scholarly journals Precise Identification of Prostate Cancer from DWI Using Transfer Learning

Sensors ◽  
2021 ◽  
Vol 21 (11) ◽  
pp. 3664
Author(s):  
Islam R. Abdelmaksoud ◽  
Ahmed Shalaby ◽  
Ali Mahmoud ◽  
Mohammed Elmogy ◽  
Ahmed Aboelfetouh ◽  
...  
Keyword(s):  
Prostate Cancer ◽  
Transfer Learning ◽  
Sensitivity And Specificity ◽  
Benign Tumors ◽  
B Values ◽  
Cad System ◽  
Average Accuracy ◽  
Apparent Diffusion ◽  
Invasive Techniques ◽  
Fine Tune

Background and Objective: The use of computer-aided detection (CAD) systems can help radiologists make objective decisions and reduce the dependence on invasive techniques. In this study, a CAD system that detects and identifies prostate cancer from diffusion-weighted imaging (DWI) is developed. Methods: The proposed system first uses non-negative matrix factorization (NMF) to integrate three different types of features for the accurate segmentation of prostate regions. Then, discriminatory features in the form of apparent diffusion coefficient (ADC) volumes are estimated from the segmented regions. The ADC maps that constitute these volumes are labeled by a radiologist to identify the ADC maps with malignant or benign tumors. Finally, transfer learning is used to fine-tune two different previously-trained convolutional neural network (CNN) models (AlexNet and VGGNet) for detecting and identifying prostate cancer. Results: Multiple experiments were conducted to evaluate the accuracy of different CNN models using DWI datasets acquired at nine distinct b-values that included both high and low b-values. The average accuracy of AlexNet at the nine b-values was 89.2±1.5% with average sensitivity and specificity of 87.5±2.3% and 90.9±1.9%. These results improved with the use of the deeper CNN model (VGGNet). The average accuracy of VGGNet was 91.2±1.3% with sensitivity and specificity of 91.7±1.7% and 90.1±2.8%. Conclusions: The results of the conducted experiments emphasize the feasibility and accuracy of the developed system and the improvement of this accuracy using the deeper CNN.

scholarly journals Deep Transfer Learning for Improved Detection of Keratoconus using Corneal Topographic Maps

2021 ◽  
Author(s):  
Ali H. Al-Timemy ◽  
Nebras H. Ghaeb ◽  
Zahraa M. Mosa ◽  
Javier Escudero
Keyword(s):  
Transfer Learning ◽  
Learning Strategy ◽  
Clinical Decision ◽  
Topographic Maps ◽  
Computer Assisted ◽  
Clinical Knowledge ◽  
Average Accuracy ◽  
Assisted Diagnosis ◽  
Improved Accuracy ◽  
Fine Tune

Abstract Clinical keratoconus (KCN) detection is a challenging and time-consuming task. In the diagnosis process, ophthalmologists must revise demographic and clinical ophthalmic examinations. The latter include slit-lamb, corneal topographic maps, and Pentacam indices (PI). We propose an Ensemble of Deep Transfer Learning (EDTL) based on corneal topographic maps. We consider four pretrained networks, SqueezeNet (SqN), AlexNet (AN), ShuffleNet (SfN), and MobileNet-v2 (MN), and fine-tune them on a dataset of KCN and normal cases, each including four topographic maps. We also consider a PI classifier. Then, our EDTL method combines the output probabilities of each of the five classifiers to obtain a decision based on the fusion of probabilities. Individually, the classifier based on PI achieved 93.1% accuracy, whereas the deep classifiers reached classification accuracies over 90% only in isolated cases. Overall, the average accuracy of the deep networks over the four corneal maps ranged from 86% (SfN) to 89.9% (AN). The classifier ensemble increased the accuracy of the deep classifiers based on corneal maps to values ranging (92.2% to 93.1%) for SqN and (93.1% to 94.8%) for AN. Including in the ensemble-specific combinations of corneal maps’ classifiers and PI increased the accuracy to 98.3%. Moreover, visualization of first learner filters in the networks and Grad-CAMs confirmed that the networks had learned relevant clinical features. This study shows the potential of creating ensembles of deep classifiers fine-tuned with a transfer learning strategy as it resulted in an improved accuracy while showing learnable filters and Grad-CAMs that agree with clinical knowledge. This is a step further towards the potential clinical deployment of an improved computer-assisted diagnosis system for KCN detection to help ophthalmologists to confirm the clinical decision and to perform fast and accurate KCN treatment.

scholarly journals Accuracy of high b-value diffusion-weighted MRI for prostate cancer detection: a meta-analysis

2017 ◽  
Vol 59 (1) ◽  
pp. 105-113 ◽  
Author(s):  
Keith Craig Godley ◽  
Tom Joseph Syer ◽  
Andoni Paul Toms ◽  
Toby Oliver Smith ◽  
Glyn Johnson ◽  
...  
Keyword(s):  
Prostate Cancer ◽  
Diagnostic Accuracy ◽  
Sensitivity And Specificity ◽  
Cancer Detection ◽  
Diagnostic Performance ◽  
Visual Assessment ◽  
B Value ◽  
B Values ◽  
Prostate Cancer Detection ◽  
High B Value

Background The diagnostic accuracy of diffusion-weighted imaging (DWI) to detect prostate cancer is well-established. DWI provides visual as well as quantitative means of detecting tumor, the apparent diffusion coefficient (ADC). Recently higher b-values have been used to improve DWI’s diagnostic performance. Purpose To determine the diagnostic performance of high b-value DWI at detecting prostate cancer and whether quantifying ADC improves accuracy. Material and Methods A comprehensive literature search of published and unpublished databases was performed. Eligible studies had histopathologically proven prostate cancer, DWI sequences using b-values ≥ 1000 s/mm2, less than ten patients, and data for creating a 2 × 2 table. Study quality was assessed with QUADAS-2 (Quality Assessment of diagnostic Accuracy Studies). Sensitivity and specificity were calculated and tests for statistical heterogeneity and threshold effect performed. Results were plotted on a summary receiver operating characteristic curve (sROC) and the area under the curve (AUC) determined the diagnostic performance of high b-value DWI. Results Ten studies met eligibility criteria with 13 subsets of data available for analysis, including 522 patients. Pooled sensitivity and specificity were 0.59 (95% confidence interval [CI], 0.57–0.61) and 0.92 (95% CI, 0.91–0.92), respectively, and the sROC AUC was 0.92. Subgroup analysis showed a statistically significant ( P = 0.03) improvement in accuracy when using tumor visual assessment rather than ADC. Conclusion High b-value DWI gives good diagnostic performance for prostate cancer detection and visual assessment of tumor diffusion is significantly more accurate than ROI measurements of ADC.

scholarly journals Evaluation of Exponential ADC (eADC) and Computed DWI (cDWI) for the Detection of Prostate Cancer

2018 ◽  
Vol 190 (08) ◽  
pp. 758-766 ◽  
Author(s):  
Alois Sprinkart ◽  
Christian Marx ◽  
Frank Träber ◽  
Wolfgang Block ◽  
Daniel Thomas ◽  
...  
Keyword(s):  
Prostate Cancer ◽  
High Risk ◽  
Repeated Measures ◽  
Low Risk ◽  
Group Differences ◽  
B Values ◽  
Key Points ◽  
Apparent Diffusion ◽  
Malignant Lesions ◽  
Post Hoc

Purpose To directly compare different methods proposed for enhanced conspicuity and discriminability of prostate cancer on diffusion-weighted imaging (DWI) and to compare the results to original DWI images and conventional apparent diffusion coefficient (ADC) maps. Materials and Methods Clinical routine prostate DWI datasets (b = 0, 50, 800 s/mm², acquired at a field strength of 3 T) of 104 consecutive patients with subsequent MR-guided prostate biopsy were included in this retrospective study. For each dataset exponential ADC maps (eADC), computed DWI images (cDWI), and additionally eADC maps for computed b-values of 2000 and 3000 s/mm² were generated (c_eADC). For each of 123 lesions, the contrast (CR) and contrast-to-noise ratio (CNR) were determined. Differences in the CR and CNR of malignant lesions (n = 83) between the different image types and group differences between benign (n = 40), low-risk (n = 53) and high-risk (n = 30) lesions were assessed by repeated measures ANOVA and one-way ANOVA with post-hoc tests. The ability to differentiate between benign and malignant and between low-risk and high-risk lesions was assessed by receiver operating characteristic (ROC) curve analyses. Results The CR and CNR were higher for computed DWI and related c_eADC at b = 3000 s/mm² and 2000 s/mm² compared to original DWI, conventional ADC and standard eADC. For differentiation of benign and malignant lesions, conventional ADC and CR of conventional ADC were best suited. For discrimination of low-risk from high-risk lesions, the CR of c_eADC was best suited followed by the CR of cDWI. Conclusion Computed cDWI or related c_eADC maps at b-values between 2000 and 3000 s/mm2 were superior to the original DWI, conventional ADC and eADC in the detection of prostate cancer. Key Points  Citation Format

scholarly journals Radiomics Driven Diffusion Weighted Imaging Sensing Strategies for Zone-Level Prostate Cancer Sensing

Sensors ◽  
2020 ◽  
Vol 20 (5) ◽  
pp. 1539 ◽  
Author(s):  
Chris Dulhanty ◽  
Linda Wang ◽  
Maria Cheng ◽  
Hayden Gunraj ◽  
Farzad Khalvati ◽  
...  
Keyword(s):  
Prostate Cancer ◽  
Sensitivity And Specificity ◽  
Diffusion Weighted Imaging ◽  
B Value ◽  
Invasive Approach ◽  
Non Invasive ◽  
High B Value ◽  
Diffusion Weighted ◽  
Apparent Diffusion ◽  
Critical Regions

Prostate cancer is the most commonly diagnosed cancer in North American men; however, prognosis is relatively good given early diagnosis. This motivates the need for fast and reliable prostate cancer sensing. Diffusion weighted imaging (DWI) has gained traction in recent years as a fast non-invasive approach to cancer sensing. The most commonly used DWI sensing modality currently is apparent diffusion coefficient (ADC) imaging, with the recently introduced computed high-b value diffusion weighted imaging (CHB-DWI) showing considerable promise for cancer sensing. In this study, we investigate the efficacy of ADC and CHB-DWI sensing modalities when applied to zone-level prostate cancer sensing by introducing several radiomics driven zone-level prostate cancer sensing strategies geared around hand-engineered radiomic sequences from DWI sensing (which we term as Zone-X sensing strategies). Furthermore, we also propose Zone-DR, a discovery radiomics approach based on zone-level deep radiomic sequencer discovery that discover radiomic sequences directly for radiomics driven sensing. Experimental results using 12,466 pathology-verified zones obtained through the different DWI sensing modalities of 101 patients showed that: (i) the introduced Zone-X and Zone-DR radiomics driven sensing strategies significantly outperformed the traditional clinical heuristics driven strategy in terms of AUC, (ii) the introduced Zone-DR and Zone-SVM strategies achieved the highest sensitivity and specificity, respectively for ADC amongst the tested radiomics driven strategies, (iii) the introduced Zone-DR and Zone-LR strategies achieved the highest sensitivities for CHB-DWI amongst the tested radiomics driven strategies, and (iv) the introduced Zone-DR, Zone-LR, and Zone-SVM strategies achieved the highest specificities for CHB-DWI amongst the tested radiomics driven strategies. Furthermore, the results showed that the trade-off between sensitivity and specificity can be optimized based on the particular clinical scenario we wish to employ radiomic driven DWI prostate cancer sensing strategies for, such as clinical screening versus surgical planning. Finally, we investigate the critical regions within sensing data that led to a given radiomic sequence generated by a Zone-DR sequencer using an explainability method to get a deeper understanding on the biomarkers important for zone-level cancer sensing.

scholarly journals Synthetic apparent diffusion coefficient for high b-value diffusion weighted MRI in Prostate

2019 ◽  
Author(s):  
Prativa Sahoo ◽  
Russell Rockne ◽  
Jung Alexander ◽  
Pradeep K Gupta ◽  
Rakesh K Gupta
Keyword(s):  
Prostate Cancer ◽  
Diffusion Coefficient ◽  
Apparent Diffusion Coefficient ◽  
B Value ◽  
T Test ◽  
B Values ◽  
Apparent Diffusion ◽  
Adc Values ◽  
Malignant Lesions ◽  
Log Linear

AbstractPurposeIt has been reported that diffusion weighted imaging (DWI) with ultrahigh b-value increases the diagnostic power of prostate cancer. DWI imaging with higher b-values is challenging as it commonly suffers from low signal to noise ratio (SNR), distortion and longer scan time. The aim of our study was to develop a technique for quantification of apparent diffusion coefficient (ADC) for higher b-values from lower b-value DW images.Materials and MethodsFifteen patient (7 malignant, 8 benign) with prostate cancer were included in this study retrospectively with the institutional ethical committee approval. All images were acquired at 3T MR scanner. The ADC values were calculated using mono-exponential model. Synthetic ADC (sADC) for higher b-value were computed using a log-linear model. Contrast ratio (CR) between prostate lesion and normal tissue on synthetic DWI (sDWI) was computed and compared with original DWI and ADC images.ResultsNo significant difference was observed between actual ADC and sADC for b-2000 in all prostate lesions. However; CR increased significantly (p=0.002, paired t-test) in sDWI as compared to DWI. Malignant lesions showed significantly lower sADC as compared to benign lesion (p=0.0116, independent t-test). Mean (±standard deviation) of sADC of malignant lesions was 0.601±0.06 and for benign lesions was 0.92 ± 0.09 (10−3mm2/s).Discussion / ConclusionOur initial investigation suggests that the ADC values corresponding to higher b-value can be computed using log-linear relationship derived from lower b-values (b≤1000). Our method might help clinician to decide the optimal b-value for prostate lesion identification.

Apparent Diffusion Coefficient for Prostate Cancer Imaging: Impact of b Values

2014 ◽  
Vol 202 (3) ◽  
pp. W247-W253 ◽  
Author(s):  
Yahui Peng ◽  
Yulei Jiang ◽  
Tatjana Antic ◽  
Ila Sethi ◽  
Christine Schmid-Tannwald ◽  
...  
Keyword(s):  
Prostate Cancer ◽  
Diffusion Coefficient ◽  
Apparent Diffusion Coefficient ◽  
Cancer Imaging ◽  
B Values ◽  
Apparent Diffusion

Improved Sensitivity and Specificity for Detection of Prostate Cancer

2007 ◽  
Author(s):  
Rao P. Gullapalli ◽  
Michael Naslund ◽  
John Papasdimitrou ◽  
Elliot Siegel
Keyword(s):  
Prostate Cancer ◽  
Sensitivity And Specificity

Diagnostic Value of Non-Invasive Prenatal Screening of β-thalassemia by Cell Free Fetal DNA and Fetal NRBC

2019 ◽  
Vol 19 (2) ◽  
pp. 105-111
Author(s):  
Nadia Shafei ◽  
Mohammad Saeed Hakhamaneshi ◽  
Massoud Houshmand ◽  
Siavash Gerayeshnejad ◽  
Fardin Fathi ◽  
...  
Keyword(s):  
Sensitivity And Specificity ◽  
Multiple Displacement Amplification ◽  
Diagnostic Value ◽  
Diagnostic Methods ◽  
Beta Thalassemia ◽  
Non Invasive ◽  
Fetal Dna ◽  
Prenatal Diagnostic ◽  
Cell Free Fetal Dna ◽  
Invasive Techniques

Background: Beta thalassemia is a common disorder with autosomal recessive inheritance. The most prenatal diagnostic methods are the invasive techniques that have the risk of miscarriage. Now the non-invasive methods will be gradually alternative for these invasive techniques. Objective: The aim of this study is to evaluate and compare the diagnostic value of two non-invasive diagnostic methods for fetal thalassemia using cell free fetal DNA (cff-DNA) and nucleated RBC (NRBC) in one sampling community. Methods: 10 ml of blood was taken in two k3EDTA tube from 32 pregnant women (mean of gestational age = 11 weeks), who themselves and their husbands had minor thalassemia. One tube was used to enrich NRBC and other was used for cff-DNA extraction. NRBCs were isolated by MACS method and immunohistochemistry; the genome of stained cells was amplified by multiple displacement amplification (MDA) procedure. These products were used as template in b-globin segments PCR. cff-DNA was extracted by THP method and 300 bp areas were recovered from the agarose gel as fetus DNA. These DNA were used as template in touch down PCR to amplify b-globin gen. The amplified b-globin segments were sequenced and the results compared with CVS resul. Results: The data showed that sensitivity and specificity of thalassemia diagnosis by NRBC were 100% and 92% respectively and sensitivity and specificity of thalassemia diagnosis by cff-DNA were 100% and 84% respectively. Conclusion: These methods with high sensitivity can be used as screening test but due to their lower specificity than CVS, they cannot be used as diagnostic test.

scholarly journals Diagnosis of transition zone prostate cancer by multiparametric MRI: added value of MR spectroscopic imaging with sLASER volume selection

2021 ◽  
Vol 28 (1) ◽  
Author(s):  
Neda Gholizadeh ◽  
Peter B. Greer ◽  
John Simpson ◽  
Jonathan Goodwin ◽  
Caixia Fu ◽  
...  
Keyword(s):  
Prostate Cancer ◽  
High Risk ◽  
Transition Zone ◽  
Sensitivity And Specificity ◽  
Cancer Detection ◽  
Diagnostic Performance ◽  
Spectroscopic Imaging ◽  
Multiparametric Mri ◽  
Low Risk ◽  
Intermediate Risk

Abstract Background Current multiparametric MRI (mp-MRI) in routine clinical practice has poor-to-moderate diagnostic performance for transition zone prostate cancer. The aim of this study was to evaluate the potential diagnostic performance of novel 1H magnetic resonance spectroscopic imaging (MRSI) using a semi-localized adiabatic selective refocusing (sLASER) sequence with gradient offset independent adiabaticity (GOIA) pulses in addition to the routine mp-MRI, including T2-weighted imaging (T2WI), diffusion-weighted imaging (DWI) and quantitative dynamic contrast enhancement (DCE) for transition zone prostate cancer detection, localization and grading. Methods Forty-one transition zone prostate cancer patients underwent mp-MRI with an external phased-array coil. Normal and cancer regions were delineated by two radiologists and divided into low-risk, intermediate-risk, and high-risk categories based on TRUS guided biopsy results. Support vector machine models were built using different clinically applicable combinations of T2WI, DWI, DCE, and MRSI. The diagnostic performance of each model in cancer detection was evaluated using the area under curve (AUC) of the receiver operating characteristic diagram. Then accuracy, sensitivity and specificity of each model were calculated. Furthermore, the correlation of mp-MRI parameters with low-risk, intermediate-risk and high-risk cancers were calculated using the Spearman correlation coefficient. Results The addition of MRSI to T2WI + DWI and T2WI + DWI + DCE improved the accuracy, sensitivity and specificity for cancer detection. The best performance was achieved with T2WI + DWI + MRSI where the addition of MRSI improved the AUC, accuracy, sensitivity and specificity from 0.86 to 0.99, 0.83 to 0.96, 0.80 to 0.95, and 0.85 to 0.97 respectively. The (choline + spermine + creatine)/citrate ratio of MRSI showed the highest correlation with cancer risk groups (r = 0.64, p < 0.01). Conclusion The inclusion of GOIA-sLASER MRSI into conventional mp-MRI significantly improves the diagnostic accuracy of the detection and aggressiveness assessment of transition zone prostate cancer.

Sign in / Sign up

Export Citation Format

Share Document