Evaluation of the performance of register data as indicators for dairy herds with high lameness prevalence

Abstract Background The modern dairy industry routinely generates data on production and disease. Therefore, the use of these cheap and at times even “free” data to predict a given state of welfare in a cost-effective manner is evaluated in the present study. Such register data could potentially be used in the identification of herds at risk of having animal welfare problems. The present study evaluated the diagnostic performance of four routinely registered indicators for identifying herds with high lameness prevalence among 40 Danish dairy herds. Indicators were extracted as within-herd annual means for a one-year period for cow mortality, bulk milk somatic cell count, proportion of lean cows at slaughter and the standard deviation (SD) of age at first calving. The target condition “high lameness prevalence” was defined as a within-herd prevalence of lame cows of ≥ 16% (third quartile). Diagnostic performance was evaluated by constructing and analysing Receiver Operating Characteristic curves and their area under the curve (AUC) for single indicators and indicator combinations. Sensitivity (Se) and specificity (Sp) of the indicators were assessed at the optimal cut-off based on data and compared to a set of predefined cut-off levels (national annual means or 90-percentile). Results Cow mortality had the highest AUC (0.76), while adding the three other indicators to the model did not yield significant increase in AUC. Cow mortality and SD of age at first calving had highest Se (100%, 95% confidence interval (CI): 72–100%), while highest Sp was found for the proportion of lean cows at slaughter (83%, 95% CI: 66–93%). The highest differential positive rate (DPR = 0.53) optimizing both Se and Sp was found for cow mortality. Optimal cut-off points were lower than the presently used pre-defined cut-offs. Conclusions The selected register-based indicators proved to be able to identify herds with high lameness prevalences. Optimized cut-offs improved the predictive ability and should therefore be preferred in official control schemes.

Download Full-text

Glioblastoma and primary central nervous system lymphoma: differentiation using MRI derived first-order texture analysis – a machine learning study

The Neuroradiology Journal ◽

10.1177/1971400921998979 ◽

2021 ◽

pp. 197140092199897

Author(s):

Sarv Priya ◽

Caitlin Ward ◽

Thomas Locke ◽

Neetu Soni ◽

Ravishankar Pillenahalli Maheshwarappa ◽

...

Keyword(s):

Machine Learning ◽

Central Nervous System ◽

Nervous System ◽

Diagnostic Performance ◽

Model Performance ◽

Area Under The Curve ◽

Central Nervous System Lymphoma ◽

First Order ◽

Single Slice ◽

Contrast Enhanced

Objectives To evaluate the diagnostic performance of multiple machine learning classifier models derived from first-order histogram texture parameters extracted from T1-weighted contrast-enhanced images in differentiating glioblastoma and primary central nervous system lymphoma. Methods Retrospective study with 97 glioblastoma and 46 primary central nervous system lymphoma patients. Thirty-six different combinations of classifier models and feature selection techniques were evaluated. Five-fold nested cross-validation was performed. Model performance was assessed for whole tumour and largest single slice using receiver operating characteristic curve. Results The cross-validated model performance was relatively similar for the top performing models for both whole tumour and largest single slice (area under the curve 0.909–0.924). However, there was a considerable difference between the worst performing model (logistic regression with full feature set, area under the curve 0.737) and the highest performing model for whole tumour (least absolute shrinkage and selection operator model with correlation filter, area under the curve 0.924). For single slice, the multilayer perceptron model with correlation filter had the highest performance (area under the curve 0.914). No significant difference was seen between the diagnostic performance of the top performing model for both whole tumour and largest single slice. Conclusions T1 contrast-enhanced derived first-order texture analysis can differentiate between glioblastoma and primary central nervous system lymphoma with good diagnostic performance. The machine learning performance can vary significantly depending on the model and feature selection methods. Largest single slice and whole tumour analysis show comparable diagnostic performance.

Download Full-text

The association between age at first calving and survival of first lactation heifers within dairy herds

animal ◽

10.1017/s1751731116000689 ◽

2016 ◽

Vol 10 (11) ◽

pp. 1877-1882 ◽

Cited By ~ 9

Author(s):

V.E. Sherwin ◽

C.D. Hudson ◽

A. Henderson ◽

M.J. Green

Keyword(s):

Dairy Herds ◽

Age At First Calving

Download Full-text

Diagnostic Performance and Accuracy of the MNA-SF against GLIM Criteria in Community-Dwelling Older Adults from Poland

Nutrients ◽

10.3390/nu13072183 ◽

2021 ◽

Vol 13 (7) ◽

pp. 2183

Author(s):

Aleksandra Kaluźniak-Szymanowska ◽

Roma Krzymińska-Siemaszko ◽

Marta Lewandowicz ◽

Ewa Deskur-Śmielecka ◽

Katarzyna Stachnik ◽

...

Keyword(s):

At Risk ◽

Diagnostic Performance ◽

Nutritional Assessment ◽

Short Form ◽

Area Under The Curve ◽

Mini Nutritional Assessment ◽

Community Dwelling ◽

Study Results ◽

Community Dwelling Older Adults ◽

Test Result

Up to 28% of elderly residents in Europe are at risk of malnutrition. As uniform diagnostic criteria for malnutrition have not been formulated, in autumn 2018, the Global Leadership Initiative on Malnutrition (GLIM) presented a consensus on its diagnosis. According to the consensus, the diagnosis of malnutrition requires a positive screening test result for the risk of malnutrition, and the presence of at least one etiologic and one phenotypic criterion. This study aimed to assess the diagnostic performance and accuracy of the Mini Nutritional Assessment—Short Form (MNA-SF) against GLIM criteria. The analysis involved 273 community-dwelling volunteers aged ≥ 60 years. All participants were screened for malnutrition with the MNA-SF questionnaire. Next, the GLIM phenotypic and etiologic criteria were assessed in all subjects. Based on the presence of at least one phenotypic and one etiologic criterion, malnutrition was diagnosed in more than one-third of participants (n = 103, 37.7%). According to the MNA-SF, only 7.3% of subjects had malnutrition, and 28.2% were at risk of malnutrition. The agreement between the MNA-SF score and the GLIM criteria were observed in only 22.3% of the population. The sensitivity and specificity of MNA-SF against the GLIM criteria were fair (59.2% and 78.8%, respectively). The area under the curve (AUC) was 0.77, indicating the fair ability of MNA-SF to diagnose malnutrition. Based on the present study results, the best solution may be an optional replacement of the screening tool in the first step of the GLIM algorithm with clinical suspicion of malnutrition.

Download Full-text

Comparison of prostate imaging reporting and data system v2.1 and 2 in transition and peripheral zones: evaluation of interreader agreement and diagnostic performance in detecting clinically significant prostate cancer

British Journal of Radiology ◽

10.1259/bjr.20201434 ◽

2021 ◽

pp. 20201434

Author(s):

Yasuyo Urase ◽

Yoshiko Ueno ◽

Tsutomu Tamada ◽

Keitaro Sofue ◽

Satoru Takahashi ◽

...

Keyword(s):

Prostate Cancer ◽

Diagnostic Performance ◽

Area Under The Curve ◽

Data System ◽

Transition Zones ◽

Prostate Imaging ◽

Institutional Review ◽

3.0 T ◽

Clinically Significant ◽

Interreader Agreement

Objectives: To evaluate the interreader agreement and diagnostic performance of the Prostate Imaging Reporting and Data System (PI-RADS) v2.1, in comparison with v2. Methods: Institutional review board approval was obtained for this retrospective study. Seventy-seven consecutive patients who underwent a prostate multiparametric magnetic resonance imaging at 3.0 T before radical prostatectomy were included. Four radiologists (two experienced uroradiologists and two inexperienced radiologists) independently scored eight regions [six peripheral zones (PZ) and two transition zones (TZ)] using v2.1 and v2. Interreader agreement was assessed using κ statistics. To evaluate diagnostic performance for clinically significant prostate cancer (csPC), area under the curve (AUC) was estimated. Results 228 regions were pathologically diagnosed as positive for csPC. With a cutoff ≥3, the agreement among all readers was better with v2.1 than v2 in TZ, PZ, or both zones combined (κ-value: TZ, 0.509 vs 0.414; PZ, 0.686 vs 0.568; both zones combined, 0.644 vs 0.531). With a cutoff ≥4, the agreement among all readers was also better with v2.1 than v2 in the PZ or both zones combined (κ-value: PZ, 0.761 vs 0.701; both zones combined, 0.756 vs 0.709). For all readers, AUC with v2.1 was higher than with v2 (TZ, 0.826–0.907 vs 0.788–0.856; PZ, 0.857–0.919 vs 0.853–0.902). Conclusions: Our study suggests that the PI-RADS v2.1 could improve the interreader agreement and might contribute to improved diagnostic performance compared with v2. Advances in knowledge: PI-RADS v2.1 has a potential to improve interreader variability and diagnostic performance among radiologists with different levels of expertise.

Download Full-text

Can pre-collected register data be used to identify dairy herds with good cattle welfare?

Acta Veterinaria Scandinavica ◽

10.1186/1751-0147-53-s1-s8 ◽

2011 ◽

Vol 53 (S1) ◽

Cited By ~ 30

Author(s):

Ann-Kristin Nyman ◽

Ann Lindberg ◽

Charlotte Hallén Sandgren

Keyword(s):

Dairy Herds

Download Full-text

Predictive Analytic Techniques to Identify Hidden Relationships between Training Load, Fatigue and Muscle Strains in Young Soccer Players

Sports ◽

10.3390/sports10010003 ◽

2021 ◽

Vol 10 (1) ◽

pp. 3

Author(s):

Mauro Mandorino ◽

António J. Figueiredo ◽

Gianluca Cima ◽

Antonio Tessitore

Keyword(s):

Area Under The Curve ◽

Predictive Ability ◽

Peak Height ◽

Neuromuscular Fatigue ◽

Training Load ◽

Soccer Players ◽

Height Velocity ◽

Support Vector ◽

Peak Height Velocity ◽

Injury Surveillance System

This study aimed to analyze different predictive analytic techniques to forecast the risk of muscle strain injuries (MSI) in youth soccer based on training load data. Twenty-two young soccer players (age: 13.5 ± 0.3 years) were recruited, and an injury surveillance system was applied to record all MSI during the season. Anthropometric data, predicted age at peak height velocity, and skeletal age were collected. The session-RPE method was daily employed to quantify internal training/match load, and monotony, strain, and cumulative load over the weeks were calculated. A countermovement jump (CMJ) test was submitted before and after each training/match to quantify players’ neuromuscular fatigue. All these data were used to predict the risk of MSI through different data mining models: Logistic Regression (LR), Random Forest (RF), Support Vector Machine (SVM). Among them, SVM showed the best predictive ability (area under the curve = 0.84 ± 0.05). Then, Decision tree (DT) algorithm was employed to understand the interactions identified by the SVM model. The rules extracted by DT revealed how the risk of injury could change according to players’ maturity status, neuromuscular fatigue, anthropometric factors, higher workloads, and low recovery status. This approach allowed to identify MSI and the underlying risk factors.

Download Full-text

A Rationale for Using Sonoelastography in Thyroid Nodular Pathology

Journal of radiology and nuclear medicine ◽

10.20862/0042-4676-2019-100-5-242-246 ◽

2019 ◽

Vol 100 (5) ◽

pp. 242-246

Author(s):

L. A. Timofeeva ◽

L. B. Shubin

Keyword(s):

Thyroid Cancer ◽

Roc Analysis ◽

Thyroid Nodules ◽

Area Under The Curve ◽

Predictive Ability ◽

Tissue Elasticity ◽

Stromal Component ◽

Collagen Level ◽

Probable Nature ◽

Benign Nodules

Objective. To provide a rationale for using sonoelastography (SEG) in the differential diagnosis of thyroid cancer (TC).Material and methods. Thirty patients with thyroid nodules of various morphological structures were examined. The authors studied the data of SEG and immunohistochemistry (IHC) with monoclonal antibodies against types III and IV collagen (they evaluated the degree of the expressed collagen fibers). Analysis of variance, ROC analysis, and logistic regression were used (by comparing with the expression of collagens) to assess the predictive ability of ultrasound.Results. The study showed that irregular and uneven contours, microcalcifications, and “the height greater than the width” were most significant among the ultrasound signs in the diagnosis of TC. Cool colors prevailed when performing SEG in the pattern of thyroid cancer. Purple-blue hues were predominantly recorded (p<0.05 with regard to benign nodules), green ones were less frequently. ROC analysis of compression elastography showed that the area under the curve was 0.785 (95% CI 0.740-0.826), sensitivity 78.1%, specificity 79.0%. Comparison of the data of IHC and SEG revealed a direct correlation of tissue elasticity with the degree of a stromal component and with the presence of collagen-containing structures.Conclusion. SEG may suppose the probable nature of thyroid nodules on the basis of their morphological features. The low degree of the stromal component and the low content of types III and IV collagen make follicular colloid goiter and adenoma soft, which is recorded at SEG. TC is characterized by a high collagen level attributable to the characteristics of the metabolism of cancer cells, which makes them solid in the mode of SEG.

Download Full-text

Application of artificial intelligence in a real-world research for predicting the risk of liver metastasis in T1 colorectal cancer

Cancer Cell International ◽

10.1186/s12935-021-02424-7 ◽

2022 ◽

Vol 22 (1) ◽

Author(s):

Tenghui Han ◽

Jun Zhu ◽

Xiaoping Chen ◽

Rujie Chen ◽

Yu Jiang ◽

...

Keyword(s):

Artificial Intelligence ◽

Colorectal Cancer ◽

Liver Metastasis ◽

Predictive Model ◽

Area Under The Curve ◽

Predictive Ability ◽

Subsequent Treatment ◽

Training Dataset ◽

Discriminative Ability ◽

Fundamental Parameters

Abstract Background Liver is the most common metastatic site of colorectal cancer (CRC) and liver metastasis (LM) determines subsequent treatment as well as prognosis of patients, especially in T1 patients. T1 CRC patients with LM are recommended to adopt surgery and systematic treatments rather than endoscopic therapy alone. Nevertheless, there is still no effective model to predict the risk of LM in T1 CRC patients. Hence, we aim to construct an accurate predictive model and an easy-to-use tool clinically. Methods We integrated two independent CRC cohorts from Surveillance Epidemiology and End Results database (SEER, training dataset) and Xijing hospital (testing dataset). Artificial intelligence (AI) and machine learning (ML) methods were adopted to establish the predictive model. Results A total of 16,785 and 326 T1 CRC patients from SEER database and Xijing hospital were incorporated respectively into the study. Every single ML model demonstrated great predictive capability, with an area under the curve (AUC) close to 0.95 and a stacking bagging model displaying the best performance (AUC = 0.9631). Expectedly, the stacking model exhibited a favorable discriminative ability and precisely screened out all eight LM cases from 326 T1 patients in the outer validation cohort. In the subgroup analysis, the stacking model also demonstrated a splendid predictive ability for patients with tumor size ranging from one to50mm (AUC = 0.956). Conclusion We successfully established an innovative and convenient AI model for predicting LM in T1 CRC patients, which was further verified in the external dataset. Ultimately, we designed a novel and easy-to-use decision tree, which only incorporated four fundamental parameters and could be successfully applied in clinical practice.

Download Full-text

Validation of Predictive Ability of Bobi Score in Burn Patients

Journal of Clinical Review & Case Reports ◽

10.33140/jcrc.06.04.01 ◽

2021 ◽

Vol 6 (4) ◽

Keyword(s):

Burn Injury ◽

Prognostic Score ◽

Area Under The Curve ◽

Predictive Ability ◽

Scoring Systems ◽

Good Prediction ◽

Burn Patients ◽

Severe Burns ◽

Inhalation Burn ◽

Outstanding Result

Introduction: Scoring systems have been used successfully in burn centers to predict the prognosis and take measures for careful monitoring of the burned patient. Belgium Outcome Burn Injury score is one of them which takes into consideration age, burn surface area, and presence of inhalation burn. Objectives: This presentation aims to validate the use of the BOBI prognostic score in our patients. Patients and Methods: The study is a retrospective analytical study that utilized the investigation of the medical charts of 1515 patients hospitalized with severe burns within the ICU of the Service of Burns in Tirana, Albania during 2010-2019. Results: The overall mortality of our patients was 7.06% (107 deaths in 1515 patients). Up to BOBI score 6, we have noticed better mortality than prediction while there is a very good prediction up to score 10. Area Under the Curve was 0.978 (p<0.0001) which is an outstanding result in being a classifier between deaths and survivors. Conclusions: BOBI score is a very good prediction score for mortality in burn patients.

Download Full-text

Spanish Validation of the Child Pornography Offender Risk Tool

Sexual Abuse ◽

10.1177/1079063220928958 ◽

2020 ◽

pp. 107906322092895

Author(s):

Virginia Soldino ◽

Enrique J. Carbonell-Vayá ◽

Kathryn C. Seigfried-Spellar

Keyword(s):

Sexual Interest ◽

Child Pornography ◽

Area Under The Curve ◽

Predictive Ability ◽

Preliminary Evidence ◽

Recidivism Rate ◽

Recidivism Rates ◽

Cultural Validity ◽

Sexual Offenses ◽

Spanish Validation

The current study examined the validity of the Child Pornography Offender Risk Tool (CPORT) in a sample of 304 men arrested in Spain for child pornography (CP) offenses, distinguishing between CP-exclusive offenders ( n = 255) and CP offenders with other criminal involvement ( n = 49). In our 5-year fixed follow-up analysis, we observed a 2.3% sexual recidivism rate for the whole sample (2.0% new CP offenses, 0.3% new contact sexual offenses). Receiver operating characteristic (ROC) analyses detected some relative predictive ability of the CPORT for CP recidivism outcomes when the Correlates of Admission of Sexual Interest in Children (CASIC) was used to replace missing CPORT Item 5. Specifically, both CPORT and CASIC total scores might help predict new CP offending among CP-exclusive offenders (area under the curve [AUC] = .57 and .70, respectively). Calibration analyses found that the observed recidivism rates were much lower than the expected recidivism rates presented by the tool developers, and, thus, suggest caution over the use of these norms for applied risk assessment. Our findings provide, to some extent, preliminary evidence of CPORT cross-cultural validity.

Download Full-text