Differences in Renal Tumor Size Measurements for Computed Tomography Versus Magnetic Resonance Imaging: Implications for Patients on Active Surveillance

Abstract

Purpose:

To evaluate and compare the accuracy of computed tomography (CT) and magnetic resonance imaging (MRI) in predicting the final pathologic tumor size of partial nephrectomy specimens.

Materials and Methods:

We analyzed a multi-institutional database of 807 patients who underwent robotic partial nephrectomy for a cT1a renal mass from 2006 to 2016. Patients who had a solitary tumor with complete data on the baseline imaging modality and the tumor size (baseline and pathologic) (n = 349) were included for analysis. Baseline tumor size evaluated by both imaging modalities, in addition to the difference between the measurements and final pathologic tumor size (cm) measurements, was compared between patients who received a baseline CT (n = 276, 79.1%) and those who received an MRI (n = 73, 20.9%).

Results:

There were no statistically significant differences between any baseline characteristics and receipt of a CT versus MRI. In multivariable analysis adjusting for confounders, there was no significant difference in the baseline tumor size between patients receiving an MRI and those receiving a CT (2.3 versus 2.6 cm; β = −0.13; 95% confidence interval [CI] = −0.33 to 0.07; P = .208). Tumor size on imaging was smaller from final pathology by 0.43 cm on average (P = .002). Measurement error for the measured baseline versus actual pathologic tumor size did not significantly differ for patients receiving an MRI versus those receiving a CT (0.38 versus 0.44 cm; β = −0.06; 95% CI = −0.16 to 0.04; P = .232).

Conclusion:

Baseline renal tumor size measurements were not significantly different for CT scan and MRI. Choice of imaging modality can be based on doctor and patient preference, including cost and exposure to radiation.

Introduction

The incidence of renal cell carcinoma (RCC) is increasing at an annual rate of 1.7% in the United States. This is largely due to an increase in incidental detection of renal masses by abdominal imaging for unrelated symptoms.¹ Overall, 48%–66% of newly diagnosed RCCs are small renal masses (SRM), which are defined as lesions measuring ≤4 cm in diameter.² Partial nephrectomy is the standard of care for treatment of SRMs, whereas active surveillance (AS) has proved to be an effective alternative to surgery, with only 1%–2% of patients progressing to clinically significant or metastatic disease.³ Further, a maximum of 30% of incidental SRMs are benign⁴ and about 20% of resected SRMs are benign,³ signifying overtreatment in a substantial number of cases. With low metastatic potential, AS may be a viable option for elderly patients with multiple comorbidities and those preferring to avoid treatment.

Patients are recommended to undergo an initial tumor measurement by computed tomography (CT), magnetic resonance imaging (MRI), or ultrasound (US).⁵ The current American Urological Association (AUA) guidelines do not offer a high level of evidence recommendation; however, based on patient preference and professional judgment, patients should specifically receive a CT/MRI within 6 months of beginning AS followed by annual imaging—MRI/CT/US—to assess growth rate.⁵ Growth in tumor size at follow-up may indicate malignancy and, thus, require intervention.¹ Since many patients alternate between imaging modalities, our study compared renal tumor size measurements between CT and final pathology versus MRI and final pathology to evaluate whether imaging modality impacts tumor measurement in SRMs.

Materials and Methods

Data source

A multi-institutional partial nephrectomy database was used to identify 807 patients who underwent surgery for a small renal mass between 2008 and 2016. Patients who underwent robotic partial nephrectomy (RPN) for multiple tumors (n = 29) were excluded as were patients with a clinical stage >cT1a. We excluded 128 patients with multiple forms of imaging in the database (e.g., CT + MRI + US) as we could not confirm which imaging modality the preoperative tumor measurement corresponded to, that is, the database did not associate the tumor measurements with the multiple imaging modalities specified. We also excluded 93 patients in whom it could not be confirmed whether a multiple or solitary RPN was performed. Patients with missing data on baseline and pathologic tumor size were excluded (n = 24). Patients undergoing RPN with a tumor thrombus (n = 8), metastatic disease (n = 4), or horseshoe kidney (n = 2) were excluded. Overall, there were 349 patients who met eligibility criteria, including 276 (79.1%) patients who received a preoperative CT scan and 73 (20.9%) patients who received a preoperative MRI. All patients underwent RPN from one of four surgeons.

Institutional Review Board (IRB) approval was independently obtained from each institution to allow data sharing with our institutional IRB and served as the data coordinating center for this study. All data were managed and downloaded from Research Electronic Data Capture (REDCap).

Patient characteristics

The association of each patient characteristic with receipt of a CT versus MRI was evaluated by using binary logistic regression models. Variables included age at the time of surgery, gender, body mass index, American Society of Anesthesiologists (ASA) score, patient comorbidities, age-adjusted Charlson comorbidity index (CCI), baseline estimated glomerular filtration rate (eGFR, mL/min/1.73 m²), baseline chronic kidney disease stage, history of prior abdominal surgery, and whether the patient had a solitary kidney.

Primary outcome

The primary outcome was the difference in baseline tumor size between patients who received a CT and those who received an MRI. This was evaluated by using a multivariable linear regression model while adjusting for any covariates that were associated with imaging modality and baseline tumor size. These included operating surgeon and chronic obstructive pulmonary disease (COPD).

Secondary outcome

The secondary outcome was a comparison of accuracy between CT and MRI. Accuracy was assessed by comparing the magnitude of measurement error (i.e., the difference between the imaging measurement and actual pathologic tumor size measurement) between CT and MRI.

All statistical analysis performed was done by using R, with statistical significance considered at the P < .05 level.

Results

Patient characteristics

Patient characteristics of patients receiving a CT scan or an MRI are presented in Table 1. There were no statistically significant associations of any baseline characteristics between CT and MRI, including age (61.0 versus 61.0 years, P = .325), body mass index (BMI) (30.1 versus 30.5 years, P = .427), comorbidities (P > .05), including age-adjusted CCI (3 versus 3, P = .485), or baseline eGFR (87.2 versus 83.7, P = .265).

Table 1.

Comparison of Baseline Characteristics Between Patients Receiving a Computed Tomography Versus Those Receiving Magnetic Resonance Imaging

	CT	MRI	OR (95% CI)	P
N, n (%)	276 (79.1)	73 (20.9)
Age	61.0 (50.0–68.0; 26.0–89.0)	61.0 (54.0–68.0; 26.0–82.0)	1.002 (0.99–1.01)	.325
Male, n (%)	163 (59.1)	42 (57.5)	0.99 (0.91–1.08)	.815
BMI	30.1 (27.0–34.3; 18.8–63.0)	30.5 (27.0–35.1; 20.0–53.6)	1.00 (0.99–1.01)	.427
ASA > = 3, n (%)	137 (52.7)	40 (58.8)	1.04 (0.95–.14)	.368
DM, n (%)	56 (20.4)	20 (27.4)	1.07 (0.96–1.19)	.203
HLD, n (%)	68 (24.8)	18 (24.7)	0.99 (0.90–.10)	.978
CAD, n (%)	19 (6.9)	5 (6.9)	0.99 (0.84–.18)	.980
COPD, n (%)	10 (3.7)	0 (0.0)	0.81 (0.62–.04)	.098
STONES, n (%)	47 (17.2)	10 (13.7)	0.96 (0.85–.08)	.480
Age adjusted CCI	3 (2–4; 0–9)	3 (2–4; 0–7)	1.01 (0.98–.03)	.485
Solitary kidney, n (%)	5 (1.8)	2 (2.7)	1.05 (0.92–.19)	.444
Baseline eGFR	87.2 (71.8–103.2; 26.2–214.3)	83.7 (63.6–100.3; 18.9–176.6)	0.99 (0.99–.00)	.265
Baseline CKD ≥3, n (%)	32 (11.9)	11 (15.3)	1.05 (0.95–.16)	.313
Prior abdominal surgey, n (%)	151 (57.9)	43 (63.2)	1.04 (0.95–.13)	.423
Prior nephrectomy, n (%)	17 (6.5)	4 (5.9)	0.98 (0.82–.18)	.850

ASA, American Society of Anesthesiologists; BMI, body mass index; CAD, coronary artery disease; CCI, Charlson comorbidity index; CI, confidence interval; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; CT, computed tomography; DM, diabetes mellitus; HLD, hyperlipidemia; eGFR, estimated glomerular filtration rate; MRI, magnetic resonance imaging; OR, odds ratio.

Primary outcome

Overall median baseline tumor size was 2.6 cm (mean: 2.62 cm, interquartile range [IQR]: 2.00–3.2; range: 1.00–4.00). Overall median final pathologic tumor size was 2.5 cm (mean: 2.52 cm; IQR: 1.90–3.00; range: 0.90–5.30). The median final pathologic tumor size did not differ for those who received a CT versus those who received an MRI (P = .485) and specifically was 2.50 cm (mean: 2.46 cm; IQR: 1.80–2.90; range: 1.00–5.00) for patients who received an MRI and 2.50 cm (mean: 2.53 cm; IQR: 1.90–3.10; range: 0.90–5.30) for those who received a CT.

For patients who received an MRI, the median baseline tumor size was 2.30 cm (mean: 2.50 cm; IQR: 1.90–3.10; range: 1.20–4.00) compared with 2.60 cm (mean: 2.65 cm IQR: 2.00–3.30; range: 1.00–4.00) for patients who received a CT scan, with no difference in univariable analysis (P = .529) and in multivariable analysis while controlling for final pathologic tumor size (β = −0.13; 95% confidence interval [CI] = −0.33–0.07; P = .208).

Secondary outcome

In evaluating accuracy, tumor size on baseline imaging was smaller from final pathology by 0.43 cm on average (P = .002). Measurement error for the measured baseline versus actual pathologic tumor size did not significantly differ for patients receiving an MRI versus those receiving a CT (0.38 versus 0.44 cm; β = −0.06; 95% CI = −0.16–0.04; P = .232).

Discussion

The utilization of AS for SRM remains discretionary and involves many clinical and patient factors. There are no set AS guidelines besides repeat imaging to track changes in tumor size; both timing of surveillance images and use of imaging modality are based on patient preference and minimizing patients' exposure to radiation.¹ With tumor growth rate as the primary driver for intervention, it is important to understand whether a discrepancy in measurements between imaging modalities exists, as this may erroneously indicate disease progression, and, thus, may result in overtreatment. In this study, we compared baseline tumor size of patients who received a CT scan with those who received an MRI by controlling for tumor size at final pathology. Our results showed that although both modalities underestimate actual tumor size, these measurements were similar at baseline and at final pathology for both CT and MRI. Regarding AS patients, many undergo CT as an initial study and, subsequently, prefer MRI for follow-up. Our results indicate that alternating imaging modalities to minimize radiation exposure does not compromise measurement accuracy.

Generally, SRM evaluation includes size, enhancement with contrast, presence of septa, calcification, and wall thickness. Though MRI can give additional details compared with CT—such as subtle enhancement, wall thickening, and additional septa—lesion size should remain approximately close between both imaging modalities.⁶ Our study tested this assumption for its implications in the management of SRMs. In a similar study, Mucksavage et al. reported that patients who underwent MRI and CT scanning before a radical or partial nephrectomy did not show a significant difference in lesion size between the imaging modalities.⁷ Timing between subsequent imaging evaluations was not assessed, so any difference in measurements may have reflected actual lesion growth. Contrary to our results, a number of studies found that CT scans slightly overestimated tumor size, which was attributed to tumor devascularization.^8–10 However, these studies did not limit analysis to cT1a masses.

AS has been recommended for patients with limited life expectancy and those at high risk for complications.¹¹ However, a key limitation of AS is the reported 1%–2% metastatic disease progression in a mean follow-up of nearly 3 years.^12–14 These studies may exclude a considerable amount of patients who developed metastatic cancer thereafter. Tumor growth rate is an established proxy—although not without limitations—for disease progression in AS. In fact, every 1 cm increase in tumor size has been shown to increase the likelihood of malignancy by 17%.¹⁵ Moreover, Remzi et al. showed that tumor size holds prognostic value in predicting metastatic disease. Tumors >3 cm were shown to be at 6% greater risk of developing distant metastasis compared with tumors <3 cm.¹⁶

Our results show that both CT and MRI report lower tumor size compared with final pathology. Thus, for a more accurate evaluation, additional tools such as a percutaneous biopsy may hold utility in defining tumor burden. However, success in evaluating the lesion is contingent on tumor composition. Percutaneous biopsy is well established for solid renal masses; however, it remains controversial for cystic masses.^17,18 Nevertheless, Lane et al. reported that renal mass biopsies have a 92% sensitivity for malignancy and a 90% specificity with a minimal failure rate.¹⁹ Cystic and solid tumors have presented differential prognosis and should be further evaluated. Winters et al. compared cancer-specific mortality and pathological features between cystic RCC and clear-cell RCC in a population-based tumor registry. They found that cystic RCC displayed better prognosis—improved cancer-specific survival—than clear-cell RCC for larger tumors only, suggesting that the relevance of tumor biology increases with tumor size,²⁰ thus suggesting that AS may be an appropriate avenue for selected patients with cystic renal masses.

Patients put on AS are suggested to undergo serial imaging every 3 months for 1 year, then every 6 months for 3 years, followed by annual evaluation thereafter if tumor growth is negligible.⁴ In a study with 109 patients on AS, Crispen et al. reported that 36% of patients eventually underwent surgical treatment and the median duration of AS was 26 months.²¹ SRMs, on average, grow about 0.31 cm/year.²² Experts recommend delayed intervention when growth rate is 0.4–0.5 cm/year, when a lesion becomes greater than 3–4 cm, or when lesion volume doubles.¹¹ Our results do not show a significant discrepancy in measurements between CT and MRI, so imaging modality can be alternated. However, to achieve diagnostic success, and, in turn, better clinical outcomes, urologists must explore additional options to distinguish aggressive malignant masses from indolent ones.

In the future, radiogenomics may be used for such differentiation. This approach associates imaging characteristics with underlying mutations. Karlo et al. showed that mutations in the VHL gene were significantly associated with gross appearance of intratumoral vascularity (P = .018), well-defined tumor margins (P = .013), and nodular tumor enhancement (P = .021) on contrast-enhanced CT. BAP1 and KDM5C mutations were significantly associated with apparent renal vein invasion.²³ However their cohort included advanced-staged patients, whereas AS is an option for only SRMs. Future studies are needed to validate these associations for clinical application.

Our article had limitations, including the retrospective nature of this study. Ideally, multimodal imaging evaluations should be ordered for a single patient in a prospective setting to accurately compare imaging modalities. Distribution of patients who received either a CT or an MRI was skewed, with 79.1% who received a CT and 20.9% who received an MRI. We did not examine US measurements and compare them with CT and MRI. AS patients can undergo US imaging at follow-up⁵; thus, tumor measurements on US should be compared with both CT and MRI for a more complete analysis. However, it should be noted that US is the least sensitive for detecting SRMs and can be difficult to evaluate, as it entails dynamic examination.^24,25 Because a multi-institutional database was used, our study lacked a central radiologist, which may have resulted in measurement variation.

Conclusion

MRI and CT are two methods with similar accuracy for predicting the actual tumor size of SRM. Alternating between MRI and CT while on AS does not pose a significant risk of discrepancy in tumor size. Additional tools are needed to better assess tumor aggressiveness.

Footnotes

Disclosure Statement

I.K., A.T.B., D.P., A.B., and K.K.B. declare they have no conflicts of interest. R.A. reports grants from Intuitive Surgical, grants from ConMed (Surgiquest), outside the submitted work. D.D.E. reports personal fees from Intuitive Surgical, personal fees from Medtronic, personal fees from Conmed, outside the submitted work.

References

Nayyar

, et al. Active surveillance of small renal masses: A review on the role of imaging with a focus on growth rate. J Comput Assist Tomogr, 2016; 40:517–523.

Sanchez-Martin

, et al. Small renal masses: Incidental diagnosis, clinical symptoms, and prognostic factors. Adv Urol, 2008; 2008:310694.

Bhatt

, Richard

, Jewett

MAS

. Natural history and active surveillance. In: Oya

(ed). Renal Cell Carcinoma Molecular Features and Treatment Updates. Tokyo: Springer Nature. 2017, pp. 209–220.

Ahmad

, Finelli

, Jewett

. Surveillance of small renal masses. Urology, 2016; 98:8–13.

Jacobs

, et al. Understanding criteria for surveillance of patients with a small renal mass. Urology, 2012; 79:1027–1032.

Vedanayagam

, Bhattacharya

, Sriprasad

. Magnetic resonance imaging and the use in small renal masses. Indian J Surg Oncol, 2017; 8:19–23.

Mucksavage

, et al. Is ultrasound imaging inferior to computed tomography or magnetic resonance imaging in evaluating renal mass size?. Urology, 2012; 79:28–31.

Jeffery

, et al. Discrepancy between radiological and pathological size of renal masses. BMC Urol, 2011; 11:2.

Herr

. Radiographic vs surgical size of renal tumours after partial nephrectomy. BJU Int, 2000; 85:19–21.

10.

Kurta

, et al. Contemporary imaging of patients with a renal mass: Does size on computed tomography equal pathological size?. BJU Int, 2009; 103:24–27.

11.

Lane

, Tobert

, Riedinger

. Growth kinetics and active surveillance for small renal masses. Curr Opin Urol, 2012; 22:353–359.

12.

Yamaguchi

, Simmons

, Campbell

. Small renal masses: Risk prediction and contemporary management. Hematol Oncol Clin North Am, 2011; 25:717–736.

13.

Graversen

, et al. Active surveillance of renal cortical neoplasms: A contemporary review. Postgrad Med, 2011; 123:105–113.

14.

Smaldone

, et al. Small renal masses progressing to metastases under active surveillance: A systematic review and pooled analysis. Cancer, 2012; 118:997–1006.

15.

Frank

, et al. Solid renal tumors: An analysis of pathological features related to tumor size. J Urol, 2003; 170(6 Pt 1):2217–2220.

16.

Remzi

, 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.

17.

Silverman

, et al. Renal masses in the adult patient: The role of percutaneous biopsy. Radiology, 2006; 240:6–22.

18.

Sahni

, Silverman

. Biopsy of renal masses: When and why. Cancer Imaging, 2009; 9:44–55.

19.

Lane

, et al. Renal mass biopsy—a renaissance?. J Urol, 2008; 179:20–27.

20.

Winters

, et al. Cystic renal cell carcinoma carries an excellent prognosis regardless of tumor size. Urol Oncol, 2015; 33:505 e9–e13.

21.

Crispen

, et al. Predicting growth of solid renal masses under active surveillance. Urol Oncol, 2008; 26:555–559.

22.

Volpe

. The role of surveillance in the management of small renal masses. Sci World J, 2007; 7:860–868.

23.

Karlo

, et al. Radiogenomics of clear cell renal cell carcinoma: Associations between CT imaging features and mutations. Radiology, 2014; 270:464–471.

24.

Warshauer

, et al. Detection of renal masses: Sensitivities and specificities of excretory urography/linear tomography, US, and CT. Radiology, 1988; 169:363–365.

25.

Sahni

, Silverman

. Imaging management of incidentally detected small renal masses. Semin Intervent Radiol, 2014; 31:9–19.