Classification of cancer-related death certificates using machine learning

Background Cancer monitoring and prevention relies on the critical aspect of timely notification of cancer cases. However, the abstraction and classification of cancer from the free-text of pathology reports and other relevant documents, such as death certificates, exist as complex and time-consuming activities. Aims In this paper, approaches for the automatic detection of notifiable cancer cases as the cause of death from free-text death certificates supplied to Cancer Registries are investigated. Method A number of machine learning classifiers were studied. Features were extracted using natural language techniques and the Medtex toolkit. The numerous features encompassed stemmed words, bi-grams, and concepts from the SNOMED CT medical terminology. The baseline consisted of a keyword spotter using keywords extracted from the long description of ICD-10 cancer related codes. Results Death certificates with notifiable cancer listed as the cause of death can be effectively identified with the methods studied in this paper. A Support Vector Machine (SVM) classifier achieved best performance with an overall F-measure of 0.9866 when evaluated on a set of 5,000 free-text death certificates using the token stem feature set. The SNOMED CT concept plus token stem feature set reached the lowest variance (0.0032) and false negative rate (0.0297) while achieving an F-measure of 0.9864. The SVM classifier accounts for the first 18 of the top 40 evaluated runs, and entails the most robust classifier with a variance of 0.001141, half the variance of the other classifiers. Conclusion The selection of features significantly produced the most influences on the performance of the classifiers, although the type of classifier employed also affects performance. In contrast, the feature weighting schema created a negligible effect on performance. Specifically, it is found that stemmed tokens with or without SNOMED CT concepts create the most effective feature when combined with an SVM classifier.

Treatment plan complexity metrics for predicting IMRT pre-treatment quality assurance results

The planning of IMRT treatments requires a compromise between dose conformity (complexity) and deliverability. This study investigates established and novel treatment complexity metrics for 122 IMRT beams from prostate treatment plans. The Treatment and Dose Assessor software was used to extract the necessary data from exported treatment plan files and calculate the metrics. For most of the metrics, there was strong overlap between the calculated values for plans that passed and failed their quality assurance (QA) tests. However, statistically significant variation between plans that passed and failed QA measurements was found for the established modulation index and for a novel metric describing the proportion of small apertures in each beam. The ‘small aperture score’ provided threshold values which successfully distinguished deliverable treatment plans from plans that did not pass QA, with a low false negative rate.

The genome of the domesticated apple (Malus × domestica Borkh.)

We report a high-quality draft genome sequence of the domesticated apple (Malus × domestica). We show that a relatively recent (>50 million years ago) genome-wide duplication (GWD) has resulted in the transition from nine ancestral chromosomes to 17 chromosomes in the Pyreae. Traces of older GWDs partly support the monophyly of the ancestral paleohexaploidy of eudicots. Phylogenetic reconstruction of Pyreae and the genus Malus, relative to major Rosaceae taxa, identified the progenitor of the cultivated apple as M. sieversii. Expansion of gene families reported to be involved in fruit development may explain formation of the pome, a Pyreae-specific false fruit that develops by proliferation of the basal part of the sepals, the receptacle. In apple, a subclade of MADS-box genes, normally involved in flower and fruit development, is expanded to include 15 members, as are other gene families involved in Rosaceae-specific metabolism, such as transport and assimilation of sorbitol.

Recursive actuator fault detection and diagnosis for emergency landing of UASs

This paper presents a practical recursive fault detection and diagnosis (FDD) scheme for online identification of actuator faults for unmanned aerial systems (UASs) based on the unscented Kalman filtering (UKF) method. The proposed FDD algorithm aims to monitor health status of actuators and provide indication of actuator faults with reliability, offering necessary information for the design of fault-tolerant flight control systems to compensate for side-effects and improve fail-safe capability when actuator faults occur. The fault detection is conducted by designing separate UKFs to detect aileron and elevator faults using a nonlinear six degree-of-freedom (DOF) UAS model. The fault diagnosis is achieved by isolating true faults by using the Bayesian Classifier (BC) method together with a decision criterion to avoid false alarms. High-fidelity simulations with and without measurement noise are conducted with practical constraints considered for typical actuator fault scenarios, and the proposed FDD exhibits consistent effectiveness in identifying occurrence of actuator faults, verifying its suitability for integration into the design of fault-tolerant flight control systems for emergency landing of UASs.

Methodology for developing real-time motorway traffic risk identification models using individual-vehicle data

Most of existing motorway traffic safety studies using disaggregate traffic flow data aim at developing models for identifying real-time traffic risks by comparing pre-crash and non-crash conditions. One of serious shortcomings in those studies is that non-crash conditions are arbitrarily selected and hence, not representative, i.e. selected non-crash data might not be the right data comparable with pre-crash data; the non-crash/pre-crash ratio is arbitrarily decided and neglects the abundance of non-crash over pre-crash conditions; etc. Here, we present a methodology for developing a real-time MotorwaY Traffic Risk Identification Model (MyTRIM) using individual vehicle data, meteorological data, and crash data. Non-crash data are clustered into groups called traffic regimes. Thereafter, pre-crash data are classified into regimes to match with relevant non-crash data. Among totally eight traffic regimes obtained, four highly risky regimes were identified; three regime-based Risk Identification Models (RIM) with sufficient pre-crash data were developed. MyTRIM memorizes the latest risk evolution identified by RIM to predict near future risks. Traffic practitioners can decide MyTRIM’s memory size based on the trade-off between detection and false alarm rates. Decreasing the memory size from 5 to 1 precipitates the increase of detection rate from 65.0% to 100.0% and of false alarm rate from 0.21% to 3.68%. Moreover, critical factors in differentiating pre-crash and non-crash conditions are recognized and usable for developing preventive measures. MyTRIM can be used by practitioners in real-time as an independent tool to make online decision or integrated with existing traffic management systems.

An attack-localizing watermarking scheme for natural language documents

We present a text watermarking scheme that embeds a bitstream watermark Wi in a text document P preserving the meaning, context, and flow of the document. The document is viewed as a set of paragraphs, each paragraph being a set of sentences. The sequence of paragraphs and sentences used to embed watermark bits is permuted using a secret key. Then, English language sentence transformations are used to modify sentence lengths, thus embedding watermarking bits in the Least Significant Bits (LSB) of the sentences’ cardinalities. The embedding and extracting algorithms are public, while the secrecy and security of the watermark depends on a secret key K. The probability of False Positives is extremely small, hence avoiding incidental occurrences of our watermark in random text documents. Majority voting provides security against text addition, deletion, and swapping attacks, further reducing the probability of False Positives. The scheme is secure against the general attacks on text watermarks such as reproduction (photocopying, FAX), reformatting, synonym substitution, text addition, text deletion, text swapping, paragraph shuffling and collusion attacks.

Field validation of controlled Monte Carlo data generation for statistical damage identification employing Mahalanobis squared distance

This article presents the field applications and validations for the controlled Monte Carlo data generation scheme. This scheme was previously derived to assist the Mahalanobis squared distance–based damage identification method to cope with data-shortage problems which often cause inadequate data multinormality and unreliable identification outcome. To do so, real-vibration datasets from two actual civil engineering structures with such data (and identification) problems are selected as the test objects which are then shown to be in need of enhancement to consolidate their conditions. By utilizing the robust probability measures of the data condition indices in controlled Monte Carlo data generation and statistical sensitivity analysis of the Mahalanobis squared distance computational system, well-conditioned synthetic data generated by an optimal controlled Monte Carlo data generation configurations can be unbiasedly evaluated against those generated by other set-ups and against the original data. The analysis results reconfirm that controlled Monte Carlo data generation is able to overcome the shortage of observations, improve the data multinormality and enhance the reliability of the Mahalanobis squared distance–based damage identification method particularly with respect to false-positive errors. The results also highlight the dynamic structure of controlled Monte Carlo data generation that makes this scheme well adaptive to any type of input data with any (original) distributional condition.

Asymptotic minimax robust and misspecified Lorden quickest change detection for dependent stochastic processes

The quick detection of an abrupt unknown change in the conditional distribution of a dependent stochastic process has numerous applications. In this paper, we pose a minimax robust quickest change detection problem for cases where there is uncertainty about the post-change conditional distribution. Our minimax robust formulation is based on the popular Lorden criteria of optimal quickest change detection. Under a condition on the set of possible post-change distributions, we show that the widely known cumulative sum (CUSUM) rule is asymptotically minimax robust under our Lorden minimax robust formulation as a false alarm constraint becomes more strict. We also establish general asymptotic bounds on the detection delay of misspecified CUSUM rules (i.e. CUSUM rules that are designed with post- change distributions that differ from those of the observed sequence). We exploit these bounds to compare the delay performance of asymptotically minimax robust, asymptotically optimal, and other misspecified CUSUM rules. In simulation examples, we illustrate that asymptotically minimax robust CUSUM rules can provide better detection delay performance at greatly reduced computation effort compared to competing generalised likelihood ratio procedures.

On-the-fly web content integrity check boosts users' confidence

Web servers are accessible by anyone who can access the Internet. Although this universal accessibility is attractive for all kinds of Web-based applications, Web servers are exposed to attackers who may want to alter their contents. Alterations range from humorous additions or changes, which are typically easy to spot, to more sinister tampering, such as providing false or damaging information.

Robust numeric set watermarking : numbers don’t lie

Ever since Cox et. al published their paper, “A Secure, Robust Watermark for Multimedia” in 1996 [6], there has been tremendous progress in multimedia watermarking. The same pattern re-emerged with Agrawal and Kiernan publishing their work “Watermarking Relational Databases” in 2001 [1]. However, little attention has been given to primitive data collections with only a handful works of research known to the authors [11, 10]. This is primarily due to the absence of an attribute that differentiates marked items from unmarked item during insertion and detection process. This paper presents a distribution-independent, watermarking model that is secure against secondary-watermarking in addition to conventional attacks such as data addition, deletion and distortion. The low false positives and high capacity provide additional strength to the scheme. These claims are backed by experimental results provided in the paper.

Visualization of long-duration acoustic recordings of the environment

Acoustic recordings of the environment are an important aid to ecologists monitoring biodiversity and environmental health. However, rapid advances in recording technology, storage and computing make it possible to accumulate thousands of hours of recordings, of which, ecologists can only listen to a small fraction. The big-data challenge addressed in this paper is to visualize the content of long-duration audio recordings on multiple scales, from hours, days, months to years. The visualization should facilitate navigation and yield ecologically meaningful information. Our approach is to extract (at one minute resolution) acoustic indices which reflect content of ecological interest. An acoustic index is a statistic that summarizes some aspect of the distribution of acoustic energy in a recording. We combine indices to produce false-color images that reveal acoustic content and facilitate navigation through recordings that are months or even years in duration.

Genome-wide analysis in a murine Dnmt1 knockdown model identifies epigenetically silenced genes in primary human pituitary tumors

DNA methylation at promoter CpG islands (CGI) is an epigenetic modification associated with inappropriate gene silencing in multiple tumor types. In the absence of a human pituitary tumor cell line, small interfering RNA-mediated knockdown of the maintenance methyltransferase DNA methyltransferase (cytosine 5)-1 (Dnmt1) was used in the murine pituitary adenoma cell line AtT-20. Sustained knockdown induced reexpression of the fully methylated and normally imprinted gene neuronatin (Nnat) in a time-dependent manner. Combined bisulfite restriction analysis (COBRA) revealed that reexpression of Nnat was associated with partial CGI demethylation, which was also observed at the H19 differentially methylated region. Subsequent genome-wide microarray analysis identified 91 genes that were significantly differentially expressed in Dnmt1 knockdown cells (10% false discovery rate). The analysis showed that genes associated with the induction of apoptosis, signal transduction, and developmental processes were significantly overrepresented in this list (P < 0.05). Following validation by reverse transcription-PCR and detection of inappropriate CGI methylation by COBRA, four genes (ICAM1, NNAT, RUNX1, and S100A10) were analyzed in primary human pituitary tumors, each displaying significantly reduced mRNA levels relative to normal pituitary (P < 0.05). For two of these genes, NNAT and S100A10, decreased expression was associated with increased promoter CGI methylation. Induced expression of Nnat in stable transfected AtT-20 cells inhibited cell proliferation. To our knowledge, this is the first report of array-based "epigenetic unmasking" in combination with Dnmt1 knockdown and reveals the potential of this strategy toward identifying genes silenced by epigenetic mechanisms across species boundaries.

Outbreak detection algorithms for seasonal disease data : a case study using Ross River virus disease

Background Detection of outbreaks is an important part of disease surveillance. Although many algorithms have been designed for detecting outbreaks, few have been specifically assessed against diseases that have distinct seasonal incidence patterns, such as those caused by vector-borne pathogens. Methods We applied five previously reported outbreak detection algorithms to Ross River virus (RRV) disease data (1991-2007) for the four local government areas (LGAs) of Brisbane, Emerald, Redland and Townsville in Queensland, Australia. The methods used were the Early Aberration Reporting System (EARS) C1, C2 and C3 methods, negative binomial cusum (NBC), historical limits method (HLM), Poisson outbreak detection (POD) method and the purely temporal SaTScan analysis. Seasonally-adjusted variants of the NBC and SaTScan methods were developed. Some of the algorithms were applied using a range of parameter values, resulting in 17 variants of the five algorithms. Results The 9,188 RRV disease notifications that occurred in the four selected regions over the study period showed marked seasonality, which adversely affected the performance of some of the outbreak detection algorithms. Most of the methods examined were able to detect the same major events. The exception was the seasonally-adjusted NBC methods that detected an excess of short signals. The NBC, POD and temporal SaTScan algorithms were the only methods that consistently had high true positive rates and low false positive and false negative rates across the four study areas. The timeliness of outbreak signals generated by each method was also compared but there was no consistency across outbreaks and LGAs. Conclusions This study has highlighted several issues associated with applying outbreak detection algorithms to seasonal disease data. In lieu of a true gold standard, a quantitative comparison is difficult and caution should be taken when interpreting the true positives, false positives, sensitivity and specificity.

Circulating antibodies against Plasmodium falciparum histidine-rich proteins 2 interfere with antigen detection by rapid diagnostic tests

Background Rapid diagnostic tests (RDTs) for detection of Plasmodium falciparum infection that target P. falciparum histidine-rich protein 2 (PfHRP2), a protein that circulates in the blood of patients infected with this species of malaria, are widely used to guide case management. Understanding determinants of PfHRP2 availability in circulation is therefore essential to understanding the performance of PfHRP2-detecting RDTs. Methods The possibility that pre-formed host anti-PfHRP2 antibodies may block target antigen detection, thereby causing false negative test results was investigated in this study. Results Anti-PfHRP2 antibodies were detected in 19/75 (25%) of plasma samples collected from patients with acute malaria from Cambodia, Nigeria and the Philippines, as well as in 3/28 (10.7%) asymptomatic Solomon Islands residents. Pre-incubation of plasma samples from subjects with high-titre anti-PfHRP2 antibodies with soluble PfHRP2 blocked the detection of the target antigen on two of the three brands of RDTs tested, leading to false negative results. Pre-incubation of the plasma with intact parasitized erythrocytes resulted in a reduction of band intensity at the highest parasite density, and a reduction of lower detection threshold by ten-fold on all three brands of RDTs tested. Conclusions These observations indicate possible reduced sensitivity for diagnosis of P. falciparum malaria using PfHRP2-detecting RDTs among people with high levels of specific antibodies and low density infection, as well as possible interference with tests configured to detect soluble PfHRP2 in saliva or urine samples. Further investigations are required to assess the impact of pre-formed anti-PfHRP2 antibodies on RDT performance in different transmission settings.

Alarm setting for the critically ill patient : A descriptive pilot survey of nurses’ perceptions of current practice in an Australian Regional Critical Care Unit

Aim: The aim of this survey was to assess registered nurse’s perceptions of alarm setting and management in an Australian Regional Critical Care Unit. Background: The setting and management of alarms within the critical care environment is one of the key responsibilities of the nurse in this area. However, with up to 99% of alarms potentially being false-positives it is easy for the nurse to become desensitised or fatigued by incessant alarms; in some cases up to 400 per patient per day. Inadvertently ignoring, silencing or disabling alarms can have deleterious implications for the patient and nurse. Method: A total population sample of 48 nursing staff from a 13 bedded ICU/HDU/CCU within regional Australia were asked to participate. A 10 item open-ended and multiple choice questionnaire was distributed to determine their perceptions and attitudes of alarm setting and management within this clinical area. Results: Two key themes were identified from the open-ended questions: attitudes towards inappropriate alarm settings and annoyance at delayed responses to alarms. A significant number of respondents (93%) agreed that alarm fatigue can result in alarm desensitisation and the disabling of alarms, whilst 81% suggested the key factors are those associated with false-positive alarms and inappropriately set alarms.