the same texture class with high probability. Well-known relaxation techniques are available which find the optimal label field with respect to the maximum a posteriori or the maximum posterior mode criterion. But, these techniques are usually computationally intensive because they require a large number of iterations to converge. In this paper, we propose a new Bayesian framework by modeling two-dimensional textured images as the concatenation of two one-dimensional hidden Markov autoregressive models for the lines and the columns, respectively. A segmentation algorithm, which is similar to turbo decoding in the context of error-correcting codes, is obtained based on a factor graph approach. The proposed method estimates the unknown parameters using the Expectation-Maximization algorithm

International audience; We consider the equalization of two-dimensional intersymbol interference (2D ISI) channels. We propose a message passing algorithm based on Gaussian belief propagation (GaBP) to perform joint decoding and equalization. While in existing 2D equalizers the complexity is exponential in the length of the state-space, the complexity is only polynomial for the proposed algorithm. This enables to handle 2D channels with long memory with a tractable computational complexity

Limiting numbers of peripheral-blood mononuclear cells (PBMC) from melanoma patients were stimulated with irradiated autologous tumor cells in the presence of interleukins-2 and -4 and in the absence of feeder cells. The responder cells were restimulated every week. After 2 to 4 weeks, the microcultures were tested for their lytic activity against the autologous tumor cells. Significant lysis of the tumor cells was observed with a fraction of these microcultures, whereas no lysis was observed with control microcultures seeded without stimulator melanoma cells. Because our aim was to measure the precursor frequency of CTL showing specificity for the tumor, and not that of NK-like effectors that were also capable of lysing the melanoma cells, we used cold-target inhibition with an excess of NK target K562 to inhibit the NK-like activity. Microcultures whose lysis on the tumor cells was not abolished by K562 competition were observed. The specificity of these CTL clones was confirmed by the absence of lytic activity on autologous T-cell blasts. The numbers of microcultures with anti-tumor CTL activity fitted the zero-order of the Poisson distribution equation, indicating that they resulted from the activity of single T-cell clones. The frequency of anti-tumor CTL precursor cells (CTL-P) of 7 melanoma patients ranged from 1/900 to 1/33,000. Frequencies of anti-tumoral CTL-P were higher and NK-like effectors were less frequent when sorted CD8+ T lymphocytes were used as responder cells.

International audience; We present a joint phase estimation and decoding method for convolutional turbo codes in the presence of strong phase noise. In order to overcome the problem of phase ambiguity and cycle slips, a combined state-space model for the time varying phase and the component convolutional codes is introduced. The proposed algorithm uses a Gaussian sum approach to approximate the multimodal a posteriori probability density function (pdf) of the phase in a blind context. We compare our method to the well known alternative consisting in discretizing the phase.Monte-Carlo simulations for the turbo code used in the DVB-RCS standard show that the performances of the proposed scheme are close to decoding with perfect knowledge of the phase

International audience; Blind coherent detection of convolutional turbo codes is a hard problem in the presence of strong phase noise. Since the operating signal-to-noise ratio is usually very low, phase synchronisation algorithms suffer from phase ambiguities and cycle slips. A possible remedy is to perform joint phase estimation and decoding on a combined state-space model for the time-varying phase and the component convolutional codes. We demonstrate that joint phase estimation and decoding is in fact mandatory only for one component code, while ordinary BCJR decoding can be used for the second component code. Monte Carlo simulations for the turbo code used in the DVB-RCS standard show that the performances of the proposed scheme are close to decoding with perfect knowledge of the phase

International audience; We consider blind equalization for block transmissions over the frequency selective Rayleigh fading channel. In the absence of pilot symbols, the receiver must be able to perform joint equalization and blind channel identification. Relying on a mixed discrete-continuous state-space representation of the communication system, we introduce a blind Bayesian equalization algorithm based on a Gaussian mixture parameterization of the a posteriori probability density function (pdf) of the transmitted data and the channel. The performances of the proposed algorithm are compared with existing blind equalization techniques using numerical simulations.

We present a joint phase estimation and decoding method for convolutional turbo codes in the presence of strong phase noise. In order to overcome the problem of phase ambiguity and cycle slips, a combined state-space model for the time-varying phase and the component convolutional codes is introduced. The proposed algorithm uses a Gaussian sum approach to approximate the multimodal a posteriori probability density function (pdf) of the phase in a blind context. Monte-Carlo simulations for the turbo code used in the DVB-RCS standard show that the performances of the proposed scheme are close to decoding with perfect knowledge of the phase.

International audience; We present a joint channel estimation and detection method of space-time trellis codes (STTC) in the context of an unknown flat fading multiple-input multiple-output (MIMO) channel. A combined state-space model for the space-time code and the Rayleigh fading MIMO channel is introduced, in order to use deterministic particle filtering at the receiver side. An important feature of the proposed method is that the fading rate need not be known to the receiver. Monte-carlo simulations show that the performances of the proposed scheme are close to decoding with perfect channel state information (CSI) using the Viterbi algorithm (VA)

We present an iterative joint phase/timing estimation and decoding method for GEO satellite transmission systems. A combined state-space model for the binary digits and the channel parameters is introduced, in order to use deterministic Maximum-Likelihood (ML) particle filtering at the receiver side. Monte Carlo simulations show that near-optimum performances can be obtained for BPSK modulation.

Interleaved convolutional time hopping (ICTH) is a recently proposed coding/modulation scheme for UWB impulse radio. ICTH is based upon a low-rate convolutional code, with optimized distance properties, combined with multilevel pulse-position modulation (PPM). In this context, bit-interleaved coded modulation (BICM) is used to decorrelate the bit errors and, more importantly, to obtain a flat spectrum. We present analytical bounds for the BER (bit-error-rate) performance of ICTH and an analysis of its spectral properties.

In this paper we introduce a model of the iterative decoding of product codes. For each iteration, an approximation of the bit error rate (BER) at the output of the decoders is used to evaluate the density of the extrinsic information. The decoding thresholds obtained for the additive white Gaussian noise (AWGN) and the Rayleigh fading channel are consistent with the simulation results available in the literature.

Memory CD8 T cells mediate rapid and effective immune responses against previously encountered Ags. However, these cells display considerable phenotypic and functional heterogeneity. In an effort to identify parameters that correlate with immune protection, we compared cell surface markers, proliferation, and cytokine production of distinct virus- and tumor-specific human CD8 populations. Phenotypic analysis of epitope-specific CD8 T cells showed that Ag specificity is associated with distinct CCR7/CD45RA expression profiles, suggesting that Ag recognition drives the expression of these molecules on effector/memory T cells. Moreover, the majority of central memory T cells (CD45RAlowCCR7dull) secreting cytokines in response to an EBV epitope produces both IL-2 and IFN-γ, whereas effector memory CD8 cells (CD45RAdullCCR7−) found in EBV, CMV, or Melan-A memory pools are mostly composed of cells secreting exclusively IFN-γ. However, these various subsets, including Melan-A-specific effector memory cells differentiated in cancer patients, display similar Ag-driven proliferation in vitro. Our findings show for the first time that human epitope-specific CD8 memory pools differ in IL-2 production after antigenic stimulation, although they display similar intrinsic proliferation capacity. These results provide new insights in the characterization of human virus- and tumor-specific CD8 lymphocytes.

Pseudo-chaotic time hopping (PCTH) is a recently proposed modulation scheme for UWB impulse radio. PCTH exploits concepts from symbolic dynamics to generate aperiodic spreading sequences. In this paper, we present a general analytical expression for the average bit error rate (BER) performance of a synchronous multi-access PCTH system as a function of the cross-correlation between the users' signatures in an additive white Gaussian noise (AWGN) channel. Also, it is shown that with enough users an error floor in the BER can develop.

Pseudo-chaotic time hopping (PCTH) is an encoding/modulation scheme proposed for UWB (ultra-wide band) impulse radio. PCTH exploits concepts from symbolic dynamics to generate aperiodic spreading sequences. We present a multiple access technique suitable for the pseudo-chaotic time hopping scheme.

In this paper, we investigate the distance properties of irregular LDPC codes using the ensemble construction proposed by Richardson et al (Ref. 2). By generalizing the method introduced by Gallager for regular LDPC codes [R.G. Gallager, 1963], we calculate the average weight distribution of irregular LDPC codes. Then, we interpret the stability condition of density evolution as a sufficient condition so that the error probability due to low weight codewords tends to zero under maximum-likelihood decoding.

High spectral efficiency in wavelength-division multiplexing (WDM) coherent optical communications can be achieved by lowering the channel spacing to a value close to the baud rate. As a result, such communication systems are subject to a high level of two-dimensional (2D) interference, originating from both chromatic dispersion (CD) within WDM channels and cross talk between WDM channels. We propose to digitally restore the transmitted information at the receiver side by jointly processing the WDM channels, using 2D interference cancellation methods interacting with a forward error correction decoder. Such schemes have been investigated extensively for known channel parameters; however, little is known about the error performance in the absence of channel state information. This issue is important, since in practice uncertainties about the physical parameters of the optical fiber and noise levels are unavoidable. In this paper, we propose optical channel parameter estimation schemes with limited impact on the overall complexity. First, a decision-aided carrier phase recovery, taking the presence of 2D interference into account, is designed to compensate the laser beat linewidth. Then we provide a technique to estimate the linear channel impairments for each WDM channel, based on a limited number of training symbols. The validity of the proposed method is evaluated by numerical simulations taking into account linear and nonlinear fiber impairments. In particular, we show that for high-order modulations, joint processing of spectrally overlapping WDM channels becomes mandatory to obtain satisfactory performances.

Melanoma lines MEL.A and MEL.B were derived from metastases removed from patient LB33 in 1988 and 1993, respectively. The MEL.A cells express several antigens recognized by autologous cytolytic T lymphocytes (CTL) on HLA class I molecules. The MEL.B cells have lost expression of all class I molecules except for HLA-A24. By stimulating autologous lymphocytes with MEL.B, we obtained an HLA-A24-restricted CTL clone that lysed these cells. A novel gene, PRAME, encodes the antigen. It is expressed in a large proportion of tumors and also in some normal tissues, albeit at a lower level. Surprisingly, the CTL failed to lyse MEL.A, even though these cells expressed the gene PRAME. The CTL expresses an NK inhibitory receptor that inhibits its lytic activity upon interaction with HLA-Cw7 molecules, which are present on MEL.A cells and not on MEL.B. Such CTL, active against tumor cells showing partial HLA loss, may constitute an intermediate line of anti-tumor defense between the CTL, which recognize highly specific tumor antigens, and the NK cells, which recognize HLA loss variants.

Anti-melanoma cytolytic T-lymphocyte (CTL) clones were derived from peripheral blood lymphocytes of HLA-A2 melanoma patient LB265 after stimulation with the autologous tumor cell line LB265-MEL, which showed high expression of melanocyte-lineage specific genes. Of 55 CTL clones, 46 recognized HLA-A2-restricted antigens. These 46 CTL clones were studied for their ability to specifically release tumor necrosis factor in the presence of COS cells cotransfected with the HLA-A2 gene and the cDNA of either tyrosinase, Melan-A/MART1, Pmel17/gpl00, gp75/TRP1, or MSH receptor. Six CTL clones recognized the Melan-A/MART1 antigen, whereas the remaining 40 CTL clones recognized a Pmel17/gp100 antigen. These 40 anti-PmelI7/gpl00 CTL clones were all able to lyse T2 cells pulsed with the antigenic peptide YLEPGPVTA, as previously reported. The T-cell receptor beta chain hypervariable region was sequenced and found to be identical in the 15 CTL clones analyzed. Taken together, these data show a high frequency of Pmell7/gp100-specific T cells in autologous antitumor CTL clones derived from peripheral blood of a melanoma patient.

We have identified an antigen recognized on a human melanoma by autologous cytolytic T lymphocytes. It is encoded by a gene that is expressed in many normal tissues. Remarkably, the sequence coding for the antigenic peptide is located across an exon-intron junction. A point mutation is present in the intron that generates an amino acid change that is essential for the recognition of the peptide by the anti-tumor cytotoxic T lymphocytes. This observation suggests that the T-cell-mediated surveillance of the integrity of the genome may extend to some intronic regions.

We have studied the patterns of antigens recognized by autologous cytolytic T lymphocytes (CTL) on two melanoma cell lines derived from metastases that were removed from patient LB33 at several years distance. Cell line LB33-MEL.A was obtained after surgery in 1988. A large number of CTL clones directed against LB33-MEL.A was obtained with blood lymphocytes collected from the patient in 1990. In vitro selection of melanoma cells that were resistant to these CTL clones indicated that al least five different antigens were recognized on LB33-MEL.A by autologous CTL. Four of these antigens were found to be presented by HLA-A28, B13, B44 and Cw6, respectively. The patient remained disease-free until 1993 when a metastasis was detected and was used to obtain cell line LB33-MEL.B. This cell line proved resistant to lysis by all the CTL clones directed against the LB33-MEL.A cells and showed no expression of HLA class I molecules except for HLA-A24. Using LB33-MEL.B cells to stimulate blood lymphocytes collected from the patient in 1994 we derived CTL clones that lysed these cells. All these CTL clones recognized a new antigen presented by HLA-A24. These results suggest that in patient LB33 the melanoma cells may have lost the expression of several HLA molecules under the selective pressure of an anti-tumor CTL response.

7013 Background: Previous trials with the MAGE-A3 recombinant (rec) protein formulated with an immunostimulant have shown induction of specific immune responses and signals of clinical activity in different cancer diseases. In this phase I/II study (NCT 00455572), we evaluated the safety profile of the MAGE-A3 cancer immunotherapeutic (CI), formulated with the rec MAGE-A3 protein and the AS15 immunostimulant, and the induction of specific immune response with or without adjuvant chemotherapy (CT). Methods: MAGE-A3 CI was administered i.m. 8q3w in resected MAGE-A3+ stage IB-III NSCLC patients (pts). Three cohorts (C) were evaluated: Immunization with concurrent cisplatin plus vinorelbine (C1), sequentially after the same CT (C2) or with no CT (C3). The anti-MAGE-A3 humoral and cellular immune responses were evaluated by ELISA and intracellular cytokine staining (T cells producing both IFNg and TNF) respectively. Adverse Events (AEs) were graded according to CTCAE v. 3.0. Results: A total of 38/55 treated pts received the 8 doses schedule (15/19 in C1, 14/18 in C2, 9/18 in C3). Almost all pts reported at least one AE, mostly general constitutional disorders and administration site reactions. Six patients reported related grade 3 AEs. No related grade 4-5 AE or related SAE were reported. Immunogenicity results in the total treated population are reported in the table below. Conclusions: In this trial, patients in cohorts 2 and 3 correspond to the two populations enrolled in the Phase III MAGRIT trial evaluating the same MAGE-A3 CI. The phase I/II results suggest that this CI is well tolerated and induces in all treated pts specific antibodies against MAGE-A3 after 4 doses in presence or not of concurrent or sequential adjuvant CT. About 25% of the treated pts are considered as CD4 responders in presence or not of concurrent or sequential adjuvant CT. [Table: see text]

8535 Background: Active immunization towards MAGE-A3 tumor antigen is a potential new approach to treat cancer. Current data from pilot studies have shown that a strong and persistent immune respon...

Clinical data today highlight the encouraging potential of the MAGE-A3 Antigen-Specific Cancer Immunotherapeutic (ASCI) for cancer treatment. The MAGE-A3 gene encodes a tumor-specific antigen expressed on tumor cells only. When used as a recombinant protein and combined with an immunological Adjuvant System designed to enhance the immune response to the MAGE-A3 antigen, it was shown to induce specific T-cell responses and long-lasting clinical objective responses in metastatic melanoma (Phase II NCT00086866 trial) (Kruit et al. JCO 2008 26:9065), and demonstrated proof-of-concept in Non-Small Cell Lung Cancer (NSCLC; Phase II NCT00290355 trial) (Vansteenkiste et al. JCO 2007 25:7554). In these 2 trials, gene expression profiling by microarrays was used to identify biomarkers predictive of the clinical activity of the MAGE-A3 ASCI. A first analysis carried out on samples from melanoma patients using supervised hierarchical clustering of 2 objective responders and 2 non-responders identified 2 gene clusters based on differential gene expression. This immune gene expression signature was then confirmed in an analysis of 22-patient samples, and independently validated on an additional 30-patient testing set; furthermore these analyses confirmed the association of clinical benefit and the molecular signature. Most of the identified genes are immune-related, defining a particular biological context present in the tumor environment before immunization. This was confirmed by selecting genes using all patients eligible for gene expression profiling. Upon crossvalidation, median overall survival was improved significantly in the population of patients whose tumor presented the gene signature (GS): 28 months in the GS+ population, 16.2 months in the GS−. The predictive value of the melanoma signature was tested in NSCLC. A subset of 49 genes discovered in the melanoma Phase II trial was assayed by qPCR on biopsies taken prior to any ASCI treatment from 137 patients of the lung Phase II trial. Applied to NSCLC patients, this biomarker showed that the relative reduction in the risk of recurrence upon MAGE-A3 ASCI treatment is increased by about 2 fold in the patients with the predictive gene signature as compared to the overall population: from 25% relative improvement in the overall population to 53% in patients whose tumor presents the predictive gene signature. In conclusion, two clinical proof-of-concepts have been obtained with the MAGE-A3 ASCI in 2 different types of tumors, NSCLC and melanoma. More importantly, we have identified a gene expression signature predictive of clinical activity of the MAGE-A3 ASCI treatment. The initiation of Phase III studies in NSCLC (MAGRIT) and melanoma (DERMA) is a unique opportunity to validate prospectively these biomarkers with the ultimate goal to select patients that are the most likely to benefit from the MAGE-A3 ASCI therapy. Citation Information: Mol Cancer Ther 2009;8(12 Suppl):A37.

7501 Background: A phase II randomized trial in 182 patients (pts) with completely resected stage IB or II MAGE-A3 (+) NSCLC comparing postoperative MAGE-A3 recombinant protein combined with an Adjuvant System to placebo (q3w x 5, followed by q3m x 8) showed a strong positive signal for activity (J Clin Oncol 25 Suppl 18:398S, 2007). We report here the analysis of gene expression profiling of tumors prior to treatment to identify a predictive signature that correlates with clinical activity of MAGE-A3 ASCI treatment. Methods: Microarray analysis (Affymetrix) was performed on 159 fresh-frozen tumor biopsies (MAGE-A3: 108; placebo: 51; stage IB: 109; stage II: 50). Clinical data on relapse are based on a median follow-up of 44 months. Results: Study of tumor samples from pts in the placebo arm led to identification of a gene signature associated with a high risk of postoperative relapse. The absence of this signature correlated with a very low relapse rate in stage IB pts: placebo 0%; MAGE-A3: 4%. In the MA...

9065 Background: The MAGE-A3 antigen, expressed in ±60% of metastatic melanomas, is a target of anti-tumor CD4+ and CD8+ T-cells. Previous data showed a few clinical responses of MAGE-A3 Antigen-Sp...

9045 Background: Gene expression profiling by microarrays was used to identify markers predictive of the clinical activity of the MAGE-A3 Antigen-Specific Cancer Immunotherapeutic (ASCI) recorded in a phase II study in metastatic melanoma. Methods: 75 patients with progressive, unresectable stage III or stage IV M1a MAGE-A3 (+) melanomas, were randomized as 1st-line therapy between immunization with MAGE-A3 protein combined with GSK proprietary Adjuvant Systems AS15 or AS02B. Gene expression profiling was performed on tumor biopsies taken prior to any immunization. Results: Initial analysis using supervised hierarchical clustering of 2 objective responders (OR) with 7 non-responders identified 2 gene clusters based on differential expression. The correlation of this gene expression signature (GS) and objective response was further confirmed on 22 patients. Interestingly, mixed response (MR) and stable disease > 6 months (SD) clustered with OR, suggesting a strong association between the identified signatu...