MRI-Based Grading of Clear Cell Renal Cell Carcinoma Using a Machine Learning Classifier

ObjectiveTo develop a machine learning (ML)-based classifier for discriminating between low-grade (ISUP I-II) and high-grade (ISUP III-IV) clear cell renal cell carcinomas (ccRCCs) using MRI textures.Materials and MethodsWe retrospectively evaluated a total of 99 patients (with 61 low-grade and 38 high-grade ccRCCs), who were randomly divided into a training set (n = 70) and a validation set (n = 29). Regions of interest (ROIs) of all tumors were manually drawn three times by a radiologist at the maximum lesion level of the cross-sectional CMP sequence images. The quantitative texture analysis software, MaZda, was used to extract texture features, including histograms, co-occurrence matrixes, run-length matrixes, gradient models, and autoregressive models. Reproducibility of the texture features was assessed with the intra-class correlation coefficient (ICC). Features were chosen based on their importance coefficients in a random forest model, while the multi-layer perceptron algorithm was used to build a classifier on the training set, which was later evaluated with the validation set.ResultsThe ICCs of 257 texture features were equal to or higher than 0.80 (0.828–0.998. Six features, namely Kurtosis, 135dr_RLNonUni, Horzl_GLevNonU, 135dr_GLevNonU, S(4,4)Entropy, and S(0,5)SumEntrp, were chosen to develop the multi-layer perceptron classifier. A three-layer perceptron model, which has 229 nodes in the hidden layer, was trained on the training set. The accuracy of the model was 95.7% with the training set and 86.2% with the validation set. The areas under the receiver operating curves were 0.997 and 0.758 for the training and validation sets, respectively.ConclusionsA machine learning-based grading model was developed that can aid in the clinical diagnosis of clear cell renal cell carcinoma using MRI images.

Download Full-text

Preoperative Predicting the WHO/ISUP Nuclear Grade of Clear Cell Renal Cell Carcinoma by Computed Tomography-Based Radiomics Features

Journal of Personalized Medicine ◽

10.3390/jpm11010008 ◽

2020 ◽

Vol 11 (1) ◽

pp. 8

Author(s):

Claudia-Gabriela Moldovanu ◽

Bianca Boca ◽

Andrei Lebovici ◽

Attila Tamas-Szora ◽

Diana Sorina Feier ◽

...

Keyword(s):

Computed Tomography ◽

Renal Cell Carcinoma ◽

Renal Cell ◽

Clear Cell ◽

Predictive Ability ◽

Nuclear Grade ◽

Training Set ◽

Cell Renal Cell Carcinoma ◽

Validation Set ◽

Sensitivity Specificity

Nuclear grade is important for treatment selection and prognosis in patients with clear cell renal cell carcinoma (ccRCC). This study aimed to determine the ability of preoperative four-phase multiphasic multidetector computed tomography (MDCT)-based radiomics features to predict the WHO/ISUP nuclear grade. In all 102 patients with histologically confirmed ccRCC, the training set (n = 62) and validation set (n = 40) were randomly assigned. In both datasets, patients were categorized according to the WHO/ISUP grading system into low-grade ccRCC (grades 1 and 2) and high-grade ccRCC (grades 3 and 4). The feature selection process consisted of three steps, including least absolute shrinkage and selection operator (LASSO) regression analysis, and the radiomics scores were developed using 48 radiomics features (10 in the unenhanced phase, 17 in the corticomedullary (CM) phase, 14 in the nephrographic (NP) phase, and 7 in the excretory phase). The radiomics score (Rad-Score) derived from the CM phase achieved the best predictive ability, with a sensitivity, specificity, and an area under the curve (AUC) of 90.91%, 95.00%, and 0.97 in the training set. In the validation set, the Rad-Score derived from the NP phase achieved the best predictive ability, with a sensitivity, specificity, and an AUC of 72.73%, 85.30%, and 0.84. We constructed a complex model, adding the radiomics score for each of the phases to the clinicoradiological characteristics, and found significantly better performance in the discrimination of the nuclear grades of ccRCCs in all MDCT phases. The highest AUC of 0.99 (95% CI, 0.92–1.00, p < 0.0001) was demonstrated for the CM phase. Our results showed that the MDCT radiomics features may play a role as potential imaging biomarkers to preoperatively predict the WHO/ISUP grade of ccRCCs.

Download Full-text

Fructose 1,6-Bisphosphatase 1 Expression Reduces 18F-FDG Uptake in Clear Cell Renal Cell Carcinoma

Contrast Media & Molecular Imaging ◽

10.1155/2019/9463926 ◽

2019 ◽

Vol 2019 ◽

pp. 1-6 ◽

Cited By ~ 1

Author(s):

Ruohua Chen ◽

Xiang Zhou ◽

Gang Huang ◽

Jianjun Liu

Keyword(s):

Renal Cell Carcinoma ◽

Primary Tumor ◽

Renal Cell ◽

Clear Cell ◽

Low Grade ◽

High Grade ◽

Cell Renal Cell Carcinoma ◽

Fdg Uptake ◽

Fructose 1,6 Bisphosphatase ◽

The Relationship

Purpose. To determine the relationship between fructose 1,6-bisphosphatase 1 (FBP1) expression and fluorine 18 (18F) fluorodeoxyglucose (FDG) uptake in patients with clear cell renal cell carcinoma (ccRCC), and to investigate how 18F-FDG uptake and FBP1 expression are related to tumor metabolism and tumor differentiation grade. Materials and Methods. A total of 54 patients with ccRCC underwent 18F-FDG combined positron emission tomography and computed tomography (PET/CT) before tumor resection. The maximum standardized uptake value (SUVmax) for the primary tumor was calculated from the 18F-FDG uptake. The relationship between SUVmax of primary tumor and the expression of FBP1, hexokinase 2 (HK2), and glucose transporter 1 (GLUT1) was analyzed via immunohistochemical analysis. Results. We identified an inverse relationship between FBP1 expression and SUVmax (P=0.031). SUVmax was higher in patients with high-grade ccRCC (mean, 11.6 ± 5.0) than in those with low-grade ccRCC (mean, 3.8 ± 1.6, P<0.001). FBP1 expression was significantly lower in patients with high-grade ccRCC (mean, 0.23 ± 0.1) than in those with low-grade ccRCC (mean, 0.57 ± 0.08; P=0.018). FBP1 status could be predicted with an accuracy of 66.7% when a SUVmax cutoff value of 3.55 was used. GLUT1 expression in ccRCC was positively correlated with 18F-FDG uptake and FBP1 status, whereas HK2 expression was not. Conclusion. SUVmax in patients with ccRCC is inversely associated with the expression of FBP1, and FBP1 may inhibit 18F-FDG uptake via regulating GLUT1. SUVmax is higher in patients with high-grade ccRCC than in those with low-grade ccRCC, which could be the result of lower FBP1 expression in patients with high-grade ccRCC.

Download Full-text

Differentiation of low grade from high grade clear cell renal cell carcinoma neoplasms using a CAD algorithm on four-phase CT.

Journal of Clinical Oncology ◽

10.1200/jco.2016.34.15_suppl.4564 ◽

2016 ◽

Vol 34 (15_suppl) ◽

pp. 4564-4564 ◽

Cited By ~ 1

Author(s):

Heidi Coy ◽

Michael Douek ◽

Jonathan Young ◽

Matthew S. Brown ◽

Jonathan Goldin ◽

...

Keyword(s):

Renal Cell Carcinoma ◽

Cell Carcinoma ◽

Renal Cell ◽

Clear Cell ◽

Low Grade ◽

High Grade ◽

Cell Renal Cell Carcinoma

Download Full-text

Profiles of overall survival-related gene expression-based risk signature and their prognostic implications in clear cell renal cell carcinoma

Bioscience Reports ◽

10.1042/bsr20200492 ◽

2020 ◽

Vol 40 (9) ◽

Author(s):

Zihao He ◽

Tuo Deng ◽

Xiaolu Duan ◽

Guohua Zeng

Keyword(s):

Renal Cell Carcinoma ◽

Overall Survival ◽

Cell Carcinoma ◽

Family Member ◽

Renal Cell ◽

Cox Regression ◽

Clear Cell ◽

Training Set ◽

Cell Renal Cell Carcinoma ◽

Validation Set

Abstract The present work aimed to evaluate the prognostic value of overall survival (OS)-related genes in clear cell renal cell carcinoma (ccRCC) and to develop a nomogram for clinical use. Transcriptome data from The Cancer Genome Atlas (TCGA) were collected to screen differentially expressed genes (DEGs) between ccRCC patients with OS > 5 years (149 patients) and those with <1 year (52 patients). In TCGA training set (265 patients), seven DEGs (cytochrome P450 family 3 subfamily A member 7 (CYP3A7), contactin-associated protein family member 5 (CNTNAP5), adenylate cyclase 2 (ADCY2), TOX high mobility group box family member 3 (TOX3), plasminogen (PLG), enamelin (ENAM), and collagen type VII α 1 chain (COL7A1)) were further selected to build a prognostic risk signature by the least absolute shrinkage and selection operator (LASSO) Cox regression model. Survival analysis confirmed that the OS in the high-risk group was dramatically shorter than their low-risk counterparts. Next, univariate and multivariate Cox regression revealed the seven genes-based risk score, age, and Tumor, lymph Node, and Metastasis staging system (TNM) stage were independent prognostic factors to OS, based on which a novel nomogram was constructed and validated in both TCGA validation set (265 patients) and the International Cancer Genome Consortium cohort (ICGC, 84 patients). A decent predictive performance of the nomogram was observed, the C-indices and corresponding 95% confidence intervals of TCGA training set, validation set, and ICGC cohort were 0.78 (0.74–0.82), 0.75 (0.70–0.80), and 0.70 (0.60–0.80), respectively. Moreover, the calibration plots of 3- and 5 years survival probability indicated favorable curve-fitting performance in the above three groups. In conclusion, the proposed seven genes signature-based nomogram is a promising and robust tool for predicting the OS of ccRCC, which may help tailor individualized therapeutic strategies.

Download Full-text

High-grade clear cell renal cell carcinoma has a higher angiogenic activity than low-grade renal cell carcinoma based on histomorphological quantification and qRT–PCR mRNA expression profile

British Journal of Cancer ◽

10.1038/sj.bjc.6603796 ◽

2007 ◽

Vol 96 (12) ◽

pp. 1888-1895 ◽

Cited By ~ 47

Author(s):

M M Baldewijns ◽

V L Thijssen ◽

G G Van den Eynden ◽

S J Van Laere ◽

A M Bluekens ◽

...

Keyword(s):

Renal Cell Carcinoma ◽

Cell Carcinoma ◽

Renal Cell ◽

Clear Cell ◽

Low Grade ◽

High Grade ◽

Mrna Expression Profile ◽

Cell Renal Cell Carcinoma ◽

Angiogenic Activity ◽

Qrt Pcr

Download Full-text

Correlation of tumor enhancement and imaging features on multiphasic multidetector CT with microvessel density as a step toward a minimally invasive method to predict Fuhrman nuclear grade in patients with clear cell renal cell carcinoma.

Journal of Clinical Oncology ◽

10.1200/jco.2017.35.15_suppl.e16049 ◽

2017 ◽

Vol 35 (15_suppl) ◽

pp. e16049-e16049

Author(s):

Heidi Coy ◽

Jonathan Young ◽

Michael Douek ◽

Clara Magyar ◽

Anthony Sisk ◽

...

Keyword(s):

Renal Cell Carcinoma ◽

Logistic Regression ◽

Regression Analysis ◽

Microvessel Density ◽

Renal Cell ◽

Clear Cell ◽

Low Grade ◽

High Grade ◽

Cell Renal Cell Carcinoma ◽

Invasive Method

e16049 Background: Clear cell renal cell carcinoma (ccRCC) is the most common tumor of the kidney. Up to 70% are incidentally detected on multiphasic CT. The prognosis for patients with ccRCC is related to Fuhrman grade (FG) and is diagnosed by biopsy or excision. There is a great need for a non-invasive method to assess tumor grade which may help inform clinical decision-making. The purpose of our study is to determine if contrast enhancement on CT predicts FG and microvessel density (MVD) of ccRCC lesions and to assess which combination of quantitative and qualitative radiological features and clinical features predict high FG ccRCC lesions. Methods: With IRB approval for this HIPAA-compliant retrospective study, our pathology and imaging databases were queried to obtain a cohort of ccRCC with a preoperative multiphasic (unenhanced (U), corticomedullary (C), nephrographic (N), and excretory (E)) CT scan. Tumors were stained with CD4 to quantify % MVD. Spearman’s rank correlation was calculated to test the strength of the association between CT enhancement, %MVD and FG. Stepwise logistic regression analysis was performed to determine the quantitative and/or qualitative feature with the highest performance in predicting high FG tumor. The multivariate logistic regression analysis was evaluated using ROC curves and AUCs. Results: Our cohort had 127 patients with 89 low-grade tumors and 43 high-grade tumors. The %wash-in of enhancement from the U to the C phase showed a significant correlation with %MVD of the tumor (R2= 0.181,p < . 001) as did enhancement of the tumor in the early C phase with %MVD (R2= 0.159,p < . 001). There was a significant inverse correlation with %MVD and FG (R2= 0.137,p < . 001). T-stage, tumor size, %MVD and presence of renal vein invasion were determined to be significant independent predictors of high grade lesions with an AUC of .838 (95% CI .748-.927). Conclusions: Lower grade ccRCC tumors have a higher MVD. Therefore, contrast material washes in at a faster rate from the UN to the CM phase, enabling low grade to be discriminated from high-grade tumors on CT.

Download Full-text

A Risk Score Model Based on Nine Differentially Methylated mRNAs for Predicting Prognosis of Patients with Clear Cell Renal Cell Carcinoma

Disease Markers ◽

10.1155/2021/8863799 ◽

2021 ◽

Vol 2021 ◽

pp. 1-11

Author(s):

Jingmin Zhou ◽

Guanghua Liu ◽

Xingcheng Wu ◽

Zhien Zhou ◽

Jialin Li ◽

...

Keyword(s):

Renal Cell Carcinoma ◽

Risk Score ◽

Renal Cell ◽

Clear Cell ◽

Good Prognosis ◽

Training Set ◽

Cell Renal Cell Carcinoma ◽

Mrna Methylation ◽

Score Model ◽

Validation Set

Purpose. DNA methylation alterations play important roles in initiation and progression of clear cell renal cell carcinoma (ccRCC). In this study, we attempted to identify differentially methylated mRNA signatures with prognostic value for ccRCC. Methods. The mRNA methylation and expression profiling data of 306 ccRCC tumors were downloaded from The Cancer Genome Atlas (TCGA) to screen differentially methylated lncRNAs and mRNAs (DMLs and DMMs) between bad and good prognosis patients. Uni- and multivariable Cox regression analyses and LASSO Cox-PH regression analysis were used to select prognostic lncRNAs and mRNAs. Corresponding risk scores were calculated and compared for predictive performance in the training set using Kaplan-Meier OS and ROC curve analyses. The optimal risk score was then identified and validated in the validation set. Function enrichment analysis was conducted. Results. This study screened 461 DMMs and 63 DMLs between good prognosis and bad prognosis patients, and furthermore, nine mRNAs and six lncRNAs were identified as potential prognostic molecules. Compared to nine-mRNA status risk score model, six-lncRNA methylation risk score model, and six-lncRNA status risk score model, the nine-mRNA methylation risk score model showed superiority for prognosis stratification of ccRCC patients in the training set. The prognostic ability of the nine-mRNA methylation risk score model was validated in the validation set. The nine prognostic mRNAs were functionally associated with neuroactive ligand receptor interaction and inflammation-related pathways. Conclusion. The nine-mRNA methylation signature (DMRTA2, DRGX, FAM167A, FGGY, FOXI2, KRTAP2-1, TCTEX1D1, TTBK1, and UBE2QL1) may be a useful prognostic biomarker and tool for ccRCC patients. The present results would be helpful to elucidate the possible pathogenesis of ccRCC.

Download Full-text

Assessment of intratumor heterogeneity using imaging texture features in clear cell renal cell carcinoma.

Journal of Clinical Oncology ◽

10.1200/jco.2019.37.7_suppl.663 ◽

2019 ◽

Vol 37 (7_suppl) ◽

pp. 663-663

Author(s):

Durga Udayakumar ◽

Durgesh Dwivedi ◽

Ze Zhang ◽

Yin Xi ◽

Tao Wang ◽

...

Keyword(s):

Gene Expression ◽

Renal Cell Carcinoma ◽

Cell Carcinoma ◽

Renal Cell ◽

Clear Cell ◽

Texture Features ◽

Tumor Grade ◽

High Grade ◽

Metabolic Processes ◽

Cell Renal Cell Carcinoma

663 Background: Intratumoral heterogeneity (ITH) relates to aggressiveness in clear cell renal cell carcinoma (ccRCC), the most common and aggressive subtype of kidney cancer. Percutaneous biopsies have high diagnostic accuracy. However, ITH lowers their reliability in larger, heterogeneous tumors. Haralick texture features extracted from a gray level co-occurrence matrix (GLCM) is a robust method to assess intrinsic tumor imaging characteristics. Some of these features, including entropy as a measure of ITH, have recently been used in differentiating malignant from benign tumors in various organs. We aim to understand how tumor entropy extracted from magnetic resonance (MR) imaging correlate with tumor grade (aggressiveness) and gene expression heterogeneity in ccRCC. Methods: This IRB-approved, prospective study included T2-weighted (T2W) and arterial spin labeled (ASL) MR images of 62 patients with ccRCC. The GLCM was constructed for regions-of interest (ROI) within the tumor and 13 Haralick texture features were estimated. Correlations between texture features and tumor grade were evaluated by logistic regression and quantified by the area under the receiving operating characteristic (ROC) curve (AUC). RNA sequencing of 182 tumor samples in 49 resected tumors was performed. Entropy was correlated with standard deviation (SD) of normalized gene expression levels in multiple samples from the same tumor. Spearman correlation (rho) was computed for each gene. False discovery rate q values < 0.05 were considered statistically significant. Results: Entropy was higher in high-grade than low-grade tumors (11.28 ± 0.52 vs. 10.95 ± 0.65) on T2W (q = 0.028) and ASL (10.45 ± 1.15 vs. 9.65 ± 1.29) (q = 0.013). Entropy had an AUC of 0.70 (T2) for high-grade prediction and was weakly correlated with tumor size (R2 = 0.2). Higher T2 and ASL entropy correlated with higher SD of gene expression. Gene ontology analysis of top correlated genes revealed strong enrichment of genes in metabolic processes. Conclusions: Higher MRI entropy predicts high tumor grade and correlates with increased heterogeneity in gene expression of metabolic processes.

Download Full-text