Direct Comparison of Diagnostic Accuracy of Fast Kilovoltage Switching Dual-Energy Computed Tomography and Magnetic Resonance Imaging for Detection of Enhancement in Renal Masses : Journal of Computer Assisted Tomography

Secondary Logo

Journal Logo

Abdominopelvic Imaging: Genitourinary

Direct Comparison of Diagnostic Accuracy of Fast Kilovoltage Switching Dual-Energy Computed Tomography and Magnetic Resonance Imaging for Detection of Enhancement in Renal Masses

McGrath, Trevor A. MD; Ahmad, Faraz MD; Sathiadoss, Paul MD; Haroon, Mohammad MD; McInnes, Matthew DF MD, PhD†,‡; Bossuyt, Patrick MM PhD§; Schieda, Nicola MD, FRCPC∗,‡

Author Information
Journal of Computer Assisted Tomography: 11/12 2022 - Volume 46 - Issue 6 - p 862-870
doi: 10.1097/RCT.0000000000001361

Abstract

Renal masses are often encountered as incidental findings when patients undergo imaging examinations for other indications.1,2 When an incidentally discovered renal mass cannot be diagnosed as a benign cyst with certainty at time of detection, further characterization with a renal protocol computed tomography (CT) or magnetic resonance imaging (MRI) is typically performed.3–5 The fundamental imaging criterion that differentiates a benign cyst from a cystic or solid renal mass is the presence of contrast enhancement. The Bosniak classification version 2019 defines enhancement as an unequivocal increase in CT attenuation or MRI signal intensity (SI) after intravenous (IV) contrast administration that is perceived visually or when there is a quantitative change that meets one of the following thresholds: a 20-HU or greater increase in attenuation within the mass comparing CT images obtained before and after iodinated IV contrast administration or, with MRI, a 15% or greater SI increase on post–gadolinium-enhanced T1-weighted (T1W) images compared with precontrast images, when both acquisitions are obtained with identical acquisition parameters.1,2,6

Detection of enhancement can be challenging in low-enhancing renal masses, such as papillary renal cell carcinoma (RCC).7,8 In one series, up to 50% of papillary RCC did not enhance by 20 HU or greater at corticomedullary (CM) phase contrast-enhanced CT8 and nearly 20% of papillary RCC did not enhance by 20 HU or greater with dedicated nephrographic (NG) phase contrast-enhanced CT.7,8 In these instances, it has been shown that enhancement in papillary RCC may be better depicted with contrast-enhanced MRI (CE-MRI) using subjective or quantitative evaluation.7,8 A limitation of MRI when evaluating for the presence of enhancement is the need for adequately coregistered data sets without motion, which is not possible in some patients,9–11 such as those with inability to breath hold. This is particularly important in masses, which show intrinsically increased SI on precontrast T1W images, a common finding in papillary RCC, where adequately registered subtraction images (contrast-enhanced T1W – precontrast T1W) are critical to detect enhancement.2,9 Investigators have explored using prospective and retrospective free-breathing techniques to improve image subtraction in renal MRI; however, results remain preliminary and lack validation.12,13

Dual-energy CT (DECT) is a technique that has several genitourinary applications, including evaluation of renal masses. Dual-energy CT has been shown to be as accurate as conventional CT for renal mass diagnosis,14 with several additional potential advantages, including better correction of beam-hardening artifacts (which may mitigate pseudoenhancement in benign cysts15), the ability to generate iodine image sets (which may improve visual detection of enhancement16,17), and a hypothesized increased sensitivity to iodine (which may improve detection of enhancement in low-enhancing renal masses such as papillary RCC, compared with conventional CT18). Dual-energy CT iodine concentration has shown promise in differentiating clear cell and papillary RCC.19 Another advantage of DECT, compared with MRI, relates to the speed with which DECT images are acquired using modern DECT systems, which substantially reduces or eliminates motion artifacts across multiphase protocols.20

We hypothesized that DECT, given previously described benefits in renal mass imaging, may be equal to or possibly better for detection of enhancement in renal masses when compared with MRI. The purpose of this study was therefore to compare subjective and quantitative diagnostic accuracy of DECT and MRI for detection of enhancement in renal masses.

MATERIALS AND METHODS

Ethics approval for this study was obtained from our institutional research ethics board (OHSN-REB Protocol # 20150932-01H). Study protocol is available from the authors. This study report was prepared in accordance with the Standards for Reporting Diagnostic accuracy studies 2015 checklist.21,22 Because this is an exploratory study, we used a convenience sample without a priori sample size calculation.

Study Design and Participants

We retrospectively searched our institutional picture archiving and communications system (PACS) for consecutive patients who underwent abdominal DECT between December 1, 2015, and June 1, 2020. Patients were eligible for inclusion if they were adults, had a renal lesion of 10 mm or greater present on renal DECT protocol, underwent CE-MRI within 12 months of the DECT examination, and had received an acceptable reference standard (details hereinafter). Dual-energy CT is the standard of care at our institution for characterization of renal masses, and patients do not routinely undergo both DECT and MRI in short interval. Within this convenience sample, patients may have received both DECT and MRI for reasons, such as radiologist recommendation or clinician preference.

A maximum of 4 masses or cysts were eligible for inclusion per patient, maximum 2 per kidney. A size cutoff of 10 mm was selected based on the American College of Radiology criteria, which advise “many of the renal masses on enhanced CT that are too small to characterise (TSTC) are either benign or clinically insignificant. Most of these TSTC masses are visibly much lower than the enhanced renal parenchyma and no further imaging is needed; if not, then MRI or CT without and with IV contrast is suggested within 6 to 12 months.”3

Patients were excluded if they were found to have a renal mass containing macroscopic fat, defined as area within the mass, which visually appears as bulk fat quantitatively measuring less than −10 HU, without calcification, which is diagnostic of renal angiomyolipoma with rare exception.3 A radiology resident (T.A.M.) with 4 years of experience in abdominal CT and MRI reviewed original reports and examinations for possible exclusions, which were verified by an abdominal radiologist with 15 years of experience in abdominal CT and MRI (N.S.).

For the reference standard, histological confirmation (through surgery or 18-Gauge core-needle biopsy) was used for diagnosis of Bosniak greater than or equal to 2F cystic or solid (≥25% enhancing component) renal masses.2 A Bosniak class 2 cyst was diagnosed by using unenhanced T2W MRI if the cyst was entirely homogeneous, with internal SI as bright as cerebrospinal fluid with a smooth and thin wall.2,23 A Bosniak class 2 hemorrhagic cyst was diagnosed using unenhanced fat-suppressed T1W MRI if the cystic mass was entirely homogeneous with markedly increased T1W SI that was 2.5 times or greater than that of adjacent renal parenchyma.2,24,25 Because of these sequences were not directly used to determine enhancement (the index test in question), their use as the reference standard minimizes incorporation bias.

Renal lesions were the unit of analysis in this study. A true positive (TP) is considered an enhancing renal mass diagnosed on DECT or CE-MRI confirmed by the pathological reference standard. A true negative (TN) is considered a benign Bosniak 2 cystic renal mass without enhancement on DECT or CE-MRI confirmed by the reference standard. A false positive (FP) is considered a diagnosis of an enhancing renal mass at DECT or CE-MRI found to be a benign cyst by the reference standard, and a false negative (FN) is considered a diagnosis of a nonenhancing renal mass at DECT or CE-MRI classified as a solid renal mass by the pathological reference standard.

Imaging Techniques

All patients included in the study underwent renal mass protocol DECT using a single rapid kilovoltage potential (kVp) switch DECT platform with 64-slice multidetector CT systems (Discovery; GE Healthcare). True unenhanced images were acquired. The iodine overlay image was derived using material basis pair iodine-water analysis, as described previously,26 which enables calculation of the iodine content or concentration of a renal mass in milligrams per milliliter. Material basis pairs were generated using projection-based algorithm for material-decomposition using low- and high-energy projections for creation of material-specific projections enabling mass attenuation measurements to be calculated for 2 preselected materials. Reconstruction of material basis pair images was performed in the projection domain, improving the confounding effect of beam hardening, which is described as improving the accuracy of the material attenuation measurements.27 Derived images were reconstructed at a 2.5-mm thickness and stored in our institutional PACS for subsequent analysis. A summary of the DECT protocol used is provided in Supplementary Tables 1 and 2, https://links.lww.com/RCT/A144.26

Magnetic resonance imaging was performed using 1.5T or 3T clinical MRI systems and required axial or coronal single-shot T2W fast/turbo spin echo, precontrast, and postcontrast fat-suppressed T1W gradient recalled echo images with subtraction data sets for the CM phases and NG phases of enhancement. Dynamic precontrast and postcontrast imaging was performed in the axial plane in all patients. Subtraction images were derived at time of scanning by the imaging technologist and automatically transferred to our PACS. A full description of the MRI protocols used is provided in Supplementary Table 3, https://links.lww.com/RCT/A144.

Subjective Image Interpretation

Image interpretation was performed independently by 2 radiologists (P.S. and M.H.) with 7 and 8 years of experience in abdominal radiology, respectively, in 2 sessions minimum 14 days apart. Both were blinded to the reference standard and to clinical information. Dual-energy CT and MRI examinations were evaluated separately. A mixed order of modality interpretation was used to reduce bias between interpretation sessions.

For DECT examinations, radiologists evaluated: mass size (maximal dimension in millimeters), mass homogeneity (5-point Likert scale: completely heterogeneous, mostly heterogeneous, mixed areas of homogeneous and heterogeneous, mostly homogeneous, completely homogeneous),8 subjective enhancement (5-point scale Likert scale: definitely enhancing, likely enhancing, indeterminate, likely not enhancing, definitely not enhancing),8 image quality of iodine overlay data set (5-point Likert scale: nondiagnostic, poor, average, above average, excellent) based on the radiologists experience interpreting DECT images and, lastly, image artifact of iodine overlay data set (5-point scale Likert scale: severe, moderate, mild, minimal, none). Artifacts evaluated on iodine overlay data set included misregistered iodine content signals seen on iodine maps, residual beam hardening on virtual monochromatic data sets, and image quality degradation occurring on dual-energy data sets in patient with elevated body mass index.28

For MRI examinations, radiologists also evaluated: mass size (maximal dimension in millimeters), mass homogeneity (5-point Likert scale: completely heterogeneous, mostly heterogeneous, mixed areas of homogeneous and heterogeneous, mostly homogeneous, completely homogeneous),8 subjective enhancement using CM phase and NG phase subtraction images, which were derived automatically by subtracting the precontrast fat-suppressed T1W data set from the postcontrast fat-suppressed T1W data set with the difference in the 2 data sets representing the subtraction image (5-point scale: definitely enhancing, likely enhancing, indeterminate, likely not-enhancing, definitely not enhancing), image quality of subtraction images (5-point scale: nondiagnostic, poor, average, above average, excellent),12 and image artifact (motion, blur) of subtraction (5-point scale: severe, moderate, mild, minimal, none).12

Quantitative Analysis

At time of subjective analysis of DECT or MRI, radiologists also derived quantitative attenuation or SI measurements, respectively. For DECT, radiologists placed region of interest (ROI) within the masses on axial images from precontrast, CM phase, and NG phase (using PACS, Horizon Medical Imaging version 13. 1; McKesson Corporation) and recorded the mean attenuation in Hounsfield units. For homogeneous masses, a circular ROI was placed within the mass encompassing two thirds of its area avoiding the edges of the renal mass where it interfaced with renal parenchyma, renal sinus, or retroperitoneal fat. For heterogeneous lesions, the area subjectively showing the lowest attenuation at non-enhanced computed tomography was selected and an ROI was placed in the most hyperattenuating area (measuring at least 5 mm in diameter), copied and reproduced on the matching image sets. In this fashion, 3 measurements of attenuation were made for each mass in total, one for each phase of the examination. The circular ROI was then copied and reproduced on the matching image sets. Three ROI measurements were performed for each image and the average obtained.

Radiologists placed circular ROIs on renal masses on iodine-water images and measured iodine concentration in milligrams per milliliter. For homogeneous masses, a circular ROI was placed within the mass on the axial image where it appeared the largest encompassing two thirds of its area avoiding the edges of the renal mass where it interfaced with renal parenchyma, renal sinus, or retroperitoneal fat. Three ROI measurements were performed, and the average obtained. For heterogeneous lesions, the area subjectively showing the highest enhancement was selected and an ROI was placed in the most enhancing area (measuring at least 5 mm in diameter) in accordance with the study by Dai et al29 who showed that measuring the area of maximal enhancement on iodine overlay images was most accurate compared with other techniques including measuring the maximal enhancing area on individual slices or by using volumetric analysis.

For MRI, for homogeneously enhancing tumors, an ROI encompassing two thirds of the circumference of the enhancing portion of the mass was placed on the CM and NG phase fat-suppressed T1W-enhanced images and copied onto the precontrast fat-suppressed T1W images. For heterogeneously enhancing tumors, an ROI was placed encompassing two thirds of the area subjectively showing maximal enhancement on the CM or NG phase–enhanced images measuring at least 5 mm in diameter. This ROI was then copied onto the NG and precontrast phase images. Using the ROI measurements from the precontrast, CM and NG phase images, SI ratio as described by Ho et al6 was derived as follows: the SI of the precontrast phase was subtracted from the SI of the postcontrast phase (CM or NG), divided by the SI of the precontrast phase, multiplied by 100 to yield a percentage.

Data Analysis

Diagnostic accuracy of each index test was expressing as the area under the receiver operator curve (AUC), estimated for each reader with 95% confidence intervals (CIs). Differences in AUC between index tests were evaluated for statistical significance using the DeLong test.

Estimates of sensitivity and specificity, with 95% CIs were calculated at prespecified thresholds. Likert scale scores greater than or equal to 3 of subjective enhancement for both DECT and MRI were considered enhancing for the purpose of data analysis as these lesions would be most likely to undergo further management (eg, surgery, ablation, surveillance) or follow-up in clinical practice. For quantitative analysis, test positivity thresholds of 15 and 20 HU were used for conventional CT,2,26 signal, iodine concentration of 1.2 mg/mL or greater or 2.0 mg/mL or greater for DECT,17,27,30 and SI ratio of 15% or greater for MRI.6

Interobserver agreement between readers for subjective enhancement for both DECT and MRI was assessed using Fleiss κ. Image quality scores of DECT iodine overlay images and subtraction MRI images were compared using a paired t test. Statistical analysis was performed using R version 3.6.2 (R Project for Statistical Computing, Vienna, Austria), using the “pROC” package in R.

RESULTS

Patients

Twenty-four patients with 41 renal masses met inclusion criteria and received an acceptable reference standard. Reasons for patient exclusion are detailed in Figure 1. Three included patients were female, 21 were male; the mean age was 68.6 years (range, 49–88 years). Summary of patient demographics is included in Table 1.

F1
FIGURE 1:
Flow diagram depicting patient selection. *Picture archiving and communication system.
TABLE 1 - Patient Demographics
Characteristic n %
Age
 Mean 68.6
 Range 49–88
Sex
 Female 3 13
 Male 21 87
Mass type
 Solid 17 41
 Cyst 24 59
Target condition
 Clear cell RCC 6 15
 Papillary RCC 9 22
 Chromophobe RCC 1 2
 Oncocytoma 1 2
 Hemorrhagic cyst 4 10
 Simple cyst 20 49
Mass size
 Median, 30 mm
 IQR, 20–37 mm
 11–30 mm 23 56
 31–50 mm 10 25
 51–70 mm 4 10
 71–90 mm 3 7
 >90 mm 1 2
IQR indicates interquartile range.

The 24 included patients had 17 solid masses in total: 6 clear cell RCC, 9 papillary RCC, 1 chromophobe RCC, and 1 oncocytoma. Twenty-four Bosniak 1 or 2 cysts were included: 20 simple Bosniak 1 cysts, and 4 hemorrhagic or proteinaceous Bosniak 2 cysts.

Index Test Results

The AUCs of DECT and MRI, for both subjective and quantitative assessment, are summarized in Table 2. There was no difference in subjective enhancement comparing subtraction MRI and iodine image set derived from DECT for either reader 1 (0.99 [95% CI, 0.99–100] vs 0.99 [95% CI, 0.97–1.00], P = 0.38) or reader 2 (1.00 [95% CI, 0.99–1.00] vs 0.94 [95% CI, 0.85–1.00], P = 0.12). Estimates of sensitivity and specificity of subjective enhancement using subtraction MRI and iodine image set derived from DECT are provided in Table 3. Examples of subjective enhancement are shown for DECT in Figure 2 and MRI in Figure 3. Interobserver agreement of subjective enhancement was κ = 0.61 for DECT and κ = 0.71 for MRI.

TABLE 2 - Area Under the Receiver Operator Curve of DECT and MRI for Detection of Enhancement in Renal Masses
Index Test Reader 1, AUC (95% CI) Reader 2, AUC (95% CI)
Subtraction MRI subjective enhancement 0.99 (0.99–1.00) 1.00 (0.99–1.00)
DECT subjective enhancement 0.99 (0.97–1.00) 0.94 (0.85–1.00)
MRI quantitative SI, ΔSI 0.94 (0.87–1.00) 0.97 (0.94–1.00)
CT quantitative attenuation, ΔHU 0.96 (0.88–1.00) 0.96 (0.89–1.00)
DECT iodine concentration, mg/mL 0.94 (0.88–1.00) 0.92 (0.85–1.00)

TABLE 3 - Sensitivity and Specificity of MRI and DECT for Detection of Enhancement in Renal Masses at Prespecified Thresholds
Index Test Sensitivity (95% CI) Specificity (95% CI) TP FN FP TN
Reader 1
 Subtraction MRI subjective enhancement 0.92 (0.71–0.98) 0.98 (0.83–1.00) 16 1 0 24
 DECT subjective enhancement 0.92 (0.71–0.98) 0.94 (0.78–0.99) 16 1 1 23
 MRI quantitative SI, ΔSI 15% 0.86 (0.64–0.96) 0.74 (0.55–0.87) 15 2 6 18
 CT quantitative attenuation, Δ 15 HU 0.92 (0.71–0.98) 0.98 (0.83–1.00) 16 1 0 24
 CT quantitative attenuation, Δ 20 HU 0.86 (0.64–0.96) 0.98 (0.83–1.00) 15 2 0 24
 DECT iodine concentration, 1.2 mg/mL 0.81 (0.58–0.93) 0.94 (0.78–0.99) 14 3 1 23
 DECT iodine concentration, 2.0 mg/mL 0.42 (0.22–0.64) 0.98 (0.83–1.00) 7 10 0 24
Reader 2
 Subtraction MRI subjective enhancement 0.97 (0.78–1.00) 0.98 (0.83–1.00) 17 0 0 24
 DECT subjective enhancement 0.86 (0.64–0.96) 0.94 (0.78–0.99) 15 2 1 23
 MRI quantitative SI, ΔSI 15% 0.97 (0.78–1.00) 0.74 (0.55–0.87) 17 0 6 18
 CT quantitative attenuation, Δ 15 HU 0.75 (0.52–0.89) 0.98 (0.83–1.00) 13 4 0 24
 CT quantitative attenuation, Δ 20 HU 0.75 (0.52–0.89) 0.98 (0.83–1.00) 13 4 0 24
 DECT iodine concentration, 1.2 mg/mL 0.69 (0.46–0.86) 0.94 (0.78–0.99) 12 5 1 23
 DECT iodine concentration, 2.0 mg/mL 0.42 (0.22–0.64) 0.98 (0.83–1.00) 7 10 0 24

F2
FIGURE 2:
Dual-energy computed tomography images illustrating the 5-point homogeneity scale, subjective enhancement, and image quality in a 49-year-old man with histopathology-proven RCC. Axial unenhanced (A) and postcontrast CM phase (B) DECT images in the same patient depict the mass with mixed homogeneous and heterogeneous areas (grade 3 on homogeneity scale), confirmed on the greyscale iodine map (C) image. Definite enhancing mural nodules are seen on CM phase and iodine map images (grade 5). Greyscale iodine map image demonstrates above-average image quality (grade 4). Better visualization of subjective enhancement could be achieved using color map representation (D) of the iodine map image using image-J software, with enhancing mural nodules appearing green-yellow (arrowheads) against a background of blue nonenhancing areas. Figure 2 can be viewed online in color at www.jcat.org.
F3
FIGURE 3:
Magnetic resonance images illustrating the 5-point homogeneity scale, subjective enhancement, and image quality in a 49-year-old man with histopathology-proven RCC. Axial unenhanced axial T1W-FS (A) and postgadolinium CM (b) and NG (d) phase T1W-FS images show a 26-mm right interpolar renal mass (arrows) with mural nodules of varying intensities and enhancement interpreted as a completely heterogeneous mass (grade 5 on homogeneity Likert scale) with definite enhancement (grade 5). Corticomedullary phase subtraction (C) image is of excellent quality (grade 5).

For both reader 1 and reader 2, there were no differences in accuracy in pairwise AUC comparisons between quantitative MRI enhancement (ΔSI), quantitative CT enhancement (ΔHU), and iodine concentration. Results of pairwise accuracy comparisons for each index test for each reader are summarized in Table 4. Estimates of sensitivity and specificity for each index test for each reader, at prespecified thresholds, are summarized in Table 3. Examples of quantitative assessment for both DECT and MRI are shown in Figures 4 and 5, respectively.

TABLE 4 - Area Under the Receiver Operator Curve Comparison of MRI and DECT to Detect Enhancement in Renal Masses
Comparison Reader 1, P Reader 2, P
Subtraction MRI vs DECT subjective enhancement 0.38 0.12
Δ MRI quantitative SI vs DECT iodine concentration 0.88 0.16
Δ MRI quantitative SI vs Δ CT quantitative attenuation 0.54 0.64
DECT iodine concentration vs Δ CT quantitative attenuation 0.66 0.37

F4
FIGURE 4:
Dual-energy CT images of papillary RCC in a 53-year-old man identified as TP on both DECT and CE-MRI. Axial unenhanced (A) and postcontrast CM phase (B) DECT images in the same patient depict the mass with intense heterogeneous enhancement consistent with RCC. Given the homogeneous appearance of the mass on unenhanced CT, a circular ROI encompassing two thirds of its area was used (circle in a) and then copied on to the matching images in the CM (circle in B) and NG phases. Iodine concentration was measured by placing a circular ROI within the maximally enhancing area in the mass (circle in C). Better visualization of subjective enhancement could be achieved using color map representation (D) of the iodine map image using image-J software, with enhancing areas appearing red-orange (arrowhead) comparable with adjacent renal cortical enhancement. Figure 4 can be viewed online in color at www.jcat.org.
F5
FIGURE 5:
Magnetic resonance images of papillary RCC in a 53-year-old man identified as TP on both DECT and CE-MRI. Axial unenhanced axial T1W-FS (A) and postgadolinium CM (B) and NG (D) phase T1W-FS MR images show a heterogeneously enhancing 30-mm right interpolar renal mass (arrows) consistent with RCC. Given the heterogeneous enhancement, a circular ROI was placed in the area of maximal enhancement in the CM phase image (arrowhead in C, circle in B) and then copied on to the unenhanced and NG phase images for measurement of SI. Figure 5 can be viewed online in color at www.jcat.org.

There was no significant difference in image quality scores for DECT versus MRI for either reader. The mean image quality score was 4.34 for DECT and 4.14 for MRI for reader 1 (P = 0.29). The mean quality score was 4.02 for DECT and 4.04 for MRI for reader 2 (P = 0.14).

Among the solid renal masses incorrectly classified as nonenhancing (FN), nearly all were papillary RCC. Two cases of papillary RCC were subjectively scored as “likely not enhancing” by reader 1, one on DECT and one on MRI. Two additional cases of papillary RCC were subjectively scored as “likely not enhancing” by reader 2, both on DECT. The 4 cases of papillary RCC subjectively scored as “likely not enhancing” by one reader on one modality were all scored as “indeterminate” by the other reader. Images of a papillary RCC correctly identified as enhancing on DECT but scored as “likely not enhancing” on MRI are shown in Figures 6 and 7, respectively. Images of a papillary RCC scored as “likely not enhancing” on DECT but correctly identified as enhancing on MRI are shown in Figures 8 and 9, respectively.

F6
FIGURE 6:
Dual-energy CT images illustrating a discrepancy between DECT and CE-MRI in a 78-year-old man with RCC—TP on DECT and FN on CE-MRI. Axial unenhanced DECT (A) image demonstrated a questionable hyperattenuating component (arrowhead) within the lesion, which was confirmed to be a definitive enhancing focus on CM phase DECT (B) and NG phase-based iodine map (C) images (arrowheads in F and G), better visualized using a color-processed image (D) as a green area (arrowhead) in a background of blue nonenhancing areas. Figure 6 can be viewed online in color at www.jcat.org.
F7
FIGURE 7:
Magnetic resonance images illustrating a discrepancy between DECT and CE-MRI in a 78-year-old man with RCC—TP on DECT and FN on CE-MRI. Axial unenhanced T1W-FS MRI image (A) demonstrated a 63-mm left lower pole renal mass (white arrow) with extensive T1W hyperintense areas (black arrows) and septations. Postgadolinium CM (B) and NG (C) phase T1W-FS and CM phase subtraction (D) images do not demonstrate any areas of definitive enhancement above the test positive threshold. False-negative interpretation on CE-MRI may have been related to the confounding effect of baseline T1W hyperintensity from extensive intralesional hemorrhage.
F8
FIGURE 8:
Dual-energy CT images illustrating a discrepancy between DECT and CE-MRI in a 60-year-old man with papillary RCC—FN on DECT and TP on CE-MRI. Axial unenhanced DECT (A) shows a mildly hyperattenuating 13-mm lesion in the right renal anterior interpolar region (arrow). Although intralesional enhancement on postcontrast CM phase (B) image appears to be indeterminate, there is no iodine uptake on the iodine map (C), confirmed by the lack of enhancing areas on the color-processed iodine map image (D). Figure 8 can be viewed online in color at www.jcat.org.
F9
FIGURE 9:
Magnetic resonance images illustrating a discrepancy between DECT and CE-MRI in a 60-year-old man with papillary RCC—FN on DECT and TP on CE-MRI. Axial unenhanced T1W-FS (A) and postgadolinium CM phase (B) and NG phase (C) T1W-FS MR images demonstrate a T1W hypointense lesion, with a questionable enhancing mural nodule in the CM phase. A progressively enhancing mural nodule is definitively seen in the NG phase, confirmed using an NG phase subtraction image (D), suggestive of a papillary RCC.

The only FN results in quantitative evaluation that were not papillary RCC occurred on DECT at the test positivity threshold for iodine concentration of 2.0 mg/mL. A clear cell RCC had an iodine concentration of 1.2 mg/mL in measurements by both reader 1 and reader 2. A chromophobe RCC had iodine concentration of 1.8 and 1.5 mg/mL in measurements by reader 1 and reader 2, respectively. Using the more sensitive, previously described, DECT test positivity threshold for iodine concentration of 1.2 mg/mL resulted in no FN results that were not papillary RCC.

DISCUSSION

This study demonstrates no significant difference in diagnostic accuracy between MRI and fast kVp switching DECT when evaluating enhancement of renal masses, using either subjective or quantitative assessment, by 2 independent abdominal radiologists. Among masses that were incorrectly classified as nonenhancing, nearly all were papillary RCC, which are known to be low-enhancing lesions, which can present diagnostic difficulty7,8; however, our study shows that FN interpretations occurred using both DECT and MRI. Similar to previous literature, we found that an iodine concentration threshold of 1.2 mg/mL yielded fewer FN results for detection of enhancement than a threshold of 2.0 mg/mL.26

In our study, papillary RCC represented 22% of the overall sample and 53% of the solid renal masses included. This is greater than would be expected from a consecutive sample of solid renal masses.31 It is known that papillary RCC represent more challenging masses encountered on imaging (to establish enhancement) and are more likely to undergo both MRI and DECT within a short timeframe. Thus, the diagnostic accuracy estimates within this study may be underestimated when compared with routine clinical practice due to an overrepresentation of low-enhancing masses, although our reported accuracies compare favorably to the literature.7,8,14

Previous studies have shown that MRI consistently outperforms conventional CT for assessment of enhancement in renal masses, with advantages noted in hemorrhagic cysts and low-enhancing papillary RCC.7,8,11 Dual-energy computed tomography offers advantages compared with conventional CT including the ability to evaluate iodine overlay images, improved diagnosis of pseudoenhancement, and a potentially heightened sensitivity to iodine and, thus, enhancement.16 Moreover, DECT is faster than MRI and theoretically could offer an alternative diagnostic pathway for detection of enhancement in patients who are not able to adequately breath-hold or cooperate with lengthy MRI examination times resulting in reduced quality of precontrast and postcontrast and generated subtraction image sets, although no difference in image quality scores were seen in our study.

Although no significant difference in accuracy between MRI and DECT was documented in our study, our sample size was admittedly limited, with only 24 patients and 41 renal masses meeting inclusion criteria. While we may be underpowered to detect small differences in accuracy between DECT and MRI, given the similar AUC with relatively narrow CIs, a large difference in accuracy is unlikely. Our sample size also precluded meaningful subgroup analysis, such as comparing endophytic and exophytic masses. With the emergence of active surveillance for small renal masses, fewer patients with a pathological reference standard were available for inclusion, making obtaining robust sample sizes more difficult. Lack of pathological reference standard was our most frequent reason for patient exclusion (n = 25). The diagnostic accuracy of MRI and DECT are most important in cT1a (≤4 cm) renal masses in which active surveillance is a possible treatment strategy, which excluded potentially eligible masses in our study.

In conclusion, this preliminary study demonstrated no significant difference in accuracy comparing DECT and MRI for detection of enhancement in renal masses using subjective and quantitative analysis. False-negative instances were noted in papillary RCC using both DECT and MRI, with no systematic bias toward performance comparing either modality. A limitation of our study is a relatively modest sample size, necessitating validation in larger samples.

REFERENCES

1. Silverman SG, Israel GM, Herts BR, et al. Management of the incidental renal mass. Radiology. 2008;249:16–31.
2. Silverman SG, Pedrosa I, Ellis JH, et al. Bosniak classification of cystic renal masses, version 2019: an update proposal and needs assessment. Radiology. 2019;292:475–488.
3. Herts BR, Silverman SG, Hindman NM, et al. Management of the incidental renal mass on CT: a white paper of the ACR Incidental Findings Committee. J Am Coll Radiol. 2018;15:264–273.
4. Siddaiah M, Krishna S, McInnes MDF, et al. Is ultrasound useful for further evaluation of homogeneously hyperattenuating renal lesions detected on CT?AJR Am J Roentgenol. 2017;209:604–610.
5. Wang ZJ, Nikolaidis P, Khatri G, et al. ACR Appropriateness Criteria® indeterminate renal mass. J Am Coll Radiol. 2020;17(S 11):S415–S428.
6. Ho VB, Allen SF, Hood MN, et al. Renal masses: quantitative assessment of enhancement with dynamic MR imaging. Radiology. 2002;224:695–700.
7. Egbert ND, Caoili EM, Cohan RH, et al. Differentiation of papillary renal cell carcinoma subtypes on CT and MRI. AJR Am J Roentgenol. 2013;201:347–355.
8. Dilauro M, Quon M, McInnes MD, et al. Comparison of contrast-enhanced multiphase renal protocol CT versus MRI for diagnosis of papillary renal cell carcinoma. AJR Am J Roentgenol. 2016;206:319–325.
9. Ramamurthy NK, Moosavi B, McInnes MD, et al. Multiparametric MRI of solid renal masses: pearls and pitfalls. Clin Radiol. 2015;70:304–316.
10. Israel GM, Bosniak MA. Pitfalls in renal mass evaluation and how to avoid them. Radiographics. 2008;28:1325–1338.
11. Israel GM, Bosniak MA. How I do it: evaluating renal masses. Radiology. 2005;236:441–450.
12. Tu W, Alzahrani A, Currin S, et al. Evaluation of a free-breathing respiratory-triggered (navigator) 3-D T1-weighted (T1W) gradient recalled echo sequence (LAVA) for detection of enhancement in cystic and solid renal masses. Eur Radiol. 2019;29:2507–2517.
13. Duffy PB, Stemmer A, Callahan MJ, et al. Free-breathing radial stack-of-stars three-dimensional Dixon gradient echo sequence in abdominal magnetic resonance imaging in sedated pediatric patients. Pediatr Radiol. 2021;51:1645–1653.
14. Salameh JP, McInnes MDF, McGrath TA, et al. Diagnostic accuracy of dual-energy CT for evaluation of renal masses: systematic review and meta-analysis. AJR Am J Roentgenol. 2019;1–6.
15. Mileto A, Nelson RC, Samei E, et al. Impact of dual-energy multi-detector row CT with virtual monochromatic imaging on renal cyst pseudoenhancement: in vitro and in vivo study. Radiology. 2014;272:767–776.
16. Song KD, Kim CK, Park BK, et al. Utility of iodine overlay technique and virtual unenhanced images for the characterization of renal masses by dual-energy CT. AJR Am J Roentgenol. 2011;197:W1076–W1082.
17. Kaza RK, Caoili EM, Cohan RH, et al. Distinguishing enhancing from nonenhancing renal lesions with fast kilovoltage-switching dual-energy CT. AJR Am J Roentgenol. 2011;197:1375–1381.
18. Alanee S, Dynda DI, Hemmer P, et al. Low enhancing papillary renal cell carcinoma diagnosed by using dual energy computerized tomography: a case report and review of literature. BMC Urol. 2014;14:102.
19. Mileto A, Marin D, Alfaro-Cordoba M, et al. Iodine quantification to distinguish clear cell from papillary renal cell carcinoma at dual-energy multidetector CT: a multireader diagnostic performance study. Radiology. 2014;273:813–820.
20. Heye T, Nelson RC, Ho LM, et al. Dual-energy CT applications in the abdomen. AJR Am J Roentgenol. 2012;199(suppl 5):S64–S70.
21. Bossuyt PM, Reitsma JB, Bruns DE, et al. STARD 2015: an updated list of essential items for reporting diagnostic accuracy studies. Radiology. 2015;277:826–832.
22. Hong PJ, Korevaar DA, McGrath TA, et al. Reporting of imaging diagnostic accuracy studies with focus on MRI subgroup: adherence to STARD 2015. J Magn Reson Imaging. 2018;47:523–544.
23. Nelson SM, Oettel DJ, Lisanti CJ, et al. Incidental renal lesions on lumbar spine MRI: who needs follow-up?AJR Am J Roentgenol. 2019;212:130–134.
24. Davarpanah AH, Spektor M, Mathur M, et al. Homogeneous T1 hyperintense renal lesions with smooth borders: is contrast-enhanced MR imaging needed?Radiology. 2016;281:326.
25. Kim CW, Shanbhogue KP, Schreiber-Zinaman J, et al. Visual assessment of the intensity and pattern of T1 hyperintensity on MRI to differentiate hemorrhagic renal cysts from renal cell carcinoma. AJR Am J Roentgenol. 2017;208:337–342.
26. Sadoughi N, Krishna S, Macdonald DB, et al. Diagnostic accuracy of attenuation difference and iodine concentration thresholds at rapid-kilovoltage-switching dual-energy CT for detection of enhancement in renal masses. AJR Am J Roentgenol. 2019;213:619–625.
27. Marin D, Davis D, Roy Choudhury K, et al. Characterization of small focal renal lesions: diagnostic accuracy with single-phase contrast-enhanced dual-energy CT with material attenuation analysis compared with conventional attenuation measurements. Radiology. 2017;284:737–747.
28. Mileto A, Nelson RC, Paulson EK, et al. Dual-energy MDCT for imaging the renal mass. AJR Am J Roentgenol. 2015;204:W640–W647.
29. Dai C, Cao Y, Jia Y, et al. Differentiation of renal cell carcinoma subtypes with different iodine quantification methods using single-phase contrast-enhanced dual-energy CT: areal vs. volumetric analyses. Abdom Radiol (New York). 2018;43:672–678.
30. Zarzour JG, Milner D, Valentin R, et al. Quantitative iodine content threshold for discrimination of renal cell carcinomas using rapid kV-switching dual-energy CT. Abdom Radiol (New York). 2017;42:727–734.
31. Remzi M, Ozsoy M, Klingler HC, et al. Are small renal tumors harmless? Analysis of histopathological features according to tumors 4 cm or less in diameter. J Urol. 2006;176:896–899.
Keywords:

kidney; cyst; renal mass; magnetic resonance imaging; dual-energy computed tomography

Supplemental Digital Content

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.