Pekka Kumpulainen

Detection of prostate cancer by an electronic nose: a proof of principle study.

PURPOSE We evaluate the ability of an electronic nose to discriminate prostate cancer from benign prostatic hyperplasia using urine headspace, potentially offering a clinically applicable noninvasive and rapid diagnostic method. MATERIALS AND METHODS The ChemPro® 100-eNose was used to discriminate prostate cancer from benign prostatic hyperplasia using urine sample headspace. Its performance was tested with 50 patients with confirmed prostate cancer and 24 samples from 15 patients with benign prostatic hyperplasia (15 patients provided urine preoperatively and 9 patients provided samples 3 months postoperatively) scheduled to undergo robotic assisted laparoscopic radical prostatectomy or transurethral resection of prostate, respectively. The patients provided urine sample preoperatively and those with benign prostatic hyperplasia also provided samples 3 months postoperatively to be used as a pooled control sample population. A discrimination classifier was identified for eNose and subsequently, sensitivity and specificity values were determined. Leave-one-out cross-validation was performed. RESULTS Using leave-one-out cross-validation the eNose reached a sensitivity of 78%, a specificity of 67% and AUC 0.77. CONCLUSIONS The electronic nose is capable of rapidly and noninvasively discriminating prostate cancer and benign prostatic hyperplasia using urine headspace in patients undergoing surgery.

Local anomaly detection for mobile network monitoring

Huge amounts of operation data are constantly collected from various parts of communication networks. These data include measurements from the radio connections and system logs from servers. System operators and developers need robust, easy to use decision support tools based on these data. One of their key applications is to detect anomalous phenomena of the network. In this paper we present an anomaly detection method that describes the normal states of the system with a self-organizing map (SOM) identified from the data. Large deviation in the data samples from the SOM nodes is detected as anomalous behavior. Large deviation has traditionally been detected using global thresholds. If variation of the data occurs in separate parts of the data space, the global thresholds either fail to reveal anomalies or reveal false anomalies. Instead of one global threshold, we can use local thresholds, which depend on the local variation of the data. We also present a method to find an adaptive threshold using the distribution of the deviations. Our anomaly detection method can be used both in exploration of history data or comparison of unforeseen data against a data model derived from history data. It is applicable to wide range of processes that produce multivariate data. In this paper we present examples of this method applied to server log data and radio interface data from mobile networks.

Rapid and Accurate Detection of Urinary Pathogens by Mobile IMS-Based Electronic Nose: A Proof-of-Principle Study

Urinary tract infection (UTI) is a common disease with significant morbidity and economic burden, accounting for a significant part of the workload in clinical microbiology laboratories. Current clinical chemisty point-of-care diagnostics rely on imperfect dipstick analysis which only provides indirect and insensitive evidence of urinary bacterial pathogens. An electronic nose (eNose) is a handheld device mimicking mammalian olfaction that potentially offers affordable and rapid analysis of samples without preparation at athmospheric pressure. In this study we demonstrate the applicability of ion mobility spectrometry (IMS) –based eNose to discriminate the most common UTI pathogens from gaseous headspace of culture plates rapidly and without sample preparation. We gathered a total of 101 culture samples containing four most common UTI bacteries: E. coli, S. saprophyticus, E. faecalis, Klebsiella spp and sterile culture plates. The samples were analyzed using ChemPro 100i device, consisting of IMS cell and six semiconductor sensors. Data analysis was conducted by linear discriminant analysis (LDA) and logistic regression (LR). The results were validated by leave-one-out and 5-fold cross validation analysis. In discrimination of sterile and bacterial samples sensitivity of 95% and specificity of 97% were achieved. The bacterial species were identified with sensitivity of 95% and specificity of 96% using eNose as compared to urine bacterial cultures. In conclusion: These findings strongly demonstrate the ability of our eNose to discriminate bacterial cultures and provides a proof of principle to use this method in urinanalysis of UTI.

Detection of smell print differences between nonmalignant and malignant prostate cells with an electronic nose

AIM To determine whether an electronic nose can differentiate cultured nonmalignant and malignant prostatic cells from each other and whether the smell print is secreted to the surrounding medium. MATERIALS & METHODS Prostatic nonmalignant (EP-156T and controls) and malignant (LNCaP) cell lines, as well as conditioned and unconditioned media, were collected. The smell prints of the samples were analyzed by a ChemPro(®) 100 electronic nose device. The data were normalized and dimension reduction was conducted. The samples were classified and misclassification rates were calculated. RESULTS The electronic nose differentiated the nonmalignant and malignant cell lines from each other, achieving misclassification rates of 2.9-3.6%. Cells did not differ from the conditioned medium but differed from the unconditioned medium (misclassification rates: 0.0-25.6%). CONCLUSION Malignant and nonmalignant prostatic cell lines have distinct smell prints. Prostatic cancer cells seem to modify the smell print of their medium.

Correlation approach for the detection of the heartbeat intervals using force sensors placed under the bed posts

Abstract This study proposes a method for detecting the heartbeat intervals of a person lying on a bed from ballistocardiographic signals recorded unobtrusively with four dynamic force sensors located under the bed posts. The method does not recognize individual heartbeats, but the intervals where the correlation between two consecutive signal segments maximizes. This study evaluated the performance of the method with nine subjects in 1-h long recordings and achieved 91% beat-to-beat interval (BBI) recognition coverage; 98.6% of the detected BBIs differed less than 50 ms from the values calculated from a reference electrocardiogram signal. This study also evaluated the reliability of two parameters of heart rate variability that have been used in sleep quality assessment in several studies and are usually calculated for 30 s epochs. The results suggest that the method is able to provide sufficient reliability for using the data in evaluation of sleep quality.

Finding 3G Mobile Network Cells with Similar Radio Interface Quality Problems

A mobile network provides a continuous stream of data describing the performance of its cells. Most of the data describes cells with acceptable performance. Detecting and analysing mobile network cells with quality problems from the data stream is a tedious and continuous problem for network operators. Anomaly detection can be used to identify cells, whose performance deviates from the average and which are potentially having some sub-optimal configuration or are in some error condition. In this paper we provide two methods to detect such anomalously behaving cells. The first method estimates the distance from a cell to an optimal state and the second one is based on detecting the support of the data distribution using One-Class Support Vector Machine (OC-SVM). We use the methods to analyse a data sample from a live 3G network and compare the analysis results. We also show how clustering of found anomalies can be used to find similarly behaving cells that can benefit from the same corrective measures.

Urinary polyamines as biomarkers for ovarian cancer

Objectives Elevated concentrations of polyamines have been found in urine of patients with malignant tumors, including ovarian cancer. Previous research has suffered from poorly standardized detection methods. Our liquid chromatography–tandem mass spectrometry (LC-MS/MS) method is capable of simultaneous standardized analysis of most known polyamines. Liquid chromatography–tandem mass spectrometry has not previously been used in the differential diagnostics of ovarian tumors in postmenopausal women. Materials and Methods In this prospective study, postmenopausal women (n = 71) presenting with an adnexal mass and, as controls, women with genital prolapse or urinary incontinence scheduled for surgery (n = 22) were recruited in the study. For analysis of the polyamines, a morning urine sample was obtained before surgery. Preoperative serum CA125 concentrations were determined in the study group. Results Twenty-three women with benign and 37 with malignant ovarian tumors were eligible. Of all analyzed polyamines, only urinary N1,N12-diacetylspermine showed statistically significant differences between all groups except controls versus benign tumors. N1,N12-diacetylspermine was elevated in malignant versus benign tumors (P < 0.001), in high-grade versus low malignant potential tumors (P < 0.001), in stage III to IV versus stage I to II cancers (P < 0.001), and even in early-stage cancer (stage I–II) versus benign tumors (P = 0.017). N1,N12-diacetylspermine had better sensitivity (86.5%) but lower specificity (65.2%) for distinguishing benign and malignant ovarian tumors than CA125 with a cut-off value of 35 kU/L (sensitivity, 75.7%; specificity, 69.6%). Conclusions Urinary N1,N12-diacetylspermine seems to be able to distinguish benign and malignant ovarian tumors as well as early and advanced stage, and low malignant potential and high-grade ovarian cancers from each other, respectively.

Analysing 3G radio network performance with fuzzy methods

In comparison to the earlier telecommunications networks, present-day 3rd generation (3G) networks are able to provide more complex and detailed performance data, such as distributions of channel quality indicators. However, the operators lack proper methods and tools to efficiently utilize these data in monitoring and analysis of the networks. In this article, we apply fuzzy computing to channel quality measurement distributions to get the network elements (cells) clustered into groups of similar behavior. Groups and their descriptors provide valuable information for a radio expert, who is responsible for hundreds or thousands of elements. We introduce a fuzzy inference system based on features extracted from the distributional data and provide interpretation of the found categories to demonstrate their usability on network monitoring. Additionally we present how fuzzy clustering can be used in network performance monitoring and anomaly detection. Finally, we introduce further analysis on how time dimension is an interesting perspective to analyze network element behavior. All the achieved results were discussed with radio network performance experts who found them informative and useful.

Electronic Nose in the Detection of Wound Infection Bacteria from Bacterial Cultures: A Proof-of-Principle Study

Background: Soft tissue infections, including postoperative wound infections, result in a significant burden for modern society. Rapid diagnosis of wound infections is based on bacterial stains, cultures, and polymerase chain reaction assays, and the results are available earliest after several hours, but more often not until days after. Therefore, antibiotic treatment is often administered empirically without a specific diagnosis. Methods: We employed our electronic nose (eNose) system for this proof-of-concept study, aiming to differentiate the most relevant bacteria causing wound infections utilizing a set of clinical bacterial cultures on identical blood culture dishes, and established bacterial lines from the gaseous headspace. Results: Our eNose system was capable of differentiating both methicillin-sensitive Staphylococcus aureus (MSSA) and methicillin-resistant Staphylococcus aureus (MRSA), Streptococcus pyogenes, Escherichia coli, Pseudomonas aeruginosa, and Clostridium perfringens with an accuracy of 78% within minutes without prior sample preparation. Most importantly, the system was capable of differentiating MRSA from MSSA with a sensitivity of 83%, a specificity of 100%, and an overall accuracy of 91%. Conclusions: Our results support the concept of rapid detection of the most relevant bacteria causing wound infections and ultimately differentiating MRSA from MSSA utilizing gaseous headspace sampling with an eNose.

Characterizing Mobile Network Daily Traffic Patterns by 1-Dimensional SOM and Clustering

Mobile network traffic produces daily patterns. In this paper we show how exploratory data analysis can be used to inspect the origin of the daily patterns. We use a 1-dimensional self-organizing map to characterize the patterns. 1-dimensional map enables compact visualization that is especially suitable for data where the variables are not independent but form a pattern. We introduce a stability index for analyzing the variation of the daily patterns of network elements along the days of the week. We use clustering to construct profiles for the network elements to study the stability of the traffic patterns within each element. We found out that the day of the week is the main explanation for the traffic patterns on weekends. On weekdays the traffic patterns are mostly specific to groups of networks elements, not the day of the week.