scholarly journals Feature Selection of Non-Dermoscopic Skin Lesion Images for Nevus and Melanoma Classification

Computation ◽  
2020 ◽  
Vol 8 (2) ◽  
pp. 41 ◽  
Author(s):  
Felicia Anisoara Damian ◽  
Simona Moldovanu ◽  
Nilanjan Dey ◽  
Amira S. Ashour ◽  
Luminita Moraru
Keyword(s):  
Fourier Transform ◽  
Area Under The Curve ◽  
Roc Curves ◽  
Noise Removal ◽  
Skin Lesions ◽  
Asymmetry Index ◽  
Color Distribution ◽  
Atypical Feature ◽  
Melanoma Classification ◽  
Selection Of

(1) Background: In this research, we aimed to identify and validate a set of relevant features to distinguish between benign nevi and melanoma lesions. (2) Methods: Two datasets with 70 melanomas and 100 nevi were investigated. The first one contained raw images. The second dataset contained images preprocessed for noise removal and uneven illumination reduction. Further, the images belonging to both datasets were segmented, followed by extracting features considered in terms of form/shape and color such as asymmetry, eccentricity, circularity, asymmetry of color distribution, quadrant asymmetry, fast Fourier transform (FFT) normalization amplitude, and 6th and 7th Hu’s moments. The FFT normalization amplitude is an atypical feature that is computed as a Fourier transform descriptor and focuses on geometric signatures of skin lesions using the frequency domain information. The receiver operating characteristic (ROC) curve and area under the curve (AUC) were employed to ascertain the relevance of the selected features and their capability to differentiate between nevi and melanoma. (3) Results: The ROC curves and AUC were employed for all experiments and selected features. A comparison in terms of the accuracy and AUC was performed, and an evaluation of the performance of the analyzed features was carried out. (4) Conclusions: The asymmetry index and eccentricity, together with F6 Hu’s invariant moment, were fairly competent in providing a good separation between malignant melanoma and benign lesions. Also, the FFT normalization amplitude feature should be exploited due to showing potential in classification.

57Co-Bleomycin Scintigraphy in Staging of Carcinoma of the Cervix

1985 ◽  
Vol 24 (01) ◽  
pp. 48-51 ◽  
Author(s):  
K. Dursk ◽  
T. Pertyński
Keyword(s):  
Cervical Cancer ◽  
Roc Curves ◽  
Selection Of

Summary65 patients with cervical cancer at various stages of the disease were studied scintigraphically after 57Co-bleomycin administration, active urine being flushed out from the bladder continuously. The patients were submitted to other auxiliary tests and the interpretation of scintigraphic images could thus be verified. Scintigraphic data were classified semiquantitatively by scoring in the range from zero to 5. Results were evaluated in terms of ROC-curves. This method of analysis permitted selection of cut-off levels for binary decisions that yielded satisfactory sensitivity and acceptable specificity of the test.

Evaluation of the clot burden score (cbs) for acute ischemic stroke (ais) in intent-to-treat patients.

2019 ◽  
Vol 1 (1) ◽  
pp. 11-15 ◽  
Author(s):  
Sarah Yaziz ◽  
Ahmad Sobri Muda ◽  
Wan Asyraf Wan Zaidi ◽  
Nik Azuan Nik Ismail
Keyword(s):  
Ischemic Stroke ◽  
Clinical Outcome ◽  
Acute Ischemic Stroke ◽  
Predictive Value ◽  
Area Under The Curve ◽  
Rank Correlation ◽  
Treatment Decision ◽  
Roc Curves ◽  
Significant Negative Correlation ◽  
Clot Burden

Background : The clot burden score (CBS) is a scoring system used in acute ischemic stroke (AIS) to predict patient outcome and guide treatment decision. However, CBS is not routinely practiced in many institutions. This study aimed to investigate the feasibility of CBS as a relevant predictor of good clinical outcome in AIS cases. Methods:  A retrospective data collection and review of AIS patients in a teaching hospital was done from June 2010 until June 2015. Patients were selected following the inclusion and exclusion criteria. These patients were followed up after 90 days of discharge. The Modified Rankin scale (mRS) was used to assess their outcome (functional status). Linear regression Spearman Rank correlation was performed between the CBS and mRS. The quality performance of the correlations was evaluated using Receiver operating characteristic (ROC) curves. Results: A total of 89 patients with AIS were analysed, 67.4% (n=60) male and 32.6% (n=29) female. Twenty-nine (29) patients (33.7%) had a CBS ?6, 6 patients (6.7%) had CBS <6, while 53 patients (59.6%) were deemed clot free. Ninety (90) days post insult, clinical assessment showed that 57 (67.6%) patients were functionally independent, 27 (30.3%) patients functionally dependent, and 5 (5.6%) patients were deceased. Data analysis reported a significant negative correlation (r= -0.611, p<0.001). ROC curves analysis showed an area under the curve of 0.81 at the cut-off point of 6.5. This showed that a CBS of more than 6 predicted a good mRS clinical outcome in AIS patients; with sensitivity of 98.2%, specificity of 53.1%, positive predictive value (PPV) of 76%, and negative predictive value (NPV) of 21%. Conclusion: CBS is a useful additional variable for the management of AIS cases, and should be incorporated into the routine radiological reporting for acute ischemic stroke (AIS) cases.

scholarly journals Magnetic resonance cisternography imaging findings related to the leakage of Gadolinium into the subarachnoid space

2021 ◽  
Author(s):  
Rei Nakamichi ◽  
Toshiaki Taoka ◽  
Hisashi Kawai ◽  
Tadao Yoshida ◽  
Michihiko Sone ◽  
...  
Keyword(s):  
Magnetic Resonance ◽  
Subarachnoid Space ◽  
Area Under The Curve ◽  
Rank Correlation ◽  
Predictive Performance ◽  
Roc Curves ◽  
Imaging Findings ◽  
Spearman’S Rank Correlation ◽  
Sagittal Sinus ◽  
Magnetic Resonance Cisternography

Abstract Purpose To identify magnetic resonance cisternography (MRC) imaging findings related to Gadolinium-based contrast agent (GBCA) leakage into the subarachnoid space. Materials and methods The number of voxels of GBCA leakage (V-leak) on 3D-real inversion recovery images was measured in 56 patients scanned 4 h post-intravenous GBCA injection. Bridging veins (BVs) were identified on MRC. The numbers of BVs with surrounding cystic structures (BV-cyst), with arachnoid granulations protruding into the superior sagittal sinus (BV-AG-SSS) and the skull (BV-AG-skull), and including any of these factors (BV-incl) were recorded. Correlations between these variables and V-leak were examined based on the Spearman’s rank correlation coefficient. Receiver-operating characteristic (ROC) curves were generated to investigate the predictive performance of GBCA leakage. Results V-leak and the number of BV-incl were strongly correlated (r = 0.609, p < 0.0001). The numbers of BV-cyst and BV-AG-skull had weaker correlations with V-leak (r = 0.364, p = 0.006; r = 0.311, p = 0.020, respectively). The number of BV-AG-SSS was not correlated with V-leak. The ROC curve for contrast leakage exceeding 1000 voxels and the number of BV-incl had moderate accuracy, with an area under the curve of 0.871. Conclusion The number of BV-incl may be a predictor of GBCA leakage and a biomarker for waste drainage function without using GBCA.

Factors affecting progression to permanent atrial fibrillation in an unselected population of patients with non-permanent form of atrial fibrillation

2020 ◽  
Vol 41 (Supplement_2) ◽  
Author(s):  
V.L Malavasi ◽  
E Fantecchi ◽  
V Tordoni ◽  
L Melara ◽  
A Barbieri ◽  
...  
Keyword(s):  
Atrial Fibrillation ◽  
Cox Regression ◽  
Area Under The Curve ◽  
Roc Curves ◽  
Factors Affecting ◽  
Cox Proportional Hazard ◽  
Significant Difference ◽  
History Of ◽  
Unselected Population ◽  
Cox Proportional Hazard Regression

Abstract Background Natural history of atrial fibrillation (AF) shows a progression of arrhythmia from non-permanent to permanent AF. Permanent AF was found associated with a worse prognosis than non-permanent one. Aim To assess the factors associated with progression to permanent AF in an unselected population of AF patients with non-permanent AF. Methods In this prospective study we enrolled in- as well as out-patients with non-permanent AF and age ≥18 years, with at least one episode of ECG-documented AF within 1 year. The patients were followed-up at 1 month and every 6 months thereafter. Results Out of 523 patients, 314 (60%) were in non-permanent AF (80 [25.5%] paroxysmal AF, 165 [52.5%] persistent AF, 69 [2%] first diagnosed AF), mostly male (188, 59.9%), median age 71 years (IQ range 62–77), median CHA2DS2VASc 3 (1–4), median HATCH score 1 (1–2). After a median follow-up of 701 (IQ range 437–902) days, 66 patients (21%) showed permanent AF. CHA2DS2VASc and HATCH scores were incrementally associated to progression to permanent AF (CHA2DS2VASc χ2 p=0.001; HATCH χ2 p=0.017; p for trend CHA2DS2VASc &lt;0.001, HATCH p=0.001). At multivariable Cox proportional hazard regression the following variables were significantly associated with AF progression: age (hazard ratio [HR] 1.041; 95% CI: 1.004–1.079; p=0.028), at least moderate left atrial (LA) enlargement (&gt;42 ml/m2) (HR 2.092; 95% CI: 1.132–3.866; p=0.018), antiarrhythmics drugs after the enrollment (HR 0.087; 95% CI: 0.011–0.662; p=0.018), EHRA score &gt;2 (HR 0.351; 95% CI: 0.158–0.779; p=0.010) and Valvular HD (HR 2.161; 95% CI: 1.057–4.420; p=0.035). Adding LA dilation to HATCH score (HATCH-LA) and assigning 2 points based on multivariable Cox regression, HATCH-LA was statistically better in ROC curves in prediction of AF progression vs HATCH score (area under the curve 0.695 vs 0.636; DeLong p=0.0225). Survival-free curves on freedom from permanent AF using as discriminator HATCH-LA score ≤2 vs &gt;2 led to a statistically significant difference (χ2=16.080 p&lt;0.001), but the same was not found for HATCH score (χ2 =3.099; p=0.078). Conclusions In patients without permanent AF, progression of AF was independentely related to age, LA dilation, AF symptoms severity, antiarrhythmic drugs and Valvular HD. HATCH score predicted AF progression and adding to it LA dilation (at least moderate) improved patients stratification for the risk of evolution to permanent AF. Funding Acknowledgement Type of funding source: None

Gestational Age Dependency of Umbilical Cord Serum IL-6 Levels for Detecting Fetal Inflammation

2020 ◽  
Author(s):  
Sota Iwatani ◽  
Takao Kobayashi ◽  
Sachiko Matsui ◽  
Akihiro Hirata ◽  
Miwa Yamamoto ◽  
...  
Keyword(s):  
Umbilical Cord ◽  
Gestational Age ◽  
Area Under The Curve ◽  
Roc Curves ◽  
Significant Negative Correlation ◽  
Cord Serum ◽  
Very Preterm ◽  
Key Points ◽  
Extremely Preterm ◽  
Preterm Newborns

Objective The fetal inflammatory response syndrome (FIRS) is characterized by elevated concentrations of inflammatory cytokines in fetal blood, with preterm delivery and morbidity. Umbilical cord serum interleukin-6 (UC-s-IL-6) is an ideal marker for detecting FIRS. However, the effect of gestational age (GA) on UC-s-IL-6 levels has not been reported. This study aimed to determine the relationship between GA and UC-s-IL-6 levels, and GA-dependent cutoff values of UC-s-IL-6 levels for detecting fetal inflammation. Study Design UC-s-IL-6 concentrations were measured in 194 newborns (44 extremely preterm newborns (EPNs) at 22–27 weeks' GA, 68 very preterm newborns (VPNs) at 28–31 weeks' GA, and 82 preterm newborns (PNs) at 32–34 weeks' GA). Linear regression analyses were used to correlate GA and UC-s-IL-6 levels. Receiver operating characteristic (ROC) curves analyses were performed for detecting the presence of funisitis, as the histopathological counterpart of FIRS. Results A significant negative correlation between GA and UC-s-IL-6 levels was found in newborns with severe funisitis (r s =  − 0.427, p = 0.004) and those with mild funisitis (r s =  − 0.396, p = 0.025). ROC curve analyses revealed the area under the curve for detecting funisitis were 0.856, 0.837, and 0.622 in EPNs, VPNs, and PNs, respectively. The UC-s-IL-6 cutoff value in EPNs (28.1 pg/mL) exceeded those in VPNs and PNs (3.7 and 3.0 pg/mL, respectively). Conclusion UC-s-IL-6 levels were inversely correlated with GA especially in newborns with funisitis. Such GA dependency of UC-s-IL-6 should be considered for detecting fetal inflammation. Key Points

scholarly journals A novel predictor of clinical progression in patients on active surveillance for prostate cancer

2019 ◽  
Vol 13 (8) ◽  
Author(s):  
Guan Hee Tan ◽  
Antonio Finelli ◽  
Ardalan Ahmad ◽  
Marian Wettstein ◽  
Alexandre Zlotta ◽  
...  
Keyword(s):  
Prostate Cancer ◽  
Active Surveillance ◽  
Cox Regression ◽  
Area Under The Curve ◽  
Progression Free Survival ◽  
Roc Curves ◽  
Standard Of Care ◽  
Low Risk ◽  
Clinical Progression

Introduction: Active surveillance (AS) is standard of care in low-risk prostate cancer (PC). This study describes a novel total cancer location (TCLo) density metric and aims to determine its performance in predicting clinical progression (CP) and grade progression (GP).     Methods: This was a retrospective study of patients on AS after confirmatory biopsy (CBx). We excluded patients with Gleason ≥7 at CBx and <2 years follow-up. TCLo was the number of locations with positive cores at diagnosis (DBx) and CBx. TCLo density was TCLo / prostate volume (PV). CP was progression to any active treatment while GP occurred if Gleason ≥7 was identified on repeat biopsy or surgical pathology. Independent predictors of time to CP or GP were estimated with Cox regression. Kaplan-Meier analysis compared progression-free survival curves between TCLo density groups. Test characteristics of TCLo were explored with receiver operating characteristic (ROC) curves.     Results: We included 181 patients who had CBx between 2012-2015, and met inclusion criteria. The mean age of patients was 62.58 years (SD=7.13) and median follow-up was 60.9 months (IQR=23.4). A high TCLo density score (>0.05) was independently associated with time to CP (HR 4.70, 95% CI: 2.62-8.42, p<0.001), and GP (HR 3.85, 95% CI: 1.91-7.73, p<0.001). ROC curves showed TCLo density has greater area under the curve than number of positive cores at CBx in predicting progression.     Conclusion: TCLo density is able to stratify patients on AS for risk of CP and GP. With further validation, it could be added to the decision-making algorithm in AS for low-risk localized PC.

ESR and CRP Diagnostic Thresholds for Prosthetic Joint Infection in Hip Hemiarthroplasty

2020 ◽  
Vol 4 (04) ◽  
pp. 187-192
Author(s):  
Jared A. Warren ◽  
Oliver Scotting ◽  
Hiba K. Anis ◽  
James Bircher ◽  
Alison K. Klika ◽  
...  
Keyword(s):  
Prosthetic Joint Infection ◽  
Joint Infection ◽  
Area Under The Curve ◽  
Roc Curves ◽  
Musculoskeletal Infection ◽  
Hip Hemiarthroplasty ◽  
Diagnostic Thresholds ◽  
Diagnostic Threshold ◽  
Prosthetic Joint ◽  
The Mean

AbstractDiagnostic thresholds used to standardize the definition for prosthetic joint infection (PJI) have largely focused on total joint arthroplasty (TJA). Established PJI thresholds exist for serum erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) in TJA; however, they do not exist for revision hip hemiarthroplasty (rHHA). The purpose of this study was to establish thresholds for (1) ESR and (2) CRP to diagnose PJI in rHHA. Data were collected on a prospective cohort of 69 rHHA patients undergoing orthopaedic surgery between 1/2017 and 2/2019 in a single health care system. Procedures were categorized as septic or aseptic revisions using Musculoskeletal Infection Society (MSIS) criteria (2013). There were 44 ESRs (n = 28 aseptic, n = 16 septic) and 46 CRPs (n = 29 aseptic, n = 17 septic) available for analysis. Two tailed t-tests were performed to compare the mean ESR and CRP in aseptic and septic cases. Receiver operator characteristic (ROC) curves were generated to obtain diagnostic cutoff thresholds using the Youden's Index (J) for ESR and CRP. The mean ESR was 50.3 ± 30.6 mm/h versus 15.4 ± 17.7 mm/h (p < 0.001), while the mean CRP was 29.9 ± 24.8 mg/L versus 4.1 ± 8.2 mg/L (p < 0.001) for septic and aseptic revisions, respectively. The diagnostic threshold for PJI determined by the ROC curve was 44 mm/h for ESR (sensitivity = 56.3%; specificity = 100.0%; J = 0.563; area under the curve (AUC) = 0.845), while it was 12.5 mg/L for CRP (sensitivity = 70.6%; specificity = 96.6%; J = 0.672; AUC = 0.896). For patients with HHA, an ESR of 44 mm/h was and a CRP of 12.5 mg/L was highly specific for PJI. The thresholds are similar to the MSIS thresholds currently published. Larger prospective trials are needed to establish more robust and conclusive diagnostic criteria for PJI in HHA, including investigations not only of ESR and CRP but synovial white blood cell count and synovial polymorphonuclear leukocytes % as well.

scholarly journals Deep learning for accurately recognizing common causes of shoulder pain on radiographs

2021 ◽  
Author(s):  
Nils F. Grauhan ◽  
Stefan M. Niehues ◽  
Robert A. Gaudin ◽  
Sarah Keller ◽  
Janis L. Vahldiek ◽  
...  
Keyword(s):  
Shoulder Pain ◽  
Sensitivity And Specificity ◽  
Area Under The Curve ◽  
Roc Curves ◽  
Clinical Situation ◽  
Assistive Device ◽  
Test Dataset ◽  
Joint Dislocation ◽  
Independent Expert ◽  
Common Causes

Abstract Objective Training a convolutional neural network (CNN) to detect the most common causes of shoulder pain on plain radiographs and to assess its potential value in serving as an assistive device to physicians. Materials and methods We used a CNN of the ResNet-50 architecture which was trained on 2700 shoulder radiographs from clinical practice of multiple institutions. All radiographs were reviewed and labeled for six findings: proximal humeral fractures, joint dislocation, periarticular calcification, osteoarthritis, osteosynthesis, and joint endoprosthesis. The trained model was then evaluated on a separate test dataset, which was previously annotated by three independent expert radiologists. Both the training and the test datasets included radiographs of highly variable image quality to reflect the clinical situation and to foster robustness of the CNN. Performance of the model was evaluated using receiver operating characteristic (ROC) curves, the thereof derived AUC as well as sensitivity and specificity. Results The developed CNN demonstrated a high accuracy with an area under the curve (AUC) of 0.871 for detecting fractures, 0.896 for joint dislocation, 0.945 for osteoarthritis, and 0.800 for periarticular calcifications. It also detected osteosynthesis and endoprosthesis with near perfect accuracy (AUC 0.998 and 1.0, respectively). Sensitivity and specificity were 0.75 and 0.86 for fractures, 0.95 and 0.65 for joint dislocation, 0.90 and 0.86 for osteoarthrosis, and 0.60 and 0.89 for calcification. Conclusion CNNs have the potential to serve as an assistive device by providing clinicians a means to prioritize worklists or providing additional safety in situations of increased workload.

scholarly journals Validation of the Zwolle score for selection of very low-risk STEMI patients treated with primary angioplasty

2020 ◽  
Vol 41 (Supplement_2) ◽  
Author(s):  
A Cordero ◽  
B Cid ◽  
P Monteiro ◽  
J.M Garcia-Acuna ◽  
M Rodriguez-Manero ◽  
...  
Keyword(s):  
Hospital Mortality ◽  
Risk Score ◽  
Mortality Risk ◽  
Area Under The Curve ◽  
Coronary Intervention ◽  
Low Risk ◽  
Individual Risk ◽  
Care Organization ◽  
Grace Score ◽  
Selection Of

Abstract Background The Zwolle risk score was designed to stratify the actual in-hospital mortality risk of ST-elevation myocardial infarction (STEMI) patients treated with primary percutaneous coronary intervention (p-PCI) but, also, for decision-making related to patients location in an intensive care unit or not. Since the GRACE score continues being the gold-standard for individual risk assessment in STEMI in most institutions we assessed the specificity of both scores for in-hospital mortality. Methods We assessed the accuracy of Zwolle risk score for in-hospital mortality estimation as compared to the GRACE score in all patients admitted for STEMI in 3 tertitary hospitals. Patients with Zwolle risk score &lt;3 would qualify as “low risk”, 3–5 as “intermediate risk” and ≥6 as “high risk”. Patients with GRACE score &lt;140 were classified as low-risk. Specificity, sensitivity and classification were assessed by ROC curves and the area under the curve (AUC). Results We included 4,446 patients, mean age 64.7 (13.6) years, 24% women and 39% with diabetes. Mean GRACE score was 157.3 (4.9) and Zwolle was 2.8 (3.3). In-hospital mortality was 10.6% (471 patients). Patients who died had higher GRACE score (218.4±4.9 vs. 149.6±37.5; p&lt;0.001) and Zwolle score (7.6±4.3 vs. 2.3±2.18; p&lt;0.001); a statistically significant increase of in-hospital mortality risk, adjusted adjusted by age, gender and revascularization, was observed with both scores (figure). A total of 1,629 patients (40.0%) were classified as low risk by the GRACE score and 2,962 (66.6%) by the Zwolle score; in-hospital mortality was 1.6% and 2.7%, respectively. Moreover, the was a significant increase of in-hospital mortality rate according to Zwolle categories (2.7%; 13.0%; 41.6%)The AUC of both score was the same (p=0.49) but the specificity of GRACE score &lt;140 was 43.1% as compared to 72.6% obtained by Zwolle score &lt;3; patients accurately classified was also lower with the GRACE score threshold (48.8% vs. 73.7%). Conclusions Selection of low-risk STEMI patients treated with p-PCI based on the Zwolle risk score has higher specificity than the GRACE score and might be useful for the care organization in clinical practice. Funding Acknowledgement Type of funding source: None

scholarly journals Texture-Based Analysis of 18F-Labeled Amyloid PET Brain Images

2021 ◽  
Vol 11 (5) ◽  
pp. 1991
Author(s):  
Alexander P. Seiffert ◽  
Adolfo Gómez-Grande ◽  
Eva Milara ◽  
Sara Llamas-Velasco ◽  
Alberto Villarejo-Galende ◽  
...  
Keyword(s):  
Area Under The Curve ◽  
Standardized Uptake Value ◽  
Texture Features ◽  
Roc Curves ◽  
Significant Feature ◽  
Study Cohort ◽  
Amyloid Pet ◽  
Reference Region ◽  
Brain Images ◽  
Positron Emission

Amyloid positron emission tomography (PET) brain imaging with radiotracers like [18F]florbetapir (FBP) or [18F]flutemetamol (FMM) is frequently used for the diagnosis of Alzheimer’s disease. Quantitative analysis is usually performed with standardized uptake value ratios (SUVR), which are calculated by normalizing to a reference region. However, the reference region could present high variability in longitudinal studies. Texture features based on the grey-level co-occurrence matrix, also called Haralick features (HF), are evaluated in this study to discriminate between amyloid-positive and negative cases. A retrospective study cohort of 66 patients with amyloid PET images (30 [18F]FBP and 36 [18F]FMM) was selected and SUVRs and 6 HFs were extracted from 13 cortical volumes of interest. Mann–Whitney U-tests were performed to analyze differences of the features between amyloid positive and negative cases. Receiver operating characteristic (ROC) curves were computed and their area under the curve (AUC) was calculated to study the discriminatory capability of the features. SUVR proved to be the most significant feature among all tests with AUCs between 0.692 and 0.989. All HFs except correlation also showed good performance. AUCs of up to 0.949 were obtained with the HFs. These results suggest the potential use of texture features for the classification of amyloid PET images.

Sign in / Sign up

Export Citation Format

Share Document