CytoProcessorTM: A New Cervical Cancer Screening System for Remote Diagnosis

BACKGROUND

Current automated cervical cytology screening systems still heavily depend on manipulation of glass slides. We developed a new system called CytoProcessorTM (DATEXIM, Caen, France), which increases sensitivity and takes advantage of virtual slide technology to simplify the workflow and save worker time. We used an approach based on artificial intelligence to identify abnormal cells among the tens of thousands in a cervical preparation.

OBJECTIVES

We set out to compare the diagnostic sensitivity and specificity of CytoProcessorTM and the ThinPrep Imaging System (HOLOGIC, Marlborough, MA, USA).

METHODS

A representative population of 1,352 cases was selected from the routine workflow in a private laboratory. Diagnoses were established using the ThinPrep Imaging System and CytoProcessorTM. All discordances were resolved by a consensus committee.

RESULTS

Compared to the ThinPrep Imaging System, CytoProcessorTM significantly improves diagnostic sensitivity without compromising specificity. The sensitivity of detection of "atypical squamous cells of undetermined significance (ASC-US) and more severe" and "low-grade squamous intraepithelial lesion and more severe" was significantly higher using CytoProcessorTM. Considering that cases with a truth diagnosis of ASC-US or more severe required clinical follow-up, 1.5% of the cases (21/1,360) would have been missed if the CytoProcessorTM diagnosis had been used for clinical decision-making. In contrast, 4% of the cases (54/1,360) were missed when the ThinPrep Imaging System diagnosis was used for clinical decision-making. There were 2.6 times fewer false negatives using CytoProcessorTM. The CytoProcessorTM workflow was 1.5 times faster in terms of worker time.

CONCLUSIONS

CytoProcessorTM is the first of a new generation of automated screening systems, demonstrating improved sensitivity and yielding significant gains in processing time. In addition, the fully digital nature of slide presentation in CytoProcessorTM allows the remote diagnosis of Papanicolaou tests for the first time.

Multilingual event extraction for epidemic detection

OBJECTIVE

This paper presents a multilingual news surveillance system applied to tele-epidemiology. It has been shown that multilingual approaches improve timeliness in detection of epidemic events across the globe, eliminating the wait for local news to be translated into major languages. We present here a system to extract epidemic events in potentially any language, provided a Wikipedia seed for common disease names exists.

METHODS

The Daniel system presented herein relies on properties that are common to news writing (the journalistic genre), the most useful being repetition and saliency. Wikipedia is used to screen common disease names to be matched with repeated characters strings. Language variations, such as declensions, are handled by processing text at the character-level, rather than at the word level. This additionally makes it possible to handle various writing systems in a similar fashion.

MATERIAL

As no multilingual ground truth existed to evaluate the Daniel system, we built a multilingual corpus from the Web, and collected annotations from native speakers of Chinese, English, Greek, Polish and Russian, with no connection or interest in the Daniel system. This data set is available online freely, and can be used for the evaluation of other event extraction systems.

RESULTS

Experiments for 5 languages out of 17 tested are detailed in this paper: Chinese, English, Greek, Polish and Russian. The Daniel system achieves an average F-measure of 82% in these 5 languages. It reaches 87% on BEcorpus, the state-of-the-art corpus in English, slightly below top-performing systems, which are tailored with numerous language-specific resources. The consistent performance of Daniel on multiple languages is an important contribution to the reactivity and the coverage of epidemiological event detection systems.

CONCLUSIONS

Most event extraction systems rely on extensive resources that are language-specific. While their sophistication induces excellent results (over 90% precision and recall), it restricts their coverage in terms of languages and geographic areas. In contrast, in order to detect epidemic events in any language, the Daniel system only requires a list of a few hundreds of disease names and locations, which can actually be acquired automatically. The system can perform consistently well on any language, with precision and recall around 82% on average, according to this paper's evaluation. Daniel's character-based approach is especially interesting for morphologically-rich and low-resourced languages. The lack of resources to be exploited and the state of the art string matching algorithms imply that Daniel can process thousands of documents per minute on a simple laptop. In the context of epidemic surveillance, reactivity and geographic coverage are of primary importance, since no one knows where the next event will strike, and therefore in what vernacular language it will first be reported. By being able to process any language, the Daniel system offers unique coverage for poorly endowed languages, and can complete state of the art techniques for major languages.

Document Model and Prototyping Methods for Web Engineering

This paper proposes models for managing documents in a web engineering context. First, it proposes a document model to better manage multilingual composite documents. The approach, inspired by the FRBR report, is to group all versions, translations, formats, etc. of a document in a unique document tree, putting document data and metadata at the same level. Then it proposes a model for prototyping applications, using a combination of class-based inheritance and prototype programming principles. This model applies to document models, documents views and actions. Finally, it proposes a metadata management model, laying foundations for easier integration and management of information in web applications. The proposed models are implemented in the framework Sydonie and several applications are built with the model and framework.