Identification and Validation of Radiographic Enhancement for Reliable Differentiation of CD117(+) Benign Renal Oncocytoma and Chromophobe Renal Cell Carcinoma

Convolutional neural networks for the differentiation between benign and malignant renal tumors with a multicenter international computed tomography dataset

OBJECTIVES

To use convolutional neural networks (CNNs) for the differentiation between benign and malignant renal tumors using contrast-enhanced CT images of a multi-institutional, multi-vendor, and multicenter CT dataset.

METHODS

A total of 264 histologically confirmed renal tumors were included, from US and Swedish centers. Images were augmented and divided randomly 70%:30% for algorithm training and testing. Three CNNs (InceptionV3, Inception-ResNetV2, VGG-16) were pretrained with transfer learning and fine-tuned with our dataset to distinguish between malignant and benign tumors. The ensemble consensus decision of the three networks was also recorded. Performance of each network was assessed with receiver operating characteristics (ROC) curves and their area under the curve (AUC-ROC). Saliency maps were created to demonstrate the attention of the highest performing CNN.

RESULTS

Inception-ResNetV2 achieved the highest AUC of 0.918 (95% CI 0.873-0.963), whereas VGG-16 achieved an AUC of 0.813 (95% CI 0.752-0.874). InceptionV3 and ensemble achieved the same performance with an AUC of 0.894 (95% CI 0.844-0.943). Saliency maps indicated that Inception-ResNetV2 decisions are based on the characteristics of the tumor while in most tumors considering the characteristics of the interface between the tumor and the surrounding renal parenchyma.

CONCLUSION

Deep learning based on a diverse multicenter international dataset can enable accurate differentiation between benign and malignant renal tumors.

CRITICAL RELEVANCE STATEMENT

Convolutional neural networks trained on a diverse CT dataset can accurately differentiate between benign and malignant renal tumors.

KEY POINTS

• Differentiation between benign and malignant tumors based on CT is extremely challenging. • Inception-ResNetV2 trained on a diverse dataset achieved excellent differentiation between tumor types. • Deep learning can be used to distinguish between benign and malignant renal tumors.

Smoking enhanced the expression of c-kit in chromophobe renal cell carcinoma

INTRODUCTION

Smoking is an important risk factor for inducing renal cell carcinoma (RCC), but its specific mechanism affecting the development of RCC remains to be elucidated. Chromophobe RCC (ChRCC) is a subtype of RCC. Many studies have shown smoking is closely associated with RCC occurrence and c-kit plays a critical role in the progression of RCC, however, few studies focus on ChRCC. This study investigated the molecular mechanism between smoking and the c-kit pathway in ChRCC.

METHODS

Differentially expressed genes (DEGs) were obtained from The Cancer Genome Atlas (TCGA) in ChRCC and the expression of KIT in ChRCC was analyzed through the TCGA database combined with Gene Expression Omnibus (GEO) and oncomine databases. Moreover, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses and Protein Protein Interaction (PPI) network analysis were performed to explore the function of KIT and correlated DEGs as well as its co-expression genes in ChRCC. Finally, ChRCC patient samples were used to verify the effect of smoking on the c-kit expression.

RESULTS

The results showed that KIT is one of the DEGs and plays a vital role in ChRCC tumorigenesis. Interestingly, the expression of c-kit in cancer tissues of 27 smoking patients was significantly higher than that of 25 non-smoking patients (p<0.05), which suggests smoking might enhance the expression of c-kit in ChRCC patients.

CONCLUSIONS

Our results demonstrate that smoking might play a pivotal role in the ChRCC tumorigenesis via a pathway related to c-kit, and provided new insight into the relationship between smoking and the c-kit pathway in ChRCC.

Deep learning techniques for imaging diagnosis of renal cell carcinoma: current and emerging trends

This study summarizes the latest achievements, challenges, and future research directions in deep learning technologies for the diagnosis of renal cell carcinoma (RCC). This is the first review of deep learning in RCC applications. This review aims to show that deep learning technologies hold great promise in the field of RCC diagnosis, and we look forward to more research results to meet us for the mutual benefit of renal cell carcinoma patients. Medical imaging plays an important role in the early detection of renal cell carcinoma (RCC), as well as in the monitoring and evaluation of RCC during treatment. The most commonly used technologies such as contrast enhanced computed tomography (CECT), ultrasound and magnetic resonance imaging (MRI) are now digitalized, allowing deep learning to be applied to them. Deep learning is one of the fastest growing fields in the direction of medical imaging, with rapidly emerging applications that have changed the traditional medical treatment paradigm. With the help of deep learning-based medical imaging tools, clinicians can diagnose and evaluate renal tumors more accurately and quickly. This paper describes the application of deep learning-based imaging techniques in RCC assessment and provides a comprehensive review.

A pilot radiometabolomics integration study for the characterization of renal oncocytic neoplasia

Differentiating benign renal oncocytic tumors and malignant renal cell carcinoma (RCC) on imaging and histopathology is a critical problem that presents an everyday clinical challenge. This manuscript aims to demonstrate a novel methodology integrating metabolomics with radiomics features (RF) to differentiate between benign oncocytic neoplasia and malignant renal tumors. For this purpose, thirty-three renal tumors (14 renal oncocytic tumors and 19 RCC) were prospectively collected and histopathologically characterised. Matrix-assisted laser desorption/ionisation mass spectrometry imaging (MALDI-MSI) was used to extract metabolomics data, while RF were extracted from CT scans of the same tumors. Statistical integration was used to generate multilevel network communities of -omics features. Metabolites and RF critical for the differentiation between the two groups (delta centrality > 0.1) were used for pathway enrichment analysis and machine learning classifier (XGboost) development. Receiver operating characteristics (ROC) curves and areas under the curve (AUC) were used to assess classifier performance. Radiometabolomics analysis demonstrated differential network node configuration between benign and malignant renal tumors. Fourteen nodes (6 RF and 8 metabolites) were crucial in distinguishing between the two groups. The combined radiometabolomics model achieved an AUC of 86.4%, whereas metabolomics-only and radiomics-only classifiers achieved AUC of 72.7% and 68.2%, respectively. Analysis of significant metabolite nodes identified three distinct tumour clusters (malignant, benign, and mixed) and differentially enriched metabolic pathways. In conclusion, radiometabolomics integration has been presented as an approach to evaluate disease entities. In our case study, the method identified RF and metabolites important in differentiating between benign oncocytic neoplasia and malignant renal tumors, highlighting pathways differentially expressed between the two groups. Key metabolites and RF identified by radiometabolomics can be used to improve the identification and differentiation between renal neoplasms.

A Brief Review of Artificial Intelligence in Genitourinary Oncological Imaging

Genitourinary (GU) system is among the most commonly involved malignancy sites in the human body. Imaging plays a crucial role not only in diagnosis of cancer but also in disease management and its prognosis. However, interpretation of conventional imaging methods such as CT or MR imaging (MRI) usually demonstrates variability across different readers and institutions. Artificial intelligence (AI) has emerged as a promising technology that could improve the patient care by providing helpful input to human readers through lesion detection algorithms and lesion classification systems. Moreover, the robustness of these models may be valuable in automating time-consuming tasks such as organ and lesion segmentations. Herein, we review the current state of imaging and existing challenges in GU malignancies, particularly for cancers of prostate, kidney and bladder; and briefly summarize the recent AI-based solutions to these challenges.

Stability of renal parenchymal volume and function during active surveillance of renal oncocytoma patients

INTRODUCTION AND OBJECTIVE

To evaluate whether significant loss in ipsilateral renal parenchymal volume (IRPV) and renal function occurs during active surveillance (AS) of renal oncocytoma (RO) patients.

METHODS

Renal function (estimated glomerular filtration rate, eGFR) dynamics were retrospectively analyzed in 32 consecutive biopsy-diagnosed RO patients managed with AS at a National Comprehensive Cancer Network institute. Three-dimensional kidney and tumor reconstructions were generated and IRPV was calculated using volumetry software (Myrian®) for all patients with manually estimated RO growth >+10 cm3. GFR and IRPV were compared at AS initiation vs. the last follow-up using 2-sided paired t-tests. The correlation between change in IRPV and change in RO size or GFR was tested using a Spearman coefficient.

RESULTS

With median follow-up of 37 months, there was no significant change between initial vs. last eGFR (median 71.0 vs. 70.5 ml/min/1.73 m2, P = 0.50; median change -3.0 ml/min/1.73 m2). Among patients (n = 17) with RO growth >+10 cm3 during AS (median growth +28.6 cm3, IQR +16.9- + 46.5 cm3), IRPV generally remained stable (median change +0.5%, IQR -1.2%- + 1.2%), with only 2 cases surpassing 5% loss. No IRPV loss was detected among any patient within the top tertile of RO growth magnitude. RO growth magnitude did not correlate with loss of either IRPV (ρ = -0.30, P = 0.24) or eGFR (ρ = -0.16, P = 0.40), including among patient subsets with lower initial eGFR. Study limitations include a lack of long-term follow-up.

CONCLUSIONS

Volumetry is a promising novel tool to measure kidney and tumor tissue changes during AS. Our study using volumetry indicates that clinically significant loss of IRPV or eGFR is uncommon and unrelated to tumor growth among untreated RO patients with intermediate follow-up. These findings support that AS is in general functionally safe for RO patients, however longer study is needed to determine safety durability, particularly among uncommon ≥cT2 RO variants.

[Artificial intelligence in urology-opportunities and possibilities]

The use of artificial intelligence (AI) in urology can contribute to a significant improvement with regard to individualization of diagnostics and therapy as well as healthcare cost reduction. The potential applications and advantages of AI in medicine are often underestimated or incompletely understood. This makes it difficult to conceptually solve relevant medical problems using AI. With current advances in computer science, multiple, highly complex nonmedical processes have already been studied and optimized in an automated fashion. The development of AI models, if applied correctly, can lead to more effective processing and analysis of patient-related data and correspondingly optimized diagnosis and therapy of urological patients. In this review, the current status on the application of AI in medicine and its opportunities and possibilities in urology are presented from a conceptual perspective using practical examples.

Tumorigenesis Mechanisms Found in Hereditary Renal Cell Carcinoma: A Review

Renal cell carcinoma is a heterogenous cancer composed of an increasing number of unique subtypes each with their own cellular and tumor behavior. The study of hereditary renal cell carcinoma, which composes just 5% of all types of tumor cases, has allowed for the elucidation of subtype-specific tumorigenesis mechanisms that can also be applied to their sporadic counterparts. This review will focus on the major forms of hereditary renal cell carcinoma and the genetic alterations contributing to their tumorigenesis, including von Hippel Lindau syndrome, Hereditary Papillary Renal Cell Carcinoma, Succinate Dehydrogenase-Deficient Renal Cell Carcinoma, Hereditary Leiomyomatosis and Renal Cell Carcinoma, BRCA Associated Protein 1 Tumor Predisposition Syndrome, Tuberous Sclerosis, Birt-Hogg-Dubé Syndrome and Translocation RCC. The mechanisms for tumorigenesis described in this review are beginning to be exploited via the utilization of novel targets to treat renal cell carcinoma in a subtype-specific fashion.

Differentiating Oncocytic Renal Tumors from Chromophobe Renal Cell Carcinoma: Comparison of Peak Early-phase Enhancement Ratio to Clinical Risk Factors and Rater Predictions

Background

Most surgically resected benign renal tumors are found to be oncocytomas or indolent hybrid oncocytic tumors, which are difficult to differentiate from chromophobe renal cell carcinoma (chRCC) on renal mass biopsy. Both often exhibit CD117⁺ staining.

Objective

To evaluate the ability of the peak early-phase enhancement ratio (PEER) to distinguish oncocytomas from chRCC and compare its discrimination to traditional clinical risk factors and blinded clinical raters.

Design setting and participants

This was a diagnostic case-control study of patients (2006-2020) with oncocytoma or chRCC according to surgical pathology.

Intervention

Partial or radical nephrectomy.

Outcome measurements and statistical analysis

Three clinical raters blinded to histology measured the PEER and the presence of stellate scar and predicted the final histology for each tumor. Averaged and individual PEER values were compared to surgical pathology and assessed for interobserver variability. Subanalyses were conducted for patients with confirmed CD117⁺ status.

Results and limitations

For the 76 patients identified, PEER was higher among the 32 (42.1%) oncocytomas than among the 44 (57.9%) chRCCs (median 0.81 vs 0.43; p < 0.001), with high correlation across raters (correlation coefficients ≥0.85). A PEER cutoff of <0.60 was strongly associated with identification of chRCC (OR 95.7 (95% CI 19.9-460.8), p < 0.001). In the overall and CD117⁺ cohorts, sensitivity was 93.2% and 97.0%, the negative predictive value was 90.3% and 95.5%, and the area under the receiver operating characteristic curve (AUC) on multivariable modeling was 95.0% and 98.1%, respectively. PEER outperformed models with clinical risk factors alone (AUC 70.4%) and histology predictions by three raters (AUC 51.6%, 62.5%, and 63.1%). Limitations include reliance on surgical pathology and inclusion of a mix of early contrast-enhanced phases.

Conclusions

PEER reliably differentiated benign renal oncocytomas and indolent hybrid tumors from malignant chRCC with excellent diagnostic performance. A diagnostic pathway with biopsy, CD117 staining, and PEER deserves further study to potentially avoid unnecessary surgery for oncocytic renal tumors.

Patient summary

We assessed a measurement called PEER on computed tomography (CT) scans and found higher values for benign and lower values for malignant kidney masses, so we were able to tell these apart. PEER was reliable for identifying tumors with positive staining for the CD117 protein biomarker as well as in the overall patient group. Our results show that PEER could be considered for use with biopsy and CD117 staining to potentially avoid unnecessary surgery for benign kidney masses.

Apparent Diffusion Coefficient Map-Based Texture Analysis for the Differentiation of Chromophobe Renal Cell Carcinoma from Renal Oncocytoma

Preoperative imaging differentiation between ChRCC and RO is difficult with conventional subjective evaluation, and the development of quantitative analysis is a clinical challenge. Forty-nine patients underwent partial or radical nephrectomy preceded by MRI and followed by pathological diagnosis with ChRCC or RO (ChRCC: n = 41, RO: n = 8). The whole-lesion volume of interest was set on apparent diffusion coefficient (ADC) maps of 1.5T-MRI. The importance of selected texture features (TFs) was evaluated, and diagnostic models were created using random forest (RF) analysis. The Mean Decrease Gini as calculated through RF analysis was the highest for mean_ADC_value. ChRCC had a significantly lower mean_ADC_value than RO (1.26 vs. 1.79 × 10−3 mm2/s, p < 0.0001). Feature selection by the Boruta method identified the first-quartile ADC value and GLZLM_HGZE as important features. ROC curve analysis showed that there was no significant difference in the classification performances between the mean_ADC_value-only model and the Boruta model (AUC: 0.954 vs. 0.969, p = 0.236). The mean ADC value had good predictive ability for the distinction between ChRCC and RO, comparable to that of the combination of TFs optimized for the evaluated cohort. The mean ADC value may be useful in distinguishing between ChRCC and RO.

Differential diagnosis of renal oncocytoma and chromophobe renal cell carcinoma using CT features: a central scar-matched retrospective study

BACKGROUND

Renal oncocytoma (RO) and chromophobe renal cell carcinoma (chRCC) have a common cellular origin and different clinical management and prognosis.

PURPOSE

To explore the utility of computed tomography (CT) in the differentiation of RO and chRCC.

MATERIAL AND METHODS

Twenty-five patients with RO and 73 patients with chRCC presenting with the central scar were included retrospectively. Two experienced radiologists independently reviewed the CT imaging features, including location, tumor size, relative density ratio, segmental enhancement inversion (SEI), necrosis, and perirenal fascia thickening, among others. Interclass correlation coefficient (ICC, for continuous variables) or Kappa coefficient test (for categorical variables) was used to determine intra-observer and inter-observer bias between the two radiologists.

RESULTS

The inter- and intra-reader reproducibility of the other CT imaging parameters were nearly perfect (>0.81) except for the measurements of fat (0.662). RO differed from chRCC in the cortical or medullary side (P = 0.005), relative density ratio (P = 0.020), SEI (P < 0.001), and necrosis (P = 0.045). The logistic regression model showed that location (right kidney), hypo-density on non-enhanced CT, SEI, and perirenal fascia thickening were highly predictive of RO. The combined indicators from logistic regression model were used for ROC analysis. The area under the ROC curve was 0.923 (P < 0.001). The sensitivity and specificity of the four factors combined for diagnosing RO were 88% and 86.3%, respectively. The correlation coefficient between necrosis and tumor size in all tumors including both of RO and chRCC was 0.584, indicating a positive correlation (P < 0.001).

CONCLUSION

The CT imaging features of location (right kidney), hypo-density on non-enhanced CT, SEI, and perirenal fascia thickening were valuable indicators in distinguishing RO from chRCC.

A CT-based radiomics nomogram for differentiation of renal oncocytoma and chromophobe renal cell carcinoma with a central scar-matched study

OBJECTIVE

Pre-operative differentiation between renal oncocytoma (RO) and chromophobe renal cell carcinoma (chRCC) is critical due to their different clinical behavior and different clinical treatment decisions. The aim of this study was to develop and validate a CT-based radiomics nomogram for the pre-operative differentiation of RO from chRCC.

METHODS

A total of 141 patients (84 in training data set and 57 in external validation data set) with ROs (n = 47) or chRCCs (n = 94) were included. Radiomics features were extracted from tri-phasic enhanced-CT images. A clinical model was developed based on significant patient characteristics and CT imaging features. A radiomics signature model was developed and a radiomics score (Rad-score) was calculated. A radiomics nomogram model incorporating the Rad-score and independent clinical factors was developed by multivariate logistic regression analysis. The diagnostic performance was evaluated and validated in three models using ROC curves.

RESULTS

Twelve features from CT images were selected to develop the radiomics signature. The radiomics nomogram combining a clinical factor (segmental enhancement inversion) and radiomics signature showed an AUC value of 0.988 in the validation set. Decision curve analysis revealed that the diagnostic performance of the radiomics nomogram was better than the clinical model and the radiomics signature.

CONCLUSIONS

The radiomics nomogram combining clinical factors and radiomics signature performed well for distinguishing RO from chRCC.

ADVANCES IN KNOWLEDGE

Differential diagnosis between renal oncocytoma (RO) and chromophobe renal cell carcinoma (chRCC) is rather difficult by conventional imaging modalities when a central scar was present.A radiomics nomogram integrated with the radiomics signature, demographics, and CT findings facilitates differentiation of RO from chRCC with improved diagnostic efficacy.The CT-based radiomics nomogram might spare unnecessary surgery for RO.

Radiomic Features and Machine Learning for the Discrimination of Renal Tumor Histological Subtypes: A Pragmatic Study Using Clinical-Routine Computed Tomography

Simple Summary

This study evaluates how advanced image analyses (radiomic features) and machine learning algorithms can help to distinguish subtypes of kidney tumors in computed tomography (CT) images, which is important for further patient treatment. For 201 patients, the image analyses showed a moderate performance, but robustly performed across various imaging centers and even in cases with suboptimal image quality. In particular, distinguishing one specific subtype of kidney tumor (oncocytomas) from other subtypes proves to be challenging. The algorithms presented in this study can help in the clinical decision-making process for kidney tumor patients, for example, to decide whether to perform kidney surgery or not.

Abstract

This study evaluates the diagnostic performance of radiomic features and machine learning algorithms for renal tumor subtype assessment in venous computed tomography (CT) studies from clinical routine. Patients undergoing surgical resection and histopathological assessment of renal tumors at a tertiary referral center between 2012 and 2019 were included. Preoperative venous-phase CTs from multiple referring imaging centers were segmented, and standardized radiomic features extracted. After preprocessing, class imbalance handling, and feature selection, machine learning algorithms were used to predict renal tumor subtypes using 10-fold cross validation, assessed as multiclass area under the curve (AUC). In total, n = 201 patients were included (73.7% male; mean age 66 ± 11 years), with n = 131 clear cell renal cell carcinomas (ccRCC), n = 29 papillary RCC, n = 11 chromophobe RCC, n = 16 oncocytomas, and n = 14 angiomyolipomas (AML). An extreme gradient boosting algorithm demonstrated the highest accuracy (multiclass area under the curve (AUC) = 0.72). The worst discrimination was evident for oncocytomas vs. AML and oncocytomas vs. chromophobe RCC (AUC = 0.55 and AUC = 0.45, respectively). In sensitivity analyses excluding oncocytomas, a random forest algorithm showed the highest accuracy, with multiclass AUC = 0.78. Radiomic feature analyses from venous-phase CT acquired in clinical practice with subsequent machine learning can discriminate renal tumor subtypes with moderate accuracy. The classification of oncocytomas seems to be the most complex with the lowest accuracy.

Automated differentiation of benign renal oncocytoma and chromophobe renal cell carcinoma on computed tomography using deep learning

OBJECTIVES

To develop and evaluate the feasibility of an objective method using artificial intelligence (AI) and image processing in a semi-automated fashion for tumour-to-cortex peak early-phase enhancement ratio (PEER) in order to differentiate CD117(+) oncocytoma from the chromophobe subtype of renal cell carcinoma (ChRCC) using convolutional neural networks (CNNs) on computed tomography imaging.

METHODS

The CNN was trained and validated to identify the kidney + tumour areas in images from 192 patients. The tumour type was differentiated through automated measurement of PEER after manual segmentation of tumours. The performance of this diagnostic model was compared with that of manual expert identification and tumour pathology with regard to accuracy, sensitivity and specificity, along with the root-mean-square error (RMSE), for the remaining 20 patients with CD117(+) oncocytoma or ChRCC.

RESULTS

The mean ± sd Dice similarity score for segmentation was 0.66 ± 0.14 for the CNN model to identify the kidney + tumour areas. PEER evaluation achieved accuracy of 95% in tumour type classification (100% sensitivity and 89% specificity) compared with the final pathology results (RMSE of 0.15 for PEER ratio).

CONCLUSIONS

We have shown that deep learning could help to produce reliable discrimination of CD117(+) benign oncocytoma and malignant ChRCC through PEER measurements obtained by computer vision.

Clear cell renal cell carcinoma bone metastasis: What should be considered in prognostic evaluation

INTRODUCTION

Knowledge of clear cell renal cell carcinoma bone metastasis (ccRCC-BM) remains scarce. This study depicts clinical, pathological and outcome features of the disease and provides suggestions to establish prognosis prediction system more appropriate for ccRCC-BM.

MATERIALS AND METHODS

Patients with ccRCC-BM had clinical, pathological data collected. Kaplan-Meier survival analysis was used for outcome profiles. Prognostic risks were evaluated using MSKCC/Motzer score. Univariate and multivariate logistic regression were performed to investigate association between clinical, pathological features and prognosis.

RESULTS

In the series containing 106 ccRCC-BM patients with 4:1 male predominance, 44.3% of them had synchronous bone metastasis and 28.3% had multi-organ metastasis. Axial bone was prone to bone metastasis and the incidence of severe skeletal-related events was 54.7%. Curative bone lesion resection was performed in 70.7% patients. The median overall survival (mOS) time was 45 months for all and 32 months for those in unfavorable risk stratification. Shorter time to bone metastasis (TTBM) [OR 1.019, 95% CI (1.007, 1.031)], elderly age [OR 1.040, 95% CI (1.001, 1.080)], concomitant multi-organ metastasis [OR 3.883, 95% CI (1.375, 10.967)] and carbonic anhydrase (CA)-IX expression loss [OR 58.824, 95% CI (2.653, 1000)] were associated with poor prognosis.

CONCLUSION

The outcome of ccRCC-BM remained poor in unfavorable risk stratification. Bone lesion resection accompanied by systematic therapy for selected patient could improve prognosis. Shorter TTBM, elderly age, concomitant multi-organ metastasis and the expression loss of CA-IX along with gender-bias, feasibility for surgical treatment are suggested to be incorporated in modified ccRCC-BM-specific prognosis prediction system.