|
Testis is the main organ of the male reproductive system. A testis consists of two parts, the testis parenchyma and intercellular substances. The parenchyma is composed of long and coiled tubes called seminiferous tubules. As the seminiferous tubules comprise approximately 80% of the testicular mass, the morphological measurements of seminiferous tubules are important in the studies of testis tissues.1-4
Currently, the main quantitative method for tissue structure study is computer-assisted planimetry.5,6 The major limitation of this method is time-consuming. Fully automated planimetry, which can save much time, is possible when contrast at the boundary of structures is sufficiently high. But in the measurement of testis tissue the automatic delineation of the tubule/intercellular substance interface is hard to define by algorithms.7
Stereology is a quantitative measurement method. The methodology of stereology is based mainly on the observations made on sections or subsamples of tissues, applying test probes such as points, lines or planes and counting the interactions of the probes with the structures under study.8 Although stereological method is generally applied manually, it is extremely efficient and used widely in the medical field.9
In this paper, we presented an image processing program for stereological measurement which was developed with Delphi. We then measured the diameter, area and volume of the rats′ seminiferous tubules with the program and compared the area results with planimetry, the generally accepted technique for the measurement of cross-sectional areas. Furthermore, we analyzed the repeatability of the stereological method by two observers to examine the inter-observer variability.
METHODS
Materials
Fifteen healthy cleaning male Wistar rats (more than 8 weeks and 250–270 g) were purchased from the Institute of Laboratory Animal Sciences, CAMS & PUMC (Beijing, China). Sections of testis tissue were taken from these Wistar rats. Then 15 to 20 fields of view were selected uniformly and randomly from each section to be measured (n=247). The customized image processing program was used to calculate the diameter, area and volume of the seminiferous tubule of the rat based on the stereological theory. All the above measurements were performed by two independent observers (one in Peking Union Medical College and the other in Tsinghua University). For the comparison, the areas of 160 seminiferous tubules selected from 4 rats (n=160) were measured by both stereological method and planimetry (Image Pro Plus 6.0, Media Cybernetics, Inc. USA).
Stereological image processing program
The program in the present study was developed with Borland Delphi 2007. The parameters such as the number, the size and the place of measurement frames or crossed measurement points could be adjusted by the observer according to the region of interest. The program also provided the function of distance measurement, including transforming the pixels pitch to the real distance automatically by input of the coefficient. After the measurement, the program could save the altered images and output the measurement results to “.csv” files which could be opened by spreadsheet software such as Microsoft Excel for further analysis (Figure 1).
Diameter measurement of seminiferous tubule
The diameter of a seminiferous tubule was defined as the shortest distance between two parallel tangent lines of the outer edge of the tubule. Dozens of tubules were needed to estimate the average diameter of the whole structure. Rectangular measurement frames were placed on each field of view to select the interested tubules. The tubules within the measurement frame or intersectant with the right or upper edges were selected while those intersectant with the left or bottom edges, or with the extended lines of the right and left edges were not.10,11 The mean value of the diameters of the tubules selected from a group of fields was calculated and considered to be the estimated average diameter of seminiferous tubules of the testis tissue (Figure 2).
Area and volume measurement of seminiferous tubule
Uniformly distributed measurement points (cross lines) were superimposed on each field of view, to count the measurement points in the seminiferous tubules. A positive count was defined as a hit when the upper-right hand corner of a point, which was defined by two perpendicular cross-lines, entirely covered the seminiferous tubules.7,12
To measure the area of an individual tubule the measurement points should cover the whole image and the number should be more than 200 (Figure 3A). By counting the points hitting the tubule of interest (Pi), the unbiased estimation of the area of this tubule was given by Ai=PiA/T, where A is the area of the whole image and T is the total number of measurement points.13
When estimating the volume of the whole structure of the seminiferous tubules, a group of fields were needed and no more than 20 measurement points were superimposed on each one. The the valid points (which were not located in the blank area) in the seminiferous tubules of each field were counted (Pi) and added ( ). The ratio of the number of valid points in the seminiferous tubules to the sum of all valid points, r= , was the estimated volume fraction of the seminiferous tubules in the testis tissue.14,15 The volume of seminiferous tubules in the testis could be calculated by multiplying this volume fraction by the volume of the testis that could be obtained by its weight and density.
Planimetry
The same tubules were also measured by means of planimetry using software Image Pro Plus 6.0 (Figure 3B). Using a hand-held mouse, the cursor traced around the boundary of the interest tubule. The software calculated the pixel dimension enclosed within the traced area.16,17 Then the pixel dimension was multiplied by the coefficient of pixels pitch and the real distance to obtain the area of the tubule.
 view in a new window |
Figure 1. The results of the diameter and volume measurement were recorded and then saved as a ”.csv” file.
Figure 2. The diameter measurement. Two rectangular measurement frames were placed on a field of view and the smaller one was used to select tubules. Tubules A and B were intersectant with the right or upper edges and thence were selected. The arrow lines indicate the diameters of these two tubules.
Figure 3. The area measurement: A: superimposition of a square point grid (20×15) for stereology. The white crosses illustrated positive counts hitting the interest tubule. B: tracing the tubule border (white) for planimetry. |
Statistical analysis
Statistical analyses were carried out using R for Windows (Version 2.9.0). The diameter results were expressed mean ± standard deviation (SD), and coefficients of variation (CV= (SD/mean)×100%) of those mean values. The area values of 160 seminiferous tubules were calculated and used for the comparison of planimetry and stereology. The differences of the estimated area obtained by two different approaches were compared using a paired t test to check the methodological differences and P value less than 0.05 was accepted as being statistically different. The volume results of two observers were compared statistically to evaluate inter-observer variability using a paired t test and P <0.05 was considered statistically significant. To assess the agreement between the two approaches and the two observers a linear regression analysis were applied and Bland-Altman graphs were created.
RESULTS
Diameter of seminiferous tubule
We selected 836 seminiferous tubules from 247 fields of view which were taken from 15 rat testis tissues. The mean diameter was (331.3±29.1) μm and CV was 8.8%; the difference in the diameters of seminiferous tubules of Wistar rats was small.
Area of seminiferous tubule and method comparison
The area values of 160 seminiferous tubules from 4 rats that were calculated by planimetry and stereology are shown in Table 1. Unlike the diameter results, the variation of the areas of different tubules was much larger. The area values obtained by two different methods were not statistically different (P=0.846). The agreement between the methods is presented in Figure 4. The two methods were well correlated with each other (r=0.982). Bland-Altman plots indicated that the difference between the two methods was within 95% limits of agreement.
 view in a new window |
Table 1. The comparison of planimetry and stereology to esimate the area of seminiferous tubules |
As shown in Table 1, stereological measurements took less time than planimetry for estimating the area of seminiferous tubules. It took 44.9% less time.
Volume of seminiferous tubule and inter-observer variability
The volumes of the seminiferous tubule of 15 rats were calculated by two individual observers in the Peking Union Medical College and in Tsinghua University (Table 2). There was no significant difference in volume measurement observed (P=0.6404). The agreement between the two observers is presented in Figure 5.
 view in a new window |
Table 2. The comparison of two observers measuring the volume of seminiferous tubule |
 view in a new window |
Figure 4. Method correlation was illustrated by linear regression curve (A) and Bland-Altman plots (B) for area measurements. Linear regression dotted curve gave the line of unity (perfect agreement). Bland-Altman graphs plotted the difference against the mean of the two results. The central dotted line represented the mean difference and was almost coincident with a zero value (good conformity). The other two dotted lines represented the 95% limits of agreement.
Figure 5. Comparison of volume measurements between two observers, illustrated by Linear regression curve (A) and Bland-Altman plots (B). |
DISCUSSION
This study was designed to test whether stereology was applicable for the geometric measurement of the seminiferous tubule. To this end, a stereological image processing program was written and the areas of the tubules obtained by stereology were compared with those obtained by the standard method, planimetry.
The stereological image processing program proved to be of great help when used in stereological quantitative analysis. The geometric parameters could be measured and estimated conveniently. The results of measurements could be saved for further analysis.
The results of the diameter measurement indicated that the differences in the diameters of different seminiferous tubules were small. Measuring about 50 tubules and obtaining the mean value was sufficient for the estimation of the average diameter in the rat. To control the number of the interest tubules, we placed two rectangular measurement frames on each field for the diameter measurement. The smaller frame was used under the condition of the tubules selected by the bigger frame being much more than necessary. It also ensured the profiles of selected tubules were complete in the image.
Point counting was the basic procedure for the area and volume measurement. We did some research into the optimal number of the test points for the measurement. More points meant a higher accuracy in the estimation, but it also took much more time. We used point girds from 3×3 to 8×8 to estimate the volume of the seminiferous tubule of a rat, and the statistical analysis of results indicated that there was no significant difference between the different numbers of test points. We just ensured that the sum of the points from all fields was more than 200. In this study, the grid of test points we used was 4×4. On the other hand, in the area measurement, the number of points needed in each image was at least 200. In order to remove the influence of tubule shape, the points′ distribution should also be in direct proportion to the pixel size of the image to guarantee that the vertical basic distance of the points was equal to the horizontal distance. Since the pixel size of a field of view was 1024×768, we used the grid of 20×15 test points in the area measurement.
Good agreement was found between the results obtained by stereology and planimetry. The major difference distinguishing the two methods was time. The point counting approach was less time-consuming with an average gain of 45%. This event could be explained by planimetry depending on the precision of the eye-brain-hand complex. Tracing a thin irregular curve by hand could cause errors due to the mechanical and dynamic limitations of the hand, whereas stereological point counting only demanded the detection of a positive hit scored by the eye-brain complex.7
Finally, the reproducibility of stereological measurements was examined. There was a significant agreement between two observers′ estimation. In summary, the stereological method described in this study represented a simple, practical, high-efficient, and inexpensive technique to estimate the geometric parameters of seminiferous tubules.
REFERENCES
1. Sakamoto H, Yajima T, Nagata M, Okumura T, Suzuki K, Ogawa Y. Relationship between testicular size by ultrasonography and testicular function: measurement of testicular length, width, and depth in patients with infertility. Int J Urol 2008; 15: 529-533.
2. Hsieh ML, Huang ST, Huang HC, Chen Y, Hsu YC. The reliability of ultrasonographic measurements for testicular volume assessment: comparison of three common formulas with true testicular volume. Asian J Androl 2009; 11: 261-265.
3. Prakash S, Prithiviraj E, Suresh S. Developmental changes of seminiferous tubule in prenatal, postnatal and adult testis of bonnet monkey (Macaca radiata). Anat Histol Embryol 2008; 37: 19-23.
4. Jiang YG, Peng T, Luo Y, Li MC, Lin YH. Resveratrol reestablishes spermatogenesis after testicular injury in rats caused by 2, 5-hexanedione. Chin Med J 2008; 121: 1204-1209.
5. Ganzer R, Blana A, Gaumann A, Stolzenburg JU, Rabenalt R, Bach T, et al. Topographical anatomy of periprostatic and capsular nerves: quantification and computerised planimetry. Eur Urol 2008; 54: 353-360.
6. Kakuta T, Usui M, Coats WD Jr, Currier JW, Numano F, Faxon DP. Arterial remodeling at the reference site after angioplasty in the atherosclerotic rabbit model. Arterioscler Thromb Vasc Biol 1998; 18: 47-51.
7. Van Vre EA, van Beusekom HM, Vrints CJ, Bosmans JM, Bult H, Van der Giessen WJ. Stereology: a simplified and more time-efficient method than planimetry for the quantitative analysis of vascular structures in different models of intimal thickening. Cardiovasc Pathol 2007; 16: 43-50.
8. Kubinova L, Janacek J, Karen P, Radochova B, Difato F, Krekule I. Confocal stereology and image analysis: methods for estimating geometrical characteristics of cells and tissues from three-dimensional confocal images. Physiol Res 2004; 53 Suppl 1: s47-s55.
9. Roberts N, Puddephat MJ, McNulty V. The benefit of stereology for quantitative radiology. Br J Radiol 2000; 73: 679-697.
10. Muhlfeld C, Nyengaard JR, Mayhew TM. A review of state-of-the-art stereology for better quantitative 3D morphology in cardiac research. Cardiovasc Pathol 2009; Epub ahead of print.
11. Charleston JS, Hansen KR, Thyer AC, Charleston LB, Gougeon A, Siebert JR, et al. Estimating human ovarian non-growing follicle number: the application of modern stereology techniques to an old problem. Hum Reprod 2007; 22: 2103-2110.
12. Mazonakis M, Pagonidis K, Schlosser T, Hunold P, Damilakis J, Barkhausen J, et al. Stereological estimation of left-ventricular volumetric and functional parameters from multidetector-row computed tomography data. Eur Radiol 2008; 18: 1338-1349.
13. Weibel ER, Hsia CC, Ochs M. How much is there really? Why stereology is essential in lung morphometry. J Appl Physiol 2007; 102: 459-467.
14. Bas O, Acer N, Mas N, Karabekir HS, Kusbeci OY, Sahin B. Stereological evaluation of the volume and volume fraction of intracranial structures in magnetic resonance images of patients with Alzheimer′s disease. Ann Anat 2009; 191: 186-195.
15. Albrechtova J, Janacek J, Lhotakova Z, Radochova B, Kubinova L. Novel efficient methods for measuring mesophyll anatomical characteristics from fresh thick sections using stereology and confocal microscopy: application on acid rain-treated Norway spruce needles. J Exp Bot 2007; 58: 1451-1461.
16. Acer N, Sahin B, Usanmaz M, Tatoglu H, Irmak Z. Comparison of point counting and planimetry methods for the assessment of cerebellar volume in human using magnetic resonance imaging: a stereological study. Surg Radiol Anat 2008; 30: 335-339.
17. Aydin F, Senturk N, Sahin B, Bek Y, Yuksel EP, Turanli AY. A practical method for the estimation of vitiligo surface area: a comparison between the point counting and digital planimetry techniques. Eur J Dermatol 2007; 17: 30-32.
|
PDF(392K)
Rapid Response |