Final Results of a Prospective Evaluation of the Predictive Value of Interim Positron Emission Tomography in Patients With Diffuse Large B-Cell Lymphoma Treated With R-CHOP-14 (SAKK 38/07)

PURPOSE Our main objective was to prospectively determine the prognostic value of [(18)F]fluorodeoxyglucose positron emission tomography/computed tomography (PET/CT) after two cycles of rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisone given every 14 days (R-CHOP-14) under standardized treatment and PET evaluation criteria. PATIENTS AND METHODS Patients with any stage of diffuse large B-cell lymphoma were treated with six cycles of R-CHOP-14 followed by two cycles of rituximab. PET/CT examinations were performed at baseline, after two cycles (and after four cycles if the patient was PET-positive after two cycles), and at the end of treatment. PET/CT examinations were evaluated locally and by central review. The primary end point was event-free survival at 2 years (2-year EFS). RESULTS Median age of the 138 evaluable patients was 58.5 years with a WHO performance status of 0, 1, or 2 in 56%, 36%, or 8% of the patients, respectively. By local assessment, 83 PET/CT scans (60%) were reported as positive and 55 (40%) as negative after two cycles of R-CHOP-14. Two-year EFS was significantly shorter for PET-positive compared with PET-negative patients (48% v 74%; P = .004). Overall survival at 2 years was not significantly different, with 88% for PET-positive versus 91% for PET-negative patients (P = .46). By using central review and the Deauville criteria, 2-year EFS was 41% versus 76% (P < .001) for patients who had interim PET/CT scans after two cycles of R-CHOP-14 and 24% versus 72% (P < .001) for patients who had PET/CT scans at the end of treatment. CONCLUSION Our results confirmed that an interim PET/CT scan has limited prognostic value in patients with diffuse large B-cell lymphoma homogeneously treated with six cycles of R-CHOP-14 in a large prospective trial. At this point, interim PET/CT scanning is not ready for clinical use to guide treatment decisions in individual patients.

Accelerated magnetic resonance diffusion tensor imaging of the median nerve using simultaneous multi-slice echo planar imaging with blipped CAIPIRINHA.

PURPOSE To investigate the feasibility of MR diffusion tensor imaging (DTI) of the median nerve using simultaneous multi-slice echo planar imaging (EPI) with blipped CAIPIRINHA. MATERIALS AND METHODS After federal ethics board approval, MR imaging of the median nerves of eight healthy volunteers (mean age, 29.4 years; range, 25-32) was performed at 3 T using a 16-channel hand/wrist coil. An EPI sequence (b-value, 1,000 s/mm(2); 20 gradient directions) was acquired without acceleration as well as with twofold and threefold slice acceleration. Fractional anisotropy (FA), mean diffusivity (MD) and quality of nerve tractography (number of tracks, average track length, track homogeneity, anatomical accuracy) were compared between the acquisitions using multivariate ANOVA and the Kruskal-Wallis test. RESULTS Acquisition time was 6:08 min for standard DTI, 3:38 min for twofold and 2:31 min for threefold acceleration. No differences were found regarding FA (standard DTI: 0.620 ± 0.058; twofold acceleration: 0.642 ± 0.058; threefold acceleration: 0.644 ± 0.061; p ≥ 0.217) and MD (standard DTI: 1.076 ± 0.080 mm(2)/s; twofold acceleration: 1.016 ± 0.123 mm(2)/s; threefold acceleration: 0.979 ± 0.153 mm(2)/s; p ≥ 0.074). Twofold acceleration yielded similar tractography quality compared to standard DTI (p > 0.05). With threefold acceleration, however, average track length and track homogeneity decreased (p = 0.004-0.021). CONCLUSION Accelerated DTI of the median nerve is feasible. Twofold acceleration yields similar results to standard DTI. KEY POINTS • Standard DTI of the median nerve is limited by its long acquisition time. • Simultaneous multi-slice acquisition is a new technique for accelerated DTI. • Accelerated DTI of the median nerve yields similar results to standard DTI.

Automatic identification of discourse markers in dialogues: An in-depth study of like and well

The lexical items like and well can serve as discourse markers (DMs), but can also play numerous other roles, such as verb or adverb. Identifying the occurrences that function as DMs is an important step for language understanding by computers. In this study, automatic classifiers using lexical, prosodic/positional and sociolinguistic features are trained over transcribed dialogues, manually annotated with DM information. The resulting classifiers improve state-of-the-art performance of DM identification, at about 90% recall and 79% precision for like (84.5% accuracy, κ = 0.69), and 99% recall and 98% precision for well (97.5% accuracy, κ = 0.88). Automatic feature analysis shows that lexical collocations are the most reliable indicators, followed by prosodic/positional features, while sociolinguistic features are marginally useful for the identification of DM like and not useful for well. The differentiated processing of each type of DM improves classification accuracy, suggesting that these types should be treated individually.

Discourse-level Annotation over Europarl for Machine Translation: Connectives and Pronouns

This paper describes methods and results for the annotation of two discourse-level phenomena, connectives and pronouns, over a multilingual parallel corpus. Excerpts from Europarl in English and French have been annotated with disambiguation information for connectives and pronouns, for about 3600 tokens. This data is then used in several ways: for cross-linguistic studies, for training automatic disambiguation software, and ultimately for training and testing discourse-aware statistical machine translation systems. The paper presents the annotation procedures and their results in detail, and overviews the first systems trained on the annotated resources and their use for machine translation.

How comparable are parallel corpora? Measuring the distribution of general vocabulary and connectives

In this paper, we question the homogeneity of a large parallel corpus by measuring the similarity between various sub-parts. We compare results obtained using a general measure of lexical similarity based on χ2 and by counting the number of discourse connectives. We argue that discourse connectives provide a more sensitive measure, revealing differences that are not visible with the general measure. We also provide evidence for the existence of specific characteristics defining translated texts as opposed to non-translated ones, due to a universal tendency for explicitation.

Shallow Dialogue Processing Using Machine Learning Algorithms (or not)

This paper presents a shallow dialogue analysis model, aimed at human-human dialogues in the context of staff or business meetings. Four components of the model are defined, and several machine learning techniques are used to extract features from dialogue transcripts: maximum entropy classifiers for dialogue acts, latent semantic analysis for topic segmentation, or decision tree classifiers for discourse markers. A rule-based approach is proposed for solving cross-modal references to meeting documents. The methods are trained and evaluated thanks to a common data set and annotation format. The integration of the components into an automated shallow dialogue parser opens the way to multimodal meeting processing and retrieval applications.

Towards Automatic Identification of Discourse Markers in Dialogs: the Case of 'Like'

This article discusses the detection of discourse markers (DM) in dialog transcriptions, by human annotators and by automated means. After a theoretical discussion of the definition of DMs and their relevance to natural language processing, we focus on the role of like as a DM. Results from experiments with human annotators show that detection of DMs is a difficult but reliable task, which requires prosodic information from soundtracks. Then, several types of features are defined for automatic disambiguation of like: collocations, part-of-speech tags and duration-based features. Decision-tree learning shows that for like, nearly 70% precision can be reached, with near 100% recall, mainly using collocation filters. Similar results hold for well, with about 91% precision at 100% recall.