How reliable are imaging techniques in aging males with bladder cancer?

Abstract

Introduction

The aim of this study was to evaluate consistency between clinical lymph node positivity and pathological lymph node positivity in patients undergoing open radical cystectomy and pelvic lymph node dissection due to bladder cancer.

Material and method

A total of 135 patients who had open radical cystectomy, extended lymph node dissection, and clear preoperative contrast-enhanced abdominopelvic computed tomography (CT) or magnetic resonance imaging (MRI) images were included in the study. Positive clinical lymph nodes and positive pathological lymph nodeswere recorded. The largest positive clinical and pathological lymph nodeswere recorded. In terms of clinical lymph node involvement, compatibility between radiological findings and pathological results was evaluated.

Results

In the CT group, the sensitivity was 25.81%, specificity was 95.45%, positive predictive value (PPV) was 66.67%, negative predictive value (NPV) was 78.50%, and accuracy was 77.31%. In the MRI group, the sensitivity was 50.00%, specificity was 100%, PPV was 100%, NPV was 76.92%, and accuracy was 81.25%. For consistency between pathological lymph nodes and clinical lymph nodes according to the imaging type, there was no statistically significant difference in the sensitivity, specificity, NPV, and accuracy rates between the imaging techniques (p > 0.05). However, the PPV was significantly higher in the MRI group than the CT group (100% vs.66.67%, respectively; p = 0.014).

Conclusion

Positive lymph nodes play a critical role in the prognosis of patients with bladder cancer and the sensitivity of contrast-enhanced abdominopelvic CT and MRI used routinely in clinical practice is low in lymph node detection. MRI seems more reliable than CT in lymph node detection.

Keywords:

• Bladder cancer
• computed tomography
• magnetic resonance imaging

Introduction

Bladder cancer is the seventh most common cancer in men worldwide, and it is the eleventh most common cancer in both sexes. The incidence of bladder cancer is 9/100,000 per year for men and 2/100,000per year for women [1]. Approximately 25% of bladder cancers invade the bladder muscle [2]. Radical cystectomy combined with pelvic lymph node dissection is the standard treatment modality for localized muscle-invasive bladder cancer (MIBC) [3]. Radical cystectomy provides a significant degree of survival and a significant increase in the quality of life and has an acceptable complication rates with experienced surgeon in elderly patients [4,5].

Clinical studies in recent years have shown that the bladder-sparing approach (radiotherapy and chemotherapy) has comparable survival rates with selected radical cystectomy [5–7]. Therefore, clinical staging before radical cystectomy decision is critical to tailor the treatment algorithm. In clinical practice, contrast-enhanced thoracoabdominal tomography (CT) and magnetic resonance imaging (MRI) are common imaging techniques for clinical staging in MIBC [8]. Local tumor invasion, lymph node involvement, tumor spread to the upper urinary system, and distant metastasis can be evaluated using imaging methods. Detection of clinical lymph-positive disease is important to decide the treatment algorithm for each individual patient.

In the literature, there are studies reporting that extended lymph node dissection has a curative effect and dramatically improves the 5-year survival rates [9,10]. Therefore, preoperative imaging methods are helpful for the surgeon to decide which patient will benefit most from extended lymph node dissection. However, the sensitivity of current imaging methods (CT and MRI) is low in detecting lymph node metastases. In addition, CT and MRI have similar results in detecting lymph node metastasis [11,12]. Although CT and MRI have similar results in lymph node detection, in our patients MRI seems more reliable than CT. Therefore, in the present study, we aimed to evaluate the consistency between clinical lymph node positivity and pathological lymph node positivity and to compare CT and MRI in patients undergoing open radical cystectomy and pelvic lymph node dissection due to bladder cancer.

Materials and methods

Patient selection

In this retrospective study, a total of 146 patients who underwent open radical cystectomy and extended pelvic lymph node dissection due to MIBC or high-risk non-muscle invasive bladder cancer between January 2014 and November 2019 were analyzed. Age, preoperative imaging for clinical staging, and pathology reports of the patients were reviewed. Finally, 135 male patients who had open radical cystectomy combined with extended lymph node dissection and clear preoperative images (contrast-enhanced abdominopelvic CT or MRI) were included in the study. All patients had complete transurethral resection (TUR)-bladder tumor (BT). Patients who received neoadjuvant chemotherapy, laparoscopic radical cystectomy, and preoperative imaging performed in external centers, and those with artifact on CT or MRI images were excluded from the study.

Preoperative clinical staging was performed within one month before the operation. According to contrast-enhanced abdominopelvic CT and MRI examinations, pelvic lymph nodes greater than 8 mm and abdominal lymph nodes greater than 1 cm were considered clinically positive, while pelvic lymph nodes below of <8 mm and abdominal lymph nodes of <1 cm were considered clinically negative lymph nodes. Pathological tumor (pT) staging was performed according to 2017 8th Edition Tumor, Node, Metastasis (TNM) classification [13]. T0 indicates no evidence of tumor; T1 indicates subepithelial connective tissue invasion; T2a superficial muscle tissue involvement; T2b indicates deep muscle tissue involvement; T3a indicates microscopic perivesical adipose tissues; T3b indicates macroscopic perivesical adipose tissues; and T4 indicates prostatic stroma, seminal and vesicle and abdominal or pelvic sidewall involvement. In addition, patients with one lymph node involvement in the pelvis were considered N1, patients with multiple regional lymph node involvement were considered N2, and patients with common iliac lymph node involvement were accepted as N3. Those with distant metastases were considered M1. The total number of lymph nodes removed during the extended lymph node dissection was recorded. Positive clinical lymph nodes and positive pathological lymph nodes were documented. The largest positive clinical and pathological lymph nodes were also noted. In terms of clinical lymph node involvement, compatibility between radiological findings and pathological results was evaluated.

Radiological evaluation

All CT examinations were performed using a 128-detector CT scanner (Optima CT660, GE Healthcare, Chicago, IL, USA). The CT scanning parameters were as follows: tube voltage 120 kV, 100–400 mAs, pitch 1.5, gantry rotation time 0.6 s, and matrix size 512 × 512 with intravenous injection of iodinated contrast agent (Iohexol) using biphasic protocol. After non-enhanced CT scanning, contrast-enhanced CT scanning was performed, and the late portal phase scans were obtained at 50 to 60 s after bolus tracking. The slice thickness used to reconstruct images for retrospective review was 1.3 mm.

MRI examinations were performed with 1.5 T MRI system (Optima MR450w; GE Healtcare Medical Systems, Waukesha, WI, USA). Institutional abdominal MRI protocol, T2-weighted MRI and T2 FS sequence sequence, T1-weighted MRI and T1 FS sequence, transverse diffusion-weighted (DW) imaging with two b-values (0, 800 s/mm²), was acquired. Contrastenhanced scanning was performed when necessary. Data were collected using a picture archiving and communication system, PACS Workstation (GE Medical Systems, Waukesha, WI, USA).

Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy rates of contrast-enhanced abdominopelvic CT or MRI were calculated.

Statistical analysis

Statistical analysis was performed using the Number Cruncher Statistical System (NCSS) version 2007 software (NCSS LLC, Kaysville, UT, USA). Descriptive data were expressed in mean ± standard deviation (SD), median (min–max), or number and frequency. The suitability of quantitative data to normal distribution was tested using the Shapiro–Wilk test and graphical evaluations. For the comparison of qualitative data, the Fisher–Freeman–Halton and McNemartests were used. A value of p < 0.05 was considered statistically significant.

Results

A total of 135 patients were included in this study. Demographic and clinical characteristics of patients are listed in Table 1.

Table 1. Demographic and clinical characteristics of patients.

	n (%)
Age (year)
Min–Max (median)	48–86 (66)
Mean ± SD	65.48 ± 7.82
Tumor size (cm)
Min–Max (median)	0–15 (4)
Mean ± SD	3.87 ± 2.76
Grade
Low grade	12 (8.9)
High grade	123 (91.1)
pT stage
T 0	3 (2.2)
T a	5 (3.7)
T 1	17 (12.6)
T 2a	17 (12.6)
T 2b	16 (11.9)
T 3a	27 (20.0)
T 3b	27 (20.0)
T 4	23 (17.0)
pN stage
N0	98 (72.6)
N1	12 (8.9)
N2	19 (14.1)
N3	6 (4.4)
Histological type
Urothelial ca	129 (95.6)
Sarcomatoid ca	3 (2.2)
Squamous ca	2 (1.5)
Adenocarcinoma	1 (0.7)
Metastasis	3 (2.2)
CIS	47 (34.8)
Surgical margin (+)	17 (12.6)
Perineural invasion	80 (59.3)
Lymphovascular invasion	46 (34.1)
Imaging technique
CT	119 (88.1)
MRI	16 (11.9)

pT: pathological tumor, pN: pathological node, CIS: carcinoma in situ, CT: computed tomography, MRI: magnetic resonance imaging.

The median number of lymph nodes removed was 15 (range, 10–22). The pathological lymph node was negative in 72.6% (n = 98) of the patients and positive in 27.4% (n = 37) of the patients. The mean number of the pathologically positive lymph nodes was 3.81 ± 3.70 (range, 1–15) and the mean lymph node size was 15.32 ± 5.78 (range, 1–30) mm. While the ratio of patients with a clinically negative lymph node was 88.9% (n = 120), 11.1% (n = 15) of the patients were positive. The mean number of clinically positive lymph nodes was 2.66 ± 1.91 (range, 1–7), and the mean lymph node size was 11.40 ± 5.47 (range, 8–30) mm (Table 2).

Table 2. Lymph node characteristics.

	n (%)
Pathological lymph node
Negative	98 (72.6)
Positive	37 (27.4)
No. of pathological lymph nodes (n = 37)
Min–Max (median)	1–15 (2)
Mean ± SD	3.81 ± 3.70
Pathological lymph node size (n = 37)
Min–Max (median)	1–30 (15)
Mean ± SD	15.32 ± 5.78
Clinical lymph node
Negative	120 (88.9)
Positive	15 (11.1)
No. of clinical lymph node (n = 15)
Min–Max (median)	1–7 (2)
Mean ± SD	2.66 ± 1.91
Clinical lymph node size (n = 15)
Min–Max (median)	8–30 (10)
Mean ± SD	11.40 ± 5.47

In the CT group, pathological lymph node was negative in 73.9% (n = 88) and positive in 26.1% (n = 31) of the patients. Clinical lymph node was negative in 89.9% (n = 107) and positive in 10.1% (n = 12) of the patients. Of 88 patients who had a negative pathological lymph node, 84 had also a negative clinical lymph node, while four had a positive clinical lymph node. Of 31 patients with positive pathological lymph node, eight had also a positive clinical lymph node, while 23 had a negative clinical lymph node. There was a statistically significant difference between the clinical lymph node positivity and pathological lymph node positivity; however, there was no consistency between them (p = 0.001 and p < 0.01, respectively). Accordingly, the sensitivity was 25.81%, specificity was 95.45%, PPV was 66.67%, NPV was 78.50%, and accuracy was 77.31% (Table 3).

Table 3: Comparison of pathological lymph nodes and clinical lymph nodes in CT group.

CT group (n = 119)	Pathological lymph node			^ap Value
	Negative	Positive	Total
Clinical lymph node
Negative	84 (70.6)	23 (19.3)	107 (89.9)	0.001**
Positive	4 (3.4)	8 (6.7)	12 (10.1)
Total	88 (73.9)	31 (26.1)	119 (100)
Sensitivity	25.81
Specificity	95.45
PPV.	66.67
NPV.	78.50
Accuracy	77.31

^a McNemartest **p < 0.01.

CT: computed tomography; PPV: positive predictive value; NPV: negative predictive value.

In the MRI group, the pathological lymph node was negative in 62.5% (n = 10) and positive in 37.5% (n = 6) of the patients. In 81.3% (n = 13) of the patients, clinical lymph node was negative, while it was positive in 18.8% (n = 3). All 10 patients with a negative pathological lymph node had also a negative clinical lymph node. Three of six patients with a positive pathological lymph node had also a positive clinical lymph node, and three of them had a negative clinical lymph node. There was no statistically significant difference between the clinical lymph node positivity and pathological lymph node positivity (p > 0.05). Accordingly, the sensitivity was 50.00%, specificity was 100%, PPV was 100%, NPV was 76.92%, and accuracy was 81.25% (Table 4).

Table 4. Comparison of pathological lymph nodes and clinical lymph nodes in MRI group.

MRI group (n = 16)	Pathological lymph node			^ap Value
	Negative	Positive	Total
Clinical lymph node
Negative	10 (62.5)	3 (18.8)	13 (81.3)	0.250
Positive	0 (0)	3 (18.8)	3 (18.8)
Total	10 (62.5)	6 (37.5)	16 (100)
Sensitivity	50.00
Specificity	100.00
PPV	100.00
NPV	76.92
Accuracy	81.25

^a McNemar test.

MRI: magnetic resonance imaging; PPV: positive predictive value; NPV: negative predictive value.

For consistency between pathological lymph nodes and clinical lymph nodes according to the imaging type, there was no statistically significant difference between CT group and MRI group in the sensitivity (25.81% vs.50.00%), specificity (95.45% vs. 100%), NPV (78.50% vs. 76.92%), and accuracy (77.31% vs. 81.25%) rates (p > 0.05). However, the PPV was significantly higher in the MRI group than the CT group (100% vs. 66.67%, respectively; p = 0,014). There was also a statistically significant difference between the positive pathological lymph node rates according to the pT stage (p = 0.001 and p < 0.01, respectively). The positive lymph node rate in the patients with T0, Ta, T1, T2a, and T2b was lower than those with T3b. The positive lymph node rate in the patients with T0, Ta, and T1 was also lower than those with T4. However, there was no statistically significant difference in the positive clinical lymph node rates according to the pT stage (p > 0.05) (Table 5).

Table 5. Pathological lymph node and clinical lymph node evaluation according to pT stage.

	n	Pathological lymph node (+)	Clinical lymph node (+)
pTstage
T 0	3	0 (0)	0 (0)
T a	5	0 (0)	0 (0)
T 1	17	0 (0)	1 (5.9)
T 2a	17	3 (17.6)	0 (0)
T 2b	16	3 (18.8)	1 (6.3)
T 3a	27	7 (25.9)	3 (11.1)
T 3b	27	14 (51.9)	5 (18.5)
T 4	23	10 (47.6)	5 (23.8)
^ap Value		0.001**	0.437

^a Fisher–Freeman–Haltontest **p < 0.01.

pT: pathological tumor.

Discussion

Regional lymph nodes are critical for the spread of pelvic tumors. Nodal metastases are an important prognostic factor in bladder cancer. Nodal metastases often occur in the obturator and internal iliac lymph nodes. In the absence of metastasis in these lymph node groups, it is unlikely to spread to more cranial lymph nodes [14]. In bladder cancer, lymph node involvement adversely affects the 5-year survival rates and prognosis of patients with an increased number of positive lymph nodes and an increased metastatic lymph node size and capsule penetration [15]. Since lymph node involvement is a strong prognostic marker, diagnosing lymph node-positive disease in patients with bladder cancer is of great importance to tailor the treatment algorithm in clinical practice [16]. Preoperative imaging methods are helpful for the surgeon and give information about the local spread of the tumor, lymph node status, and distant metastasis [8].

Detection of lymph node involvement in contrast-enhanced abdominopelvic CT usually depends on the size and shape of the lymph node. Pelvic lymph nodes greater than 8 mm and abdominal lymph nodes greater than 1 cm are considered clinically positive [8]. However, this may result in unnoticed metastatic lymph nodes smaller than 8 mm or over diagnosing reactive lymph nodes greater than 10 mm. In a study by Pichler et al. [17], the cut-off value for metastatic lymph node involvement was 8 mm in contrast-enhanced abdominopelvic CT with a sensitivity and specificity of 45.5% and 91.5%, respectively. In another study by Li et al. [18], the cut-off value for metastatic lymph node involvement was 6.8 mm in contrast-enhanced abdominopelvic CT (the receiver operating characteristic [ROC] curve: 0.815). In the aforementioned study which included 191 patients, 51 of 82 clinically positive lymph nodes were pathologically confirmed. In addition, 51 (28%) of 184 metastatic lymph nodes were detected during imaging studies. In our study, we considered the pelvic lymph nodes greater than 8 mm and abdominal lymph nodes greater than 1 cm clinically positive. We found a statistically significant difference between the clinical lymph node positivity and pathological lymph node positivity; however, there was no consistency between them (p = 0.001 and p < 0.01, respectively). Accordingly, the sensitivity was 25.81%, specificity was 95.45%, PPV was 66.67%,NPV was 78.50%, and accuracy was 77.31%. In our study, the specificity of contrast-enhanced abdominopelvic CT was comparable with the literature; however, the sensitivity was relatively low in our study.

On the contrary, CT has certain limitations in bladder cancer staging. Its major disadvantage is the lack of functional data and heavy reliance on morphology for staging of nodal metastases, which often leads to under staging of cancers with small foci of metastases. Previous studies have shown that the sensitivity of CT in lymph node detection depends on the size of the lymph node and the use of low cut-off values may increase the false-negative results [17,18]. Despite these limitations, CT is still a reliable and cost-effective method for the initial staging and follow-up of patients with MIBC undergoing radical cystectomy.

Contrast-enhanced abdominopelvic MRI is another imaging tool used in detection of clinical lymph node positivity in bladder cancer. It shows pelvic anatomy in more detail and soft tissue contrast resolution is better than CT. Also, diffusion-weighted MRI (DW-MRI) and dynamic contrast-enhanced MRI (DCE-MRI) can provide functional information. In a retrospective study involving 45 patients, van der Pol et al. [19] reported that sensitivity and specificity of T2-weighted imaging (T2WI) + Diffusion-weighted imaging (DWI) apparent diffusion coefficient (ADC) + DCE-MRI sequence in lymph node detection were 45% and 90%, respectively. In a prospective study including 122 patients by Daneshmand et al. [20], the sensitivity and specificity of DCE imaging MRI sequence in clinical lymph node detection were 41% and 92%, respectively. In our study, there was no statistically significant difference between the clinical lymph node positivity and pathological lymph node positivity (p > 0.05) and the sensitivity was 50.00%, specificity was 100%, PPV was 100%, NPV was 76.92%, and accuracy was 81.25%. In our study, the sensitivity and specificity of contrast-enhanced thoraco- abdominal MRI is consistent with the literature; however, we believe that the reason for relatively high sensitivity and specificity is due to the low number of patients with MRI in our study.

Currently, there are several alternatives to conventional gadolinium-based contrast agents. One of these agents is ultra-small superparamagnetic particles of iron oxide (USPIO) [21]. In a prospective study including 75 patients by Birkhauser et al. [22], the sensitivity and specificity of the USPIO-DWI-MRI sequence in detecting lymph node metastasis were found to be 65–75% and 93–96%, respectively. The USPIO seems to have a higher sensitivity in detecting metastatic lymph nodes in bladder cancer, compared to conventional CT and other MRI sequences, although further studies are still needed to establish a definite conclusion.

In addition to CT and MRI, which are frequently used in clinical practice, positron emission tomography (PET)-CT is another imaging method for bladder cancer staging. The most common radiotracers in clinical practice are fluorine-18 2-fluoro-2-deoxy-D-glucose (FDG), C-11 choline, and C-11 acetate. In a study of 61 patients who underwent radical cystectomy and lymph node dissection by Jeong et al. [23], the sensitivity and specificity of FDG PET-CT and conventional CT in bladder cancer were found to be 47% versus 29% and 93% versus 98%, respectively. The authors concluded that the combined FDG PET-CT did not increase the diagnostic accuracy of conventional CT in the detection of lymph nodes in bladder cancer. In a meta-analysis including 14 studies, Ha et al. [24] showed that FDG PET-CT had a low sensitivity and high specificity in the detection of metastatic lymph nodes in bladder cancer. On the contrary, Kim et al. [25] reported that C-11 choline and C-11 acetate PET-CT had also a low sensitivity and moderate specificity in detecting metastatic lymph nodes. Although FDG, C-11 choline, and acetate PET-CT are used for bladder cancer staging, there is still no recommendation for in routine clinical practice due to the lack of adequate clinical data in the current European Urology guidelines [26].

It has been well-established that the lymph node involvement rate increases with increasing pT stage in bladder cancer patients. In a study by Li et al.[16], the rate of lymph node metastasis was the lowest (3.6%) in pT1 patients and the highest (75%) in pT4 patients (p < 0.001). In another study including 583 patients who had radical cystectomy, the lymph node involvement rate was reported as 53.5% in pT4 stage [27]. In our study, a statistically significant difference was found between the positive pathological lymph node rates according to the pT stage (p = 0.001). In the patients with T0, Ta, T1, T2a, and T2b disease, the positive lymph node rate was lower than those with T3b. The positive lymph node rate in the patients with T0, Ta, and T1 was also lower than those with T4. In addition, 47.6% of the pT4 cases had pathological lymph node involvement. On the contrary, there was no statistically significant difference in the positive clinical lymph node rates according to the pT stage (p > 0.05).

Reliable preoperative imaging is crucial in bladder cancer. We compared CT and MRI for lymph node detection and according to our results MRI had more sensitivity and PPV was higher in MRI when compared to CT. It seems that MRI is more reliable than CT. In clinical evaluation and during follow up of bladder cancer, usually more than one imaging is needed. In some patients there is orifice related bladder tumor needed TUR-Orifices and because of that, renal functions can be compromised. In addition, some patients need neo or adjuvant chemotherapy, which is also threat for renal functions. Patients of bladder cancer are usually elderly male patients and contrast- enhanced CT can be problematic in this patient group because of high radiation doses and contrast nephropathy. Although CT is more commonly used modality in staging of bladder cancer as in our study MRI can be considered in elderly patient group since it seems more reliable and safe for renal functions.

Nonetheless, there are some limitations to this study. First, the study has a retrospective design which may have led to bias. In addition, in clinical evaluation, the size of clinical negative lymph nodes was unable to be evaluated. Therefore, the cut-off value was unable to be used for the clinical positive lymph nodes. Also, clinical positive lymph nodes could not be evaluated separately according to the anatomical regions (i.e. internal iliac, obturator, etc.). In the pathological evaluation, the size of the metastatic lymph node foci was unable to be calculated and compared with the clinical lymph node size. Also, extracapsular extension could not be examined in the pathological positive lymph nodes. Although the relationship between the pT stage and clinical and pathological lymph node was evaluated, the relationship between the primary bladder cancer size and lymph node status was unable to be analyzed. The number of patients undergoing MRI examination was relatively low. Finally, the clinical imaging of the patients was evaluated by multiple radiologists and no inter-observer agreement was calculated.

Conclusion

In conclusion, imaging methods are of utmost importance in the detection of lymph node metastasis in bladder cancer patients. Positive lymph nodes play a critical role in the prognosis of patients with bladder cancer and the sensitivity of contrast-enhanced abdominopelvic CT and MRI used routinely in clinical practice is low in lymph node detection. Although in literature CT and MRI have similar results in lymph node detection, in our study MRI seems more reliable. Further large-scale, prospective, clinical studies are still needed to draw a conclusion.

Ethical approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Disclosure statement

No potential conflict of interest was reported by the author(s).