Retrospective evaluation of ultrasound-indeterminate renal multilocular cystic masses by using neutrophil-lymphocyte ratio and computed tomography

Urologic Oncology: Seminars and Original Investigations ] (2016) ∎∎∎–∎∎∎

Original article

Retrospective evaluation of ultrasound-indeterminate renal multilocular cystic masses by using neutrophil-lymphocyte ratio and computed tomography Hanfeng Wang, M.D.a,1, Liangyou Gu, M.D.a,1, Rui Jia, M.D.b, Jing Zeng, M.D.c, Xiaoxia Liu, M.D.d, Dawei Zhang, M.D.e, Yongjie Wu, M.D.f, Guangda Luo, M.D.a, Xu Zhang, M.D.a,* a

Department of Urology, State Key Laboratory of Kidney Diseases, Chinese PLA General Hospital, PLA Medical School, Beijing, China b Department of Radiology, Chinese PLA General Hospital, Beijing, China c Institute of Geriatrics, Chinese PLA General Hospital, Beijing, China d Department of Ultrasonics, Chinese PLA General Hospital, Beijing, China e Department of Urology, Chinese PLA 309 Hospital, Beijing, China f Department of General Surgery, Chinese PLA 264 Hospital, Taiyuan, China Received 19 February 2016; received in revised form 27 July 2016; accepted 22 August 2016

Abstract Purpose: To evaluate the clinical usefulness of neutrophil-lymphocyte ratio (NLR) in differentiating the ultrasound-indeterminate renal multilocular cystic masses (RMCM) in comparison with computed tomography (CT) and whether NLR has additional benefits to CT on sensitivity of detecting the malignant. Materials and methods: Overall, 93 patients who underwent normal ultrasound with a conclusion of indeterminate RMCM were examined by NLR and CT within 30 days before surgery or follow-up from March to September 2014 at PLA General Hospital and enrolled in this retrospective study. Logistic regression model was performed to find independent predictors for differentiating true nature of RMCM; differences in the validity parameters and diagnostic power of CT, NLR, and their combination were compared using McNemar tests and AUC model, respectively. Results: The final diagnoses of the 93 patients consisted of 36 patients with benign complex cysts, 16 with multilocular cystic renal cell carcinoma, 9 with multilocular cystic nephroma, and 32 with clear cell renal cell carcinoma. Higher NLR were strongly associated with malignant masses. Multivariate logistic regression analysis revealed that NLR could be an independent predictor for differentiating true nature of these masses (OR ¼ 3.617; 95% CI: 1.219–10.727; P ¼ 0.020). For detecting the malignant masses, the sensitivity, specificity, and accuracy were 71.9%, 80.6%, and 75.3% for CT and 57.9%, 88.9%, and 69.9% for NLR under cutoff value of 2.31, respectively, whereas those of CT þ NLR were 89.5%, 69.4%, and 81.7%. No significant difference was found between CT and NLR in sensitivity (P ¼ 0.185), specificity (P ¼ 0.549), and accuracy (P ¼ 0.428). But the sensitivity of CT þ NLR was significantly higher than those of CT (P ¼ 0.002) and NLR (P o 0.001), respectively; AUC model analysis indicated that CT þ NLR got the largest area of 0.795 (P o 0.001, 95% CI: 0.693–0.896) in comparison with those of CT (area ¼ 0.795, P o 0.001, 95% CI: 0.661–0.864) and NLR (area ¼ 0.734, P o 0.001, 95% CI: 0.631–0.836). Conclusions: Given that NLR, under cutoff value of 2.31, had no diagnostic difference with CT in evaluating the ultrasoundindeterminate RMCM. However, combination of CT and NLR could increase the sensitivity of detecting malignant masses and acquire the best diagnostic power. Prospectively larger cohort and multicenter studies are still necessary. r 2016 Elsevier Inc. All rights reserved.

Keywords: Renal multilocular cystic masses; Ultrasound; Neutrophil-lymphocyte ratio; CT; Diagnosis

1

Equally contributed to this article. Corresponding author. Tel.: þ86-10-6693-8008; fax: þ86-10-6822-3575. E-mail address: [email protected] (X. Zhang). *

http://dx.doi.org/10.1016/j.urolonc.2016.08.011 1078-1439/r 2016 Elsevier Inc. All rights reserved.

2

H. Wang et al. / Urologic Oncology: Seminars and Original Investigations ] (2016) 1–8

1. Introduction Most renal multilocular cystic masses (RMCM) are asymptomatic [1] and are usually found incidentally during regular physical examinations. The Bosniak classification system is routinely used to define these masses through imaging studies [2]. However, this classification system is limited by several anthropic factors and nonanthropic factors in clinical work. For example, in ultrasound tests, differences in the criteria adopted, ages and skills of physicians, and sensitivity levels for recognizing pathological appearances, particularly for similar multilocular complex masses, result in challenges in identifying whether a mass is benign or malignant [3]. Even if computed tomography (CT) can improve diagnostic rate and discriminate parts of RMCM according to their CT values, enhanced, and unenhanced phases, it cannot clearly detect a slight enhancement in low-grade renal cell carcinoma (RCC); some small simple cysts may present pseudoenhancement, and some benign hemorrhagic and inflammatory cysts may exhibit complex appearances resembling those of malignant masses [4–6]. In addition, CT is still unobtainable for most primary community medical facilities. Therefore, another convenient and efficient predictor is imperative to improve the evaluation of the true nature of RMCM and avoid treatment delays. According to Coussens and Werb [7], and Whiteside [8], systemic inflammatory responses are closely related to the carcinogenesis and to the development and prognosis of cancers; they can also be used to independently predict the clinical state and outcome of various neoplasms. In particular, neutrophil-lymphocyte ratio (NLR) has been proposed as a reliable and easily accessible marker for evaluating and differentiating between benign masses and malignant masses, presenting such advantages as low cost and absence of radiation and nephrotoxicity [9,10]. Viers et al. [11] conducted a study involving 2,402 patients with RCC and found that an elevated NLR was associated with RCC pathology, high-grade tumors, and aggressive histological subtypes of cancer and could be useful in predicting malignancy and guiding cancer management. Sejima et al. [12] also found that the NLRs of the late recurrence group were significantly lower than those of the early recurrence group, indicating that NLRs were higher among patients with advanced RCC; they also found that patients with low NLRs tended to have a late recurrence and relatively longer survival. However, the study of Viers et al. [11] was unable to accurately rule out other influential factors that could significantly affect the inflammatory profiles of the patients. Moreover, only few studies have directly compared the diagnostic difference among NLR, CT, and even their combination, which could provide information of RMCM from totally different perspectives and help us know and diagnose these masses more accurately. Therefore, the aim of our study was to assess the clinical usefulness of NLR in differentiating the ultrasoundindeterminate RMCM in comparison with CT and whether

NLR has additional benefits to CT on sensitivity of detecting the malignant.

2. Materials and methods 2.1. Patients This study is a single-institution retrospective study approved by the institutional review board of PLA General Hospital. In this study, the following selection criteria were adopted: (1) the patient was RMCM-specific; (2) the lesion of patient was incidentally found by physical examination, or the patient had focal malaise around the kidneys; (3) the patient had no acute or chronic inflammation in any system within latest 2 weeks; (4) the patient had not taken antibiotics within latest 2 weeks; (5) the patient had no immune diseases; and (6) the patient had no hematological diseases. We searched the radiological and pathological databases and medical records from the Urology Department of the PLA General Hospital to identify 122 patients with original indeterminate diagnoses on ultrasound findings. Final diagnosis obtained by histology or contrast enhanced ultrasound plus CT with a follow-up at least 12 months. However, 1 patient with missing pathological slide, 5 patients without contrast-enhanced CT images, 2 patients with acute nonspecific inflammation, 2 patients with hematologic system diseases, 6 patients without regular blood tests, and 13 patients without complete follow-up and conclusive diagnoses were excluded from this study. Therefore, only 93 patients were enrolled and all of them have sporadic, unilateral masses; regular blood tests and CT examinations were done within 30 days before partial nephrectomy or followup. All regular blood tests and CT results were provided by a single laboratory and department of radiology, respectively. The pathological features of the patients were studied according to the staging system of the American Joint Committee on Cancer [13] and reviewed by a single pathologist/genitourinary pathology specialist who was blinded to the radiological outcomes and NLR results.

2.2. Imaging techniques 2.2.1. Ultrasound technique Ultrasound was performed with a Philips iU Elite or iU22 ultrasound machine. The baseline of the kidneys was presented in the fundamental mode and in grayscale with a multifrequency C5-1 convex array probe with harmonics (scanning parameters: 3–5 MHz; MI ¼ 1.1; TIS ¼ 1.2) to identify the nodules. Color Doppler examination (scanning parameters: color frequency ¼ 2.5 MHz, color gain ¼ 80%) was performed to evaluate intratumoral vascularity.

H. Wang et al. / Urologic Oncology: Seminars and Original Investigations ] (2016) 1–8

2.2.2. CT technique CT was performed with a 16-MDCT scanner (Sensation 16, Siemens Healthcare) and a 64-MDCT scanner (Sensation 64, Siemens). The following scanning parameters were adopted: 120 kVp, 200 to 300 mA, section thickness ¼ 5 mm, rotation time ¼ 0.8 second, and reconstruction kernel ¼ B40. Images were obtained using a renal mass protocol that included unenhanced images followed by nephrographic phase images captured 120 s after IV administration of a nonionic contrast medium (iohexol 300 mg I/ml, Omnipaque, GE Healthcare) at 5 ml/s and volume of 2 ml/kg. 2.3. Image analysis Two radiologists separately and independently reviewed the ultrasound and CT images. They were blinded to the clinical histories, histological results, and original conclusions. For the ultrasound, anechoic masses with hairlinethin walls and septa (r1 mm) without color Doppler signals were considered benign. Multilocular cystic masses were considered indeterminate if they satisfied the following criteria: (1) complex cyst with homogeneous content and irregular wall or septum and without a color Doppler signal; (2) complex cyst with a thickened septum or wall or solid mural nodules and without apparent color Doppler signal (Bosniak IIF and III); (3) cyst with intracystic hyperechoic/heterogeneous content and without a color Doppler signal. The remaining masses with solid components or heterogeneous content, thickened walls and septa, apparent color Doppler signals, and multicalcification were defined as potentially malignant. For CT, the characteristic features of benign masses are as follows: masses with cystic density (CT attenuation approximately equal to 15 HU) and smooth internal septa or peripheral rims of calcification in the walls or septa without enhancement or more calcium in the walls or thicker septa, and hyperdense content without enhancement. Masses were considered potentially malignant if they are complex cystic structures with irregular and thickened septa with slight enhancement, irregular calcification, heterogeneous content, nonuniform wall or septum thickening, and heterogeneous content with obvious enhancement.

3

corresponded to the maximum value of sensitivity plus specificity. Differences in sensitivity, specificity, positive predictive value, negative predictive value, and accuracy between any 2 methods comprised by NLR, CT, and CT þ NLR were calculated for diagnosing the malignant RMCM by the McNemar test. We defined that the group of CT þ NLR was positive when either CT results or NLR outcomes indicated the masses were malignant. To evaluate the diagnostic power, AUC model analyses were used. All statistical analyses were performed using the Statistical Package for Social Sciences (SPSS) version 22.0 (SPSS Inc., Chicago, USA). Two-sided P o 0.05 was considered statistically significant.

3. Results 3.1. Baseline characteristics Table 1 shows the clinical and pathological characteristics of the enrolled patients. Among the 93 patients, 36 had benign complex cysts (8 patients were diagnosed by histology and the remaining 28 were diagnosed during a follow-up after 12 months by CEUS plus CT), and 57 had malignant cysts (16 patients had multilocular cystic RCC, 9 had multilocular cystic nephroma, and 32 had clear cell RCC) and were diagnosed by histology. In terms of surgical approach, the patients except for 28 patients with follow-up underwent partial nephrectomy. The mean age of all the patients was 51.66 ⫾ 13.378 years. The mean size of all the masses was 4.77 ⫾ 1.82 cm, with the minimum being 2 cm and the maximum being 14 cm. Among the patients, 55 (59.1%) were men. The following variables were similar between 2 groups: age distribution (P ¼ 0.590), sex (P ¼ 0.459), site (P ¼ 0.576), size (P ¼ 0.391), smoking history (P ¼ 0.334), location (P ¼ 0.565), diabetes (P ¼ 0.735), presentation (P ¼ 0.803), BMI (P ¼ 0.107), neutrophil count (P ¼ 0.197), and lymphocyte count (P ¼ 0.861). However, the NLRs of malignant masses were significantly higher than those of benign masses (P ¼ 0.001). The cutoff NLR value was 2.31 with AUC ¼ 0.623 (corresponding sensitivity of 0.456 and specificity of 0.833) (Fig. 1).

2.4. Statistical analysis All continuous data with normal distribution were presented as means ⫾ standard deviation (SD), and the median and interquartile range was used under nonnormal distribution. Student's t-test and chi-square test were used to compare continuous and categorical variables, respectively. The NLR was calculated by the formula of NLR ¼ neutrophil count/lymphocyte count. Univariate and multivariate logistic regression analyses were performed to identify independent potential predictors for differentiating true nature of RMCM. We used receiver operating characteristic curve to calculate the cutoff value of NLR which

3.2. Univariate and multivariate logistic regression analyses The clinicopathological variables for logistic regression included age, sex, size, NLR, BMI, and smoking history. The univariate regression analysis revealed that NLR (P ¼ 0.046) and BMI (P ¼ 0.045) were statistically significant. Therefore, both NLR and BMI entered into multivariate regression analysis, which finally demonstrated that only NLR (OR ¼ 3.617; 95% CI: 1.219–10.726; P ¼ 0.045) could be the independent predictor of

H. Wang et al. / Urologic Oncology: Seminars and Original Investigations ] (2016) 1–8

4

Table 1 Patient demographics and pathological characteristics Variables

Benign masses n ¼ 36

Malignant masses n ¼ 57

P value

Age, year, mean ⫾ SD Sex, no. (%) Men Women

51.25 ⫾ 13.75

51.91 ⫾ 13.27

0.590 0.459

23 (63.9) 13 (36.1)

32(56.1) 25 (43.9)

16 (44.4) 20 (55.6)

22 (38.6) 35 (61.4)

Mass size, cm, mean ⫾ SD Neutrophil count, 109/l, mean ⫾ SD Lymphocyte count, 109/l, mean ⫾ SD Neutrophil count/lymphocyte count (NLR), % Smoking history (%) Yes No

4.49 ⫾ 1.46 0.597 ⫾ 0.066 0.315 ⫾ 0.062 1.86 ⫾ 0.43

4.94 ⫾ 2.01 0.607 ⫾ 0.073 0.294 ⫾ 0.058 2.43 ⫾ 0.75

14 (38.9) 22 (61.1)

28 (49.1) 29 (50.9)

Body mass index (BMI), kg/m2, mean ⫾ SD Presentation, no. (%) Incidental Symptomatic

23.240 ⫾ 2.285

23.947 ⫾ 2.567

31 (86.1) 5 (13.9)

48 (84.2) 9 (15.8)

Diabetes, no. (%) Absent Present

33 (91.7) 3 (8.3)

50 (87.7) 7 (12.3)

Location, no. (%) Lower Medium Upper

14 (38.9) 15 (41.7) 7 (19.4)

19 (33.3) 30 (52.6) 8 (14.1)

Mass site, no. (%) Left Right

0.576

0.391 0.197 0.861 0.001 0.334

0.107 0.803

0.735

0.565

discriminating the nature of the ultrasound-indeterminate RMCM in our study (Table 2).

3.3. Diagnostic performances of NLR, CT, and CT þ NLR

Area 0.623

P 0.046

95% C.I. 0.510-0.736

Fig. 1. ROC curve for calculating cutoff value of NLR (with corresponding sensitivity of 0.456 and specificity of 0.833). ROC ¼ receiver operating characteristic. (Color version of figure is available online.)

For evaluating the true nature of RMCM, as highlighted in Tables 3 and 4, CT þ NLR got a significantly higher sensitivity of 51/57 (89.5%) in comparison with that of 41/57 (71.9%) from CT (P ¼ 0.002) and that of 33/57 (57.9%) form NLR (P o 0.001), whereas there was no statistical difference on sensitivity between CT and NLR (P ¼ 0.185); NLR showed the highest specificity of 32/36 (88.9%), whereas CT had a specificity of 29/36 (80.6%) (P ¼ 0.549) and CT þ NLR got a specificity of 25/36 (P ¼ 0.016). The overall accuracy of NLR, CT, and CT þ NLR were 65/93 (69.9%), 70/93 (75.3%), and 76/93 (81.7%), respectively. Except the comparison group between CT þ NLR and NLR (P ¼ 0.047), no significant differences between group of CT with NLR (P ¼ 0.428) and group of CT þ NLR with CT (P ¼ 0.228) on accuracy. Positive predictive value and NPR were 85.4% and 64.4% for CT, 89.1% and 57.1% for NLR, 82.3% and 80.6% for CT þ NLR accordingly.

H. Wang et al. / Urologic Oncology: Seminars and Original Investigations ] (2016) 1–8

5

Table 2 Univariate and multivariate logistic regression of clinicopathological variables Variables

Univariate P value

Age, year o50 Z50, o70 Z70

Multivariate OR

95% CI for EXP (B)

P value

Sex Size, cm r4 44, r7 47

0.979 0.766

NLR, (%) o2.31 Z2.31

0.046

BMI, kg/m2 o18.5 Z18.5, o25 Z25

0.045

Smoking history

0.467 0.8

95% CI for EXP (B)

1 (Referent) 3.617

1.219–10.726

–

0.901 0.718 0.713

OR

1 (Referent) 1.198 1.374

0.449–3.193 0.252–7.497

0.986

0.329–2.954

1 (Referent) 1.431 1.283

0.544–3.764 0.187–8.784

– –

0.020 1 (Referent) 3.555

1.022–12.365 0.061

0.511 0.083

1 (Referent) 2.125 8.93

0.224–20.127 0.750–106.351

0.152

2.193

0.748–6.425

Given the AUC model analysis revealed that CT þ NLR showed the largest area of 0.795 (P o 0.001, 95% CI: 0.693–0.896) in comparison with that of 0.762 (P o 0.001, 95% CI: 0.661–0.864) for CT and that of 0.734 (P o0.001, 95% CI: 0.631–0.836) for NLR (Fig. 2).

4. Discussion The low-accuracy rate of ultrasound in this study could be explained by the atypical characteristics of benign and malignant multilocular cystic masses and the lack of more precise images of color Doppler signals. Nearly half of the cysts with echogenic content and thin or thickened walls and septa without color Doppler signals were classified as indeterminate. Thus, under this condition, further examination, such as CT, is necessary. Under normal conditions, detection of renal masses by CT usually requires the performance of unenhanced and enhanced phases, given that pseudoenhancement is exhibited in some small cysts and slight or even insignificant enhancement is

–

demonstrated on the walls and septa of multilocular cystic masses. Moreover, hemorrhagic and infectious cysts also exhibit complex appearances resembling those of malignant cysts, thereby increasing the difficulty in evaluating the enhancement [1,5,14,15]. Therefore, CT still possesses certain limitations in distinguishing between benign and malignant masses and reaching a satisfactory accuracy rate and acceptable diagnostic rate. Studies on the link between cancer and inflammation reported that numerous inflammatory cells, which mainly indicate an inflammatory response, gathered in and around the tumors as confirmed by pathological examinations; inflammation was considered a key contributor to carcinogenesis among several kinds of cancers [10,16]. Inflammatory cells could damage normal cells and tissues both directly and indirectly by generating and releasing inflammatory agents. Inversely, tumor cells could also produce several kinds of proinflammatory cytokines and chemokines, such as TNF-α, IL-6, IL-12, IL-10, TGF-β, CXC, and CC, which induce inflammatory cells to infiltrate and surround tumor tissues, subsequently forming a vicious

Table 3 Results of each of 3 methods (CT, NLR, and CT þ NLR) for estimating the true nature of multilocular cystic masses Pathology

Malignant Benign Total

CT

Total

Potentially malignant

Benign

41 7 48

16 29 45

57 36 93

NLR

Total

Potentially malignant

Benign

33 4 37

24 32 56

57 36 93

CTþNLR

Total

Potentially malignant

Benign

51 11 62

6 25 31

57 36 93

H. Wang et al. / Urologic Oncology: Seminars and Original Investigations ] (2016) 1–8

6

Table 4 Diagnostic performances of each of 3 methods (CT, NLR, and CT þ NLR) for detection of the malignant masses Comparison group 1

Sensitivity Specificity PPV NPV Accuracy

Comparison group 2

Comparison group 3

CT, %

NLR, %

P value

CT þ NLR, %

CT, %

P value

CT þ NLR, %

NLR, %

P value

71.9 80.6 85.4 64.4 75.3

57.9 88.9 89.1 57.1 69.9

0.185 0.549

89.5 69.4 82.3 80.6 81.7

71.9 80.6 85.4 64.4 75.3

0.002 0.125

89.5 69.4 82.3 80.6 81.7

57.9 88.9 89.1 57.1 69.9

o0.001 0.016

0.428

0.228

0.047

PPV ¼ positive predictive value; NPV ¼ negative predictive value.

cycle [14,17]. A study conducted by Jensen et al. [18] suggested that an elevated amount of tumor-infiltrating neutrophils is related to poor RCC outcomes, and lymphocytes, which form a class of leukocytes, could contribute to cancer occurrence, growth, and even spread, as well as immunosuppression, which is associated with malignant masses. CD4þCD25þFoxp3 T cells, a particular set of Treg cells, and activated helper T cells are significantly associated with survival among patients with RCC [19]. According to the study of Liotta et al. [20], PD-L1, also known as B7-H1 and CD274, is significantly up-regulated in tumor tissues but not in normal tissues after exposure to IFN-γ, which is a typical proinflammatory agent released by T cells. With the combination of its ligand PD-1, cytotoxic T cells were mediated to promote lysis, enhance tumor cells, and escape from immune surveillance; high PD-1 expressed in the tissue of advanced RCC was verified to be

Variables NLR CT CT+NLR

Area 0.734 0.762 0.795

P <0.001 <0.001 <0.001

95% C.I. 0.631-0.836 0.661-0.864 0.693-0.896

Fig. 2. AUC model analysis of 3 methods (NLR, CT, and CT þ NLR). (Color version of figure is available online.)

associated with poor outcomes [21]. Moreover, proinflammatory cytokines and chemokines could enhance the catabolism of tumors and reactively up-regulate the level of metabolism of the entire body, leading to an irreversible loss of nutrition, accelerating the progress of cachexia, and ultimately decreasing survival rate [22]. As a systemic inflammatory index, NLR obtained from peripheral blood could reactively represent the status of the inflammatory microenvironment within and around the solid tumor. NLR has drawn much attention owing to its advantages of convenience, low cost, and absence of radiation. The worse the tumor differentiation is, the more severe the inflammatory response within and around tumors [23]. Elevated NLR strongly indicates poor outcomes for several malignant tumors, and it can serve as a biomarker to predict and discriminate benign from malignant masses, particularly tumors in the digestive, respiratory, gynecological, and urinary systems [24]. The study of Viers et al. [11] reported significant differences among the NLRs of RCC subtypes, with cystic clear cell RCC having the lowest NLR and the collecting-duct RCC having the highest NLR. Furthermore, the difference between NLRs of benign and malignant renal masses is statistically significant. A large European cohort study of 678 patients with RCC indicated that an increased NLR could represent an independent risk factor for prognosis and suggested that adding NLR to fullfledged prognostic models could improve the predictive power of these models [25]. A meta-analysis conducted by Hu et al. [26] demonstrated that NLR should be monitored in RCC for rational risk stratification and treatment individualization. Moreover, previous studies demonstrated the value of NLR in differentiating between benign and malignant thyroid disorders and its capability of serving as a biomarker for colorectal cancer in screening the general population and for postoperative follow-up [9,27]. Our study extended the application of NLR to renal masses, particularly to RMCM. We found that significantly elevated NLRs were associated with malignant RMCM. We also assessed the diagnostic performances of NLR in evaluating the true nature of these masses and found that it had no statistical difference in comparison with those of CT. As CT and NLR were totally different examinations that possessed different advantages and perspectives in providing information of RMCM to help us diagnose these

H. Wang et al. / Urologic Oncology: Seminars and Original Investigations ] (2016) 1–8

kind of masses more accurately. In this study, CT showed a relatively high sensitivity and NLR showed a relatively high specificity. Accordingly, we were trying to combine them together and found that NLR could significantly improve the sensitivity of CT after their combination. Besides, CT þ NLR showed the best diagnostic power in comparison with CT and NLR alone. Undoubtedly, this kind of combination provide us a new perspective to treat RMCM with less vascularization, more liquids, less blood signal on US and nonapparent enhanced phase on CT, which resembled multilocular cysts during our clinical work. As with all retrospective studies, our study have some limitations, such as incomplete retrospective data collection, single center, small cohort size because of the extremely low morbidity in some types of RCC, and design defects; thus, the applicability of our conclusions remains restricted. Although higher NLR is strongly associated with malignant masses, the specificity of NLR has the feature of easy to be influenced by active infection, inflammatory diseases, smoking behavior, or stress at the time of blood collection [28]. In this study, we were trying to rule out factors that would affect the real status of NLR as many as possible. Our result on smoking behavior conflicts with that of Alkhouri et al. [28], which suggested that more stratification models for comparison were necessary. Moreover, in the study of Bazzi et al. [29], 1,004 patients with r4 cm small renal masses were included, and the authors found no association between increased NLR and malignancy risk or differentiation among RCC histological subtypes. However, the finding of the current study on the NLRs of r4 cm multilocular renal cystic masses was opposite to that of the study of Bazzi et al. [29]; such a deviation may be caused by comparison of benign multilocular cysts and malignant RMCM with more vascularized, inflammatory cell infiltration and inflammatory responses; as a result, the specificity of NLR was significantly promoted. Compared with several previous studies on the prognostic value of NLR, our cohort size is relatively smaller, resulting in a possible selection bias and sample errors. Furthermore, the histological diagnosis of benign multilocular cystic masses was not obtained, even with a follow-up with at least 12 months. Some RCC types may demonstrate an extremely slow growth; thus, a follow-up after at least 3 years is recommended to rule out malignancy. Therefore, multicenter studies and prospectively large cohort are still necessary to validate our findings.

5. Conclusion Our study demonstrated that NLR could be a potential independent predictor for evaluating ultrasoundindeterminate RMCM. It had no statistical difference in comparison with CT. However, combination of CT and NLR could increase the sensitivity of detecting malignant masses significantly and acquire the best diagnostic power,

7

which provided a new perspective for us to differentiate ultrasound-indeterminate RMCM more efficiently and accurately. Prospectively, larger cohort and multicenter studies are still necessary. Acknowledgments Statistical assistance was provided by Peng Duan (Department of Cardiovascular, Chinese PLA General Hospital, Beijing, China) and Dan Feng (Hospital Management Institute, Medical Statistic Division, Chinese PLA General Hospital, Beijing, China). Pathologic review was provided by Xiaocong Wang (Department of Pathology, Chinese PLA General Hospital, Beijing, China). Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/ j.urolonc.2016.08.011.

References [1] Hindman NM, Bosniak MA, Rosenkrantz AB, Lee-Felker S, Melamed J. Multilocular cystic renal cell carcinoma: comparison of imaging and pathologic findings. AJR Am J Roentgenol 2012;198: W20–6. [2] Bosniak MA. The use of the Bosniak classification system for renal cysts and cystic tumors. J Urol 1997;157:1852–3. [3] Suzigan S, Lopez-Beltran A, Montironi R, Drut R, Romero A, Hayashi T, et al. Multilocular cystic renal cell carcinoma: a report of 45 cases of a kidney tumor of low malignant potential. Am J Clin Pathol 2006;125:217–22. [4] Egbert ND, Caoili EM, Cohan RH, Davenport MS, Francis IR, Kunju LP, et al. Differentiation of papillary renal cell carcinoma subtypes on CT and MRI. Am J Roentgenol 2013;201:347–55. [5] Tappouni R, Kissane J, Sarwani N, Lehman EB. Pseudoenhancement of renal cysts: influence of lesion size, lesion location, slice thickness, and number of MDCT detectors. Am J Roentgenol 2012;198:133–7. [6] Nicolau C, Bunesch L, Pano B, Salvador R, Ribal MJ, Mallofre C, et al. Prospective evaluation of CT indeterminate renal masses using US and contrast-enhanced ultrasound. Abdom Imaging 2015; 40:542–51. [7] Coussens LM, Werb Z. Inflammation and cancer. Nature 2002; 420:860–7. [8] Whiteside TL. The tumor microenvironment and its role in promoting tumor growth. Oncogene 2008;27:5904–12. [9] Kocer D, Karakukcu C, Karaman H, Gokay F, Bayram F. May the neutrophil/lymphocyte ratio be a predictor in the differentiation of different thyroid disorders? Asian Pac J Cancer Prev 2015;16:3875–9. [10] Yildirim M, Demir Cendek B, Filiz Avsar A. Differentiation between benign and malignant ovarian masses in the preoperative period using neutrophil-to-lymphocyte and platelet-to-lymphocyte ratios. Mol Clin Oncol 2015;3:317–21. [11] Viers BR, Thompson RH, Lohse CM, Cheville JC, Leibovich BC, Boorjian SA, et al. Pre-treatment neutrophil-to-lymphocyte ratio predicts tumor pathology in newly diagnosed renal tumors. World J Urol 2016. [12] Sejima T, Iwamoto H, Morizane S, Hinata N, Yao A, Isoyama T, et al. The significant immunological characteristics of peripheral

8

[13]

[14] [15]

[16] [17] [18]

[19]

[20]

[21]

H. Wang et al. / Urologic Oncology: Seminars and Original Investigations ] (2016) 1–8 blood neutrophil-to-lymphocyte ratio and Fas ligand expression incidence in nephrectomized tumor in late recurrence from renal cell carcinoma. Urol Oncol 2013;31:1343–9. Edge SB, Compton CC. The American Joint Committee on Cancer: the 7th edition of the AJCC cancer staging manual and the future of TNM. Ann Surg Oncol 2010;17:1471–4. Balkwill F, Mantovani A. Inflammation and cancer: back to Virchow? Lancet 2001;357:539–45. Johnson PT, Horton KM, Fishman EK. How not to miss or mischaracterize a renal cell carcinoma: protocols, pearls, and pitfalls. AJR Am J Roentgenol 2010;194:W307–15. Grivennikov SI, Greten FR, Karin M. Immunity, inflammation, and cancer. Cell 2010;140:883–99. Candido J, Hagemann T. Cancer-related inflammation. J Clin Immunol 2013;33(Suppl. 1):S79–84. Jensen HK, Donskov F, Marcussen N, Nordsmark M, Lundbeck F, von der Maase H. Presence of intratumoral neutrophils is an independent prognostic factor in localized renal cell carcinoma. J Clin Oncol 2009;27:4709–17. Siddiqui SA, Frigola X, Bonne-Annee S, Mercader M, Kuntz SM, Krambeck AE, et al. Tumor-infiltrating Foxp3CD4þ CD25þ T cells predict poor survival in renal cell carcinoma. Clin Cancer Res 2007;13:2075–81. Liotta F, Gacci M, Frosali F, Querci V, Vittori G, Lapini A, et al. Frequency of regulatory T cells in peripheral blood and in tumourinfiltrating lymphocytes correlates with poor prognosis in renal cell carcinoma. BJU Int 2011;107:1500–6. Thompson RH, Dong H, Lohse CM, Leibovich BC, Blute ML, Cheville JC, et al. PD-1 is expressed by tumor-infiltrating immune

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

cells and is associated with poor outcome for patients with renal cell carcinoma. Clin Cancer Res 2007;13:1757–61. Hasselgren PO, Fischer JE. Muscle cachexia: current concepts of intracellular mechanisms and molecular regulation. Ann Surg 2001;233:9–17. Cho JS, Park MH, Ryu YJ, Yoon JH. The neutrophil to lymphocyte ratio can discriminate anaplastic thyroid cancer against poorly or well differentiated cancer. Ann Surg Treat Res 2015;88:187–92. Roxburgh CS, McMillan DC. Role of systemic inflammatory response in predicting survival in patients with primary operable cancer. Future Oncol 2010;6:149–63. Pichler M, Hutterer GC, Stoeckigt C, Chromecki TF, Stojakovic T, Golbeck S, et al. Validation of the pre-treatment neutrophil-lymphocyte ratio as a prognostic factor in a large European cohort of renal cell carcinoma patients. Br J Cancer 2013;108:901–7. Hu K, Lou L, Ye J, Zhang S. Prognostic role of the neutrophillymphocyte ratio in renal cell carcinoma: a meta-analysis. BMJ Open 2015;5:e006404. Kilincalp S, Coban S, Akinci H, Hamamci M, Karaahmet F, Coskun Y, et al. Neutrophil/lymphocyte ratio, platelet/lymphocyte ratio, and mean platelet volume as potential biomarkers for early detection and monitoring of colorectal adenocarcinoma. Eur J Cancer Prev 2015; 24:328–33. Alkhouri N, Morris-Stiff G, Campbell C, Lopez R, Tamimi TA, Yerian L, et al. Neutrophil to lymphocyte ratio: a new marker for predicting steatohepatitis and fibrosis in patients with nonalcoholic fatty liver disease. Liver Int 2012;32:297–302. Bazzi WM, Dejbakhsh SZ, Bernstein M, Russo P. Neutrophil-lymphocyte ratio in small renal masses. ISRN Urol 2014;2014:759253.