Dataset Information

Machine learning classifiers can predict Gleason pattern 4 prostate cancer with greater accuracy than experienced radiologists.

ABSTRACT: OBJECTIVE:The purpose of this study was: To test whether machine learning classifiers for transition zone (TZ) and peripheral zone (PZ) can correctly classify prostate tumors into those with/without a Gleason 4 component, and to compare the performance of the best performing classifiers against the opinion of three board-certified radiologists. METHODS:A retrospective analysis of prospectively acquired data was performed at a single center between 2012 and 2015. Inclusion criteria were (i) 3-T mp-MRI compliant with international guidelines, (ii) Likert ??3/5 lesion, (iii) transperineal template ± targeted index lesion biopsy confirming cancer ? Gleason 3?+?3. Index lesions from 164 men were analyzed (119 PZ, 45 TZ). Quantitative MRI and clinical features were used and zone-specific machine learning classifiers were constructed. Models were validated using a fivefold cross-validation and a temporally separated patient cohort. Classifier performance was compared against the opinion of three board-certified radiologists. RESULTS:The best PZ classifier trained with prostate-specific antigen density, apparent diffusion coefficient (ADC), and maximum enhancement (ME) on DCE-MRI obtained a ROC area under the curve (AUC) of 0.83 following fivefold cross-validation. Diagnostic sensitivity at 50% threshold of specificity was higher for the best PZ model (0.93) when compared with the mean sensitivity of the three radiologists (0.72). The best TZ model used ADC and ME to obtain an AUC of 0.75 following fivefold cross-validation. This achieved higher diagnostic sensitivity at 50% threshold of specificity (0.88) than the mean sensitivity of the three radiologists (0.82). CONCLUSIONS:Machine learning classifiers predict Gleason pattern 4 in prostate tumors better than radiologists. KEY POINTS:• Predictive models developed from quantitative multiparametric magnetic resonance imaging regarding the characterization of prostate cancer grade should be zone-specific. • Classifiers trained differently for peripheral and transition zone can predict a Gleason 4 component with a higher performance than the subjective opinion of experienced radiologists. • Classifiers would be particularly useful in the context of active surveillance, whereby decisions regarding whether to biopsy are necessitated.

SUBMITTER: Antonelli M

PROVIDER: S-EPMC6682575 | biostudies-literature | 2019 Sep

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Machine learning classifiers can predict Gleason pattern 4 prostate cancer with greater accuracy than experienced radiologists.

Antonelli Michela M Johnston Edward W EW Dikaios Nikolaos N Cheung King K KK Sidhu Harbir S HS Appayya Mrishta B MB Giganti Francesco F Simmons Lucy A M LAM Freeman Alex A Allen Clare C Ahmed Hashim U HU Atkinson David D Ourselin Sebastien S Punwani Shonit S Punwani Shonit S

European radiology 20190611 9

<h4>Objective</h4>The purpose of this study was: To test whether machine learning classifiers for transition zone (TZ) and peripheral zone (PZ) can correctly classify prostate tumors into those with/without a Gleason 4 component, and to compare the performance of the best performing classifiers against the opinion of three board-certified radiologists.<h4>Methods</h4>A retrospective analysis of prospectively acquired data was performed at a single center between 2012 and 2015. Inclusion criteria ...[more]

PMID: 31187216

Similar Datasets

Project description:Purpose To determine the interobserver reproducibility of the Prostate Imaging Reporting and Data System (PI-RADS) version 2 lexicon. Materials and Methods This retrospective HIPAA-compliant study was institutional review board-approved. Six radiologists from six separate institutions, all experienced in prostate magnetic resonance (MR) imaging, assessed prostate MR imaging examinations performed at a single center by using the PI-RADS lexicon. Readers were provided screen captures that denoted the location of one specific lesion per case. Analysis entailed two sessions (40 and 80 examinations per session) and an intersession training period for individualized feedback and group discussion. Percent agreement (fraction of pairwise reader combinations with concordant readings) was compared between sessions. κ coefficients were computed. Results No substantial difference in interobserver agreement was observed between sessions, and the sessions were subsequently pooled. Agreement for PI-RADS score of 4 or greater was 0.593 in peripheral zone (PZ) and 0.509 in transition zone (TZ). In PZ, reproducibility was moderate to substantial for features related to diffusion-weighted imaging (κ = 0.535-0.619); fair to moderate for features related to dynamic contrast material-enhanced (DCE) imaging (κ = 0.266-0.439); and fair for definite extraprostatic extension on T2-weighted images (κ = 0.289). In TZ, reproducibility for features related to lesion texture and margins on T2-weighted images ranged from 0.136 (moderately hypointense) to 0.529 (encapsulation). Among 63 lesions that underwent targeted biopsy, classification as PI-RADS score of 4 or greater by a majority of readers yielded tumor with a Gleason score of 3+4 or greater in 45.9% (17 of 37), without missing any tumor with a Gleason score of 3+4 or greater. Conclusion Experienced radiologists achieved moderate reproducibility for PI-RADS version 2, and neither required nor benefitted from a training session. Agreement tended to be better in PZ than TZ, although was weak for DCE in PZ. The findings may help guide future PI-RADS lexicon updates. (©) RSNA, 2016 Online supplemental material is available for this article.

Project description:(1) Background:One of the most common cancers that affect North American men and men worldwide is prostate cancer. The Gleason score is a pathological grading system to examine the potential aggressiveness of the disease in the prostate tissue. Advancements in computing and next-generation sequencing technology now allow us to study the genomic profiles of patients in association with their different Gleason scores more accurately and effectively. (2) Methods: In this study, we used a novel machine learning method to analyse gene expression of prostate tumours with different Gleason scores, and identify potential genetic biomarkers for each Gleason group. We obtained a publicly-available RNA-Seq dataset of a cohort of 104 prostate cancer patients from the National Center for Biotechnology Information's (NCBI) Gene Expression Omnibus (GEO) repository, and categorised patients based on their Gleason scores to create a hierarchy of disease progression. A hierarchical model with standard classifiers in different Gleason groups, also known as nodes, was developed to identify and predict nodes based on their mRNA or gene expression. In each node, patient samples were analysed via class imbalance and hybrid feature selection techniques to build the prediction model. The outcome from analysis of each node was a set of genes that could differentiate each Gleason group from the remaining groups. To validate the proposed method, the set of identified genes were used to classify a second dataset of 499 prostate cancer patients collected from cBioportal. (3) Results: The overall accuracy of applying this novel method to the first dataset was 93.3%; the method was further validated to have 87% accuracy using the second dataset. This method also identified genes that were not previously reported as potential biomarkers for specific Gleason groups. In particular, PIAS3 was identified as a potential biomarker for Gleason score 4 + 3 = 7, and UBE2V2 for Gleason score 6. (4) Insight: Previous reports show that the genes predicted by this newly proposed method strongly correlate with prostate cancer development and progression. Furthermore, pathway analysis shows that both PIAS3 and UBE2V2 share similar protein interaction pathways, the JAK/STAT signaling process.

Project description:PURPOSE:Very high-risk prostate cancer (PC) is associated with poor response to local and systemic treatments; however, few cases have been molecularly profiled. We studied clinical outcomes and molecular profiles of patients with clinically localized primary Gleason pattern 5 PC. PATIENTS AND METHODS:Clinicopathologic features, targeted somatic and germline sequencing, and PTEN, TP53, and ERG status by immunohistochemistry were assessed in patients undergoing surgery from 2005 to 2015; 60 consecutive patients were identified with Gleason score 5 + 4 = 9 or 5 + 5 = 10 PC after radical prostatectomy with available tissue and clinical follow-up. Clinicopathologic and genomic parameters were correlated with biochemical relapse, metastasis-free survival, time to castration resistance, and overall survival using Cox proportional hazards models. RESULTS:Of patients with somatic sequencing data and clinical follow-up, 34% had DNA repair gene mutations, including 22% (11 of 49) with homologous recombination and 12% (six of 49) with mismatch repair gene alterations. Homologous recombination mutations were germline in 82% (nine of 11) of patients. In addition, 33% (16 of 49) had TP53 mutation, and 51% (29 of 57) had PTEN loss. Overall, 43% developed metastasis, with a time to castration resistance of 12 months. On multivariable analysis of clinicopathologic variables, only ductal/intraductal histology (hazard ratio, 4.43; 95% CI, 1.76 to 11.15; P = .002) and seminal vesicle invasion (hazard ratio, 5.14; 95% CI, 1.83 to 14.47; P = .002) were associated with metastasis. Among genomic alterations, only TP53 mutation and PTEN loss were associated with metastasis on univariable analysis, and neither remained significant in multivariable analyses. These data are retrospective and hypothesis generating. CONCLUSION:Potentially actionable homologous recombination and mismatch repair alterations are observed in a significant proportion of patients with very high-risk PC at the time of radical prostatectomy. These findings could inform the design of prospective trials in this patient population.

Dataset Information

Machine learning classifiers can predict Gleason pattern 4 prostate cancer with greater accuracy than experienced radiologists.

Publications

Machine learning classifiers can predict Gleason pattern 4 prostate cancer with greater accuracy than experienced radiologists.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets