Search

Bacterial pathogen identification, which is critical for human health, has historically relied on culturing organisms from clinical specimens. More recently, the application of machine learning (ML) to whole-genome sequences (WGSs) has facilitated pathogen identification. However, relying solely on genetic information to identify emerging or new pathogens is fundamentally constrained, especially if novel virulence factors exist. In addition, even WGSs with ML pipelines are unable to discern phenotypes associated with cryptic genetic loci linked to virulence. Here, we set out to determine if ML using phenotypic hallmarks of pathogenesis could assess potential pathogenic threat without using any sequence-based analysis. This approach successfully classified potential pathogenetic threat associated with previously machine-observed and unobserved bacteria with 99% and 85% accuracy, respectively. This work establishes a phenotype-based pipeline for potential pathogenic threat assessment, which we term PathEngine, and offers strategies for the identification of bacterial pathogens.

In the field of phytohormone defense, the general perception is that salicylate (SA)-mediated defense is induced against biotrophic pathogens while jasmonate (JA)-mediated defense functions against necrotrophic pathogens. Our goals were to observe the behavior of the necrotrophic pathogen Rhizoctonia solani in the vicinity, on the surface, and within the host tissue after priming the host with SA or JA, and to see if priming with these phytohormones would affect the host defense differently upon infection. It was observed for the first time, that R. solani could not only distinguish between JA versus SA-primed tomato plants from a distance, but surprisingly avoided SA-primed plants more than JA-primed plants. To corroborate these findings, early infection events were monitored and compared through microscopy, Scanning Electron Microscopy, and Confocal Laser Scanning Microscopy using transformed R. solani expressing green fluorescence protein gene (gfp). Different histochemical and physiological parameters were compared between the unprimed control, JA-primed, and SA-primed plants after infection. The expression of a total of fifteen genes, including the appressoria-related gene of the pathogen and twelve marker genes functioning in the SA and JA signaling pathways, were monitored over a time course during early infection stages. R. solani being traditionally designated as a necrotroph, the major unexpected observations were that Salicylate priming offered better tolerance than Jasmonate priming and that it was mediated through the activation of SA-mediated defense during the initial phase of infection, followed by JA-mediated defense in the later phase. Hence, the present scenario of biphasic SA-JA defense cascades during R. solani infection, with SA priming imparting maximum tolerance, indicate a possible hemibiotrophic pathosystem that needs to be investigated further.

The modern era of artificial intelligence is mostly driven by Deep Neural Network (DNN). As a result most of the intelligent/smart apps running on edge devices (mobile phones, televisions etc.) use DNN for their predictive ability. DNN based prediction suffers with operational overheads, which is the summation of model loading latency and inference latency. Model loading latency affects the first response of the DNN powered apps, whereas inference latency affects the subsequent responses. As apps switching has become a common practice among edge device users, so it is of utmost interest to make the switching smooth by reducing the model loading latency. In this paper asymmetrical DNN modularization is proposed as a potential solution. The proposed method solves two distinct problems (a). Improves the model loading latency by parallel loading of all the modules (child models) of given DNN model. (b). Helps in on-device training by keeping the live gradients only for last child model. The decision about modularization index and their corresponding positions are taken by reinforcement learning unit (RLU). RLU takes into account the available hardware resources on-device (eg. h/w threads), loading latency of each layer on dedicated compute units. In response, it provides the best modularization index k and their corresponding positions $\vec{p}$ specific to the DNN model and device, where $P=(p_{1},p_{2},\ \ldots,p_{k})$ and $p_{i}$ is the end position of child $i$ . The proposed method has shown significant loading time improvement (up to 7X) on popular DNNs, used for camera use-cases. Along with improving the loading latency the proposed modularization method facilitates for On-device personalization by separating the module with trainable layers and loading them particularly while training on-device.

Counting the number of true instances of a clause is arguably a major bottleneck in relational probabilistic inference and learning. We approximate counts in two steps: (1) transform the fully grounded relational model to a large hypergraph, and partially-instantiated clauses to hypergraph motifs; (2) since the expected counts of the motifs are provably the clause counts, approximate them using summary statistics (in/outdegrees, edge counts, etc). Our experimental results demonstrate the efficiency of these approximations, which can be applied to many complex statistical relational models, and can be significantly faster than state-of-the-art, both for inference and learning, without sacrificing effectiveness.

Summary Background VEGF and VEGF receptor 2 (VEGFR-2)-mediated signalling and angiogenesis can contribute to the pathogenesis and progression of gastric cancer. We aimed to assess whether the addition of ramucirumab, a VEGFR-2 antagonist monoclonal antibody, to first-line chemotherapy improves outcomes in patients with metastatic gastric or gastro-oesophageal junction adenocarcinoma. Methods For this double-blind, randomised, placebo-controlled, phase 3 trial done at 126 centres in 20 countries, we recruited patients aged 18 years or older with metastatic, HER2-negative gastric or gastro-oesophageal junction adenocarcinoma, an Eastern Cooperative Oncology Group (ECOG) performance status of 0 or 1, and adequate organ function. Eligible patients were randomly assigned (1:1) with an interactive web response system to receive cisplatin (80 mg/m2, on the first day) plus capecitabine (1000 mg/m2, twice daily for 14 days), every 21 days, and either ramucirumab (8 mg/kg) or placebo on days 1 and 8, every 21 days. 5-Fluorouracil (800 mg/m2 intravenous infusion on days 1–5) was permitted in patients unable to take capecitabine. The primary endpoint was investigator-assessed progression-free survival, analysed by intention to treat in the first 508 patients. We did a sensitivity analysis of the primary endpoint, including a central review of CT scans. Overall survival was a key secondary endpoint. This study is registered with ClinicalTrials.gov, number NCT02314117. Findings Between Jan 28, 2015, and Sept 16, 2016, 645 patients were randomly assigned to receive ramucirumab plus fluoropyrimidine and cisplatin (n=326) or placebo plus fluoropyrimidine and cisplatin (n=319). Investigator-assessed progression-free survival was significantly longer in the ramucirumab group than the placebo group (hazard ratio [HR] 0·753, 95% CI 0·607–0·935, p=0·0106; median progression-free survival 5·7 months [5·5–6·5] vs 5·4 months [4·5–5·7]). A sensitivity analysis based on central independent review of the radiological images did not corroborate the investigator-assessed difference in progression-free survival (HR 0·961, 95% CI 0·768–1·203, p=0·74). There was no difference in overall survival between groups (0·962, 0·801–1·156, p=0·6757; median overall survival 11·2 months [9·9–11·9] in the ramucirumab group vs 10·7 months [9·5–11·9] in the placebo group). The most common grade 3–4 adverse events were neutropenia (85 [26%] of 323 patients in the ramucirumab group vs 85 [27%] of 315 in the placebo group), anaemia (39 [12%] vs 44 [14%]), and hypertension (32 [10%] vs 5 [2%]). The incidence of any-grade serious adverse events was 160 (50%) of 323 patients in the ramucirumab group and 149 (47%) of 315 patients in the placebo group. The most common serious adverse events were vomiting (14 [4%] in the ramucirumab group vs 21 [7%] in the placebo group) and diarrhoea (11 [3%] vs 19 [6%]). There were seven deaths in each group, either during study treatment or within 30 days of discontinuing study treatment, which were the result of treatment-related adverse events. In the ramucirumab group, these adverse events were acute kidney injury, cardiac arrest, gastric haemorrhage, peritonitis, pneumothorax, septic shock, and sudden death (n=1 of each). In the placebo group, these adverse events were cerebrovascular accident (n=1), multiple organ dysfunction syndrome (n=2), pulmonary embolism (n=2), sepsis (n=1), and small intestine perforation (n=1). Interpretation Although the primary analysis for progression-free survival was statistically significant, this outcome was not confirmed in a sensitivity analysis of progression-free survival by central independent review, and did not improve overall survival. Therefore, the addition of ramucirumab to cisplatin plus fluoropyrimidine chemotherapy is not recommended as first-line treatment for this patient population. Funding Eli Lilly and Company.

Recently, deep models have been successfully adopted in several applications, especially where low-level representations are needed. However, sparse, noisy samples and structured domains (with multiple objects and interactions) are some of the open challenges in most deep models. Column Networks, a deep architecture, can succinctly capture domain structure and interactions, but may still be prone to sub-optimal learning from sparse and noisy samples. Inspired by the success of human-knowledge guided learning in AI, especially in data-scarce domains, we propose Knowledge-augmented Column Networks that leverage human advice/knowledge for better learning with noisy/sparse samples. Our experiments demonstrate that our approach leads to either superior overall performance or faster convergence (i.e., both effective and efficient).

Background: Leukemia inhibitory factor (LIF) is an immunosuppressive cytokine linked to tumor growth and metastasis. LIF overexpression correlates with poor prognosis and chemoresistance in multiple tumor types including PDAC. Preclinical data show that LIF promotes an immunosuppressive tumor microenvironment, hindering cytotoxic CD8+ T cell recruitment; low T cell infiltration also correlates with mortality in patients with PDAC. Preclinical studies show LIF-blocking antibodies sensitize tumors to PD-1/PD-L1 inhibition, inhibit epithelial-mesenchymal transition, and prolong survival in combination with CT. In patients with PDAC, higher LIF levels correlate with more aggressive pathology, elevated CA19-9 (a PDAC biomarker), and lower response and survival rates. AZD0171 (formerly MSC-1), a first-in-class, humanized, IgG1 monoclonal antibody, binds specifically and potently to LIF, preventing downstream signaling. In a phase 1 dose escalation study (NCT03490669), AZD0171 monotherapy had manageable safety and led to stable disease (SD) in 34.2% of patients with advanced solid tumors across all dose levels. AZD0171 + CT may improve survival outcomes vs CT alone. Based on preclinical data, AZD0171 may function to stimulate antitumor immune response, and combination with the PD-L1 inhibitor durvalumab could prolong that response and overcome peripheral tolerance in patients with metastatic PDAC. Methods: This is a phase 2, open-label, single-arm, multicenter study of AZD0171 + durvalumab and CT in treatment-naive patients with metastatic PDAC (NCT04999969). Eligible patients must have an ECOG performance status 0 or 1, a Gustave Roussy Immune Score 0 or 1, ≥1 measurable target lesion per RECIST v1.1, and confirmed presence of tumoral CD8+ T cells. Patients with central nervous system metastasis, history of leptomeningeal disease or cord compression, a thromboembolic event ≤3 months prior to study treatment, unresolved grade ≥2 toxicities from prior therapy, or a sensitizing mutation or tumor characteristic for which there is a preferred treatment, are excluded. As PDAC is poorly immune infiltrated, a novel clinical trial assay will be used to select for patients who have existing resident CD8+ T cells and may therefore be more likely to respond. About 115 patients will receive intravenous AZD0171, durvalumab and CT until disease progression or unacceptable toxicity. The primary endpoints are safety and overall survival (OS) rate at 12 months. Secondary endpoints include objective response rate, disease control rate (confirmed response or SD ≥16 weeks), duration of response, median progression-free survival (PFS), PFS rate at 4 months, median OS, pharmacokinetics, pharmacodynamics (changes in serum CA19-9 level and tumor CD8+ T cell infiltration) and immunogenicity. The trial is currently recruiting. Citation Format: Grainne O'Kane, Teresa Macarulla, Fiyinfolu Balogun, Antoine Hollebecque, Matthew J. Reilley, Sreenivasa R. Chandana, Jim Eyles, Oluwaseun Ojo, Philip Overend, Douglas C. Palmer, Nadia Luheshi, Mayukh Das, Antoine Italiano, Joan Seoane. A phase 2 trial of first-line AZD0171 + durvalumab and chemotherapy (CT) in patients with metastatic pancreatic ductal adenocarcinoma (PDAC) and CD8+ T cell infiltration [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr CT126.

Introduction/Background The immune checkpoints NKG2A and programmed cell death-1 (PD-1) are expressed on both tumour-infiltrating natural killer (NK) and CD8+ T cells in several cancer types, and are implicated in reducing antitumor immune response. To evaluate whether dual targeting of non-redundant checkpoint pathways (NKG2A/HLA-E and PD-1/PD-L1) may enhance antitumour immunity, the combination of monalizumab and durvalumab is being assessed in a Phase 1b expansion study in multiple solid tumours (NCT02671435). Here we report safety and efficacy results in patients with ovarian, cervical or microsatellite-stable (MSS) endometrial cancer. Methodology Eligible patients had advanced recurrent or metastatic high-grade serous epithelial ovarian cancer or cervical cancer (patients in each cohort could have received up to 2 prior lines of systemic therapy) or MSS endometrial cancer (patients could have received up to 3 prior lines of systemic therapy), with Eastern Cooperative Oncology Group performance status 0–1. Patients received monalizumab 750 mg Q2W and durvalumab 1500 mg Q4W for up to 3 years until unacceptable toxicity or confirmed progression. The primary endpoint was safety and tolerability; secondary endpoints included antitumour activity. Results Between March 2017 and March 2019, 40 patients with ovarian cancer (age range, 42–75 years), 16 patients with cervical cancer (age range, 32–79 years) and 40 patients with MSS endometrial cancer (age range, 45–79 years) were enrolled. Rates of treatment-related adverse events (AEs) were generally similar across cohorts (table 1). There were no grade 5 AEs and no events leading to discontinuation of monalizumab or durvalumab. Objective responses were seen only in ovarian cancer (table 2): among the 37 evaluable patients, 2 (5.4%) had confirmed partial responses and 10 (27.0%) had stable disease (SD) including 6 (16.2%) with disease control at 24 weeks (DCR24). Median progression-free survival (mPFS) was 1.8 months and median overall survival (mOS) was 16.7 months. Six (37.5%) of the 16 evaluable patients with cervical cancer had SD; mPFS was 2.0 months and mOS was 8.6 months. Fifteen (38.5%) of the 39 evaluable patients with MSS endometrial cancer had SD, including 5 (12.8%) with DCR24; mPFS was 1.8 months and mOS was 10.7 months. Conclusion Monalizumab plus durvalumab treatment had manageable safety in all cohorts. Modest clinical activity was demonstrated in recurrent ovarian cancer, whereas activity in cervical and MSS endometrial cancers was minimal. Further understanding of dual immune-checkpoint targeting is required. Disclosures Susana Banerjee has received institution research grants from AstraZeneca, Tesaro, and GlaxoSmithKline, and honoraria for advisory boards/lectures from AstraZeneca, Amgen, Clovis, Genmab, GlaxoSmithKline, Immunogen, Mersana, Merck Serono, Merck Sharp & Dohme, Pfizer, Roche, Seattle Genetics, and Tesaro. Ana Oaknin has served on advisory boards for Clovis Oncology, AstraZeneca, Genmab/Seattle Genetics, ImmunoGen, PharmaMar, Roche, GSK/Tesaro, Merck/Merck Sharp & Dohme, Deciphera and Mersana; has received support for travel or accommodation from Clovis Oncology, AstraZeneca, PharmaMar, and Roche; and reports institutional research grant support from Clovis Oncology, AbbVie Deutschland, Ability Pharmaceuticals, Advaxis, Aeterna Zentaris, Amgen, Aprea Therapeutics, Eisai, ImmunoGen, Merck/Merck Sharp & Dohme, Millennium Pharmaceuticals, PharmaMar, Roche, and Tesaro. IS-S: No conflicts to disclose ACS: No conflicts to disclose Sandip Pravin Patel receives scientific advisory income from AstraZeneca, Bristol Myers Squibb, Eli Lilly, Illumina, Rakuten, Paradigm and Tempus, and research funding from Bristol Myers Squibb, Eli Lilly, Incyte, AstraZeneca, Merck, Pfizer, Roche/Genentech, Xcovery, Fate Therapeutics, Genocea and Iovance. AOz: No conflicts to disclose Mayukh Das is an employee of and stockholder in AstraZeneca. Panagiotis Kourtesis is an employee of and stockholder in AstraZeneca. Maria Libera Ascierto is an employee of and stockholder in AstraZeneca. Jennifer R. Diamond reports research funding from AstraZeneca, Immunomedics, OnKure, Bristol Myers Squibb, Merck, Deciphera, Bayer, Taiho and Takeda and consulting work for Immunomedics.

We consider the problem of learning generalized first-order representations of concepts from a small number of examples. We augment an inductive logic programming learner with two novel contributions. First, we define a distance measure between candidate concept representations that improves the efficiency of search for target concept and generalization. Second, we leverage richer human inputs in the form of advice to improve the sample-efficiency of learning. We prove that the proposed distance measure is semantically valid and use that to derive a PAC bound. Our experiments on diverse learning tasks demonstrate both the effectiveness and efficiency of our approach.

In this preset work bio-sourced hydroxyapatite (HAP) from waste egg shells and synthetic hydroxyapatite from calcium hydroxide were synthesized. A comparative analysis and characterization were done using XRD, FTIR, green density, porosity, cytotoxicity and hemolysis studies. It was found that both types of materials are almost similar in their properties and highly biocompatible. The biologically derived egg shell hydroxyapatite can be an ideal substitute for synthetically derived hydroxyapatite.

We develop a pipeline to mine complex drug interactions by combining different similarities and interaction types (molecular, structural, phenotypic, genomic etc). Our goal is to learn an optimal kernel from these heterogeneous similarities in a supervised manner. We formulate an extensible framework that can easily integrate new interaction types into a rich model. The core of our pipeline features a novel kernel-learning approach that tunes the weights of the heterogeneous similarities, and fuses them into a Similarity-based Kernel for Identifying Drug-Drug interactions and Discovery, or SKID3. Experimental evaluation on the DrugBank database shows that SKID3 effectively combines similarities generated from chemical reaction pathways (which generally improve precision) and molecular and structural fingerprints (which generally improve recall) into a single kernel that gets the best of both worlds, and consequently demonstrates the best performance.

One of the key advantages of Inductive Logic Programming systems is the ability of the domain experts to provide background knowledge as modes that allow for efficient search through the space of hypotheses. However, there is an inherent assumption that this expert should also be an ILP expert to provide effective modes. We relax this assumption by designing a graphical user interface that allows the domain expert to interact with the system using Entity Relationship diagrams. These interactions are used to construct modes for the learning system. We evaluate our algorithm on a probabilistic logic learning system where we demonstrate that the user is able to construct effective background knowledge on par with the expert-encoded knowledge on five data sets., Comment: 8 pages. Published in Proceedings of the Knowledge Capture Conference, 2017

Arsenic (As) being present in the groundwater of many regions has significant detrimental effects on the health of the local people. Moreover, in current scenario iron ore mining residues have detrimental effects on the environment besides occupying space. Reject of iron ore slime (RIOS) has high adsorption properties irrespective of temperature and pH. The simulation study has confirmed that As adsorption by RIOS follows Langmuir isotherm and after 24 h residence time more than 90% As can be removed. Moreover, the effect of variation both As and RIOS has been understood. Based on the simulation study, a desired RIOS concentrated slurry was injected into groundwater reservoir for in-situ As removal and after 20 days nearly 90% As was entrapped in RIOS. Thus, the present study ensured that RIOS may be used as an alternative adsorbent for geotechnical cleaning of As from groundwater.

The present study aims at finding out the implications of seasonal precipitation in bringing about habitat variability in Sundarbans, the world’s largest mangrove ecoregion. This work takes into consideration datasets of 2010 and 2013 from a long-term dataset that is presently being collected for Sundarbans mangrove ecoregion (SBOTS—Sundarbans Biological Observatory Time Series). The purpose of selecting these two datasets was mainly because of the differences in seasonal precipitation which was significantly higher in 2013 as compared to 2010. The study area was selected at the confluence of a creek and estuary where tidal variations were pronounced due to the proximity of the Bay of Bengal. The two stations were specifically selected to spatially segregate the effects of seasonal precipitation and tidal influences especially with regard to habitat variability. Analysis of abiotic variables and nutrient status by a combination of quantitative estimation and statistical methods clearly suggested that habitat variability of the study area was brought about by seasonal precipitation mainly with regard to nitrate concentrations and pH. However, salinity variations remained independent of seasonal precipitation, underlying the importance of tidal fluctuations in this area. The biotic community (phytoplankton) was dominated by diatoms with intermittent availability of chlorophytes and dinoflagellates showing both spatial and temporal changes. Thus, seasonal precipitation along with anthropogenic activities played important role in bringing about habitat variability of the study area with a concomitant effect on spatial and temporal patterns in phytoplankton population in Sundarbans.

The xylem-limited bacterium Xylella fastidiosa is the causal agent of several plant diseases, most notably Pierce's disease of grape and citrus variegated chlorosis. We report the isolation and characterization of the first virulent phages for X. fastidiosa, siphophages Sano and Salvo and podophages Prado and Paz, with a host range that includes Xanthomonas spp. Phages propagated on homologous hosts had observed adsorption rate constants of ~4 × 10(-12) ml cell(-1) min(-1) for X. fastidiosa strain Temecula 1 and ~5 × 10(-10) to 7 × 10(-10) ml cell(-1) min(-1) for Xanthomonas strain EC-12. Sano and Salvo exhibit >80% nucleotide identity to each other in aligned regions and are syntenic to phage BcepNazgul. We propose that phage BcepNazgul is the founding member of a novel phage type, to which Sano and Salvo belong. The lysis genes of the Nazgul-like phage type include a gene that encodes an outer membrane lipoprotein endolysin and also spanin gene families that provide insight into the evolution of the lysis pathway for phages of Gram-negative hosts. Prado and Paz, although exhibiting no significant DNA homology to each other, are new members of the phiKMV-like phage type, based on the position of the single-subunit RNA polymerase gene. The four phages are type IV pilus dependent for infection of both X. fastidiosa and Xanthomonas. The phages may be useful as agents for an effective and environmentally responsible strategy for the control of diseases caused by X. fastidiosa.

Full-text article is free to read on the publisher website. INTRODUCTION In retrospective analyses of patients with nonsquamous non-small-cell lung cancer treated with pemetrexed, low thymidylate synthase (TS) expression is associated with better clinical outcomes. This phase II study explored this association prospectively at the protein and mRNA-expression level. METHODS Treatment-naive patients with nonsquamous non-small-cell lung cancer (stage IIIB/IV) had four cycles of first-line chemotherapy with pemetrexed/cisplatin. Nonprogressing patients continued on pemetrexed maintenance until progression or maximum tolerability. TS expression (nucleus/cytoplasm/total) was assessed in diagnostic tissue samples by immunohistochemistry (IHC; H-scores), and quantitative reverse-transcriptase polymerase chain reaction. Cox regression was used to assess the association between H-scores and progression-free/overall survival (PFS/OS) distribution estimated by the Kaplan-Meier method. Maximal χ analysis identified optimal cutpoints between low TS- and high TS-expression groups, yielding maximal associations with PFS/OS. RESULTS The study enrolled 70 patients; of these 43 (61.4%) started maintenance treatment. In 60 patients with valid H-scores, median (m) PFS was 5.5 (95% confidence interval [CI], 3.9-6.9) months, mOS was 9.6 (95% CI, 7.3-15.7) months. Higher nuclear TS expression was significantly associated with shorter PFS and OS (primary analysis IHC, PFS: p < 0.0001; hazard ratio per 1-unit increase: 1.015; 95%CI, 1.008-1.021). At the optimal cutpoint of nuclear H-score (70), mPFS in the low TS- versus high TS-expression groups was 7.1 (5.7-8.3) versus 2.6 (1.3-4.1) months (p = 0.0015; hazard ratio = 0.28; 95%CI, 0.16-0.52; n = 40/20). Trends were similar for cytoplasm H-scores, quantitative reverse-transcriptase polymerase chain reaction and other clinical endpoints (OS, response, and disease control). CONCLUSIONS The primary endpoint was met; low TS expression was associated with longer PFS. Further randomized studies are needed to explore nuclear TS IHC expression as a potential biomarker of clinical outcomes for pemetrexed treatment in larger patient cohorts. © 2013 by the International Association for the Study of Lung Cancer.

Over the past decade, exploiting relations and symmetries within probabilistic models has been proven to be surprisingly effective at solving large scale data mining problems. One of the key operations inside these lifted approaches is counting be it for parameter/structure learning or for efficient inference. Typically, however, they just count exploiting the logical structure using adhoc operators. This paper investigates whether ‘Compilation to Graph Databases’ could be a practical technique for scaling lifted probabilistic inference and learning methods. We demonstrate that the proposed approach achieves reasonable speed-ups for both inference and learning, without sacrificing performance.

Objectives: In countries where cholera is endemic, Vibrio cholerae O1 bacteriophages have been detected in sewage water. These have been used to serve not only as strain markers, but also for the typing of V. cholerae strains. Vibriophage N4 (ATCC 51352-B1) occupies a unique position in the new phage-typing scheme and can infect a larger number of V. cholerae O1 biotype El Tor strains. Here we characterized the complete genome sequence of this typing vibriophage. Methods: The complete DNA sequence of the N4 genome was determined by using a shotgun sequencing approach. Results: Complete genome sequence explored that phage N4 is comprised of one circular, double-stranded chromosome of 38,497 bp with an overall GC content of 42.8%. A total of 47 open reading frames were identified and functions could be assigned to 30 of them. Further, a close relationship with another vibriophage, VP4, and the enterobacteriophage T7 could be established. DNA-DNA hybridization among V. cholerae O1 and O139 phages revealed homology among O1 vibriophages at their genomic level. Conclusions: This study indicates two evolutionary distinctive branches of the possible phylogenetic origin of O1 and O139 vibriophages and provides an unveiled collection of information on viral gene products of typing vibriophages.

Serious pandemics of cholera have occurred throughout the known history of mankind, especially in India, which is a motherland for cholera disease. For the last 20 years several DNA-based typing methods have been employed to study the clonal relatedness between Vibrio cholerae isolates irrespective of their geographical locations. Traditional typing methods, such as biochemical tests, phage typing, serotyping, biotyping and antimicrobial susceptibility tests, have produced reliable and informative data regarding V. cholerae for a long time. Gradually molecular typing techniques have taken the place of traditional typing methods because they produce the same results upon repeat testing of V. cholerae strain. In this article we focus on the discriminatory power of different DNA fingerprint techniques that are generally used to know the homogeneity and heterogeneity among different V. cholerae isolates.

4044Background: The global, randomized, placebo (PL)-controlled phase 3 RAINFALL trial evaluated if addition of ramucirumab (RAM) to first-line (1L) cisplatin plus fluoropyrimidine (chemo) improved...

5 Background: Ramucirumab, a VEGFR-2 IgG1 human monoclonal antibody, is the only biologic with proven efficacy as both a single-agent and in combination with paclitaxel in the second-line treatment of G-GEJ adenocarcinoma. RAINFALL (NCT02314117) is a global, double-blind, placebo-controlled randomized clinical trial addressing the hypothesis that adding ramucirumab to first-line Cis plus Cape or 5-fluorouracil (5FU) produces significant clinical benefit. Methods: Metastatic G-GEJ cancer patients eligible for first-line chemotherapy with ECOG performance status 0-1 were randomized 1:1 to receive either RAM (8 mg/kg iv D1, D8) or placebo (PL), every 21d. All patients received Cape (or 5FU)+Cis. Cis was given for up to 6 cycles. Cape+RAM/placebo was continued until progressive disease, toxicity or other discontinuation criteria. The primary endpoint was progression-free survival (PFS) for the first 508 patients; overall survival (OS) for ITT population was a powered secondary endpoint. Results: 645 patients were randomized to receive RAM+Cape/Cis (n=326) or PL+Cape/Cis (n=319). PFS was significantly prolonged in patients treated with RAM+Cape/Cis versus PL+Cape/Cis (HR, 0.75; 95% CI 0.61–0.94; p=0.011; median, 5.7 vs 5.4 mos), meeting the primary endpoint. There was no survival benefit for patients treated with RAM+Cape/Cis versus PL+Cape/Cis (HR, 0.96; 95% CI 0.80–1.16; p=0.68; median, 11.2 vs 10.7 mos). ORR in the ITT population was 41.1% in the RAM arm (95% CI 35.8–46.4) and 36.4% (95% CI 31.1–41.6) in the PL arm. Grade ≥3 adverse events in ≥10% of patients in the RAM arm were: neutropenia (26.3% RAM; 27.0% PL), anemia (12.1% RAM; 14.0% PL), and hypertension (9.9% RAM; 1.6% PL). No new safety signals were observed. Conclusions: In treatment-naïve patients with metastatic G-GEJ adenocarcinoma, the addition of ramucirumab to first-line chemotherapy conferred a significant 25% reduction in the risk of disease progression or death in the primary endpoint of PFS; however, ramucirumab was not associated with an improved OS. Clinical trial information: NCT02314117.

Objectives: We have isolated a total of five newer cholera phages which are novel broad host range to incorporate with the existing phage typing schemes for an extended typing scheme. Materials and Methods: These newly isolated phages were well characterized including the electron micrograph. A total of 300 Vibrio cholerae strains were isolated from the different endemic region in India were included in phage typing study. Results: These phages were found different from the existing phages. Electron microscopic results showed that the phages belonged to myophage and podophage group. Characterization of the phages based on pH, temperature, and organic solvent sensitivity showed differences among the phages used in this study. All the strains of Vibrio O1 were typeable (100%) with the five set of cholera phages. Of these, 40% strains were clustered under Type-1. Conclusion: The newer Vibrio phages are novel and broad host range and will be useful to incorporate with the existing phage typing system for more precisely discriminate the strains of Vibrio cholerae.

Cholera continues to be an important public health problem in developing countries, including India. This study concerns the feasibility of possible exploitation of bacteriophages as a biocontrol agent to eliminate the pathogen Vibrio cholerae from the gut using the removable intestinal tie-adult rabbit diarrhoea (RITARD) model. A control rabbit challenged with 10(9) colony-forming units (CFU)/mL of V. cholerae MAK 757 developed Grade II to IV diarrhoea, but the phage-treated rabbit that received 10(9)CFU/mL MAK 757 and 10(8) plaque-forming units (PFU)/mL cocktail phages produced only Grade II diarrhoea. Histological results revealed that in the control rabbit (MAK 757-treated), villi lost their normal shape and showed more inflammatory cellular infiltration in the lamina propria compared with the experimental rabbit. Our data suggest that phages could be valuable as prophylaxis against V. cholerae infection.

Cholera is a public health threat in all developing countries. Kolkata, a city in eastern India, is an endemic zone for cholera. During the course of a comprehensive investigation on the distribution of phages of Vibrio cholerae O1 and O139 in freshwater bodies in Kolkata, we were able to isolate the phages of V. cholerae O1 and O139. Vibrio cholerae O1 phages were found at all the sites and exhibited a distinct seasonal cycle, with a primary peak (13.6–17.2 PFU mL−1) during monsoon (June to August) in both 2006 and 2007. Vibrio cholerae O139 phages were present in the environment and were predominant during monsoon in the year 2006, except for late winter and early summer from February to April. In contrast, in the year 2007, the O139 phages could be isolated only during July to December, with the highest counts of 12.0 PFU mL−1 determined in August. The multiplex PCR results showed that 90 samples were positive for wbe of V. cholerae O1, 32 samples for O139 ( wbf ) and 18 samples for both. This study shows that surveillance of vibriophages indicates the presence of V. cholerae O1 and O139 in water bodies in and around Kolkata and could therefore serve as a powerful biomonitoring agent.

Chemoresistance is mediated, in part, by the inhibition of apoptosis in tumor cells. Survivin is an antiapoptotic protein that blocks chemotherapy-induced apoptosis. To investigate whether blocking survivin expression enhances docetaxel-induced apoptosis in patients with non–small-cell lung cancer (NSCLC), we compared the antitumor activity of the survivin inhibitor LY2181308 plus docetaxel with docetaxel alone. We used change in tumor size (CTS) as a primary endpoint to assess its use in early decision-making for this and future studies of novel agents in NSCLC. Patients (N = 162) eligible for second-line NSCLC treatment (stage IIIB/IV) with an Eastern Cooperative Oncology Group performance status of 0 to 1 were randomized 2:1 to receive LY2181308 (750 mg intravenously, weekly) and docetaxel (75 mg/m2 intravenously, day 1) or docetaxel alone every 21 days. CTS from baseline to the end of cycle 2 was compared between the two treatment arms. The mean (SD) tumor size ratio for LY2181308/docetaxel and docetaxel was 1.05 (0.21) and 1.00 (0.15) (p = 0.200), respectively, suggesting no significant improvement in antitumor activity between the arms. Because there was also no significant difference between the two arms for progression-free survival (PFS) (2.83 months with LY2181308/docetaxel and 3.35 months with docetaxel [p = 0.191]), both arms were combined. Using the combined arms, CTS correlated with PFS (PFS = 4.63 months in patients with decreased CTS compared with 2.66 months in patients with increased CTS), supporting its use in early decision-making in phase II studies.

Xylella fastidiosa subsp. fastidiosa (Xff) is the causal agent of Pierce's Disease (PD) of grapevines and is vectored by the glassy-winged sharpshooter (GWSS, Homalodisca vitripennis). Previously we have reported the development of a bacteriophage (phage) based biocontrol system for PD, but no information on insect transmission of phages has been reported. Here we communicate that laboratory reared GWSSs fed on cowpea plants (Vigna unguiculata subsp. unguiculata) harboring the virulent phage Paz were able to uptake of phage efficiently when the phage was present in high concentration, but were inefficient in transfer to plants.

TPS178 Background: Ramucirumab, a human IgG1 monoclonal antibody directed to the ectodomain of VEGFR-2, prevents ligand binding to the receptor, blocking activation of downstream receptor-mediated pathways. Ramucirumab has demonstrated significant improvement in overall survival (OS) and progression-free survival (PFS) in 2 phase III registration studies (REGARD, RAINBOW) in patients in second-line treatment of gastric cancer. This global phase III trial will compare PFS in patients with HER2-negative, metastatic gastric or GEJ adenocarcinoma receiving ramucirumab with cisplatin/capecitabine (or 5-FU) versus placebo with cisplatin/capecitabine (or 5-FU) as first-line treatment. The trial is conducted in 137 sites in the Americas, Europe and Japan and is currently open to enrollment. Methods: Eligible patients will be randomized to receive ramucirumab (8mg/kg on days 1 and 8, based upon population pharmacokinetic modelling) or placebo with cisplatin/capecitabine every 21-day cycle until disease progression, unacceptable toxicity, or other withdrawal criteria are met. The primary endpoint is PFS; OS is the key secondary endpoint. Efficacy will be considered at 3 analysis points: futility analysis for PFS, primary analysis of PFS & final analysis of OS. A gatekeeping strategy will be used to assess PFS and OS. The OS endpoint will only be tested if the PFS test is significant to control Type I error at 5% across both endpoints. An exposure/safety analysis will be done after 60 patients have started the 3rd cycle. The study has 90% power to demonstrate a PFS advantage assuming HR = 0.70 and 80% power to demonstrate an OS advantage assuming HR = 0.77. Other secondary endpoints include PFS2 (the time from randomization to disease progression after the start of additional systemic anticancer treatment, or death from any cause, whichever occurs first), objective response rate, safety and quality of life. As of 9/11/2015, 128 patients have been enrolled in 19 countries. The 1st exposure/safety analysis is underway. Clinical trial information: NCT02314117.

The growing incidences of secondary malignancies in long-term survivors of childhood leukemia following allogeneic bone marrow transplantation (alloBMT) are increasingly being reported. Among the late complications of conventional myeloablative alloBMT, the occurrence of secondary malignant solid tumors is of major concern. Secondary malignant and benign brain tumors such as astrocytoma, meningioma and glioblastoma have been described in long-term survivors of conventional myeloablative alloBMT. Here we report a case of secondary anaplastic oligoastrocytoma that developed 7 years after matched unrelated alloBMT for relapsing childhood acute myeloid leukemia (AML) with CNS involvement. Although isolated CNS relapse of primary leukemia following alloBMT is not uncommon, it is important to identify and define potential risk factors that may lead to the development of secondary brain tumors in children who received high-dose chemotherapy and irradiation prior to alloBMT presenting with progressive neurological symptoms and to differentiate them from leukemia relapse with CNS involvement.

Granulocytic sarcoma (GS) in the absence of bone marrow involvement at diagnosis is unusual. Treatment of primary (aleukemic) GS should follow intensive acute myeloid leukemia (AML)-style chemotherapy protocols, as patients otherwise appear destined to rapidly declare frank medullary disease. However, the role of hematopoietic stem cell transplantation (HSCT) is less clear. Here we present an unusual case of isolated primary GS involving small bowel treated with chemotherapy consolidated by HSCT from an unrelated donor. Full donor chimerism was achieved and maintained, without relapse of the original disease. However, 28 months post-transplant our patient developed a distinct and rapidly progressive medullary AML, apparently derived from transplanted cells of donor origin. Donor cell leukemia (DCL) is a fascinating and increasingly recognized phenomenon, with at least 50 cases reported to date. Our case adds to this growing literature and highlights the ethical dilemmas concerning notification of unrelated donors in such cases, since there is little evidence that they are at increased risk of developing leukemia themselves. This also represents the first case of DCL reported after transplantation for primary GS.

Vibrio cholerae O1, O139 and occasionally non-O1/non-O139 serogroups are most often responsible for epidemic and pandemic cholera. This study used genotypic patterns of PCR-based detection of virulence-associated and regulatory protein genes, along with phage typing, to characterize 86 V. cholerae strains. Thirty-eight of 53 O1 biotype El Tor strains harboured both tcpA classical and tcpA El Tor genes, and three El Tor strains lacked the V. cholerae O1-specific gene (Vc-O1); three O139 strains contained both Vc-O1 and Vc-O139 genes and seven out of ten non-O1/non-O139 strains possessed the Vc-O1 gene. The latter strains all harboured the virulence-associated genes ctxA, zot, ace, RS1, hlyA, ompU, rtxA and sxt. Two phage types, T27 and T25, were predominant in strains from different geographical regions of India, whereas more variation in phage susceptibility was observed for tetracycline-resistant strains from Kolkata. These results suggest that the pattern and distribution of virulence genes and phage types of V. cholerae are equally useful and discriminatory in tracing the origin of newly emerging strains.

Retrospective analysis led to the detection of twoVibrio choleraevariant O1 strains (VC51 and VC53), which were isolated in 1992 in Kolkata from clinical cases, with identical traits to 2004 Mozambique variant O1 strains. The Mozambique O1 strains that caused a huge outbreak in 2004 have been shown to have phenotypic traits of both classical and El Tor biotypes, and thereby have been reported as variant. Our study demonstrated that two O1 strains isolated in Kolkata during 1992 were of the El Tor background as evidenced by polymyxin B (50 U ml−1) resistance, positivity in Voges–Proskauer reactions and sensitivity to biotype-specific vibrio phages. With the features of classical CTX prophage, localization in the small chromosome, and an absence of RS1 and pTLC, both Mozambique and Kolkata strains appeared to be identical. Furthermore, two Kolkata strains exhibited an identical ribotype to that of the Mozambique variant, displaying ribotype pattern RI that had been assigned to KolkataV. choleraeO1 strains isolated on or before 1992.NotI pulsotype analysis indicated that these 1992 Kolkata strains along with the Mozambique variant O1 belonged to very closely related clones. Considering the chronological events, and the typical identity at the phenotypic and the genotypic level between the two O1 strains isolated during 1992 from Kolkata and during 2004 from Mozambique, we propose that some of the 1992 Kolkata O1 strains might have acted as progenitors for Mozambique variant O1 strains.

Summary Objectives To determine whether Vibrio cholerae strains with similar phage types are also genetically homogeneous or heterogeneous by molecular typing methods like PFGE and RAPD-PCR employed in this study. Methods A total of 26 V. cholerae O1 and O139 strains received from different parts of India were analyzed by using conventional phage typing method, antibiogram and molecular typing methods such as PFGE and RAPD-PCR. Results Both O1 and O139 strains were resistant against two antibiotics (Ampicillin and Furazolidone) were detected. All of these strains were clustered in a single phage type, i.e., Type 27 for V. cholerae O1 and Type 1 for V. cholerae O139. Extensive molecular characterization by RAPD and PFGE showed that six sets of O1 and O139 strains, each comprising two strains, had identical PFGE and RAPD profiles. Only one O139 strain (PL-4) had unique RAPD and PFGE profile among all the 26 V. cholerae strains used in this study. Conclusion Apart from serology, the strains of V. cholerae can be discriminated by this conventional phage typing system that offers the basic information on identification, biotyping and discrimination of strains. But, a high level of heterogeneity was observed in RAPD and PFGE profiles indicating the clonal diversity of V. cholerae O1 and O139 strains. It was concluded that these strains were phenotypically identical through phage typing system and antibiogram but genetically dissimilar, as shown in molecular typing systems.

Introduction We report exploratory gene-expression profiling data from a single-arm Phase-II-study in patients with non-squamous (ns)NSCLC treated with pemetrexed and cisplatin. Previously disclosed results indicated a significant association of low thymidylate-synthase (TS)-expression with longer progression-free and overall survival (PFS/OS). Methods Treatment-naive nsNSCLC patients (IIIB/IV) received 4 cycles of pemetrexed/cisplatin; non-progressing patients continued on pemetrexed-maintenance. Diagnostic tissue-samples were used to assess TS-expression by immunohistochemistry (IHC) and mRNA-expression array-profiling (1,030 lung cancer-specific genes). Cox proportional-hazard models were applied to explore the association between each gene and PFS/OS. Genes significantly correlated with PFS/OS were further correlated with TS-protein expression (Spearman-rank). Unsupervised clustering was applied to all evaluable samples (n = 51) for all 1,030 genes and an overlapping 870-gene subset associated with adenocarcinoma (ADC, n = 47). Results 51/70 tissue-samples (72.9%) were evaluable; 9 of 1,030 genes were significantly associated with PFS/OS (unadjusted p

Allogeneic bone marrow transplant recipients are prone to pulmonary infections caused by a wide spectrum of organisms. Since the first bone marrow trarnsplatation (BMT) done in 1983 at the Tata Memorial Hospital, we have recently seen the first case of Mycobacterium Fortuitum Chelonae complex among 117 BMT (including 90 allogeneic and 27 autologous) patients. The patient was on immunosuppressants for chronic GVHD post allogeneic BMT done for CML-CP. He developed pulmonary mycobacterial infection 13 months post BMT. Diagnosis was difficult because of the atypical presentation, negative culture reports, and the presence of multiple pathogens due to immunosuppression. In our case the diagnosis was eventually established after examination of material obtained by bronchoscopy. Patient has shown response to antituberculosis drugs after 2 months. This shows the need to consider atypical mycobacterial infection in the differential diagnosis of pulmonary illness in the post allogeneic BMT setting.

8051^ Background: Resistance to chemotherapy in progressive NSCLC is associated with overexpression of antiapoptotic proteins including survivin. Down-regulation of survivin can sensitize NSCLC to DO in vitro and in xenograft studies. On the basis of preclinical/phase I results we examined antitumor activity of DO+LY compared with DO alone. Methods: Key eligibility criteria: ECOG PS 0-1, stage IIIB/IV NSCLC all histologies, progression after first-line platinum regimen. Patients randomized (N=180) 2:1 to receive DO+LY (LY 750 mg IV loading ×3, Q1W maintenance) or DO alone (75 mg/m2 D1Q3W) until progression/toxicities. Antitumor activity was compared using CTS from baseline to end of cycle (C) 2 in each arm. This analysis, which uses tumor measurements as a continuous variable rather than a categorical endpoint based on RECIST, increases statistical efficiency and enables early assessment of clinical benefit. Secondary objectives included assessment of toxicity, PK, PFS and OS. Results: 114 patients received study drug and were included in the analyses. Baseline patient demographics were similar. No statistically significant difference in mean CTS ratio at C2 or in PFS was observed between the 2 arms: CTS was 1.07 with LY+DO (SD, 0.28) and 1.04 with DO (SD, 0.28); median PFS was 2.83 mo (95% CI, 1.84–3.65) with LY+DO and 3.35 mo (95% CI, 2.69–4.57) with DO (log-rank p=0.191). However, Cox regression revealed CTS to be a statistically significant factor for PFS: decreased CTS was associated with increased PFS (HR 0.45; 95% CI, 0.30–0.68, p=0.0001). Median OS was 7.9 mo (90% CI, 6.6–9.7) with LY+DO and 8.8 mo (90% CI, 5.7–13.8) with DO (log-rank p=0.481). There were no differences in toxicities or other secondary parameters. Incidence of grade III/IV toxicities was similar in both arms and was consistent with the known profile of LY+DO as was the PK profile. Conclusions: Addition of LY to DO did not improve the antitumor activity of DO. CTS appears to be a useful endpoint in phase II studies and should be considered further as an endpoint for early decision-making. Clinical trial information: NCT01107444.

7583 Background: Pem efficacy in non-squamous (NS) NSCLC is hypothesised to be associated with TS expression. Our primary objective was to assess the correlation between TS and PFS, prospectively. Methods: Context: A Phase II, single-arm, exploratory, multicenter, UK study with appropriate approvals and consents. Key eligibility criteria: ECOG PS 0-1, NS-NSCLC histology, stage IIIB/IV. Patients (n=70) received pem/cisplatin induction x 4. Non-progressive patients continued on maintenance pem until tumor progression or discontinuation. All patients were followed until death or study closure. TS by IHC (H-scores) and qPCR (delta CQ values normalized) were assessed on diagnostic FFPE samples. Kaplan-Meier estimates for median PFS (mPFS) and Cox regression to determine the statistical correlation between PFS and TS IHC nuclear score, (continuous variable) were performed. Maximal Chi-square analysis identified the optimal association cutpoints between PFS and low vs. high TS scores. Results: Patient demographics were unremarkable, 32/70 (45.7%) were female. 55 (78.6%) had adenocarcinoma; 15 were not otherwise specified. Maintenance pem was started in 43/70 (61.4%) patients. Evaluable patients had at least one dose of study treatment and a valid TS score (n=60). For the evaluable population, the mPFS from start of induction was 5.5 months (95% CI 3.9-6.9). A statistically significant correlation between PFS and TS IHC nuclear scores was observed [p

TPS293 Background: Nonsquamous (NS-) NSCLC is a predictive factor for response to pemetrexed (pem). Laboratory evidence suggests that this differential outcome may be related to low TS mRNA expression in NS-NSCLC. In this exploratory phase II trial the primary objective ( PO) is to test prospectively, the correlation between TS protein expression (TSPE) and progression free survival (PFS). This abstract reports the TS assay (TSA) protocol to measure TSPE using IHC. Methods: An open-label, single-arm, multicentric study (with appropriate UK approvals) was undertaken to test the PO in eligible patients (n=68) receiving pem/cisplatin induction followed by maintenance pem until tumour progression (PD). Key eligibility criteria included ECOG PS 0-1, histological diagnosis of stage IIIB/IV NS-NSCLC, and informed consent. The TSA protocol for this study was designed on a validated procedure developed by study sponsors. Diagnostic slides and FFPE blocks are centrally collected for diagnostic verification and TS s...

Monoclonal intact immunoglobulin is detectable in the serum of approximately 80% of myeloma patients while SFLC are found to be abnormal in the serum of greater than 90% of patients. One advantage of monitoring SFLC is that they have a half-life of clearance from the serum of 2–6 hours, compared with 20 days for IgG, and can therefore reveal the response to treatment more rapidly. The aim of this study is to determine whether SFLC measurements allow earlier assessment of response to treatment with bortezomib with possible implications to the design and cost of treatment strategy.Patients & Methods Serial serum samples were collected for simultaneous measurement of SFLC and serum paraprotein. Data from 8 patients receiving 2–6 cycles of bortezomib, on compassionate basis, to treat multiply treated myeloma at various stages of relapse are presented. Bortezomib was administered at a dose of 1.3mg/m2 on days 1,4,8 and11. Samples were kept frozen until thawed for SFLC measurement (using Freelite kits; Binding Site, UK on an Olympus Chemistry analyser). Results & Discussion Six of 8 patients showed a response to treatment by EBMT/Blade criteria (>25% fall in paraprotein). In 3 of these patients the tumour SFLC showed a fall and rapid recovery within 10–20 days of a treatment cycle. These patients were monitored over multiple treatment cycles and showed repeated falls in SFLC co-incident with treatment and recovery in between. This appears to correspond with the mechanism of action and biological half life of proteosome inhibition. Recovery of SFLC when seen is rapid (doubling in Conclusion: We conclude that monitoring SFLC provides a unique opportunity to follow the kinetics of tumour kill, which is obscured by the slow clearance of intact immunoglobulin monoclonal proteins. The pattern of fluctuation in the SFLC levels could suggest a temporary inhibition of tumour protein synthesis rather than tumour kill and re-growth. While extension of this study is ongoing, it does indicate that SFLC can be used as a biomarker to assess response to treatment significantly earlier allowing relevant changes of treatment strategy. This also may have major bearing on cost of treatment and utilisation of resources.

The most important consideration in modern SCT practices is the fine balance that is required to achieve adequate immunosuppression and stable engraftment thus allowing the platform to mount an effective GVL response. Though CsA facilitates this objective, it is associated with several adverse clinical manifestations due to its narrow therapeutic index (TI) and wide variability in the absorption kinetics. Weekly monitoring of trough CsA levels are used to adjust the dose. The aim of this study is to determine whether daily monitoring of CsA levels enables better achievement of these goals and yet reduce the serious toxicities. Patients & methods: Twenty three patients (M=15, F=8, median age 46yrs range 19–63) received an allogeneic transplant (sibling-15, MUD-7 mismatched unrelated-1) since March 2004 when we started to monitor daily CsA levels (Biostat TDX analyser using fluorescent immunoassay) from the day of first administration until the time of discharge. The transplants included full intensity (FI=11), reduced intensity (RI=11) or customised (1) for various haematolymphoid malignancies. Patients received uniform GVHD prophylaxis with IV CsA (RI+FI) and methotrexate (FI) as per standard loading doses. The CsA target level was adjusted to 200–220 ng/ml from Day-1 onwards and was later modified according to the treatment strategy. Oral CsA administration was started when patients were able to switch over. Results Nine of 23 (40%) patients developed acute GVH (n=7 grade1–2, n=2 grade 3–4) at a median of 23.5 days (10–28). Two of 9 patients required addition of steroids by Day+30. The median CsA level at the first occurrence of GVH was 247ng/ml (113–333) and the median CsA level during hospitalisation was 207 ng/ml (136–291). Only one patient who received T cell addback required increased dose of IV CsA to manage Grade 4 GVH. Twenty two/22 patients engrafted with a median day to neutrophil recovery being 13 days (8–35). One patient lost his graft due to a low initial cell dose and the inability of the MUD donor to donate any further cells. There was no significant difference in the baseline and Day30 creatinine levels (median highest creatinine 108μmol/l range 58–267) or bilirubin levels (median highest bilirubin 27μmol/l, range15–251). Five/22 patients developed CMV reactivation/infection and 1 developed RSV infection. No patients developed seizures and in one patient pre-existent neuropathy was exacerbated which was directly attributable to CsA. Twenty one/22 patients (95.5%) switched to oral CsA by Day+25. In 8/22 (36.36%) patients, the target CsA levels was around 100–150 ng/ml at the time of discharge. The median duration of hospital stay was 35days (24–130). Conclusions: These observations suggests that daily monitoring of CsA is safe and successful. This allowed successful engraftment and a more liberal management of Grade 1–2 GVH thus allowing us to maximise the GVL effect. Despite concomitant use of other toxic drugs, side effects were minimal and reversible. Close observation of early clinical manifestations enabled us to modify the immunosuppressive strategy at later time periods. This approach also allowed us to better manage patients suffering from CMV and EBV/PTLD (patient number 10 on day+88) by reducing the dose during active infections.

Agents that communicate back and forth with humans to help them execute non-linguistic tasks are a long sought goal of AI. These agents need to translate between utterances and actionable meaning representations that can be interpreted by task-specific problem solvers in a context-dependent manner. They should also be able to learn such actionable interpretations for new predicates on the fly. We define an agent architecture for this scenario and present a series of experiments in the Blocks World domain that illustrate how our architecture supports language learning and problem solving in this domain.