Prediction of Clinically Significant Cancer Using Radiomics Features of Pre-Biopsy of Multiparametric MRI in Men Suspected of Prostate Cancer

Background: Texture features based on the spatial relationship of pixels, known as the gray-level co-occurrence matrix (GLCM), may play an important role in providing the accurate classification of suspected prostate cancer. The purpose of this study was to use quantitative imaging parameters of pre-biopsy multiparametric magnetic resonance imaging (mpMRI) for the prediction of clinically significant prostate cancer. Methods: This was a prospective study, recruiting 200 men suspected of having prostate cancer. Participants were imaged using a protocol-based 3T MRI in the pre-biopsy setting. Radiomics parameters were extracted from the T2WI and ADC texture features of the gray-level co-occurrence matrix were delineated from the region of interest. Radical prostatectomy histopathology was used as a reference standard. A Kruskal–Wallis test was applied first to identify the significant radiomic features between the three groups of Gleason scores (i.e., G1, G2 and G3). Subsequently, the Holm–Bonferroni method was applied to correct and control the probability of false rejections. We compared the probability of correctly predicting significant prostate cancer between the explanatory GLCM radiomic features, PIRADS and PSAD, using the area under the receiver operation characteristic curves. Results: We identified the significant difference in radiomic features between the three groups of Gleason scores. In total, 12 features out of 22 radiomics features correlated with the Gleason groups. Our model demonstrated excellent discriminative ability (C-statistic = 0.901, 95%CI 0.859–0.943). When comparing the probability of correctly predicting significant prostate cancer between explanatory GLCM radiomic features (Sum Variance T2WI, Sum Entropy T2WI, Difference Variance T2WI, Entropy ADC and Difference Variance ADC), PSAD and PIRADS via area under the ROC curve, radiomic features were 35.0% and 34.4% more successful than PIRADS and PSAD, respectively, in correctly predicting significant prostate cancer in our patients (p < 0.001). The Sum Entropy T2WI score had the greatest impact followed by the Sum Variance T2WI. Conclusion: Quantitative GLCM texture analyses of pre-biopsy MRI has the potential to be used as a non-invasive imaging technique to predict clinically significant cancer in men suspected of having prostate cancer.

Download Full-text

NIMG-56. USING RADIOMIC FEATURES FROM DAILY MAGNETIC RESONANCE IMAGING TO PREDICT RESPONSE TO RADIATION THERAPY IN GLIOBLASTOMA PATIENTS: A PILOT STUDY

Neuro-Oncology ◽

10.1093/neuonc/noab196.554 ◽

2021 ◽

Vol 23 (Supplement_6) ◽

pp. vi142-vi142

Author(s):

Kaylie Cullison ◽

Garrett Simpson ◽

Danilo Maziero ◽

Kolton Jones ◽

Radka Stoyanova ◽

...

Keyword(s):

Radiation Therapy ◽

Tumor Growth ◽

Texture Feature ◽

Texture Features ◽

Gray Level ◽

Difference Matrix ◽

Patient Response ◽

Significant Difference ◽

Occurrence Matrix ◽

Surrounding Areas

Abstract A dilemma in treating glioblastoma is that MRI after chemotherapy and radiation therapy (chemoRT) shows areas of presumed tumor growth in up to 50% of patients. These areas can represent true progression (TP), tumor growth with tumors non-responsive to treatment, or pseudoprogression (PP), edema and tumor necrosis with favorable treatment response. On imaging, TP and PP are usually not discernable. Patients in this study undergo six weeks of chemoRT on a combination MRI/RT device, receiving daily MRIs. The goal of this study is to explore the correlation of radiomics features with progression. The tumor lesion and surrounding areas of growth/edema were manually outlined as regions of interest (ROIs) for each daily T2-weighted MRI scan. The ROIs were used to calculate texture features: statistical features based on the gray-level co-occurrence matrix (GLCM), the gray-level zone size matrix (GLZSM), the gray-level run length matrix (GLRLM), and the neighborhood gray-tone difference matrix (NGTDM). Each of these matrix classes describe the probability of spatial relationships of gray levels occurring within the ROI. Daily texture features were averaged per week of treatment for each patient. Patient response was retrospectively defined as no progression (NP), TP, or PP. A Kruskal-Wallis test was performed to identify texture features that correlated most strongly with patient response. Forty texture features were calculated for 12 patients (19 treated, 7 excluded due to no T2 lesion or progression status unknown, 6 NP, 3 TP, 3 PP). There was a trend of more texture features correlating significantly with response in weeks 4-6 of treatment, compared to weeks 1-3. A particular texture feature, GLSZM Small Zone Low Gray-Level Emphasis, showed increasing difference between PP and TP over time, with significant difference during week 6 of treatment (p=0.0495). Future directions include correlating early outcomes with greater numbers of patients and daily multiparametric MRI.

Download Full-text

The Primacy of High B-Value 3T-DWI Radiomics in the Prediction of Clinically Significant Prostate Cancer

Diagnostics ◽

10.3390/diagnostics11050739 ◽

2021 ◽

Vol 11 (5) ◽

pp. 739

Author(s):

Alessandro Bevilacqua ◽

Margherita Mottola ◽

Fabio Ferroni ◽

Alice Rossi ◽

Giampaolo Gavelli ◽

...

Keyword(s):

Prostate Cancer ◽

Quantitative Imaging ◽

B Value ◽

Support Vector ◽

Svm Classifier ◽

Primary Role ◽

High B Value ◽

Wilcoxon Rank Sum Test ◽

Apparent Diffusion ◽

Clinically Significant

Predicting clinically significant prostate cancer (csPCa) is crucial in PCa management. 3T-magnetic resonance (MR) systems may have a novel role in quantitative imaging and early csPCa prediction, accordingly. In this study, we develop a radiomic model for predicting csPCa based solely on native b2000 diffusion weighted imaging (DWIb2000) and debate the effectiveness of apparent diffusion coefficient (ADC) in the same task. In total, 105 patients were retrospectively enrolled between January–November 2020, with confirmed csPCa or ncsPCa based on biopsy. DWIb2000 and ADC images acquired with a 3T-MRI were analyzed by computing 84 local first-order radiomic features (RFs). Two predictive models were built based on DWIb2000 and ADC, separately. Relevant RFs were selected through LASSO, a support vector machine (SVM) classifier was trained using repeated 3-fold cross validation (CV) and validated on a holdout set. The SVM models rely on a single couple of uncorrelated RFs (ρ < 0.15) selected through Wilcoxon rank-sum test (p ≤ 0.05) with Holm–Bonferroni correction. On the holdout set, while the ADC model yielded AUC = 0.76 (95% CI, 0.63–0.96), the DWIb2000 model reached AUC = 0.84 (95% CI, 0.63–0.90), with specificity = 75%, sensitivity = 90%, and informedness = 0.65. This study establishes the primary role of 3T-DWIb2000 in PCa quantitative analyses, whilst ADC can remain the leading sequence for detection.

Download Full-text

Detection and Classification of Cancer from Microscopic Biopsy Images Using Clinically Significant and Biologically Interpretable Features

Journal of Medical Engineering ◽

10.1155/2015/457906 ◽

2015 ◽

Vol 2015 ◽

pp. 1-14 ◽

Cited By ~ 65

Author(s):

Rajesh Kumar ◽

Rajeev Srivastava ◽

Subodh Srivastava

Keyword(s):

Texture Features ◽

Histogram Equalization ◽

Gray Level ◽

Segmentation Methods ◽

Segmented Images ◽

Clinically Significant ◽

Classification Of Images ◽

Nearest Neighborhood ◽

Neighborhood Method

A framework for automated detection and classification of cancer from microscopic biopsy images using clinically significant and biologically interpretable features is proposed and examined. The various stages involved in the proposed methodology include enhancement of microscopic images, segmentation of background cells, features extraction, and finally the classification. An appropriate and efficient method is employed in each of the design steps of the proposed framework after making a comparative analysis of commonly used method in each category. For highlighting the details of the tissue and structures, the contrast limited adaptive histogram equalization approach is used. For the segmentation of background cells, k-means segmentation algorithm is used because it performs better in comparison to other commonly used segmentation methods. In feature extraction phase, it is proposed to extract various biologically interpretable and clinically significant shapes as well as morphology based features from the segmented images. These include gray level texture features, color based features, color gray level texture features, Law’s Texture Energy based features, Tamura’s features, and wavelet features. Finally, the K-nearest neighborhood method is used for classification of images into normal and cancerous categories because it is performing better in comparison to other commonly used methods for this application. The performance of the proposed framework is evaluated using well-known parameters for four fundamental tissues (connective, epithelial, muscular, and nervous) of randomly selected 1000 microscopic biopsy images.

Download Full-text

Usefulness of Grayscale Values of Hypoechoic Lesions Matched with Target Lesions Observed on Magnetic Resonance Imaging for the Prediction of Clinically Significant Prostate Cancer

10.21203/rs.3.rs-783692/v1 ◽

2021 ◽

Author(s):

Dong Gyun Kim ◽

Jeong Woo Yoo ◽

Kyo Chul Koo ◽

Byung Ha Chung ◽

Kwang Suk Lee

Keyword(s):

Magnetic Resonance Imaging ◽

Prostate Cancer ◽

Magnetic Resonance ◽

Area Under The Curve ◽

Imaging Data ◽

Resonance Imaging ◽

Scoring Method ◽

Significant Difference ◽

Picture Archiving And Communication ◽

Clinically Significant

Abstract INTRODUCTION: To analyze grayscale values for hypoechoic lesions matched with target lesions evaluated using prebiopsy magnetic resonance imaging (MRI). METHODS We collected data on 420 target lesions in patients who underwent MRI/transrectal ultrasound fusion biopsies. Images of hypoechoic lesions that matched the target lesions on MRI were stored in a picture archiving and communication system, and their grayscale values were estimated using the red/green/blue scoring method through an embedded function. We analyzed imaging data using grayscale values. RESULTS Of the 420 lesions, 261 (62.1%) were prostate cancer lesions. Grayscale ranges (42.6–91.8) were significant predictors of clinically significant prostate cancer (csPC) in multivariable logistic regression analyses. Area under the curve for detecting csPC using grayscale values along with conventional variables was 0.839, which was significantly higher than that for detecting csPC using only conventional variables (0.828; p = 0.036). Subgroup analysis revealed a significant difference for PI-RADS 3 lesions between grayscale values for benign and cancerous lesions (p = 0.008). Grayscale values were the only significant predictive factor (p = 0.005) for csPC. CONCLUSIONS Distribution of grayscale values according to PI-RAD 3 scores was useful, and the grayscale range (42.6–91.8) was an important factor for csPC diagnosis.

Download Full-text

Pancreas MRI segmentation into head, body, and tail enables regional quantitative analysis of heterogeneous disease

10.1101/2021.11.30.21266158 ◽

2021 ◽

Author(s):

Alexandre Triay Bagur ◽

Paul Aljabar ◽

Gerard R Ridgway ◽

Michael Brady ◽

Daniel Bulte

Keyword(s):

Quantitative Imaging ◽

Region Of Interest ◽

Automated Analysis ◽

Pancreatic Disease ◽

Proton Density ◽

Disease Heterogeneity ◽

Pancreas Head ◽

Pancreatic Fat ◽

Significant Difference ◽

Multiple Experts

Pancreatic disease can be spatially inhomogeneous. For this reason, quantitative imaging studies of the pancreas have often targeted the 3 main anatomical pancreatic parts, head, body, and tail, traditionally using a balanced region of interest (ROI) strategy. Existing automated analysis methods have implemented whole-organ segmentation, which provides an overall quantification, but fails to address spatial heterogeneity in disease. A method to automatically refine a whole-organ segmentation of the pancreas into head, body, and tail subregions is presented for abdominal magnetic resonance imaging (MRI). The subsegmentation method is based on diffeomorphic registration to a group average template image, where the parts are manually annotated. For a new whole-pancreas segmentation, the aligned template's part labels are automatically propagated to the segmentation of interest. The method is validated retrospectively on the UK Biobank imaging substudy (scanned using a 2-point Dixon protocol at 1.5 tesla), using a nominally healthy cohort of 100 subjects for template creation, and 50 independent subjects for validation. Pancreas head, body, and tail were annotated by multiple experts on the validation cohort, which served as the benchmark for the automated method's performance. Good intra-rater (Dice overlap mean, Head: 0.982, Body: 0.940, Tail: 0.961, N=30) as well as inter-rater (Dice overlap mean, Head: 0.968, Body: 0.905, Tail: 0.943, N=150) agreement was observed. No significant difference (Wilcoxon rank sum test, DSC, Head: p=0.4358, Body: p=0.0992, Tail: p=0.1080) was observed between the manual annotations and the automated method's predictions. Results on regional pancreatic fat assessment are also presented, by intersecting the 3-D parts segmentation with one 2-D multi-echo gradient-echo slice, available from the same scanning session, that was used to compute MRI proton density fat fraction (MRI-PDFF). Initial application of the method on a type 2 diabetes cohort showed the utility of the method for assessing pancreatic disease heterogeneity.

Download Full-text

Cancer Detection Rates of Systematic and Targeted Prostate Biopsies after Biparametric MRI

Prostate Cancer ◽

10.1155/2020/4626781 ◽

2020 ◽

Vol 2020 ◽

pp. 1-6

Author(s):

Maudy C. W. Gayet ◽

Anouk A. M. A. van der Aa ◽

Harrie P. Beerlage ◽

Bart Ph Schrier ◽

Maaike Gielens ◽

...

Keyword(s):

Prostate Cancer ◽

Cancer Detection ◽

Prostate Biopsy ◽

Detection Rates ◽

Prostate Biopsies ◽

Significant Difference ◽

Significant Pathology ◽

Study Population ◽

Clinically Significant ◽

Control Study

Objective. To compare prostate cancer detection rates (CDRs) and pathology results with targeted prostate biopsy (TB) and systematic prostate biopsy (SB) in biopsy-naive men. Methods. An in-patient control study of 82 men undergoing SB and subsequent TB in case of positive prostate MRI between 2015 and 2017 in the Jeroen Bosch Hospital, the Netherlands. Results. Prostate cancer (PCa) was detected in 54.9% with 70.7% agreement between TB and SB. Significant PCa (Gleason score ≥7) was detected in 24.4%. The CDR with TB and SB was 35.4% and 48.8%, respectively (p=0.052). The CDR of significant prostate cancer with TB and SB was both 20.7%. Clinically significant pathology upgrading occurred in 7.3% by adding TB to SB and 22.0% by adding SB to TB. Conclusions. There is no statistically significant difference between CDRs of SB and TB. Both SB and TB miss significant PCas. Moreover, pathology upgrading occurred more often by adding SB to TB than vice versa. This indicates that the omission of SB in this study population might not be justified.

Download Full-text

Quantitative Analysis of Melanosis Coli Colonic Mucosa Using Textural Patterns

Applied Sciences ◽

10.3390/app10010404 ◽

2020 ◽

Vol 10 (1) ◽

pp. 404 ◽

Cited By ~ 1

Author(s):

Chung-Ming Lo ◽

Chun-Chang Chen ◽

Yu-Hsuan Yeh ◽

Chun-Chao Chang ◽

Hsing-Jung Yeh

Keyword(s):

Quantitative Assessment ◽

Colonic Mucosa ◽

Pearson Correlation ◽

Texture Features ◽

Adenomatous Polyps ◽

Gray Level ◽

Melanosis Coli ◽

Pearson Correlation Analysis ◽

Occurrence Matrix

Melanosis coli (MC) is a disease related to long-term use of anthranoid laxative agents. Patients with clinical constipation or obesity are more likely to use these drugs for long periods. Moreover, patients with MC are more likely to develop polyps, particularly adenomatous polyps. Adenomatous polyps can transform to colorectal cancer. Recognizing multiple polyps from MC is challenging due to their heterogeneity. Therefore, this study proposed a quantitative assessment of MC colonic mucosa with texture patterns. In total, the MC colonoscopy images of 1092 person-times were included in this study. At the beginning, the correlations among carcinoembryonic antigens, polyp texture, and pathology were analyzed. Then, 181 patients with MC were extracted for further analysis while patients having unclear images were excluded. By gray-level co-occurrence matrix, texture patterns in the colorectal images were extracted. Pearson correlation analysis indicated five texture features were significantly correlated with pathological results (p < 0.001). This result should be used in the future to design an instant help software to help the physician. The information of colonoscopy and image analystic data can provide clinicians with suggestions for assessing patients with MC.

Download Full-text

Cognitive zonal fusion biopsy of the prostate: Original technique between target and saturation

Archivio Italiano di Urologia e Andrologia ◽

10.4081/aiua.2016.4.292 ◽

2016 ◽

Vol 88 (4) ◽

pp. 292 ◽

Cited By ~ 3

Author(s):

Andrea B. Galosi ◽

Guevar Maselli ◽

Giulia Sbrollini ◽

Gaetano Donatelli ◽

Lorenzo Montesi ◽

...

Keyword(s):

Prostate Cancer ◽

Transrectal Ultrasound ◽

Region Of Interest ◽

High Grade ◽

Cognitive Fusion ◽

Fusion Biopsy ◽

Fusion Technique ◽

Saturation Biopsy ◽

Standard Mapping ◽

Clinically Significant

We describe our experience in prostate biopsy using a new standardized cognitive fusion techniques, that we call “cognitive zonal fusion biopsy”. This new technique is based on two operative options: the first based on target biopsies, the Cognitive Target Biopsy (CTB) if the same target was detected with transrectal ultrasound (TRUS) and multiparametric magnetic resonance (mpMRI); the second based on saturation biopsies, the Zonal Saturation Biopsy (ZSB) on anatomical zone/s containing the region of interest if the same target was not evident with TRUS and MRI. We evaluated results of our technique compared to standard biopsy in order to identify clinically relevant prostate cancer. Methods: This is a single-center prospective study conducted in 58 pts: 25 biopsy-naïve, 25 with previous negative biopsy and in 8 with cancer in active surveillance. Based on mpMRI and transrectal ultrasonography (TRUS), all patients were scheduled for standard 12-core TRUS-guided biopsy. If mpMRI was suggestive or positive (PI-RADS 3, 4 or 5): patients underwent additional targeted 2 to 6 cores using cognitive zonal fusion technique. Results: 31/58 (53.4%) patients had a cancer. Our technique detected 80.6% (25 of 31) with clinically significant prostate cancer, leading to detection of insignificant cancer in 20%. Using standard mapping in MR negative areas we found 5 clinically significant cancer and 4 not significant cancers. MRI cancer detection rate was 18/31 (58.1%), and 9/18 (50%) in high grade tumors. Therefore MRI missed 50% of high grade cancers. The mean number of cores taken with cognitive zonal fusion biopsy was 6.1 (2-17), in addition biopsy sampling was done outside the ROI areas. Overall 15.4 cores (12-22) were taken. Cancer amount in Zonal Biopsy was larger than 7.3 mm (1-54.5) in comparison with 5.2 mm (1-23.5) in standard mapping. Largest percentage of cancer involvement with cognitive zonal fusion technique was detected in 19.4% vs 15.9%. Conclusions: Cognitive Zonal Saturation Biopsies should be used to reduce operator variability of cognitive fusion biopsy in addition to standard biopsy. Cognitive zonal biopsy based on mpMRI findings identifies clinically relevant prostate in 80%, has larger cancer extension in fusion biopsies than in random biopsies, and reduce the number of cores if compared to saturation biopsy.

Download Full-text

Changes in prostate cancer detection rate of fusion versus systematic biopsy over time: A single center experience.

Journal of Clinical Oncology ◽

10.1200/jco.2017.35.6_suppl.15 ◽

2017 ◽

Vol 35 (6_suppl) ◽

pp. 15-15

Author(s):

Brian P. Calio ◽

Abhinav Sidana ◽

Dordaneh Sugano ◽

Amit L Jain ◽

Mahir Maruf ◽

...

Keyword(s):

Prostate Cancer ◽

Cancer Detection ◽

Learning Curves ◽

Low Risk ◽

Fusion Biopsy ◽

Detection Rates ◽

Single Center Experience ◽

Significant Difference ◽

Clinically Significant ◽

Systematic Biopsy

15 Background: To determine the effect of learning curves and changes in fusion platform during 9 years of NCI’s experience with multiparametric MRI (mpMRI)/TRUS fusion biopsy. Methods: A review was performed of a prospectively maintained database of patients undergoing mpMRI followed by fusion biopsy (Fbx) and systematic biopsy (Sbx) from 2007−2016. The patients were stratified based on the timing of first biopsy in 3 groups. Cohort 1 included patients biopsied between 7/2007−12/2010, accounting for learning curve at our institution. Cohort 2 included patients biopsied from 1/2011 up to the debut of UroNav (Invivo) platform in 5/2013. Cohort 3 included patients biopsied after 5/2013. Clinically significant (CS) disease was defined as Gleason 7 (3+4) or higher. Cancer detection rates (CDR) between Sbx and Fbx during different time periods were compared using McNemar’s test. Age and PSA standardized CDRs were calculated for comparison between 3 cohorts. Results: 1528 patients were included in the study with 219, 549 and 761 patients included in 3 respective cohorts. Mean age, PSA and race distribution were similar across 3 cohorts. In cohort 1 there was no significant difference between CDR of CS disease by Fbx (24.7%) vs Sbx (21.5%), p = 0.377. Fbx was significantly better than Sbx in detection of CS disease in cohort 2 and cohort 3 (31.5% vs 25.3%, p = 0.001; 36.5% vs 30.2%, p < 0.001, respectively). There was significant decline in detection of low risk disease by Fbx compared to Sbx in the same period (cohort 2: 14.2% vs 20.9%, p < 0.001; cohort 3: 12.5% vs 19.5%, p < 0.001). Age and PSA standardized CDR of CS cancer by Fbx increased significantly between each successive cohort (cohort 1 and 2: 5.2%, 95% CI [2.1-8.5]), 2 and 3 (5.2%, 95% CI [1.8-8.6]). Conclusions: Our results show that after an early learning period using Fbx, CS prostate cancer was detected at significantly higher rates with Fbx than with Sbx, and low risk disease was detected at lower rates. Advances in software allowed for even greater detection of CS disease in the last cohort. This study shows that accuracy of Fbx is dependent on multiple factors; surgeon/radiologist experience and software improvements together produce improved accuracy.

Download Full-text

Computer Aided Interpretation of Fibrous Texture in Hepatic Magnetic Resonance Images

Advanced Materials Research ◽

10.4028/www.scientific.net/amr.647.325 ◽

2013 ◽

Vol 647 ◽

pp. 325-330 ◽

Cited By ~ 3

Author(s):

Yu Fan Zeng ◽

Xue Jun Zhang ◽

Wen Yan ◽

Li Ling Long ◽

Yu Kun Huang ◽

...

Keyword(s):

Magnetic Resonance ◽

Hepatic Fibrosis ◽

Region Of Interest ◽

Texture Features ◽

Magnetic Resonance Images ◽

Optimal Number ◽

Support Vector ◽

Mr Images ◽

Accuracy Rate ◽

Occurrence Matrix

The fibrous texture in liver is one of important signs for interpreting the chronic liver diseases in radiologists’ routines. In order to investigate the usefulness of various texture features calculated by computer algorithm on hepatic magnetic resonance (MR) images, 15 texture features were calculated from the gray level co-occurrence matrix (GLCM) within a region of interest (ROI) which was selected from the MR images with 6 stages of hepatic fibrosis. By different combination of 15 features as input vectors, the classifier had different performance in staging the hepatic fibrosis. Each combination of texture features was tested by Support Vector Machine (SVM) with leave one case out method. 173 patients’ MR images including 6 stages of hepatic fibrosis were scanned within recent two years. The result showed that optimal number of features was confirmed from 3 to 7 by investigating the classified accuracy rate between each stage/group. It is evident that angular second moment, entropy, sum average and sum entropy played the most significant role in classification.

Download Full-text