quantitative enhancement washout analysis of solid cortical renal masses using multidetector...

Quantitative Enhancement Washout Analysis of Solid CorticalRenal Masses Using Multidetector Computed TomographyHaytham M. Shebel, MD,* Khaled M. Elsayes, MD,Þ Khaled Zaky Sheir, MD,þ Heba M. Abou El Atta, MSc,*

Ahmad F. El-Sherbiny, MD,* James H. Ellis, MD,§ and Tarek A. El-Diasty, MD*

Purpose: The aim of the study was to prospectively assess the utilityof quantitative enhancement washout method in the differentiation ofbenign solid renal masses from various subtypes of malignant massesusing multidetector computed tomography.Methods: In a prospective investigation from January 2009 toMay 2010,97 patients with solid renal masses underwent CT scan examination withunenhanced, arterial, parenchymal, and delayed phases. The followingfeatures were analyzed: the maximum attenuation value in each phase,attenuation difference (enhancement) of the mass in each phase from theunenhanced phase ($H), and parenchymal and delayed phases’ washout.Of these patients, 82 (85%) underwent unilateral radical nephrectomy, 15(15%) underwent partial nephrectomy. Group comparison was performedwith the Kruskal-Wallis test and Mann-Whitney U test.Results: The masses included in our study were 45 clear cell renalcell carcinomas (CCRCCs); 18 chromophobe renal cell carcinomas,16 papillary (PRCC), 14 oncocytomas, and 4 minimal fat containingangiomyolipomas. In the arterial phase, the CCRCC was the most en-hancing type and could be differentiated from other renal masses (benignor malignant) with high sensitivity and specificity. In the parenchymalphase, the CCRCCs demonstrated the highest washout. Chromophoberenal cell carcinomas showed the second highest washout in this phase.Benign lesions and PRCCs did not exhibit significant washout inthis phase.

In the delayed phase, the malignant lesions (with the exception ofPRCCs) showed the highest washout. Benign lesions showed significantwashout but less than malignant lesions.Conclusions: Multiphasic multidetector CT utilizing arterial-phaseattenuation and quantitative enhancement washout method could help inthe preoperative differentiation of various types of solid renal masses.

Key Words: CT, solid, renal masses

(J Comput Assist Tomogr 2011;35: 337Y342)

Various solid renal masses can arise from the kidney; thesemasses can be either benign, such as oncocytoma and

angiomyolipoma (AML), or malignant such as renal cell carci-noma (RCC), which is the most common solid renal mass. Thebenign renal masses could represent a daily challenge. Thischallenge arises from the difficulty of differentiating them fromtheir malignant counterparts, particularly when these benign

masses manifest in atypical pattern as minimal fat-containingAML.1,2

The major histological subtypes of RCCs include clear cell(CCRCC), papillary cell (PRCC) (types I and II), chromophobe(ChRCC), collecting duct RCCs, medullary carcinomas, andsarcomatoid RCCs as well as unclassified variants.3,4 However,to the best of our knowledge, there are no well-establishedimaging criteria to classify the histological subtypes of thesevarious lesions. In addition, it is sometimes difficult to obtaina definitive diagnosis and distinguish between benign andmalignant masses on the basis of the biopsy findings (eg, it isoften difficult to differentiate oncocytoma from chromophobecarcinoma).5

Thus, the purpose of this prospective study was to analyzehow the various types of benign and malignant renal massesbehave in regard to the injected contrast media in different phasesof multiphasic multidetector computed tomography (MDCT)and to assess the utility of quantitative enhancement washoutanalysis in the differentiation between malignant and benignsolid renal masses.

MATERIALS AND METHODS

Patient PopulationInstitutional review board approval was obtained for the

review of subjects’ medical records, and informed consent wasobtained.

Our study population included all patients who underwentimaging by CT for solid-appearing renal masses diagnosed byultrasonography before surgical resection.

From January 2009 until May 2010, 97 patients were eli-gible for the study. This involved 76 males (78%) and 21 females(22%) from ages 23 to 67 years, with a mean of 49 years. Ofthese patients, 82 (85%) underwent unilateral total nephrectomy,and 15 patients (15%) underwent partial nephrectomy.

CT ExaminationMultiphasic CT examinations were performed using a mul-

tidetector CT scanner with 64 parallel detector rows (PhilipsBrilliance CT; PhilipsMedical Systems Nederland, Veenpluis 4-6,the Netherlands) and postprocessing by a (Brilliance; Philips)workstation V3.01.5000. Intravenous contrast material, Ultra-vist300 (Schering, Berlin, Germany) or Iopamiro300 (Bracco,Milan, Italy), was injected into an antecubital vein by using amechanical injector. The dose was 120 mL at a rate of 4.0 mL/s.The total duration of injection was 30 seconds.

Patients underwent multiphasic CT scanning that includedunenhanced, arterial, parenchymal, and delayed phases throughthe kidneys. The abdomen from diaphragm to the symphysispubis was scanned during unenhanced and parenchymal phasesonly, whereas only both kidneys were scanned in other phasesof the study.

ORIGINAL ARTICLE

J Comput Assist Tomogr & Volume 35, Number 3, May/June 2011 www.jcat.org 337

From the *Department of Radiology, Urology and Nephrology Center,Mansoura University, Mansoura, Egypt; †Department of Radiology, Uni-versity of Texas MDAnderson Cancer Center, Houston, TX; ‡Department ofUrology, Urology and Nephrology Center, Mansoura University, Mansoura,Egypt; and §Department of Radiology, University of Michigan Health Sys-tem, Ann Arbor, MI.Received for publication December 14, 2010; accepted March 11, 2011.Reprints: Khaled M. Elsayes, MD, Department of Radiology, University

of Michigan Health System, Ann Arbor, MI 48109-0030(e-mail: [email protected]).

The authors have nothing to disclose related to this study.Copyright * 2011 by Lippincott Williams & Wilkins

Copyright © 2011 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

The scanning parameters for imaging acquisition duringall phases included slice thickness = 2.5 mm; reconstructed slicethickness = 5 mm; pitch = 0.984; rotation time = 0.75 seconds;and 220 mA. The scan delay was 25 to 30 seconds for arterialphase, 80 seconds for parenchymal phase, and 300 seconds fordelayed phase.

CT Image AnalysisImages were reviewed in consensus by 2 radiologists, each

with at least 8 years of experience in abdominal/pelvic CTimaging obtained at an academic health care facility. Imageinterpretation was performed at a special workstation (Advan-tage Window 4.1; Philips). Region-of-interest (ROI) measure-ments were selected to be fully contained within the tumorconfinement. Care was taken to avoid the periphery of the lesionbecause of the risk of partial volume averaging. We measuredthe attenuation values (in Hounsfield units [HUs]) of the solidpart of the tumor in each phase including the unenhanced scan.Multiple ROIs were applied at different parts of the tumor toselect the highest 3 numeric representative attenuation values,and then we calculated the average value of the readings. Fortumors with varying densities, we selected the solid componentof the tumor and avoided areas of cystic degeneration, calcifi-cation, and the prominent vessels around and inside the tumor.The average area of ROI was 13 mm in diameter.

Data AnalysisFor each lesion, 3 features were analyzed: (a) the maximum

attenuation value (MAV) in HUs for each phase, which was

obtained by using the mean of the highest 3 representativeattenuation values within the mass; (b) the enhancement changesbefore and after tissue enhancement ($H), which was calcu-lated as follows: ($H = MAV with tissue enhancement in eachphasejMAV in the unenhanced scan); and (c) the absolute andrelative washout ratio in parenchymal and delayed phases byusing the following formula:

Absolute washout % in parenchymal or delayed phase =(attenuation value at enhanced (arterial) CT j attenuationvalue at (parenchymal or delayed) enhanced CT] / [attenuationvalue at (arterial) enhanced CT j attenuation value at unen-hanced CT]) � 100.

The relative percentage of enhancement washout was alsocalculated for all renal masses with the following equation:

Relative percentage of enhancement washout = ([attenua-tion value at (arterial) enhanced CT j attenuation value at(parenchymal or delayed) enhanced CT] / attenuation value at(arterial) enhanced CT) �100.

The 3 featured values were correlated with the final diag-nosis after histopathology results in all cases.

Statistical AnalysisData were analyzed using the SPSS software (version 15;

SPSS Inc, Chicago, Ill). Statistical comparisons were performedfor various CT findings of different renal mass types. Groupcomparison was completed with the Kruskal-Wallis test for CTattenuation values at different phases. When the test was statis-tically significant, pairwise comparisons were made with theMann-WhitneyU test. Test results were considered significant at

TABLE 1. Attenuation Values of Renal Masses

Arterial Phase Parenchymal Phase Excretory Phase

MAV $H MAV $H MAV $H

Type of malignant massesCCRCC 99Y222; M = 156 69Y199; M = 120 54Y164; M = 115 45Y126; M = 79 45Y95; M = 77 21Y69; M = 41ChRCC 77Y142; M = 112 57Y96; M = 71 69Y117; M = 94 37Y85; M = 53 39Y85; M = 73 13Y55; M = 32PRCC 35Y83; M = 42 9Y42; M = 18 42Y77; M = 55 20Y60; M = 31 43Y81; M = 61 17Y60; M = 36

Type of benign massesOncocytoma 70Y118; M = 94 40Y112; M = 69 92Y158; M = 119 50Y108; M = 94 72Y90; M = 80 27Y83; M = 55AML 60Y83; M = 79 44Y46; M = 45 64Y85; M = 78 45Y56; M = 45 15Y70; M = 62 16Y39; M = 29

M indicates mean of range of attenuation values.

FIGURE 1. Axial CT images of a 65-year-old man with CCRCC. A, Unenhanced (A), (B) arterial, (C) parenchymal, and (D) delayedphases. Computed tomographic scans show solid left renal mass arising from the mid zone of the small atrophic left kidney withattenuation values 41, 183, 147, and 83 HUs, respectively.

Shebel et al J Comput Assist Tomogr & Volume 35, Number 3, May/June 2011

338 www.jcat.org * 2011 Lippincott Williams & Wilkins


FIGURE 2. Axial CT images of a 50-year-old woman with ChRCC. A, Unenhanced, (B) arterial, (C) parenchymal, and (D) delayedphases. Computed tomographic scans show left renal mass arising from its lower pole with area of cystic degeneration; attenuationvalues 30, 98, 69, and 39 HUs, respectively.

FIGURE 3. Axial CT images of a 44-year-old woman with PRCC. A, Unenhanced, (B) arterial, (C) parenchymal, and (D) delayed phases.Computed tomographic scans show right renal mass with attenuation values 31, 35, 42, and 43 HUs, respectively.

FIGURE 4. Axial CT images of a 52-year-old manwith oncocytoma. A, Unenhanced, (B) arterial, (C) parenchymal, and (D) delayed phases.Computed tomographic scans show left renal mass with attenuation values 36, 118, 148, and 73 HUs, respectively.

J Comput Assist Tomogr & Volume 35, Number 3, May/June 2011 Solid Cortical Renal Masses Using Multidetector CT

* 2011 Lippincott Williams & Wilkins www.jcat.org 339


P e 0.05. To evaluate the diagnostic validity of the attenuationvalue in different types of renal masses (benign or malignant),we tried to calculate different cutoff points of highest sensitivity,specificity, and predictive values.

RESULTSOf the 97 lesions, 79 (81.4%) were malignant, and 18

(18.5%) were benign. Of these, 79 malignant lesions, 45 (57%)were CCRCCs; 18 (23%) were ChRCCs; and 16 (20%) werePRCCs. The 18 benign lesions included were 14 renal oncocy-tomas (78%) and 4 minimal fat-containing AMLs (22%). Thetumor size range for CCRCC was 3 to 8 cm in diameter and forChRCC was 2 to 6 cm, whereas for PRCC it was 4 to 8 cm. Incase of oncocytoma, the tumor size range was 2 to 5 cm; inAMLs, it was 2 to 3 cm. Eighty-two (85%) underwent unilateraltotal nephrectomy, and 15 patients (15%) underwent partialnephrectomy.

Table 1 shows the MAV and $H for malignant and benignrenal masses at all CT phases. Figures 1 to 5 show the MDCTimages of different types of renal masses at different phases.

Differentiation of Malignant From Benign LesionsIn the arterial phase, the CCRCCs and ChRCCs were the

most enhancing tumors. The PRCCs enhanced poorly about15% of the CCRCCs and 25% of ChRCCs (P G 0.05). Benignlesions had a lower enhancement than CCRCCs and ChRCCs:about 42% of CCRCCs (P = 0.001) and 70% of ChRCCs(P = 0.01) but higher than PRCCs (P = 0.001). Cutoff pointswere 90 and 60 HUs as values of MAV and $H in the arterialphase for differentiation between CCRCCs and ChRCC as 1group and PRCCs and benign lesions as a second group. Theyhad a sensitivity and specificity of 98% and 90%, and 92% and90%, respectively.

There were no definite cutoff values with high sensitivityand specificity in parenchymal or delayed phases for MAVor $Hto differentiate between benign and malignant masses. So thewashout analysis method was used to differentiate between thedifferent types of the renal masses (Figs. 6 and 7).

Parenchymal (Venous) WashoutClear cell RCC and ChRCC washout was significant in this

phase by 34% and 23%, respectively. Both benign tumors andPRCCs did not wash out significantly in this phase (Fig. 6).

FIGURE 7. All benign and malignant lesions except papillaryRCC show significant washout in the excretory phase.

FIGURE 6. Papillary RCC and benign lesions did not wash outthe contrast media in the parenchymal phase.

FIGURE 5. Axial CT images of a 43-year-old woman with AML. A, Unenhanced, (B) arterial, (C) parenchymal, and (D) delayed phases.Computed tomographic scans show left renal mass with attenuation values 11, 57, 64, and 14 HUs, respectively.




Cutoff points of 12% and 4%, as value of absolute and rel-ative washout, respectively, were found to have a sensitivity andspecificity of 84% and 100%, and 94% and 96%, respectively, indifferentiation between CCRCCs and ChRCCs from PRCCs andbenign lesions.

Delayed PhaseClear cell RCC and ChRCC washout was significant in this

phase by 44% and 40%, respectively. Papillary RCCs did notwash out as in venous phase. While in this phase, benign lesionwashout was significant by 38%. (Fig. 7).

Cutoff points of 14% and 30% as values of absolute andrelative washout, respectively, have a sensitivity and specificityof 100% and 97%, and 90% and 100%, respectively, in dis-crimination between CCRCCs and ChRCCs from PRCCs andbenign lesions.

For differentiation between benign lesions from PRCC inthe delayed phase, a cutoff point of 7% as value of absolute orrelative washout has a sensitivity and specificity of 93% and100%, respectively. The positive and negative predictive valuesof each cutoff point are listed in Table 2.

DISCUSSIONAmong solid cortical renal tumors, approximately 20%

are benign lesions.2 This relatively high percentage needs anapproach for accurate diagnosis to distinguish them from themore common malignant tumors, which constitute nearly 80%of renal cortical tumors and have a different biological behaviorand different prognosis.6Y8

Clear cell RCCs have the greatest metastatic potential andaccount for approximately 65% of renal cortical malignancies.Papillary RCC and ChRCC subtypes account for approximately25% of renal cortical tumors and are associated with less met-astatic potential.2,5Y7 The overall 5-year survival rates forpatients with PRCCs and ChRCCs (80%Y90%) are higher thanthose for patients with CCRCCs (50%Y60%).8,9 Because clinicalimplications and therapeutic strategies may differ for differentsubtypes of renal cortical tumors, preoperative identification ofthe subtype would be of significant clinical interest.10 However,the most important issue is to differentiate them from the solidbenign lesions (oncocytomas and minimal fat-containing AML),because in many clinical situations, preoperative differentiationis critical for patient survival and quality of life as in the case ofbilateral solid renal lesions, hereditary renal cancer syndromes,and mass in solitary functioning kidney. However, to the best ofour knowledge, there are no definite imaging criteria to differ-entiate between the 2 groups. In the last decade, many authorsinvestigated RCC to differentiate between its subtypes usingcertain imaging features.2,6,8Y12 Sheir et al7 and Kim et al11

used the morphological criteria of tumor size, calcification, andcystic degeneration. They concluded that these criteria have aminor role in the differentiation between tumor subtypes, themost reproducible findings in differentiation between RCCsubtypes were the degree of enhancement, as CCRCCs enhanceto a greater degree than other subtypes according to Zhang et al2

and Herts et al.13 Although most of these studies included onlymalignant lesions, or in some cases, subgroups of malignantlesions in their analysis, few studies included both malignant andbenign lesions per Zhang et al2 and Jinzaki et al.8 The techniqueused in these studies depended mainly on 2 postcontrast se-quences. Herts et al13 used the arterial and parenchymal phases.Sheir et al7 and Kim et al11 used the corticomedullary (arterial)and excretory phases, whereas Zhang et al2 and Ruppert-Kohlmayr et al14 used the parenchymal and excretory (delayed)phases.

In our study, we used 3 postcontrast phases, arterial, paren-chymal, and delayed phases, in addition to unenhanced scan.We analyzed only the enhancement and washout pattern ex-cluding the morphological analysis. In addition, we obtainedquantitative attenuation measurements for all renal lesions andtried to assess cutoff points with high sensitivity, specificity, andpredictive values.

Our results showed that CCRCCs and ChRCCs have thegreatest degree of enhancement in arterial phase more than theother lesions (PRCC and benign masses). In the arterial phase ofMDCT, if MAV is 90 HUs or greater or $H is 60 HUs or greater,it was suggestive of CCRCC or ChRCCwith high sensitivity andspecificity and greater than 95% positive predictive value. Othermasses with lower figures of MAV or $H will be suggestive ofPRCCs and benign lesions.

In parenchymal and delayed phases, when the analysiswashout method was used, we have a high cutoff point with highsensitivity and specificity reaching up to 100% positive predic-tive value in differentiation between the 2 groups (CCRCCs andChRCCs) from PRCCs and benign lesions. This is because ofthe characteristic feature of PRCCs and benign lesions, whichhold the contrast and do not show significant washout in theparenchymal phase. This pattern of enhancement indicates theimportance of the combination of arterial and parenchymalpostcontrast phases in examination protocol of solid renalmasses.

However, the differentiation between PRCCs and benignlesions is still in need for further differentiation, so we analyzedthem in the delayed phase. In this phase, there is a significantpoint of differentiation between PRCCs and benign lesionswhere PRCCs did not show a washout. In most of cases, delayedenhancement occurred with higher attenuation values than theprevious phases. This is a unique pattern of enhancement notedin PRCCs, whereas the benign lesions had a significant washout

TABLE 2. Predictive Values for Cut-Off Points

Phase PPV, % NPV, %

Arterial MAV for CCRCC and ChRCC (90 HUs) 95 96Arterial $H for CCRCC and ChRCC (60 HUs) 95 84Relative parenchymal washout for CCRCC and ChRCC (4%) 95 90Absolute parenchymal washout for CCRCC and ChRCC (12%) 100 75Relative delayed washout for CCRCC and ChRCC (30%) 100 81Absolute delayed washout for CCRCC and ChRCC (14%) 78 100Absolute or relative delayed washout for benign lesions (7%) 100% 94%

PPV indicates positive predictive value; NPV, negative predictive value.

J Comput Assist Tomogr & Volume 35, Number 3, May/June 2011 Solid Cortical Renal Masses Using Multidetector CT

* 2011 Lippincott Williams & Wilkins www.jcat.org 341


(P e 0.05). When 7% was used as a cutoff point for differenti-ation benign from papillary lesions, we have high sensitivity of93% and 100% specificity and positive predictive value. Thus,using delayed scan in the examination protocol of solid renalmasses is essential for diagnosing benign renal masses to cal-culate the washout of the contrast in comparison to previousparenchymal phase. So, according to our results, the combina-tion of 3 postcontrast scans is essential for discrimination ofbenign from malignant lesions.

Our study had limitations. First, the number of benignlesions in comparison to malignant lesions was slightly differentfrom the general population. Second, the tumor staging systemfor the renal masses was not assessed.

CONCLUSIONSOn the basis of our findings, we conclude that the multi-

phasicMDCTusing absolute HUs of the enhancement, as well asthe washout of contrast media in different phases of MDCTstudy, can differentiate benign renal lesions from malignantlesions and their subtypes. This differentiation can help in thepreoperative identification of the renal masses and consequentlyinfluence management strategy and surgical approach.

REFERENCES

1. Oyen RH, Verswijvel G, Vanpopel H, et al. Primary malignant renalparenchymal epithelial neoplasm. Eur Radiol. 2000;231:243Y210.

2. Zhang J, Lefkowitz RA, Ishill NM, et al. Solid renal cortical tumors:differentiation with CT. Radiology. 2007;244:494Y504.

3. Patard JJ, Leray E, Rioux-Leclercq N, et al. Prognostic value ofhistologic subtypes in renal cell carcinoma: a multicenter experience.J Clin Oncol. 2005;2763:2771.

4. Storkel S, van den Berg E. Morphological classification of renal cancer.World J Urol. 1995;13:153Y158.

5. Lechevallier E, Andre M, Barriol D, et al. Fine-needle percutaneousbiopsy of renal masses with helical CT guidance. Radiology.2000;216:506Y510.

6. Yamada T, Endo M, Tsuboi M, et al. Differentiation of pathologicsubtypes of papillary renal cell carcinoma on CT. AJR Am J Roentgenol.2008;191:1559Y1563.

7. Sheir KZ, El-Azab M, Mosbah A, et al. Differentiation of renalcell carcinoma subtypes by multislice computerized tomography. J Urol.2005;174:451Y455.

8. Jinzaki M, Tanimoto A, Mukai M, et al. Double-phase helical CT ofsmall renal parenchymal neoplasms: correlation with pathologicfindings and tumor angiogenesis. J Comput Assist Tomogr.2000;24:835Y842.

9. Bosniak MA. Problems in the radiologic diagnosis of renal parenchymaltumors. Urol Clin North Am. 1993;20:217Y230.

10. Kassouf W, Aprikian AG, Laplante M, et al. Natural history of renalmasses followed expectantly. J Urol. 2004;171:111Y113.

11. Kim JK, Kim TK, Ahn HJ, et al. Differentiation of subtypes of renalcell carcinoma on helical CT scans. AJR Am J Roentgenol. 2002;178:1499Y1506.

12. Dechet CB, Zincke H, Sebo TJ, et al. Prospective analysis ofcomputerized tomography and needle biopsy with permanent sectioningto determine the nature of solid renal masses in adults. J Urol.2003;169:71Y74.

13. Herts BR, Coll DM, Novick AC, et al. Enhancement characteristicsof papillary renal neoplasm revealed on triphasic helical CT ofthe kidneys. AJR Am J Roentgenol. 2002;178:367Y372.

14. Ruppert-Kohlmayr AJ, Uggowitzer M, Meissnitzer T, et al.Differentiation of renal clear cell carcinoma and renal papillarycarcinoma using quantitative CT enhancement parameters. AJR AmJ Roentgenol. 2004;183:1387Y1391.




quantitative enhancement washout analysis of solid cortical renal masses using multidetector...

Documents