Differential diagnosis of chromophobe renal cell carcinoma and papillary renal cell carcinoma with dual-energy spectral computed tomography

Abstract

Background

Computed tomography (CT) image features of chromophobe renal cell carcinoma (ChRCC) and papillary renal cell carcinoma (PRCC) are, occasionally, sometimes difficult to identify. However, spectral CT might provide quantitative parameters to differentiate them.

Purpose

To differentiate between ChRCC and PRCC with quantitative parameters using spectral CT.

Material and Methods

Forty cases of RCC confirmed with pathological tests were analyzed retrospectively (27 cases of PRCC and 13 cases of ChRCC). All patients underwent non-enhanced CT and dual-phase contrast-enhanced CT scans. For each lesion, the CT value of monochromatic images as well as iodine and water concentrations were measured, and the slope of spectrum curve was calculated. Data were analyzed using Student’s t-test. Sensitivity and specificity of the quantitative parameters were analyzed using the receiver operating characteristic (ROC) curve.

Results

During the cortex phase (CP) and parenchyma phase (PP), the CT value and slope of spectrum curve of ChRCC were higher than those of PRCC, and significant differences were observed at low energy levels (40–70 keV). Normalized iodine concentration of ChRCC and that of PRCC was significantly different during CP and PP (P < 0.05). The water (iodine) concentrations of ChRCC and PRCC in CP and PP were not statistically different (P > 0.05). All the ROCs for parameters were above the reference line.

Conclusion

Spectral CT may help increase the diagnostic accuracy of differentiating PRCC from ChRCC using a quantitative analysis.

Keywords

Spectral computed tomography chromophobe renal cell carcinoma papillary renal cell carcinoma

Introduction

Renal cell carcinoma (RCC) is the most common primary malignant tumor of the kidney and accounts for approximately 85%–90% of all malignant renal tumors (1). RCC includes 11 subtypes (2004 World Health Organization [WHO] histologic classification of RCC). Clear-cell RCC is the most common histologic subtype, accounting for 70% of all cases of RCC, followed by papillary RCC (PRCC), accounting for 10%–15%, and chromophobe RCC (ChRCC) accounting for < 5% of all cases of RCC (2). There is a new understanding that renal tumors are the result of complex interplay among genetics, epigenetics, and environment (3,4). Many new renal tumor subtypes with unique clinicopathological features have been widely recognized. Compared to the 2004 edition, the WHO classification of the kidney in 2016 (5) included six new RCC subtypes, and four other types of tumors, which were not fully characterized, were listed as tentative RCC subtypes. PRCC is classified into types I and II based on the nuclear grade and cell arrangement (6). Recently, eosinophilic PRCC was reported to have similar immunophenotypes and molecular pathological changes as PRCC and, therefore, was classified as a morphological subtype of PRCC (7). Meanwhile, the definition of ChRCC has not changed significantly in the latest WHO edition.

Notably, RCC subtypes are associated with different biological behaviors, prognosis, and treatment options. Currently, the common RCC subtypes can often be distinguished using different diagnostic techniques, such as, angiography, computed tomography (CT), Doppler ultrasound, and magnetic resonance imaging (MRI) (8,9). The incidences of PRCC and ChRCC increase in patients with RCC (10). Diagnosis is generally based on typical CT imaging characteristics. Imaging findings may be different: as PRCC patients are more prone to cysts, necrosis, and calcification compared to ChRCC, while some patients with ChRCC may present with spoke-wheel-like enhancement in the central tumor (11). However, PRCC and ChRCC may show less intense homogeneous and pseudocapsular enhancement. It in two-phase enhancement images. Therefore, it may be difficult to differentiate between PRCC and ChRCC in conventional CT images. While renal resection has been proven to treat patients with RCC highly effectively, while a different scope of operation is required in patients with PRCC or ChRCC; therefore, a differential diagnosis before surgery is necessary.

Clinically, the use of conventional CT to distinguish among pathological types of RCC has several limitations. Wildberger (12) identified the subtypes of renal tumors based on CT signs; however, imaging signs do not accurately indicate the subtypes of tumors. In recent years, dual-energy spectral CT using a single-source instantaneous kVp switch can provide spectral CT images and data. It can produce material decomposition images and monochromatic spectral images at energy levels in the range of 40–140 keV. Spectral CT has been found useful in several clinical applications, including detection of small lesions, optimized display of the hepatic portal vein, and differentiating hepatocellular carcinoma from angiomyolipoma of the liver (13). The purpose of this study was to preliminarily investigate the value of spectral CT imaging in differentiating PRCC from ChRCC.

Material and Methods

Patients

From July 2014 to July 2016, we enrolled 168 patients (105 men, 63 girls/women; median age = 57 years; age range = 12–82 years) who were diagnosed with or suspected to have renal tumors and, consequently, underwent non-enhanced CT or dual-phase enhanced CT imaging in the spectral mode of a high-definition CT system (Discovery CT750HD, GE Healthcare, Boston, Massa, USA). A total of 128 patients were excluded from the study because: (i) there was no renal tumor (n = 57); (ii) confirmation of histological findings was not sufficient (n = 24); (iii) there were no enhanced images and data (n = 41); and (iv) the image quality did not meet diagnostic needs (n = 6). Finally 40 patients (26 men, 14 women; median age = 51 years; age range = 24–76 years) were included in our study. Twenty-seven patients (19 men, 8 women; median age = 54 years; age range = 24–76 years) had PRCC (mean diameter = 3.8 cm, range = 1.2–7.6 cm) and 13 patients (7 men, 6 women; median age = 61 years; age range = 37–72 years) had ChRCC (mean diameter = 4.0 cm; range = 2.0–10.0 cm). All cases of ChRCC and PRCC in the 40 patients were confirmed with surgery and pathological examination. The subtypes of differentiation were determined with retrospective analysis of pathologic reports according to the 2004 WHO histologic classification of RCC. Based on the histological characteristics obtained by using hematoxylin and eosin (H&E) staining, the participants were divided into two groups: ChRCC and PRCC.

CT examinations

Non-enhanced and dual-phase contrast-enhanced CT examinations were performed using the Discovery CT750 HD CT (GE Healthcare, Waukesha, WI, USA). The scan range included the liver up to the anterior superior spine. All patients underwent imaging in the supine position craniocaudally. At tube voltage of 120 kVp, the conventional helical mode yielded scout and non-enhanced images. Subsequently, a nonionic contrast medium (iohexol, 300 mg of iodine/mL, a total of 80–100 mL, 1.2 mL/kg of body weight) was injected via an antecubital venous access and a high-pressure injector (Ulrich Medical, Ulrich GmbH & Co. KG, Ulm, Germany). The rate was 3.0–4.5 mL/s during the cortex phase (CP) and parenchyma phase (PP). In order to reduce individual differences, we determined the amount of contrast agent and injection rate based on body mass index (BMI) and some clinical conditions (age, vascular state, renal function, and medical history). Using automatic image-triggering software (SmartPrep; GE Healthcare), CP imaging could be scanned. CP imaging began 10 s after the trigger attenuation threshold (150 HU) was reached at the level of the supra-coeliac abdominal aorta. PP imaging scanning was started 40s after the end of CP imaging scanning. CP and PP imaging were performed in the spectral imaging mode with a fast tube voltage switching between 80 kVp and 140 on adjacent views in a single rotation. Other spectral imaging parameters were as follows: collimation thickness = 0.625 mm; tube current = 600 mA; rotation speed = 0.6 s; helical pitch = 0.983; and reconstruction thickness = 1.25 mm (14). The Gemstone Spectral Imaging (GSI) viewer software and a standard reconstruction kernel were used to analyze the spectral CT images: conventional polychromatic images obtained at 120 kVp; and iodine-based and water-based material decomposition images obtained at 70-keV monochromatic images. Monochromatic images were obtained at values in the range of 40–140 keV.

Quantitative analyses

With the GSI viewer, data were processed by the advanced workstation (AW4.6; GE Healthcare). Circular or elliptical regions of interest (ROIs) were set in an area of approximately 100 mm². Further, they were labeled at the lesions and aorta on the default 70-keV monochromatic images. ROIs should include as much of the high-enhancing areas of the lesions as and the following areas of focal change should be carefully avoided: lesion attenuation, necrosis, calcification, and large vessels. During the two phases, using the copy-and-paste function to ensure the size, shape, and position of ROIs were kept consistent all measurements were acquired thrice to calculate the average values. The CT values at different energy levels (40–140 keV, interval 10 keV), iodine (water) and water (iodine) concentrations of the lesions and aorta could be automatically calculated using the GSI viewer software package. The other two parameters were calculated by the following formulas: (a) normalized iodine concentration (NIC) was calculated as: NIC = IC_lesion/IC_aorta, where IC_lesion and IC_aorta are the iodine concentrations of the lesions and aorta, respectively. To minimize variations in patients, iodine concentrations in the lesions must be normalized to those of the aorta, and (b) the slope of spectrum curve was calculated as slope = ( CT_40keV–CT_100keV)/60, where CT_40keV and CT_100keV are the CT attenuation values of the tumors on the 40-keV and 100-keV monochromatic images, respectively.

Statistical analysis

Using SPSS22.0 software (Chicago, IL, USA), the data were analyzed. Quantitative values are expressed as mean ± SD. The two-sample t-test was performed to compare the quantitative parameters of the CT value on 40 keV/70 keV, NIC, water concentration, and the slope of spectrum curve between the ChRCC and PRCC groups, with P < 0.05 indicating significance. Receiver operating characteristic (ROC) curves were generated for the values with significant differences to evaluate their diagnostic efficiency and to calculate the cutoff value, sensitivity, and specificity. The diagnostic capability was determined by calculating the area under the ROC curve (AUC). The best sensitivity and specificity were achieved using the optimal thresholds. Sensitivity was defined as the number of correct diagnoses of ChRCC divided by the number of proven ChRCC cases and multiplied by 100. Specificity was defined as the number of correct diagnoses of PRCC divided by the number of proven PRCC cases and multiplied by 100. The null hypothesis test was that AUC was 0.5; the alternative was AUC > 0.5.

Results

The CT values of ChRCC and PRCC showed a certain regularity at different energy levels (40–140 keV, at intervals of 10 keV). During CP and PP, the CT values of monochromatic images at low-energy levels (40–70 keV) were higher than those at high energy levels (80–140 keV). The CT value of ChRCC was higher than that of PRCC, and significant differences were observed at low energy levels (40–70 keV; Figs. 1 and 2)

Fig. 1.

In the cortex phase, the CT value of ChRCC was higher than that of PRCC, and significant differences were observed at low energy levels (40-70 keV).

Fig. 2.

In the parenchyma phase, the CT value of ChRCC was higher than that of PRCC, and significant differences were observed at low energy levels (40-70 keV).

. Table 1 shows the comparison of values for the defined quantitative parameters (NIC, water concentration, and the slope of spectrum curve) between the patients with ChRCC and PRCC groups. NICs differed significantly between the two groups during both the CP (0.14 ± 0.07 vs. 0.25 ± 0.07, P < 0.05) and PP (0.53 ± 0.19 vs. 0.72 ± 0.14, P < 0.05). The patients with ChRCC had significantly higher slopes than those with PRCC during CP (4.65 ± 1.18 vs. 2.68 ± 1.08, P < 0.05) PP (4.44 ± 1.21 vs. 2.60 ± 0.78). The water (iodine) concentration during CP and PP showed no significant differences between the groups (P > 0.05).

Table 1.

Quantitative assessment of PRCC and ChRCC with CT spectral imaging.

Group	NIC	Water (iodine)	Slope
CP
PRCC	0.14 ± 0.07	1038.39 ± 9.40	2.68 ± 1.08
ChRCC	0.25 ± 0.07	1031.17 ± 7.14	4.65 ± 1.18
t value	–4.17	2.65	–5.27
P value	0.00	0.12	0.00
PP
PRCC	0.53 ± 0.19	1035.81 ± 6.55	2.60 ± 0.78
ChRCC	0.72 ± 0.14	1032.30 ± 6.90	4.44 ± 1.21
t value	–3.37	1.58	–5.40
P value	0.02	0.12	0.00

Values are given as mean ± SD.

ChRCC, chromophobe renal cell carcinoma; CP, cortex phase; CT, computed tomography; PP, parenchyma phase; PRCC, papillary renal cell carcinoma.

Fig. 3 presents the ROC curves of different parameters for differentiating ChRCC from PRCC. All of the ROCs for all parameters were above the reference line. The areas for the slope (0.91) during the PP were greater than other quantitative parameters for differentiating PRCC from ChRCC (Fig. 3). Using the ROCs, we determined the parameter threshold values required to optimize the sensitivity and specificity for differentiating PRCC from ChRCC (Table 2). For example, during PP, a threshold slope of 3.00 would yield a sensitivity and specificity of 76.5% (21/27 PRCCs) and 100% (all 13 ChRCCs), respectively, for differentiating PRCC from ChRCC.

Fig. 3.

Receiver operating characteristic curves for normalized iodine concentration (NIC) and the slope of spectrum curve in differentiating PRCC from ChRCC during the cortex phase and parenchyma phase.

Table 2.

Areas under the ROC curve, thresholds, sensitivities, and specificities for distinguishing PRCC from ChRCC.

Group	NIC (CP)	Slope (CP)	NIC (PP)	Slope (PP)
Areas under the ROC curve	0.83	0.89	0.78	0.91
Thresholds	0.13	2.38	0.41	3.00
Sensitivities	17 (64.7)	16 (58.8)	11 (41.2)	21 (76.5)
Specificities	13 (100)	13 (100)	13 (100)	13 (100)

Values are given as n (%).

ChRCC, chromophobe renal cell carcinoma; CP, cortex phase; CT, computed tomography; NIC, normalized iodine concentration; PP, parenchyma phase; PRCC, papillary renal cell carcinoma; ROC, receiver operating characteristic.

Discussion

The imaging manifestations of various subtypes of renal tumors have been discussed in previous studies (15). Radiogenomics uses an automated high-throughput feature extraction algorithm to obtain internal features of diseases that are invisible to the naked eye from high-quality, standardized medical images and a series of related parameters that quantify the lesions. It provides a basis for tumor typing, clinical treatment options, and prognosis. At present, radiogenomics on renal tumors is still in its infancy, and it still needs to be optimized in image acquisition, reconstruction, software selection, post-classifier algorithm, and statistical processing, particularly for image quality standardization and sample size requirements. Moreover, relevant guidelines are still needed (16,17). Some studies have shown that intratumoral hemorrhage detected using chemical shift magnetic resonance imaging and T2*-weighted T2-weighted gradient echo can be used to differentiate PRCC from fat-poor angiomyolipoma (18). When qualitative imaging findings are visually similar on conventional CT, clinicians try to use quantitative parameters to differentiate PRCC from ChRCC using spectral CT. The conventional CT has produced the averaging attenuation effect of polychromatic X-rays, which reduces the low-contrast spatial resolution between materials (19). CT imaging can obtain monochromatic images ranging from 40- to 140-keV energy levels, which eliminated the need of averaging the attenuation effects (20). The CT values at low energy levels (40–70 keV) are higher than those at high energy levels (80–140 keV). The tissue contrast performed differently, being generally smaller at high–energy levels than at low-energy levels (21). In addition, during CP and PP, the CT values of ChRCC were greater than those of PRCC at low energy levels (40–70 keV). This shows that in both phases, the iodine intake of ChRCC was higher than that of PRCC. PRCC originates in the proximal or distal convoluted tubules. The tumor can often be accompanied by bleeding, necrosis, cystic and fibrous pseudocapsules, most of the composition of the papillary structure, nipple visible fibrous vascular tissue, nipple center with foam macrophages, and hemosiderin (2). ChRCC shows less bleeding and necrosis, with tumor cells that mostly flat or have alveolar arrangements, and prominent features include a clear membrane and rich membranous vacuoles when visualized using electron microscopy (22). These microscopic findings may explain why the CT values of PRCC were lower than those of ChRCC, as the ROIs are inevitably delineated in areas of the cystic or necrotic tissues that are not visible to the naked eye (Figs. 4 and 5).

Fig. 4.

A 76-year-old man with a ChRCC in the left kidney. The iodine-based material decomposition images clearly showed a small circular mass in the left kidney under extremely during the CP (a) and PP (b). The spectrum curve of the lesion in CP (c) and PP (d). H&E staining (×200). The tumor cells were ranged along the blood vessels. Two types of cell were visible: one was a large polygon, with the transparent pulp and clear capsule; the other cell was small, cytoplasmic eosinophilic, clear enveloped, irregular nuclei, small nucleoli and with a perirenal halo (e). CP, cortex phase; PP, parenchyma phase.

Fig. 5.

A 53-year-old man with a papillary renal cell carcinoma in the right kidney. The iodine-based material decomposition images revealed an irregular mass with the necrotic area in the right kidney during the CP (a) and PP (b). The spectrum curve of the lesion in CP (c) and PP (d). H&E staining (×200). The tumor cells had more lobulated complex nipple, the central nipple was the fiber bundle, the tumor cell was cytoplasmic acidosis, the nuclear ratio increased, and nuclear membrane was clear (e). CP, cortex phase; PP, parenchyma phase.

In terms of clinical application, the basic pair for material decomposition image presentation is water and iodine, because their atomic numbers are within the range of materials generally found in medical imaging, such as approximate soft tissue and iodinated contrast materials. This results in material attenuation images that are intuitive to interpret (23 –26). According to our study, NICs varied significantly between PRCC and ChRCC during the two phases, with ChRCC showing higher values than PRCC. The iodine-based material decomposition images can be sensitive because of focal uptake of the iodinated contrast material. The iodine concentration in lesions might be an effective quantitative parameter to reflect the tumor blood supply. According to some previous studies, the blood supply types of the tumor is related to the pathological grade of the tumor (27,28). Some of the poorly differentiated RCCs contained a small number of arteries, so the iodine concentrations reflecting the blood supply were relatively low, and there was no blood supply to the necrotic area of the tumor, so its iodine concentration was lower (29 –31). The water concentration between ChRCC and PRCC showed no significant differences because water content is similar in both tumors.

The slope of the spectrum curve exhibited significant differences between PRCC and ChRCC during CP and PP. The spectrum curve reflects the material CT value varying with the energy of the X-ray and the absorption characteristics to the different energy of the X-ray. Various substances exhibit changes in chemical molecular structures, and different chemical molecules have modified energy attenuation curves (32). Thus, we can distinguish the chemical composition of substances by comparing the slope of the spectrum curve. The slope of the ChRCC group was higher than that of the PRCC group, probably because the blood supply of ChRCC is richer than that of PRCC.

In the present study, the ROC curve analysis revealed that the NIC and slope had higher specificities for differentiating PRCC from ChRCC. The best quantitative parameter was the slope in PP, and a threshold of 3.00 would yield a sensitivity and specificity of 76.5% for PRCC and 100.0% for ChRCC. Quantitative image analysis with spectral CT has a better sensitivity and specificity than conventional qualitative image analysis.

However, our study had some limitations. First, the sample size was small, so a large sample study is necessary to confirm the accuracy of the threshold values. Second, this study was focused on ChRCC and PRCC, therefore, more cases of different types of renal tumors are needed for future investigations.

In conclusion, the spectral CT parameters of monochromatic CT values, the quantitative iodine concentration, and the slope of spectrum curves can help increase the accuracy of differentiating ChRCC from PRCC.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by the National Natural Science Foundation of China (No. 81772006) Project of ‘Science and Technology Innovation of Cuiying’ in Lanzhou University Second Hospital (No. CY2018-QN07).

ORCID iD

Xiaoyu Huang

References

Lane

Kattan

MW.

Predicting outcomes in renal cell carcinoma. Curr Opin Urol 2005; 15:289–297.

Prasad

Humphrey

Catena

, et al. Common and uncommon histologic subtypes of renal cell carcinoma: imaging spectrum with pathologic correlation. Radiographics 2006; 26:1795–1806.

Bergström

Terry

Lindblad

, et al. Physical activity and risk of renal cell cancer. Int J Cancer 2001; 92:155–157.

Mucci

Hjelmborg

Harris

, et al. Familial risk and heritability of cancer among twins in Nordic countries. JAMA 2016; 315:68–76.

Moch

Cubilla

Humphrey

, et al. The 2016 WHO Classification of Tumours of the Urinary System and Male Genital Organs-Part A: Renal, Penile, and Testicular Tumours. Eur Urol 2016; 70:93–105.

Tsuda

Kinouchi

Tanikawa

, et al. Imaging characteristics of papillary renal cell carcinoma by computed tomography scan and magnetic resonance imaging. Int J Urol 2010; 12:795–800.

Srigley

Delahunt

Uncommon and recently described renal carcinomas.

Mod Pathol 2009; 22:S2–S23.

Low

Huang

, et al. Review of renal cell carcinoma and its common subtypes in radiology. World J Radiol 2016; 8:484–500.

Sheir

El-Azab

Mosbah

, et al. Differentiation of renal cell carcinoma subtypes by multislice computerized tomography. J Urology 2005; 174:451–455.

10.

Baldinger

Mehrazin

Tomaszewski

, et al. Renal cell carcinoma: management in the elderly. Curr Geri Rep 2014; 3:19–25.

11.

Kondo

Nakazawa

Sakai

, et al. Spoke-wheel-like enhancement as an important imaging finding of chromophobe cell renal carcinoma: a retrospective analysis on computed tomography and magnetic resonance imaging studies. Int J Urol 2010; 11:817–824.

12.

Wildberger JE, Adam G, Boeckmann

, et al. Computed tomography characterization of renal cell tumors in correlation with histopathology. Invest Radiol 1997; 32:596–598.

13.

Sun

, et al. Differentiating hepatocellular carcinoma from angiomyolipoma of the liver with CT spectral imaging: A preliminary study. Clin Radiol 2013; 68:e491–e497.

14.

Kang

Chandarana

Contemporary imaging of the renal mass. Curr Urol Rep 2011; 12:11–17.

15.

Woo

Kim

Cho

, et al. Differentiation between papillary renal cell carcinoma and fat-poor angiomyolipoma: a preliminary study assessing detection of intratumoral hemorrhage with chemical shift MRI and T2*-weighted gradient echo. Acta Radiol 2018; 59:627–634.

16.

Lambin

Leijenaar

RTH

Deist

, et al. Radiomics: the bridge between medical imaging and personalized medicine. Nat Rev Clin Oncol 2017; 14:749–762.

17.

Liu

Zhang

, et al. Preoperative prediction of muscular invasiveness of bladder cancer with radiomic features on conventional MRI and its high-order derivative maps. Abdom Radiol 2017; 42:1896–1905.

18.

Yin

, et al. Usefulness of dual-energy computed tomography imaging in the differential diagnosis of sellar meningiomas and pituitary adenomas: preliminary report. PLoS One 2014; 9:e90658.

19.

Lin

, et al. Differentiation of small hepatic hemangioma from small hepatocellular carcinoma: recently introduced spectral CT method. Radiology 2011; 259:720–729.

20.

Lin

Tao

, et al. Dual energy spectral CT imaging of insulinoma—Value in preoperative diagnosis compared with conventional multi-detector CT. Eur J Radiol 2012; 81:2487–2494.

21.

Lin

Chen

, et al. Hepatocellular carcinoma and focal nodular hyperplasia of the liver: differentiation with CT spectral imaging. Eur Radiol 2013; 23:1660–1668.

22.

Leroy

Moukassa

Copin

, et al. Utility of Cytokeratin 7 Distinguishing Chromophobe Renal Cell Carcinoma from Renal Oncocytoma. Eur Urol 2000; 37:484–487.

23.

Geyer

Scherr

Körner

, et al. Imaging of acute pulmonary embolism using a dual energy CT system with rapid kVp switching: Initial results. Eur J Radiol 2012; 81:3711–3718.

24.

Graser

Johnson

TRC

Chandarana

, et al. Dual energy CT: preliminary observations and potential clinical applications in the abdomen. Eur Radiol 2009; 19:13–23.

25.

Lin

Chen

, et al. Spectral CT in patients with small HCC: investigation of image quality and diagnostic accuracy. Eur Radiol 2012; 22:2117–2124.

26.

Liu

Yan

Pan

, et al. Evaluation of dual energy spectral CT in differentiating metastatic from non-metastatic lymph nodes in rectal cancer: Initial experience. Eur J Radiol 2015; 84:228–234.

27.

Kutikov

Smaldone

Egleston

, et al. Anatomic features of enhancing renal masses predict malignant and high-grade pathology: a preoperative nomogram using the Renal Nephrometry Score. Eur Urol 2011; 60:241–248.

28.

Jin Hong

Peng Qiu

Pei Jun

, et al. Dynamic CT Evaluation of Tumor Vascularity in Renal Cell Carcinoma. AJR Am J Roentgenol 2006; 186:1423–1430.

29.

Lam

Shvarts

Said

, et al. Clinicopathologic and molecular correlations of necrosis in the primary tumor of patients with renal cell carcinoma. Cancer 2005; 103:2517–2525.

30.

Lee

Byun

, et al. Significance of Macroscopic Tumor Necrosis as a Prognostic Indicator for Renal Cell Carcinoma. J Urology 2006; 176:1332–1338.

31.

Jinyan

Jianhong

Xueling

, et al. Analysis of dual energy spectral CT and pathological grading of clear cell renal cell carcinoma (ccRCC). PLoS One 2018; 13:e0195699-.

32.

Karçaaltıncaba

Aktaş

Dual-energy CT revisited with multidetector CT: review of principles and clinical applications. Diagn Interv Radiol 2011; 17:181.