British Journal of Radiology (2005) 78, 528-532
© 2005 British Institute of Radiology
doi: 10.1259/bjr/82990907
Comparative analyses of the dynamic properties of the bladder wall studied by repetitive pelvic CT scans of patients and cryo-sections of cadavers
E Dale, PhD1,
T P Hellebust, MSc1,2,
Ø S Bruland, MD, PhD1,3 and
D R Olsen, PhD1,2
1 Centre for Training and Research in Radiotherapy, 2 Department of Medical Physics and 3 Department of Oncology, The Norwegian Radium Hospital, University of Oslo, Box 20, N-0310 Oslo, Norway
 |
Abstract
|
|---|
In radiotherapy planning systems, delineation of hollow normal tissue organs, such as the bladder, is time-consuming. Automated delineation may presuppose two assumptions: (1) the bladder resembles a spherical shell and (2) the volume of bladder tissue is preserved regardless of the volume of urine (luminal volume) inside. The purpose of the present study was to test these assumptions. 22 CT scans from 7 patients were studied retrospectively. Transverse cross-sectional areas enclosed by the outer contour (Aout) and inner contour of the bladder (Ain) were recorded from the images. Hence, the transverse cross-sectional area of the wall, Awall=Aout–Ain, and the volume of bladder tissue at various luminal volumes, could be calculated. To quantify the method uncertainty, the same procedure was applied on three spherical plastic phantoms. The results were also compared with data from the Visible Human Project's photographs of cadaver cryo-sections. Assumption no. 1 stated above, implies that Awall is constant regardless of the level of intersection of the sphere. The data from cryo-sections revealed a positive correlation for Awall and Aout, in contradiction to assumption no. 1 (p<0.001). The corresponding association derived from the repetitive CT scans of patients was also statistically significant (p<0.001) although linear regression revealed a less steep slope. A relationship was found between the volume of bladder tissue and luminal volume, hence contradicting assumption no. 2 (p<0.001). In conclusion the cross-sectional wall areas of the bladder, measured from patient CT scans, increase slightly with luminal cross-sectional areas in contradiction to expected values derived from a simplistic spherical shell model. In addition, the volume of bladder tissue is related to the luminal volume. Our results may be of practical value when developing automated delineation tools in radiotherapy planning systems.
|
Introduction
|
|---|
In modern three-dimensional (3D) radiotherapy planning, based on serial CT scans, accurate definition of normal tissue organs surrounding a tumour, as well as the tumour itself, is of paramount importance. Dose–volume histograms (DVH) are obtained from the CT volumes by delineating contours around the involved structures, and calculating the dose to each subvolume within the contour. The DVHs can be used to assess the normal tissue complication probability (NTCP) and tumour control probability (TCP) associated with the dose plan in question [1]. Organs of the human body that are hollow represent special challenges. The materials filling hollow organs are irrelevant from a clinical point of view and should be excluded from the DVH [2]. The luminal subvolumes may be removed from the DVH by delineating both an outer contour enclosing the external organ border as well as an inner contour demarcating the wall tissue from the lumen. Only subvolumes between the outer and inner contours are included in the DVH calculations. However, one has to delineate the double number of contours. This makes it a time-consuming task to obtain the DVH of a hollow organ structure. Meijer et al have reported a strategy to delineate the rectal inner contours automatically, based on manually drawn outer contours [3]. They proposed a cylindrical shell model for the rectum with an elastic wall undergoing changes in shape and size as a response to variable local luminal volume. The cylinder comprised thin transverse slices, each assumed to have the same constant wall volume regardless of luminal volume. Hence, knowing the location of the outer contour only, enabled the authors to calculate the position of the inner contour assuming a constant slice (annulus) volume [3]. In a recent study, Dale et al found that the rectal wall volume has a more complex dependence with luminal volume [4]: the rectal wall volume within a fixed longitudinal length of the rectum was not constant but increased with luminal volume.
The present study expands on this observation investigating another important dose-limiting hollow organ in pelvic radiotherapy, namely the urinary bladder. Located retroperitoneally and attached to the musculature of the pelvic floor, the bladder may expand relatively freely towards the peritoneal cavity with increased luminal volume. The shape of the bladder varies with the volume of urine inside, resembling a spherical shell when distended, but more like a tetrahedron with rounded edges at smaller volumes of urine [5].
The aim of the present study was to investigate whether (1) a spherical shell model may adequately model the bladder's shape and (2) the total bladder tissue volume is preserved regardless of total luminal volume. For this purpose, repetitive CT examinations of cervix cancer patients were applied, supported by area and volume measurements from CT studies of spherical hollow phantoms of relevant size. In addition, measurements from digital photographs (and corresponding CT images) of transverse cryo-sections of a male and female cadaver (US National Library of Medicine's Visible Human Project) were included (Figure 1
).
View larger version (150K):
[in this window]
[in a new window]
|
Figure 1. Example of a photograph of a cadaver cryo-section (US National Library of Medicine's Visible Human Project) from a male pelvis (top) compared with the corresponding CT image (bottom).
|
|
|
Methods and materials
|
|---|
Repetitive CT scans of patients
22 repetitive CT examinations of 7 cervix cancer patients enrolled in a previous clinical study [6], were analysed. Two patients had two CT scans, three patients had three CT scans, and two patients had four and five CT scans each. Patients were included after informed consent was obtained, and the study was approved by a local research ethics committee. In the present report, the patient material was used to investigate the variation of cross-sectional areas and volumes of the bladder at different time points. Variations in bladder volumes were random. The CT scanner was a helical single detector, Siemens Emotion (Siemens, Erlangen, Germany), and the examinations were performed without contrast and with 3–5 mm slice thickness with an image resolution of 0.8–0.9 mm pixel–1. Minimum time interval between the first and last CT scan was 1 day and maximum was 24 days. Bladder contours were delineated by one of the authors (ED) on the radiotherapy dose planning system (PLATO, Nucletron BV) using a Silicon Graphic workstation. In each image the following data were found (Figure 2):- Aout=The transverse cross-sectional area enclosed by the outer contour of the organ, i.e. including wall tissue and lumen.
- Ain=The transverse cross-sectional area enclosed by the inner contour of the organ, i.e. the transverse area of the lumen.
- Awall=Aout–Ain, the transverse cross-sectional area of the organ wall.
The theoretical formulas for the cross-sectional areas, Aout, Ain and Awall, of a spherical shell as function of position along an axis through the sphere's centre, are outlined in Appendix 1. The main point in this model is that Awall is constant and independent of Aout, assuming the cavity of the sphere is included in the cross-section.
Volume calculations based on delineation of all bladder contours were also performed with PLATO:
- Vout=The total volume of the bladder (outer volume).
- Vin=The total volume of lumen (inner volume).
- Vwall=Vout–Vin=The total bladder tissue volume.
CT scans of spherical phantoms
The purpose of the phantom study was to estimate the uncertainties in the results from the patient study. Three spherical, hollow toy balls (36 cm3, 109 cm3 and 380 cm3) made of relatively rigid plastic (1.3 g cm–3) were used. Two small holes were made in each ball. Water was instilled through one hole, and air removed through the other hole. An attempt was made to measure the volume of instilled water, but this proved to be too difficult due to water leakage during instillation. After complete water instillation, the holes were sealed and the balls (phantoms) were put into a water tank, fixed with nylon threads and CT scanned (3 mm slice thickness). The inner and outer contours of the phantoms were delineated on the CT images by the same author as above (ED) using PLATO. Thus, the same procedure used to obtain areas and volumes from the urinary bladders of patients, was performed to obtain areas and volumes of the plastic phantoms. The three phantoms were delineated three times each in order to acquire nine sets of measurements. These CT volume data were compared with values obtained by immersing the phantoms in water and measuring the displaced volume of water. The latter volume determination method was repeated five times for each phantom with water inside. Afterwards, the phantoms were cut into pieces (the water inside was removed) and total wall volumes of each phantom were determined, again using the displacement of water principle.
Photographs of cadaver cryo-sections
Digital photographs of cadaver slices (1 male and 1 female) and corresponding CT images obtained from the US National Library of Medicine's Visible Human Project, were imported into GNU Image Manipulation Program v1.2 (GIMP – freeware by P Mattis and S Kimball) on a personal computer (PC). Outer and inner contours of the bladder were delineated every 1 mm and every 3 mm for cadaver and CT images, respectively. The delineation was performed by the same author as above (ED) and reviewed by an experienced oncologist (ØSB). The number of image pixels within each contour was obtained. The cadaver image resolution was 0.33 mm pixel–1, CT image resolution 1.0 mm pixel–1, and the area within each contour could thus be calculated. Bladder volumes were determined from the sum of the areas from all contours multiplied by image slice thickness.
Statistics
The analysis was performed with univariate linear regression using Microsoft Excel 97. A significant correlation between two variables was revealed by testing if the slope parameter differed significantly from zero comparing the appropriate statistics with the t-distribution of n–2 degrees of freedom [7]. Only two-sided tests were performed and p-values less than 0.05 were considered statistically significant.
|
Results
|
|---|
Cross-sectional area and volume measurements from repetitive CT scans of patients
Area measurements from repetitive CT scans of one patient is shown in Figure 3. The area within the outer bladder contour (Aout) varied substantially with caudocranial position (explained by the shape of the bladder) and from scan to scan (explained by variable filling).
View larger version (28K):
[in this window]
[in a new window]
|
Figure 3. Transverse cross-sectional areas of the urinary bladder as a function of caudocranial position as determined from four repetitive CT scans of the same patient. Second order polynomials have been fitted to the Aout data in agreement with theory (Appendix 1).
|
|
To compare area measurements from different scan days and patients in the same diagram, all area data were normalized by dividing by the mean area of each scan. The area of the wall tissue (Awall) increased slightly with Aout, although statistically significantly (R2=0.18, p<0.001, Figure 4). Also, the wall volume (Vwall) as a function of total volume (Vout) yielded a significant association (R2=0.72, p<0.001, Figure 5).
View larger version (27K):
[in this window]
[in a new window]
|
Figure 4. Normalized Awall (transverse cross-sectional area of the wall tissue divided by the mean Awall of each CT scan) as a function of normalized Aout (transverse cross-sectional area enclosed by the outer contour of the bladder divided by the mean Aout of the same CT scan). A linear regression line has been fitted to the cryo-section data while the patient and phantom data have been pooled (for clarity) showing mean values with 95% confidence intervals. Bin size was chosen to achieve a bin width of 0.15–0.20 along the abscissa. Linear regression analysis applied on the patient CT scan data gave y=0.21x+0.79 (R2=0.18, p<0.001, n=123), i.e. a considerably smaller, although statistically significant slope compared with the slope of the line fitted to the cryo-section data: y=0.74x+0.26 (R2=0.52, p<0.001, n=52). Linear regression applied on the phantom study provided a slight negative association: y=–0.076x+1.08 (R2=0.11, p<0.001, n=171).
|
|
View larger version (19K):
[in this window]
[in a new window]
|
Figure 5. Volume of bladder wall (Vwall) as a function of total bladder volume (Vout) from repetitive CT examinations on our own patients and from cadaver cryo-sections. Linear regression gave a statistically significant association between Vout and Vwall (R2=0.72, p<0.001, n=24).
|
|
Cross sectional area and volume measurements from CT scans of spherical phantoms
To validate the data from CT scans of patients, similar measurements via outlining on CT were performed on three hollow plastic spheres of different sizes. The measurements were repeated three times for each phantom to quantify the uncertainties involved. In accordance with the spherical shell model there is no positive correlation between Awall and Aout in contrast to the results from CT scans of patients (Figure 4
). On the contrary, there was a slight negative association, probably explained by a delineation artefact (R2=0.11, p<0.001).
The outer volume and wall volume of the plastic spheres were determined by the displacement of water principle. Within acceptable uncertainty, these volumes were in agreement with corresponding values obtained from measurements from CT images. No systematic trends in the errors between water and CT measurements (size of the objects or outer versus inner volumes) were observed (Table 1).
View this table:
[in this window]
[in a new window]
|
Table 1. Results from all volume measurements. Water: Vin has been calculated from Vin=Vout–Vwall. CT/cryo-photos: Vwall has been calculated from Vwall=Vout–Vin. NA: Not available. The bladder of the female cadaver was almost empty (Vin=8.1 ml), and it was not possible to delineate the inner contour on the CT images because of inferior contrast between wall tissue and lumen. Hence, Vin and Vwall were NA for this particular case. n=number of measurements
|
|
Cross-sectional area measurements from cadaver cryo-sections
Area measurements from The Visible Human cadaver cryo-sections corresponded well with measurements obtained from the CT images of the same cadavers (R2 in the interval 0.94–0.98 for Aout and Ain, and 0.83 for Awall), thus supporting the use of CT to measure areas and volumes of the human bladder (Table 1). Diagrams of area measurements as a function of caudocranial location were comparable with the results obtained from patient CT scans (data not shown). Linear regression applied on the relationship between Awall and Aout yielded a linear function with a steeper slope (0.74±0.20, slope coefficient±95% CI) compared with the slope obtained from the patient CT scan data (0.21±0.08, slope coefficient±95% CI) (Figure 4).
|
Discussion
|
|---|
The aim of this study was to investigate the variation of urinary bladder wall volume as a function of variable luminal volume, both locally in transverse cross-sections, and globally looking at the whole bladder. Theoretical calculations of cross-sectional wall areas (Awall) of a spherical shell showed that Awall is independent of the cross-sectional area enclosed by the outer contour (Aout). Measurements on CT scans of human bladders revealed a slight but detectable relationship between Awall and Aout in theoretical disagreement with the sphere model. A similar relationship was found from measurements on photographs of cadaver cryo-sections, although with a steeper slope. The explanation may be a larger heterogeneity in the patient study data, yielding a lower slope.
The validity of the method of delineating contours on computer images to measure cross-sectional areas is not obvious. Our findings could have been explained by a delineation artefact. Therefore, cross-sectional areas of three spherical phantoms similar to the human bladder, were measured repeatedly using the same methodology. The results from the phantom experiments show that the normalized Awall is approximately unity for all normalized Aout. This is in reasonable agreement with the spherical shell model, indicating that the delineation method is not significantly biased. However, this method is limited by the improved contrast between plastic and water compared with the contrast between bladder tissue and urine.
The measured area data from the patient CT scans were not in agreement with the spherical shell model. Correspondingly, a relationship between wall volume and luminal volume was obtained employing the CT scans of individuals with variable bladder volume at different times. The reason for variable bladder tissue volume at variable luminal volume is not evident. Dale et al have reported similar results for the rectum in a previous study using the same methodology [4]. Generally, the bladder detrusor muscle consists of smooth muscle cells with tonic activity. As the bladder is filled, intermittent micturition contractions appear, with higher frequency and force at larger luminal volumes [8]. Kershen et al found an increased blood flow in the detrusor muscle as the bladder was filled [9]. The enlarged tissue volume with increased luminal volume, found in the present study, may be explained by raised blood perfusion. This reflects the metabolic demands of the detrusor at larger luminal volumes [10].
In conclusion, we have found that the cross-sectional wall areas of the bladder differ slightly but significantly from expected values derived from a simplistic spherical shell model. The bladder tissue volume is related to luminal volume. Our results may be of practical value, for instance when developing hollow organ delineation tools in radiotherapy planning systems.
|
Appendix 1
|
|---|
Inspecting Figure 2, Aout can be calculated from the outer radius, R, of the spherical shell and the caudocranial position, z, of Aout:
Similarly, the transverse cross-sectional area of the lumen of the spherical shell is:
where r is the inner radius of the spherical shell. By subtracting Ain from Aout, Awall is obtained:
which is independent of the location, z, of the area. Aout and Ain are second order polynomials in z (Figure 3). Please note that Awall is constant for one specific luminal volume of the spherical shell. If the sphere size is increased, and the wall is made of an elastic material, it is fair to expect that Awall will decrease, since the wall of the sphere is stretched. This means that Awall will vary between different CT scans of the same patient if the volume of urine inside the bladder varies.
|
Acknowledgments
|
|---|
The authors are indebted to Professor Hans Hedlund and Consultant Urologist Trygve Talseth at Department of Urology, Rikshospitalet University Hospital for valuable discussions on bladder blood flow studies.
Received for publication May 5, 2004.
Revision received November 16, 2004.
Accepted for publication January 13, 2005.
|
References
|
|---|
- Lyman JT. Complication probability as assessed from dose-volume histograms. Radiat Res 1985;104:S13.[CrossRef]
- Lu Y, Song PY, Li S, Spelbring DR, Vijayakumar S, Haraf DJ, et al. A method of analyzing rectal surface area irradiated and rectal complication in prostate conformal radiotherapy. Int J Radiat Oncol Biol Phys 1995;33:1121.[Medline]
- Meijer GJ, Van den Brink M, Hoogeman MS, Meinders J, Lebesque JV. Dose wall histograms and normalised dose surface histograms for the rectum. A new method to analyze the dose distribution over the rectum in conformal radiotherapy. Int J Radiat Oncol Biol Phys 1999;45:1073.[Medline]
- Dale E, Hellebust TP, Bruland ØSB, Olsen DR. Comparative analyses of the dynamic properties of the rectum studied by cryo-sections of human cadavers and pelvic CT scans of patients. Br J Radiol 2003;76:104–8.[Abstract/Free Full Text]
- Moore KL, Dalley AF. Pelvis and perineum. In: Clinically oriented anatomy (4th edn). Philadelphia, PA: Lippincott, 1999:331–430.
- Hellebust TP, Dale E, Skjønsberg A, Olsen DR. Inter-fraction variations in rectum and bladder volumes and dose distributions during high dose rate brachytherapy treatment of the uterine cervix investigated by repetitive CT-examinations. Radiother Oncol 2001;60:273–80.[CrossRef][Medline]
- Larsen RJ, Marx LM. An introduction to mathematical statistics and its applications (2nd edn). New Jersey: Prentice-Hall, 1986.
- Guyton AC, Hall JE. Micturition, diuretics and kidney diseases. In: Textbook of Medical Physiology (9th edn). Philadelphia, PA: W. B. Saunders Co., 1996:405-24.
- Kershen RT, Azadzoi KM, Siroky MB. Blood flow, pressure and compliance in the male human bladder. J Urol 2002;168:121–5.[CrossRef][Medline]
- Guyton AC, Hall JE. Local control of blood flow by the tissues, and humoral regulation. In: Textbook of medical physiology (9th edn). Philadelphia, PA: W. B. Saunders Co., 1996:199–208.
Top
Abstract
Introduction
