Ahmed El-Assmy, Urology and Nephrology Centre, Mansoura 35516, Egypt. e-mail: email@example.com
Study Type – Diagnostic (exploratory cohort)
Level of Evidence 3a
What's known on the subject? and What does the study add?
Diffusion-weighted (DW) MRI is a non-invasive technique measuring the microscopic mobility of water molecules in the tissues without contrast administration. It provides information on perfusion and diffusion simultaneously in any organ, so it can be used to differentiate normal and abnormal tissue structure, and it might help in the characterization of various abnormalities. In recent years, DW-MRI has been applied in the evaluation of urinary tract lesions, such as malignant renal, prostatic and bladder tumours; however, it has not previously been tested on its ability to distinguish residual cancer from fibrotic and inflammatory changes secondary to transurethral resection (TUR) and intravesical chemotherapy, both of which manifest as bladder-wall thickening on T2-weighted MRI.
This is the first study to show the feasibility of DW-MRI in follow-up of patients with superficial bladder tumours after TUR. DW-MRI was highly reliable in differentiating post-TUR inflammatory changes from bladder tumours, with results similar to those of conventional cystoscopy. This non-invasive method could be used efficiently in future for follow-up of this patient group and may obviate the need for routine cystoscopy.
• To study the feasibility of using diffusion-weighted (DW) magnetic resonance imaging (MRI) in bladder cancer follow-up after transurethral resection (TUR).
PATIENTS AND METHODS
• Included in the study were 47 patients with a history of TUR of superficial bladder carcinoma, who were admitted to our centre between January and December 2011 for follow-up cystoscopy.
• Before cystoscopy, DW-MRI was performed and the apparent diffusion coefficient (ADC) value was measured in a circular region of interest within the carcinoma and normal bladder wall.
• Two radiologists, blinded to the results of cystoscopy, independently interpreted the DW images.
• A comparison of imaging findings with those of cystoscopy was performed using the McNemar test.
• In our 47 patients, cystoscopy identified 34 bladder lesions in 24 patients and in the remaining 23 the bladder looked normal.
• In the 24 patients with malignant bladders, DW-MRI detected 32/34 tumours with two false-negative findings of lesions in two patients.
• In 23 patients with non-malignant bladders, the DW-MRI data were accurate for 21 patients, as two patients were misdiagnosed as malignant.
• The sensitivity, specificity, accuracy, positive and negative predictive values of DW-MRI for identifying bladder tumours were 91.6% (22/24), 91.3% (21/23), 91.5% (43/47), 91.6 (22/24) and 91.3 (21/23), respectively.
• Using the McNemar test there was no significant difference between DW-MRI and cystoscopy.
• DW-MRI has a high reliability in differentiating post-TUR inflammatory changes from bladder tumours, which is similar to that of cystoscopy.
• DW-MRI could be a first-line diagnostic test in follow-up of patients after TUR.
Urinary bladder carcinoma is the second most common tumour of the urinary tract and ∼70% of such carcinomas are non-muscle-invasive (superficial) at presentation. Transurethral resection (TUR) is the standard treatment; however recurrence of the tumour is not uncommon so such patients require long-term close supervision.
Surveillance strategies for urinary bladder carcinoma recurrence have historically relied on cystoscopy, but this procedure has drawbacks, including its high cost, invasiveness and the fact that it may lead to iatrogenic bladder injury and urinary sepsis . In addition, it cannot diagnose upper tract tumours. The presence of a radiological method that could differentiate between benign and malignant lesions of the bladder would avoid the need for invasive cystoscopy, provided that upper tract tumours were excluded.
Diffusion-weighted (DW) MRI is a non-invasive technique measuring the microscopic mobility of water molecules in the tissues without contrast administration. It provides information on perfusion and diffusion simultaneously in any organ, so it can be used to differentiate normal and abnormal structures of tissues, and it might help in the characterization of various abnormalities . DW-MRI is an established method used in the diagnosis of acute stroke  and, in recent years, it has been applied in the evaluation of urinary tract lesions, such as malignant renal [4,5], prostatic [6,7] and bladder tumours [8–16].
During follow-up of patients after transurethral resection (TUR), it is difficult to distinguish residual cancer from fibrotic and inflammatory changes secondary to TUR and intravesical chemotherapy, both of which manifest as bladder wall thickening on T2-weighted MRI . We conducted our prospective study to test the value of DW-MRI as a first-line diagnostic test to differentiate tumours from benign disease of the bladder after TUR of superficial bladder tumours. To the best of our knowledge this is the first study of its kind.
MATERIALS AND METHODS
Our institutional ethical committee reviewed and approved the study protocol. Informed consent was obtained from all patients. We prospectively included 47 consecutive patients who had a history of TUR of superficial bladder tumours and who were admitted to our centre for follow-up cystoscopy between January 2011 and December 2011. The mean (range) follow-up from last TUR was 4.5 (3–8) months. Exclusion criteria included contraindication for MRI (e.g. pacemaker, claustrophobia or metallic prostheses) or cystoscopy (unfit for anaesthesia or urethral stricture). Also excluded were patients who received radio- or chemotherapy and patients refusing consent for the study.
These 47 patients were examined by MRI and subsequently conventional cystoscopy and biopsy. Within 48 h of MRI, conventional cystoscopy and biopsy was done under spinal anaesthesia using semirigid cystoscope by expert urologists who were blind to the results of MRI. Any visible lesions were resected and sent for pathological examination; in cases with no obvious lesions multiple random biopsies were taken.
The study included 41/47 men (87.2%) whose mean (sd; range) age was 60.4 (10.1; 36–81) years and 6/47 women (12.8%) whose mean (sd; range) age was 59 (19; 29–79) years. For the whole group the mean (sd; range) age was 60.2 (11.3; 29–81) years.
No anaesthesia was required for MRI. Patients were instructed to start drinking water 30 mins before the MRI study and arrive for their examination with a full bladder. In two patients with a urethral catheter; 250–400 mL sterile saline was used to distend the bladder. During the imaging procedure, fullness of the bladder was checked at localizer images and the examination was delayed if the bladder was not full.
Patients underwent MRI with a 1.5-T MRI system (SIGMA Horizon, General Electric Medical Systems, Milwaukee, WI, USA). Initially, high spatial resolution T2-weighted images of the bladder were obtained using time to repetition (TR) = 7000–8000 ms, time to echo (TE) = 90–102 ms, band width = 20–83 kHZ, 256 × 256 matrix, slice thickness of 3 mm, intersection gap of 1 mm and field of view = 20 cm.
Then, under free breathing, DW images were obtained using monodirectional gradients, multisection fast spin-echo type, echoplanar sequence in axial plane using a body coil. with; TR = 8000 ms, TE = 61.2 ms, band width = 142 kHZ, 256 × 256 matrix, slice thickness of 5 mm, intersection gap of 0 mm and field of view = 36 cm, seven excitations, water excitation with a b value of 0 and 800 s/mm2. Thirty to 54 slices were obtained in 60–120 s to cover the pelvis in each patient
Images were analysed using functional tool software (General Electric Medical Systems). To nullify inter-observer variability two radiologists (M.A. and H.R.; 11 and 9 years experience reading body MRI scans, respectively), who were blinded to the results of conventional cystoscopy, independently interpreted the DW images. Discrepancies were resolved by consensus. Both reviewers retrospectively reviewed any missed lesions. They also measured the apparent diffusion coefficient (ADC) values of bladder masses and normal bladder wall. Bladder tumours appeared on DW-MRI with a b value of 800 s/mm2 as high signal intensity relative to the bladder wall and the surrounding urine (illustrated as bright lesions). Results were considered to be negative when there was iso-intense bladder wall with no hyperintense bright areas.
The data were processed using SPSS version 16 (SPSS Inc, Chicago, IL, USA) using conventional cystoscopy and the final histopathology as the reference standard, we evaluated the sensitivity, specificity, accuracy, positive and negative predictive values for detection of bladder carcinoma for DW-MRI.
To obviate the increased influence of outliers and non-normal distributions non-parametric tests were used for all hypothesis tests. Comparison of ADC values between carcinomas and bladder wall was performed using the Mann–Whitney U-test. A P value of <0.05 was considered to indicate statistical significance.
To evaluate the performance of the two reviewers in identifying bladder tumours, we applied the κ statistic. A κ value of <0.20 was considered poor, 0.21–0.40 was considered fair, 0.41–0.60 was considered moderate, 0.61–0.80 was considered good, and 0.81–1.00 was considered excellent. A comparison of imaging findings with the results of final histopathology was subsequently performed using the McNemar test; a P value <0.05 was considered to indicate statistical significance.
In our 47 patients, conventional cystoscopy identified bladder masses in 24 patients (51%) and in the remaining 23 (49%) the bladder looked normal. Among 24 patients with bladder tumours; cystoscopy showed 34 lesions. Five patients had multiple lesions: two patients had two lesions, one patient had three lesions and two patients had four lesions. Tumour appearance was papillary 32/34 (94%) and nodular in 2/34 (6%) lesions. Mean (sd; range) tumour length was 16 (14; 2–40) mm. Tumours <10 mm, 10–30 mm and >30 mm were present in 22, seven and five patients, respectively. The locations of the 34 focal bladder lesions included the posterior wall (n= 19 lesions), right lateral wall (n= 6), left lateral wall (n= 5) and anterior wall (n= 4). The pathology from these lesions was transitional cell carcinoma not invading the bladder muscle in all patients except three in whom muscle invasion was found.
On DWI the two reviewers initially agreed on the identification of 23 normal bladders, and 22 malignant bladders. At consensus review, two additional bladders of two patients were considered malignant. Thus, after consensus review, the two reviewers identified 24 malignant and 23 non-malignant bladders. (The agreement between the two readers was excellent: κ= 0.91).
In 24 patients with malignant bladders, DW-MRI detected 32/34 tumours noted on conventional cystoscopy, with two false-negative findings of lesions in two patients (2/24 = 8.3%). All tumours identified were clearly shown on b value = 800 s/mm2 images as bright high signal intensity relative to the bladder wall and the surrounding urine (Figs 1,2). Among the 15 neoplasms <10 mm at conventional cystoscopy, two were not detected on DW-MRI. These two false-negative lesions, confirmed as transitional cell carcinomas, consisted of two polypoid lesions with diameters of 2 mm and 3 mm. These lesions were reviewed in retrospect but the reviewers could not diagnose them.
In 23 patients with non-malignant bladders, the biopsy showed non-specific cystitis in 20 patients, denuded cystitis in one and bilharzial cystitis in the remaining two. The reviewers' interpretations of DW-MRI data agreed with the interpretations of conventional cystoscopic data for 21 patients (Fig. 3). In two patients, normal bladders were misdiagnosed as malignant (2/23 = 8.7%). These two false-positive lesions, confirmed as non-specific cystitis, consisted of two small bright superficial nodules confined to the bladder wall with no muscle invasion (Fig. 4). These lesions were reviewed in retrospect and the reviewers re-diagnosed them as tumours.
At T2-weighted MRI there was irregular thickening of the bladder walls in all patients at the site of previous resections with definite bladder masses >3 cm only noted in seven patients.
Among our 47 patients, using the McNemar test there was no significant difference between DW and cystoscopy (P > 0.05). The agreement between DW-MRI and conventional cystoscopic findings was excellent for identification of malignant and normal bladders (κ= 0.87). The sensitivity, specificity, accuracy, positive and negative predictive values of DW-MRI for identifying bladder tumours were 91.6% (22/24), 91.3% (21/23), 91.5% (43/47), 91.6 (22/24) and 91.3 (21/23), respectively.
The mean (sd; range) ADC values were as follows: carcinomas 1.6 (0.57; 0.4–2.8) × 10−3 mm2/s and bladder wall 1.5 (0.71; 0.2–2.4) × 10−3 mm2/s. ADC values of the carcinomas were similar to those of the bladder wall with no significant difference. In most cases there was an overlap between the ADC values of the tumours and the bladder wall, thus, there seems to be no clear cut-off between the tumours and surrounding bladder wall.
In DW-MRI, the image contrast is influenced by the Brownian motion of water molecules. The signal intensity is high if water molecules are restricted in their motion. The mobility is then quantified by calculating the ADC. Image interpretation can be performed qualitatively by visual assessment of the DW images and the corresponding ADC map, and quantitatively by measuring the ADC value of the lesion .
Malignancies commonly have a larger cell diameter and denser cellularity than normal tissue, which restrict water diffusion and are therefore illustrated as hyperintense (bright) lesions on the high b-value images and hypointense (dark) lesions on the corresponding ADC map (low measured ADC value) .
Diffusion-weighted MRI was first applied in the brain, where it became the ‘gold standard’ for the diagnosis of acute stroke . The clinical use of body DW-MRI was limited owing to physiological motion artifacts (i.e. respiration, cardiac and bowel motion) .
Because of technical advances, including fast sequences, DW-MRI has been applied in the abdomen and pelvis. It has been used for tumour detection and characterization and to monitor treatment response in various organs including the urinary tract [4–16].
The feasibility of using DW-MRI for the detection of urinary bladder carcinoma was tested in several studies. In a study of 15 patients, Matsuki et al.  showed that all carcinomas were clearly shown as high signal intensity relative to the surrounding structure. The sensitivity and positive predictive values of DW-MRI were 100% in terms of correctly detecting the carcinomas. The ADC value of the carcinoma was lower compared with that of the normal bladder wall . Similar results were reported in a second study by El-Assmy et al. .
In a prospective study including 130 patients with gross haematuria, DW-MRI had 98.1% specificity and 92.3% sensitivity in discriminating a total of 106 malignant cases from 14 benign conditions . A smaller study including 59 patients with gross haematuria or under follow-up after previous BCa showed 97.6% specificity and 96% sensitivity in discriminating 34 malignant lesions from nine benign conditions . Notably, in many of these studies, imaging was performed before TUR [9,10]. It is well known that post-TUR inflammatory changes negatively affect conventional morphological imaging .
To the best of our knowledge, this is the first report to test the validity of DW-MRI in differentiating tumour from benign disease of the bladder during follow-up after TUR of superficial bladder tumours. In the present study, excellent agreement was found between DW-MRI and conventional cystoscopy findings. Reviewers were able to identify almost all bladder lesions and missed only two lesions which were <3 mm in diameter. The sensitivity, specificity and accuracy in the diagnosis of bladder tumours were 91.6%, 91.3%, and 91.5%, respectively.
The ADC value has been reported to be useful for quantitatively distinguishing malignancy from benign bladder lesions . Contrary to previous studies, some authors found an overlap between ADC values of malignant and benign bladder lesions, and were unable to determine a threshold for the detection of malignancy [9,13]. In the present study there was an overlap between the ADC values of the tumours and the bladder wall in most of patients. This may be attributable to the presence of inflammatory reaction from previous TUR in the surrounding bladder wall. In addition, it is difficult to define a region of interest on an ADC map when calculating the ADC value for small bladder cancer. In daily clinical practice we think that quantitative analysis of the ADC value is cumbersome and the reproducibility of this technique is questionable; therefore, we believe that qualitative analysis of DW-MRI is more appropriate.
Diffusion-weighted MRI has several advantages in follow-up of patients after TUR for bladder carcinoma. First, it avoids ionizing radiation and may obviate contrast material in those patients with renal insufficiency or contrast material allergy. Second, it can detect upper tract lesions. Third, it can be added to a routine MRI examination protocol taking another few minutes, and can be simply adopted on most current MRI scanners without any additional new equipment.
One disadvantage of DW-MRI includes failure to visualize the lumen of the urethra, as is routinely done with conventional cystoscopy. Another disadvantage is that patients with metal prostheses or implants cannot be adequately examined.
The present study has some limitations. Firstly, the number of patients is small. Secondly, our radiologists have many years of experience and this may limit the generalizability of our imaging test. Thirdly, the value of diffusion MRI in detecting upper tract urothelial tumours was not evident in the present study as all our patients had a normal upper tract. Lastly, the role of diffusion MRI in detecting carcinoma in situ was not evaluated in this series as none of our patients had this. Further studies are required to address these limitations.
In conclusion, this is the first study to show the feasibility of DW-MRI in the follow-up of patients with superficial bladder tumours after TUR. We found that DW-MRI had a high reliability in differentiating post-TUR inflammatory changes from bladder tumours, with results similar to those of conventional cystoscopy. We found that qualitative image analysis is more reliable than qualitative imaging using the ADC value in diagnosing recurrent bladder tumour among the post-resection inflammatory changes of the bladder wall. This non-invasive method could be used efficiently in future for follow-up of this patient group and may obviate the need of routine cystoscopy; however, further studies are required to confirm these results and still we are a long way from having DW-MRI supplant cystoscopy.