Charles Chih-Ho Liu

Application of Radio‐frequency Identification in Perioperative Care

Every perioperative department could benefit from having an information system that facilitates managerial function and improves efficiency in the OR. The Patient Advancement Monitoring System-Surgical implemented in a hospital in Taipei, Taiwan, is one such a system that uses radio-frequency identification technology for tracking perioperative care of patients along workflow checkpoints. This web-based medical information system can facilitate care provided throughout perioperative services by providing instant patient information to staff members in cross-functional health care teams. Manpower is not wasted on duplicating data entry because the surgical progression is displayed in real time. Satisfaction with the system has been high for both nurses and administrators.

Discovering Latent Semantics in Web Documents Using Fuzzy Clustering

Web documents are heterogeneous and complex. There exists complicated associations within one web document and linking to the others. The high interactions between terms in documents demonstrate vague and ambiguous meanings. Efficient and effective clustering methods to discover latent and coherent meanings in context are necessary. This paper presents a fuzzy linguistic topological space along with a fuzzy clustering algorithm to discover the contextual meaning in the web documents. The proposed algorithm extracts features from the web documents using conditional random field methods and builds a fuzzy linguistic topological space based on the associations of features. The associations of cooccurring features organize a hierarchy of connected semantic complexes called “CONCEPTS,” wherein a fuzzy linguistic measure is applied on each complex to evaluate 1) the relevance of a document belonging to a topic, and 2) the difference between the other topics. Web contents are able to be clustered into topics in the hierarchy depending on their fuzzy linguistic measures; web users can further explore the CONCEPTS of web contents accordingly. Besides the algorithm applicability in web text domains, it can be extended to other applications, such as data mining, bioinformatics, content-based, or collaborative information filtering, etc.

PICO element detection in medical text without metadata

Efficient identification of patient, intervention, comparison, and outcome (PICO) components in medical articles is helpful in evidence-based medicine. The purpose of this study is to clarify whether first sentences of these components are good enough to train naive Bayes classifiers for sentence-level PICO element detection. We extracted 19,854 structured abstracts of randomized controlled trials with any P/I/O label from PubMed for naive Bayes classifiers training. Performances of classifiers trained by first sentences of each section (CF) and those trained by all sentences (CA) were compared using all sentences by ten-fold cross-validation. The results measured by recall, precision, and F-measures show that there are no significant differences in performance between CF and CA for detection of O-element (F-measure=0.731±0.009 vs. 0.738±0.010, p=0.123). However, CA perform better for I-elements, in terms of recall (0.752±0.012 vs. 0.620±0.007, p<0.001) and F-measures (0.728±0.006 vs. 0.662±0.007, p<0.001). For P-elements, CF have higher precision (0.714±0.009 vs. 0.665±0.010, p<0.001), but lower recall (0.766±0.013 vs. 0.811±0.012, p<0.001). CF are not always better than CA in sentence-level PICO element detection. Their performance varies in detecting different elements.

Classification of PICO elements by text features systematically extracted from PubMed abstracts

We propose and evaluate a systematic approach to detect and classify Patient/Problem, Intervention, Comparison and Outcome (PICO) from the medical literature. The training and test corpora were generated systematically and automatically from structured PubMed abstracts. 23,472 sentences by exact pattern match of head words of P-I-O categories. Afterward, the terms with top frequencies were used as the features of Naïve Bayesian classifier. This approach achieves F-measure values of 0.91 for Patient/Problem, 0.75 for Intervention and 0.88 for Outcome, comparable to previous studied based on mixed textural, paragraphical, and semantic features. In conclusion, we show that by stricter pattern matching criteria of training set, detection and classification of PICO elements can be reproducible with minimal expert intervention. The results of this work are higher than previous studies.

AUTOMATED VOLUMETRY OF POSTOPERATIVE SKULL DEFECT ON BRAIN CT

The volume of the skull defect should be one of the most important quantitative measures for decompressive craniectomy. However, there has been no study focusing on automated estimation of the volume from postoperative computed tomography (CT). This study develops and validates three methods that can automatically locate, recover and measure the missing skull region based on symmetry without preoperative images. The low resolution estimate (LRE) method involves downsizing CT images, finding the axis of symmetry for each slice, and estimating the location and size of the missing skull regions. The intact mid-sagittal plane (iMSP) can be defined either by dimension-by-dimension (DBD) method as a global symmetry plane or by Lius method as a regression from each slices. The skull defect volume can then be calculated by skull volume difference (SVD) with respect to each iMSP. During a 48-month period between July 2006 and June 2010 at a regional hospital in northern Taiwan, we collected 30 sets of nonvolumetric CT images after craniectomies. Three board-certified neurosurgeons perform computer-assisted volumetric analysis of skull defect volume VMan as the gold standard for evaluating the performance of our algorithm. We compare the error of the three volumetry methods. The error of VLRE is smaller than that of VLiu(p < 0.0001) and VDBD(p = 0.034). The error of VDBD is significant smaller than that of VLiu(p = 0.001). The correlation coefficients between VMan and VLRE, VLiu, VDBD are 0.98, 0.88 and 0.95, respectively. In conclusion, these methods can help to define the skull defect volume in postoperative images and provide information of the immediate volume gain after decompressive craniectomies. The iMSP of the postoperative skull can be reliably identified using the DBD method.

Overview of Extremity Replantation in Taiwan by the National Health Insurance Database from 1996 to 2000

This study is the first epidemiological report of limb replantation surgery concerning the whole population of Taiwan. The materials are based on the 1:20-sampled National Healthcare Insurance (NHI) Research Database from 1996 to 2000.368 and 236 patients were extracted from the hospitalization datasets, based on the ICD procedure codes and the detailed medical orders respectively. 77.1% and 14.0% of cases received one-and two-digital replantation. The average operating times are 3.9, 5.5, 7.9, and 10.2 hours from one- to four- digit replantations. Average length of hospital stay, surgical and total fees were 12.5 days,

Special Issue: Text Mining and Information Analysis; Retrieving and Clustering Keywords in Neurosurgery Operation Reports Using Text Mining Techniques

48,585, and

Primary Cutaneous Large B-cell Lymphoma of Leg, with Unusual Presentation as a Chronic Ulcer-A Case Report

125,660 in medical centers, in comparison with 9.9 days,

Alveolar Soft Part Sarcoma: A Case Report, Literature Review, and Differentiation from Vascular Malformation

36,786-

Correction of Inverted Nipple Using Crossed X-wing Dermal Flaps

38,962, and