Usefulness of MDCT to Differentiate Between Renal Cell Carcinoma and Oncocytoma: Development of a Predictive Model

Quantitative Measurement on Contrast-Enhanced CT Distinguishes Small Clear Cell Renal Cell Carcinoma From Benign Renal Tumors: A Multicenter Study

RATIONALE AND OBJECTIVES

To evaluate the potential of quantitative measurements on contrast-enhanced CT (CECT) in differentiating small (≤4 cm) clear cell renal cell carcinoma (ccRCC) from benign renal tumors, including fat-poor angiomyolipoma (fpAML) and renal oncocytoma (RO).

MATERIALS AND METHODS

244 patients with pathologically confirmed ccRCC (n = 184) and benign renal tumors (fpAML, n = 50; RO, n = 10) were randomly assigned into training cohort (n = 193) and test cohort 1 (n = 51), while external test cohort 2 (n = 50) was from another hospital. Quantitative parameters were obtained from CECT (unenhanced phase, UP; corticomedullary phase, CMP; nephrographic phase, NP; excretory phase, EP) by measuring attenuation of renal mass and cortex and subsequently calculated. Univariable and multivariable logistic regression analyses were performed to evaluate the association between these parameters and ccRCC. Finally, the constructed models were compared with radiologists' diagnoses.

RESULTS

In univariable analysis, UP-related parameters, particularly UPC-T (cortex minus tumor attenuation on UP), demonstrated AUC of 0.766 in training cohort, 0.901 in test cohort 1, 0.805 in test cohort 2. The heterogeneity-related parameter SD (standard deviation) showed AUC of 0.781, 0.834, and 0.875 respectively. In multivariable analysis, model 1 incorporating UPC-T, NPC-T (cortex minus tumor attenuation on NP), CMPT-UPT (tumor attenuation on CMP minus UP), and SD yielded AUC of 0.866, 0.923, and 0.949 respectively. When compared with radiologists, multivariate models demonstrated higher accuracy (0.800-0.860) and sensitivity (0.794-0.971) than radiologists' assessments (accuracy: 0.700-0.720, sensitivity: 0.588-0.706).

CONCLUSION

Quantitative measurements on CECT, particularly UP- and heterogeneity-related parameters, have potential to discriminate ccRCC and benign renal tumors (fpAML, RO).

CT findings and clinical characteristics in distinguishing renal urothelial carcinoma mimicking renal cell carcinoma from clear cell renal cell carcinoma

BACKGROUND

We aimed to characterize the clinical and multiphase computed tomography (CT) features, which can distinguish renal urothelial carcinoma (RUC) mimicking renal cell carcinoma (RCC) from clear cell renal cell carcinoma (ccRCC) with collecting system invasion (CSI).

METHODS

Data from 56 patients with RUC (46 men and 10 women) and 366 patients with ccRCC (262 men and 104 women) were collected and assessed retrospectively. The median age was 65.50 (IQR: 56.25-69.75) and 53.50 (IQR: 42.25-62.5) years, respectively. Univariate and multivariate logistic regression analyses were performed on clinical and CT characteristics to determine independent factors for distinguishing RUC and ccRCC, and an integrated predictive model was constructed. Differential diagnostic performance was assessed using the area under the receiver operating characteristic curve (AUC).

RESULTS

The independent predictors for differentiating RUC from ccRCC were infiltrative growth pattern, hydronephrosis, heterogeneous enhancement, preserving reniform contour, and hematuria. The differential diagnostic performance of the integrated predictive model-1 (AUC: 0.947, sensitivity: 89.07%, specificity: 89.29%) and model-2 (AUC: 0.960, sensitivity: 92.1%, specificity: 89.3%) were both better than that of the infiltrative growth pattern (AUC: 0.830, sensitivity: 71.9%, specificity: 92.9%), heterogeneous enhancement (AUC: 0.771, sensitivity: 86.3%, specificity: 67.9%), preserving reniform contour (AUC = 0.758, sensitivity: 85.5%, specificity: 66.1%), hydronephrosis (AUC: 0.733, sensitivity: 87.7%, specificity: 58.9%), or hematuria (AUC: 0.706, sensitivity: 79.5%, specificity: 51.8%).

CONCLUSION

The CT and clinical characteristics showed extraordinary discriminative abilities in the differential diagnosis of RUC and ccRCC, which might provide helpful information for clinical decision-making.

CT findings and clinical characteristics in distinguishing renal urothelial carcinoma mimicking renal cell carcinoma from clear cell renal cell carcinoma.. [DOI: 10.21203/rs.3.rs-2655480/v1] [Citation(s) in RCA: 0] [Impact Index Per Article: 0]

Background: We aimed to characterize the clinical and multiphase computed tomography (CT) features, which can distinguish renal urothelial carcinoma (RUC) mimicking renal cell carcinoma (RCC) from clear cell renal cell carcinoma (ccRCC) with collecting system invasion (CSI). Methods: Data from 56 patients with RUC (46 men and 10 women) and 366 patients with ccRCC (262 men and 104 women) were collected and assessed retrospectively. The median age was 65.50 (IQR: 56.25–69.75) and 53.50 (IQR: 42.25–62.5) years, respectively. Univariate and multivariate logistic regression analyses were performed on clinical and CT characteristics to determine independent factors for distinguishing RUC and ccRCC, and an integrated predictive model was constructed. Differential diagnostic performance was assessed using the area under the receiver operating characteristic curve (AUC). Results: The independent predictors for differentiating RUC from ccRCC were infiltrative growth pattern, hydronephrosis, heterogeneous enhancement, preserving reniform contour, and hematuria. The differential diagnostic performance of the integrated predictive model (AUC: 0.960, sensitivity: 92.1%, specificity: 89.3%) was better than that of the infiltrative growth pattern (AUC: 0.830, sensitivity: 71.9%, specificity: 92.9%), heterogeneous enhancement (AUC: 0.771, sensitivity: 86.3%, specificity: 67.9%), preserving reniform contour (AUC=0.758, sensitivity: 85.5%, specificity: 66.1%), hydronephrosis (AUC: 0.733, sensitivity: 87.7%, specificity: 58.9%), or hematuria (AUC: 0.706, sensitivity: 79.5%, specificity: 51.8%). Conclusion: The CT and clinical characteristics showed extraordinary discriminative abilities in the differential diagnosis of RUC and ccRCC, which might provide helpful information for clinical decision-making.

CT differentiation of the oncocytoma and renal cell carcinoma based on peripheral tumor parenchyma and central hypodense area characterisation

BACKGROUND

Although the central scar is an essential imaging characteristic of renal oncocytoma (RO), its utility in distinguishing RO from renal cell carcinoma (RCC) has not been well explored. The study aimed to evaluate whether the combination of CT characteristics of the peripheral tumor parenchyma (PTP) and central hypodense area (CHA) can differentiate typical RO with CHA from RCC.

METHODS

A total of 132 tumors on the initial dataset were retrospectively evaluated using four-phase CT. The excretory phases were performed more than 20 min after the contrast injection. In corticomedullary phase (CMP) images, all tumors had CHAs. These tumors were categorized into RO (n = 23), clear cell RCC (ccRCC) (n = 85), and non-ccRCC (n = 24) groups. The differences in these qualitative and quantitative CT features of CHA and PTP between ROs and ccRCCs/non-ccRCCs were statistically examined. Logistic regression filters the main factors for separating ROs from ccRCCs/non-ccRCCs. The prediction models omitting and incorporating CHA features were constructed and evaluated, respectively. The effectiveness of the prediction models including CHA characteristics was then confirmed through a validation dataset (8 ROs, 35 ccRCCs, and 10 non-ccRCCs).

RESULTS

The findings indicate that for differentiating ROs from ccRCCs and non-ccRCCs, prediction models with CHA characteristics surpassed models without CHA, with the corresponding areas under the curve (AUC) being 0.962 and 0.914 versus 0.952 and 0.839 respectively. In the prediction models that included CHA parameters, the relative enhancement ratio (RER) in CMP and enhancement inversion, as well as RER in nephrographic phase and enhancement inversion were the primary drivers for differentiating ROs from ccRCCs and non-ccRCCs, respectively. The prediction models with CHA characteristics had the comparable diagnostic ability on the validation dataset, with respective AUC values of 0.936 and 0.938 for differentiating ROs from ccRCCs and non-ccRCCs.

CONCLUSION

The prediction models with CHA characteristics can help better differentiate typical ROs from RCCs. When a mass with CHA is discovered, particularly if RO is suspected, EP images with longer delay scanning periods should be acquired to evaluate the enhancement inversion characteristics of CHA.

Small Renal Masses without Gross Fat: What Is the Role of Contrast-Enhanced MDCT?

Increased detection of small renal masses (SRMs) has encouraged research for non-invasive diagnostic tools capable of adequately differentiating malignant vs. benign SRMs and the type of the tumour. Multi-detector computed tomography (MDCT) has been suggested as an alternative to intervention, therefore, it is important to determine both the capabilities and limitations of MDCT for SRM evaluation. In our study, two abdominal radiologists retrospectively blindly assessed MDCT scan images of 98 patients with incidentally detected lipid-poor SRMs that did not present as definitely aggressive lesions on CT. Radiological conclusions were compared to histopathological findings of materials obtained during surgery that were assumed as the gold standard. The probability (odds ratio (OR)) in regression analyses, sensitivity (SE), and specificity (SP) of predetermined SRM characteristics were calculated. Correct differentiation between malignant vs. benign SRMs was detected in 70.4% of cases, with more accurate identification of malignant (73%) in comparison to benign (65.7%) lesions. The radiological conclusions of SRM type matched histopathological findings in 56.1%. Central scarring (OR 10.6, p = 0.001), diameter of lesion (OR 2.4, p = 0.003), and homogeneous accumulation of contrast medium (OR 3.4, p = 0.03) significantly influenced the accuracy of malignant diagnosis. SE and SP of these parameters varied from 20.6% to 91.3% and 22.9% to 74.3%, respectively. In conclusion, MDCT is able to correctly differentiate malignant versus uncharacteristic benign SRMs in more than 2/3 of cases. However, frequency of the correct histopathological SRM type MDCT identification remains low.

A CT-based radiomics nomogram for differentiation of small masses (< 4 cm) of renal oncocytoma from clear cell renal cell carcinoma

PURPOSE

Renal oncocytoma (RO) is the most commonly resected benign renal tumor because of misdiagnosis as renal cell carcinoma. This misdiagnosis is generally owing to overlapping imaging features. This study describes the building of a radiomics nomogram based on clinical data and radiomics signature for the preoperative differentiation of RO from clear cell renal cell carcinoma (ccRCC) on tri-phasic contrast-enhanced CT.

METHODS

A total of 122 patients (85 in training set and 37 in external validation set) with ROs (n = 46) or ccRCCs (n = 76) were enrolled. Patient characteristics and tri-phasic contrast-enhanced CT imaging features were evaluated to build a clinical factors model. A radiomics signature was constructed by extracting radiomics features from tri-phasic contrast-enhanced CT images and a radiomics score (Rad-score) was calculated. A radiomics nomogram was then built by incorporating the Rad-score and significant clinical factors according to a multivariate logistic regression analysis. The diagnostic performance of the above three models was evaluated in training and validation sets.

RESULTS

Central stellate area and perirenal fascia thickening were selected to build the clinical factors model. Eleven radiomics features were combined to construct the radiomics signature. The AUCs of the radiomics nomogram, which was based on the selected clinical factors and Rad-score, were 0.960 and 0.898 in the training and validation sets, respectively. The decision curves of the radiomics nomogram and radiomics signature in the validation set indicated an overall net benefit over the clinical factors model.

CONCLUSION

Our radiomics nomogram can effectively predict the preoperative diagnosis of ROs and may therefore be of assistance in sparing unnecessary surgery and tailoring precise therapy. The ROC curves of the clinical model, the radiomics signature and the radiomics nomogram for the validation set. RO = Renal oncocytoma; ccRCC = Clear cell renal cell carcinoma.

Differentiation between renal oncocytomas and chromophobe renal cell carcinomas using dynamic contrast-enhanced computed tomography

PURPOSE

To evaluate the ability of inferior vena cava-lesion-attenuation-difference (ILAD) and lesion-cortex-attenuation-ratio (LCAR) to differentiate renal oncocytomas (RO) from chromophobe renal cell carcinomas (chRCC).

METHODS

Retrospective study with analysis of 84 cases of chRCC and 30 cases of RO confirmed by surgical pathology. ILAD was calculated by measuring the difference in Hounsfield units (HU) between the inferior vena cava and the lesion of interest on the same image slice on preoperative CT scan. Calculating LCAR using the CT attenuation ratio of lesion to renal cortex at the same image slice. Receiver operating characteristic (ROC) curves were plotted to analyze the diagnostic values of ILAD and LCAR for disease activity.

RESULTS

There were no statistically significant differences in demographic and lesion characteristics between patients with chRCC and RO (p > 0.05). ILAD has significant statistical differences in the identification of RO and chRCC in the arterial (p = 0.031), venous (p = 0.047), and delayed phase (p = 0.002). And LCAR showed a statistically significant difference between two lesions during the arterial (p = 0.043), venous (p = 0.026), and delayed phase (p = 0.008). When all significant variables were used in combination to build a predicting model (Mix), the AUC was 0.871 (95% CI 0.759-0.984) with 67.9% sensitivity and 100% specificity.

CONCLUSION

ILAD and LCAR at the arterial phase, venous phase and delayed phase were shown to be useful CT attenuation parameter in discriminating RO from chRCC when histologic evaluation on biopsy is indeterminate.

CT differentiation of fat-poor angiomyolipomas from papillary renal cell carcinomas: development of a predictive model

PURPOSE

To identify specific contrast-enhanced CT (CECT) findings and develop a predictive model with logistic regression to differentiate fat-poor angiomyolipomas (fpAML) from papillary renal cell carcinomas (pRCC).

METHODS

This is a single-institution retrospective study that assess CT features of histologically proven 67 pRCC and 13 fpAML. CECT variables were studied by means of univariate logistic regression. Variables included patients' demographics, tumor attenuation (unenhanced and at arterial, venous and excretory post-contrast phases), type of enhancement, morphological features (axial long and short diameters, long-short axis ratio (LSR) and tumor to kidney angle interface) and presence of visible calcifications or vessels. Those variables with a p ≤ 0.05 underwent standard stepwise logistic regression to find predictive combinations of clinical variables. Best models were evaluated by AUROC curves and were subjected to Leave-one-out cross validation to assess their robustness.

RESULTS

Odds ratio (OR) between pRCC and fpAML was statistically significant for patient's gender, tumor attenuation in arterial, venous and excretory phases, tumor's long diameter, short diameter, LSR, type of enhancement, presence of intratumoral vessels and tumor-kidney angle interface. The best predictive model resulted in an area under the curve (AUC) of 0.971 and included gender, tumor-kidney angle interface and venous attenuation with the following equation: Log(p/1 - p) = - 2.834 + 4.052 * gender +  - 0.066 * AngleInterface + 0.074 * VenousphaseHU.

CONCLUSIONS

The combination of patients' gender, tumor to kidney angle interface and venous enhancement helps to distinguish fpAML from pRCC.

Clinicopathological and radiological significance of the collateral vessels of renal cell carcinoma on preoperative computed tomography

This study aimed to investigate the clinicopathological and radiological significance of the collateral vessel of renal cell carcinoma (RCC) on preoperative computed tomography (CT). Preoperative contrast-enhanced CT of 236 consecutive patients with pathological documented RCC were retrospectively reviewed during the period of 2014. The associations of the presence of collateral vessels with perioperative clinicopathological and radiological features, as well as long term survival outcomes were analyzed. Totally, collateral vessels were detected by contrast-enhanced CT in 110 of 236 patients. The presence of collateral vessels was significantly associated with higher pathologic T stage, higher Fuhrman grade, higher overall RENAL scores, greater tumor size and enhancement, and more tumor necrosis (all P < 0.05). In patients with clear cell RCC, those harboring collateral vessels had significantly higher SSIGN scores (P < 0.001) and shorter overall survival (P = 0.01) than those without collateral vessel. The incidence of intraoperative blood loss, blood transfusion, radical nephrectomy (RN) and open surgery were also significantly higher in patients with collateral vessels (all P < 0.05). In multivariate analysis, the presence of collateral vessels was significantly associated with RN (P = 0.021) and open surgery (P = 0.012). The presence of collateral vessels was significantly associated with aggressive clinicopathological parameters and worse prognosis. It is worth paying attention to its association with the choice of RN and open surgery in clinical practice.

Imaging Characterization of Renal Masses

The detection of a renal mass is a relatively frequent occurrence in the daily practice of any Radiology Department. The diagnostic approaches depend on whether the lesion is cystic or solid. Cystic lesions can be managed using the Bosniak classification, while management of solid lesions depends on whether the lesion is well-defined or infiltrative. The approach to well-defined lesions focuses mainly on the differentiation between renal cancer and benign tumors such as angiomyolipoma (AML) and oncocytoma. Differential diagnosis of infiltrative lesions is wider, including primary and secondary malignancies and inflammatory disease, and knowledge of the patient history is essential. Radiologists may establish a possible differential diagnosis based on the imaging features of the renal masses and the clinical history. The aim of this review is to present the contribution of the different imaging techniques and image guided biopsies in the diagnostic management of cystic and solid renal lesions.

Usefulness of multidetector computed tomography to differentiate between renal cell carcinoma and oncocytoma. A model validation

OBJECTIVE

The purpose of this study is to validate a multivariable predictive model previously developed to differentiate between renal cell carcinoma (RCC) and oncocytoma using CT parameters.

METHODS AND MATERIALS

We included 100 renal lesions with final diagnosis of RCC or oncocytoma studied before surgery with 4-phase multidetector CT (MDCT). We evaluated the characteristics of the tumors and the enhancement patterns at baseline, arterial, nephrographic and excretory MDCT phases.

RESULTS

Histopathologically 15 tumors were oncocytomas and 85 RCCs. RCCs were significantly larger (median 4.4 cm vs 2.8 cm, p = 0.006). There were significant differences in nodule attenuation in the excretory phase compared to baseline (median: 31 vs 42, p = 0.015), with RCCs having lower values. Heterogeneous enhancement patterns were also more frequent in RCCs (85.9% vs 60%, p = 0.027).Multivariable analysis showed that the independent predictors of malignancy were the enhancement pattern, with oncocytomas being more homogeneous in the nephrographic phase [Odds Ratio (OR) 0.16 (95% CI 0.03 to 0.75, p = 0.02)], nodule enhancement in the excretory phase compared to baseline, with RCCs showing lower enhancement [OR 0.96 (95% CI 0.93 to 0.99, p = 0.005)], and a size > 4 cm, with RCCs being larger [OR 5.89 (95% CI 1.10 to 31.58), p = 0.038].

CONCLUSION

The multivariable predictive model previously developed which combines different MDCT parameters, including lesion size > 4 cm, lesion enhancement in the excretory phase compared to baseline and enhancement heterogeneity, can be successfully applied to distinguish RCC from oncocytoma.

ADVANCES IN KNOWLEDGE

This study confirms that multiparametric assessment using MDCT (including parameters such as size, homogeneity and enhancement differences between the excretory and the baseline phases) can help distinguish between RCCs and oncocytomas. While it is true that this multiparametric predictive model may not always correctly classify renal tumors such as RCC or oncocytoma, it can be used to determine which patients would benefit from pre-surgical biopsy to confirm that the tumor is in fact an oncocytoma, and thereby avoid unnecessary surgical treatments.

Renal Oncocytoma: An Algorithm for Diagnosis and Management

Renal oncocytoma is an uncommon tumor that exhibits numerous features which are characteristic but not necessarily unique. Percutaneous biopsy is a safe method of diagnosis. However, differentiation from other tumor subtypes often requires sophisticated analysis and is not universally feasible. This is why, surgical management can be considered as a first-line treatment or after surveillance. Potential triggers for change in management are: tumor size >3 cm, stage progression, kinetics of size progression (>5 mm/y), and clinical change in patient or tumor factors. Long-term follow-up data are lacking and greater centralization should be considered to reach adequate management.

Small renal masses (≤ 4 cm): differentiation of oncocytoma from renal clear cell carcinoma using ratio of lesion to cortex attenuation and aorta-lesion attenuation difference (ALAD) on contrast-enhanced CT

PURPOSE

We investigate the use of ratio of lesion to cortex (L/C) attenuation and aorta-lesion attenuation difference (ALAD) on multiphase contrast-enhanced CT to help distinguish oncocytoma from clear cell RCC in small renal masses (diameter < 4 cm).

METHODS

We retrospectively identified 76 patients that undergo CT before surgery for a suspicious small renal mass between January 2014 and December 2018 with pathological diagnosis of 21 oncocytomas (ROs), 25 clear cell RCCs, 7 chromophobe RCCs, 7 papillary RCCs, 7 multilocular cystic RCCs, 7 angiomyolipomas and 2 leiomyomas. CT attenuation values were obtained for the tumor, the normal renal cortex and the aorta, placing a circular region of interest (ROI) in the same slice by two radiologists, independently.

RESULTS

In the corticomedullary phase, ROs showed isodense enhancement to the renal cortex (ratio L/C 0.92 ± 0.12), while clear cell RCCs appeared hypodense to the renal cortex (ratio L/C 0.69 ± 0.20; p < 0.01) with an accuracy of 80% for diagnosing RO. In nephrographic phase, the ratio L/C attenuation was lower than the corticomedullary phase in ROs (0.78 ± 0.11) showing an early washout pattern, while the ratio L/C was similar to the corticomedullary phase in clear cell RCCs (0.69 ± 0.13; p = 0.025, with an accuracy of 65% for diagnosing RO). The ratio L/C attenuation showed considerable overlap between ROs and clear cell RCCs in the excretory phase (p = 0.27). Mean ALAD values in the nephrographic phase were 21.95 ± 16.24 for ROs and 36.96 ± 30.53 for clear cell RCCs (p = 0.049).

CONCLUSION

The ratio L/C attenuation in corticomedullary phase may be useful to differentiate RO from clear cell RCC.

Deep learning and radiomics: the utility of Google TensorFlow™ Inception in classifying clear cell renal cell carcinoma and oncocytoma on multiphasic CT

PURPOSE

Currently, all solid enhancing renal masses without microscopic fat are considered malignant until proven otherwise and there is substantial overlap in the imaging findings of benign and malignant renal masses, particularly between clear cell RCC (ccRCC) and benign oncocytoma (ONC). Radiomics has attracted increased attention for its utility in pre-operative work-up on routine clinical images. Radiomics based approaches have converted medical images into mineable data and identified prognostic imaging signatures that machine learning algorithms can use to construct predictive models by learning the decision boundaries of the underlying data distribution. The TensorFlow™ framework from Google is a state-of-the-art open-source software library that can be used for training deep learning neural networks for performing machine learning tasks. The purpose of this study was to investigate the diagnostic value and feasibility of a deep learning-based renal lesion classifier using open-source Google TensorFlow™ Inception in differentiating ccRCC from ONC on routine four-phase MDCT in patients with pathologically confirmed renal masses.

METHODS

With institutional review board approval for this 1996 Health Insurance Portability and Accountability Act compliant retrospective study and a waiver of informed consent, we queried our institution's pathology, clinical, and radiology databases for histologically proven cases of ccRCC and ONC obtained between January 2000 and January 2016 scanned with a an intravenous contrast-enhanced four-phase renal mass protocol (unenhanced (UN), corticomedullary (CM), nephrographic (NP), and excretory (EX) phases). To extract features to be used for the machine learning model, the entire renal mass was contoured in the axial plane in each of the four phases, resulting in a 3D volume of interest (VOI) representative of the entire renal mass. We investigated thirteen different approaches to convert the acquired VOI data into a set of images that adequately represented each tumor which was used to train the final layer of the neural network model. Training was performed over 4000 iterations. In each iteration, 90% of the data were designated as training data and the remaining 10% served as validation data and a leave-one-out cross-validation scheme was implemented. Accuracy, sensitivity, specificity, positive (PPV) and negative predictive (NPV) values, and CIs were calculated for the classification of the thirteen processing modes.

RESULTS

We analyzed 179 consecutive patients with 179 lesions (128 ccRCC and 51 ONC). The ccRCC cohort had a mean size of 3.8 cm (range 0.8-14.6 cm) and the ONC cohort had a mean lesion size of 3.9 cm (range 1.0-13.1 cm). The highest specificity and PPV (52.9% and 80.3%, respectively) were achieved in the EX phase when we analyzed the single mid-slice of the tumor in the axial, coronal and sagittal plane, and when we increased the number of mid-slices of the tumor to three, with an accuracy of 75.4%, which also increased the sensitivity to 88.3% and the PPV to 79.6%. Using the entire tumor volume also showed that classification performance was best in the EX phase with an accuracy of 74.4%, a sensitivity of 85.8% and a PPV of 80.1%. When the entire tumor volume, plus mid-slices from all phases and all planes presented as tiled images, were submitted to the final layer of the neural network we achieved a PPV of 82.5%.

CONCLUSIONS

The best classification result was obtained in the EX phase among the thirteen classification methods tested. Our proof of concept study is the first step towards understanding the utility of machine learning in the differentiation of ccRCC from ONC on routine CT images. We hope this could lead to future investigation into the development of a multivariate machine learning model which may augment our ability to accurately predict renal lesion histology on imaging.

CT and MR imaging for solid renal mass characterization

As our understanding has expanded that relatively large fraction of incidentally discovered renal masses, especially in small size, are benign or indolent even if malignant, there is growing acceptance of more conservative management including active surveillance for small renal masses. As for advanced renal cell carcinomas (RCCs), nonsurgical and subtype specific treatment options such as immunotherapy and targeted therapy is developing. On these backgrounds, renal mass characterization including differentiation of benign from malignant tumors, RCC subtyping and prediction of RCC aggressiveness is receiving much attention and a variety of imaging techniques and analytic methods are being investigated. In addition to conventional imaging techniques, integration of texture analysis, functional imaging (i.e. diffusion weighted and perfusion imaging) and multivariate diagnostic methods including machine learning have provided promising results for these purposes in research fields, although standardization and external, multi-institutional validations are needed.

Quantitative computer-aided diagnostic algorithm for automated detection of peak lesion attenuation in differentiating clear cell from papillary and chromophobe renal cell carcinoma, oncocytoma, and fat-poor angiomyolipoma on multiphasic multidetector computed tomography

OBJECTIVE

To evaluate the performance of a novel, quantitative computer-aided diagnostic (CAD) algorithm on four-phase multidetector computed tomography (MDCT) to detect peak lesion attenuation to enable differentiation of clear cell renal cell carcinoma (ccRCC) from chromophobe RCC (chRCC), papillary RCC (pRCC), oncocytoma, and fat-poor angiomyolipoma (fp-AML).

MATERIALS AND METHODS

We queried our clinical databases to obtain a cohort of histologically proven renal masses with preoperative MDCT with four phases [unenhanced (U), corticomedullary (CM), nephrographic (NP), and excretory (E)]. A whole lesion 3D contour was obtained in all four phases. The CAD algorithm determined a region of interest (ROI) of peak lesion attenuation within the 3D lesion contour. For comparison, a manual ROI was separately placed in the most enhancing portion of the lesion by visual inspection for a reference standard, and in uninvolved renal cortex. Relative lesion attenuation for both CAD and manual methods was obtained by normalizing the CAD peak lesion attenuation ROI (and the reference standard manually placed ROI) to uninvolved renal cortex with the formula [(peak lesion attenuation ROI - cortex ROI)/cortex ROI] × 100%. ROC analysis and area under the curve (AUC) were used to assess diagnostic performance. Bland-Altman analysis was used to compare peak ROI between CAD and manual method.

RESULTS

The study cohort comprised 200 patients with 200 unique renal masses: 106 (53%) ccRCC, 32 (16%) oncocytomas, 18 (9%) chRCCs, 34 (17%) pRCCs, and 10 (5%) fp-AMLs. In the CM phase, CAD-derived ROI enabled characterization of ccRCC from chRCC, pRCC, oncocytoma, and fp-AML with AUCs of 0.850 (95% CI 0.732-0.968), 0.959 (95% CI 0.930-0.989), 0.792 (95% CI 0.716-0.869), and 0.825 (95% CI 0.703-0.948), respectively. On Bland-Altman analysis, there was excellent agreement of CAD and manual methods with mean differences between 14 and 26 HU in each phase.

CONCLUSION

A novel, quantitative CAD algorithm enabled robust peak HU lesion detection and discrimination of ccRCC from other renal lesions with similar performance compared to the manual method.

Aorta-Lesion-Attenuation-Difference (ALAD) on contrast-enhanced CT: a potential imaging biomarker for differentiating malignant from benign oncocytic neoplasms

OBJECTIVE

To evaluate whether the Aorta-Lesion-Attenuation-Difference on contrast-enhanced CT can aid in the differentiation of malignant and benign oncocytic renal neoplasms.

MATERIALS AND METHODS

Two independent cohorts-an initial (biopsy) dataset and a validation (surgical) dataset-with oncocytomas and chromophobe renal cell carcinomas (chRCC) were included in this IRB-approved retrospective study. A region of interest was placed on the renal mass and abdominal aorta on the same CT image slice to calculate an Aorta-Lesion-Attenuation-Difference (ALAD). ROC curves were plotted for different enhancement phases, and diagnostic performance of ALAD for differentiating chRCC from oncocytomas was calculated.

RESULTS

Seventy-nine renal masses (56 oncocytomas, 23 chRCC) were analyzed in the initial (biopsy) dataset. Thirty-six renal masses (16 oncocytomas, 20 chRCC) were reviewed in the validation (surgical) cohort. ALAD showed a statistically significant difference between oncocytomas and chromophobes during the nephrographic phase (p < 0.001), early excretory phase (p < 0.001), and excretory phase (p = 0.029). The area under the ROC curve for the nephrographic phase was 1.00 (95% CI: 1.00-1.00) for the biopsy dataset and showed the narrowest confidence interval. At a threshold value of 25.5 HU, sensitivity was 100 (82.2%-100%) and specificity was 81.5 (61.9%-93.7%). When tested on the validation dataset on measurements made by an independent reader, the AUROC was 0.93 (95% CI: 0.84-1.00) with a sensitivity of 100 (80.0%-100%) and a specificity of 87.5 (60.4%-97.8%).

CONCLUSIONS

Nephrographic phase ALAD has potential to differentiate benign and malignant oncocytic renal neoplasms on contrast-enhanced CT if histologic evaluation on biopsy is indeterminate.

Oncocytoma: A Differential Consideration for an Incidentally Detected FDG-Avid Renal Mass on PET/CT

Renal oncocytoma is a benign renal neoplasm that is often discovered incidentally and closely mimics renal cell carcinoma on common imaging modalities such as computed tomography (CT) and magnetic resonance imaging (MRI). Due to the inability to reliably distinguish between these benign and malignant lesions with imaging, both are typically treated as if they are malignant. Hypermetabolic activity of renal oncocytomas is not frequently encountered because positron emission tomography (PET) is not a standard modality for imaging primary renal tumors. We present a case of a 65 year-old female with a history of thyroid cancer who had an incidentally discovered hypermetabolic renal mass on surveillance PET-CT imaging. Due to the concern for a primary renal malignancy or metastatic disease, the mass was resected and proven to be an oncocytoma on pathologic review.