Morphometry of the prostate: I. Distribution of tissue components in hyperplastic glands

RAPID C O M M U N I C A T I O N

~Q

MORPHOMETRY OF THE PROSTATE: DISTRIBUTION OF TISSUE COMPONENTS IN HYPERPLASTIC GLANDS* LEONARD S. MARKS, M.D. BRENT TREIGER, M.D. FREDERICKJ. DOREY, PH.D. Y:S. FU,M.D. JEAN B. DEKERNION, M.D. From the Departments of Surgery/Urology, Biostatistics, and Pathology, UCLA School of Medicine, Los Angeles, California

ABSTRACT--Objectives. Morphometry, or quantitative image analysis, offers great promise in characterizing the various histologic types of benign prostatic hyperplasia (BPH), but to date, a systematic study of the tissue components is lacking. Thus we employed morphometry to examine the distribution of primary BPH tissues throughout whole human prostates. Methods. The prostate glands of 20 men with BPH were removed for low-volume carcinoma and subjected to a uniform, comprehensive, systematic quantification of the primary BPH tissue Components using the technique of digitization and point-count morphometry. Results. We found the following average volumes among the 20 glands: epithelium, 19.9% (S.D. 5. ] %, range 11.7% tO 30.8%); fibromuscular stroma, 50.6% [S.D. 9.6%, range 32.2% to 74.6%); glandular lumina, 29.7% (S.D. 8.9%, range 11.9% to 67.5%]. Within the individual prostates, we found symmetry in primary BPH tissue distribution, except that the outer prostate was on average 25% richer in epithelium than the inner prostate (p < 0.05). When tissue composition was determined in simulated biopsy specimens, corrected for radial [ie, inner vs outer gland) orientation, the correlation with wholeorgan composition was statistically significant for "percentage epithelium" [r = 0.72, p < 0.01] and for "stroma/epithelial ratio" (r = 0.63, p < 0.01). Conclusions. Major differences in primary tissue composition may separate different hyperplastic prostates. Primary BPH tissues are rather symmetrically distributed within individual prostates. Quantitative histologic differences between prostates, potentially important in clinical decision-making, may be accurately diagnosed by morphometry of radially oriented biopsy specimens.

Histologic moiphometry, or quantitative tissue analysis, was applied to the human prostate by Bartsch and colleagues in 1979. ~ These authors demonstrated that, with the use of morphometry, the prostate can be classified into distinct and mea*Supported in part by the Urological Sciences Research Foundation and the American Foundation Jor Urological Diseases. Presented in part at the 89th Annual Meeting of the American Urological Association, 5an Francisco, California, May 14-18, t994. Submitted: Ja,|e 13, 1994, accepted (with revisions): June 20, t994

486

surable tissue components (ie, muscular stroma, and glandula m o r p h o m e t r y of the prostate, leagues found that symptomatic hyperplasia (BPH) is associated' crease in the stroma and a decre~ lar tissue, suggesting that BPH L, mal disease. Shapiro et al. 2 recen earlier studies of Bartsch and fl morphometry might be helpful i histologic parameters that pre alternative pharmacotherapies fc UROLOGY®/ OCTOBER199~ /

FIGURE 1. Schema for uniform processing and sectioning of prostate glands" following radical prostatectomy, as used in present study. (,4) Whole organ, illustrating transverse sections. (B) Selection of central or widest transverse section. (C) Division of central section into sextants (anterior and posterior; right, middle, and left). Arrow depicts area for morphometry (between lines) in posterior central sextant; note other lined areas for morphometry in each sextant.

G,

itatic tissue composition, as determined by hometry of biopsy specimens, may vary between different glands. For example, iand colleagues 3 found a threefold variation prostatic epithelial component (6% to 17% of }olume) among different men with symptot3PH. Shapiro et al. 4 noted an even greater ion in both epithelial and smooth muscle nents among different men with BPH. Simiservations were also made by Price et al., 5 iudied tissues obtained during enucleation or irethrat resection. Thus certain prostates ap,n biopsy specimens to be much richer than i n epithelial or muscular components. Such !differences between prostate glands, if truly entative, may be crucial in predicting an inial's treatment response to alpha-blocking ;6and perhaps to other therapies as well. ~hat extent does a limited tissue sample (ie, ) accurately reflect whole-gland prostatic hisi type? Prostatic hyperplasia has been previ:onsidered a heterogeneous process. 7-9 Conreal, qualitative histologic e x a m i n a t i o n :s to show great variation, not only between te glands but also in different parts of the gland. However, quantitative histologic extion (morphometry) has not yet been thorevaluated to determine the heterogeneity or etry of the hyperplastic process. The issue of genmty vs symmetry has major implications clinical utility of limited tissue specimens opsy specimens) to characterize the whole and potentially to guide therapy. To clarify ~portant, fundamental issue, we systematiexamined the m o r p h o m e t r y of the BPH s in whole prostates removed for treatment of ilume carcinoma. We found that, with certain itions and a predictable variability, limited

) GY~- / OCTOB~a1994 / VOLUM~4q, NUMBER4

samples do reflect quantitatively the tissue components of the entire hyperplastic process in the individual patient. MATERIAL AND METHODS PATIENTS

Twenty whole prostate glands, removed during 1991 at UCLA Hospital by a urologist (J.B.dK.) for organ-confined, low-volume carcinoma (tess than 15% of the gland involved), were the subject of this study. No patient had been pretreated with androgen deprivation. All patients who met these criteria were included. Two patients had undergone previous transurethral resection of the prostate (TURP), but the resection did not involve the areas selected for morphometric study. Final pathologic stage was A-1 in 2 patients, B-1 in 14, and B-2 in 4. Average age of the patients at operation was 62.2 years (range, 55 to 76 years). Average gland weight recorded by the pathologist was 62.1 g (range, 25 to 230 g). Prior to prostatectomy, average s e r u m prostate-specific antigen (PSA) level (Hybritech) was 7.9 ng/mL (range, 1.2 to 22 ng/mL), and average voiding symptom score (American Urological Association [AUA]-7) was 13.4. TISSUE PROCESSING AND SECTIONING

Morphometry was performed on hematoxylin and eosin-stained sections, which had all been pathologically processed in a uniform fashion (Fig. t). Following formalin fixation, each gland was transversely sectioned at 3 to 5 m m intervals (Fig. 1A). The middle or widest section was selected for morphometry (Fig. 1B), and it was further divided into sextants (Fig. 1C). in the center of each sextanl piece, parallel lines approximately 1 mm apart were drawn with a marker on the coverslip from capsule to urethra (Fig. tC, arrow). The denoted areas for 487

TABLE I. Tissue components and stroma epithelial (S/E) ratios (rank-ordered by epithelial percentage composition) Patient

FIGURE 2. Photomicrograph shows grid overlay for point-count morphometry in part of a prostatic feld. Straight arrow identifies an epithelial point or intersect; curved arrow identifies a stromal intersect; arrowhead points to a luminal intersect. Areas of carcinoma and secondary changes of BPH (ie, inflammation, infarction, cystic dilation) were excluded from morphometric analysis [hematoxylin and eosin, original magnification x63).

morphometry were thus approximately the size of a needle biopsy core, 1 m m wide and ranging in length from 5 to 14 m m (average of 9 mm), depending on gland size, with each simulated core deriving from a central sextant of the gland. Since a preliminary survey of these prostates revealed no consistent differences in BPH histologic type between the base, middle, and apex, we have taken these middle sextant cores to represent the whole prostate.

MORPHOMETRY Morphometric analysis of prostate tissue was performed by a pathologist (Y.S.E) using a computed digital imaging system consisting of a Sanyo CCD video camera mounted on a Leitz microscope, an IBM PC/AT with PCVision digitizer (640 × 480 lines, Image Technologies, Woburn, MA), a Sony color m o n i t o r (PVM-1271Q), and a Microsoft mouse. Customized software was provided by Microsciences, Inc. (Phoenix Technology, Inc., Seattle, WA). The technique has been described in detail elsewhere) ° With the use of x63 magnification, the digitized image or morphometric field, 0.94 m m wide, was overlaid with a grid of 13 vertical and 10 horizontal lines (130 intersecting points) (Fig. 2). Morphometric analysis was begun at the periphery of the prostate, excluding the capsule, and proceeded toward the urethra field by field until the core was studied in its entirety. W h e n a point fell on the border of two tissue elements, the tissue on the left side of the point was favored. In each field, the 488

PercentageTissue Composition

No.

Epithelium

Stroma

Lumen

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

11.7 13.4 13.7 13.7 15.1 t6.7 17.5 18.2 18.4 19.3 19.5 22.5 22.5 22.8 22.9 24.3 24.4 25.0 27.2 30.8

41.0 57.2 54.8 74.4 57.2 54.8 54.9 51.0 53.9 33.2 57.0 41.8 48.4 32.3 47.8 52.4 42.0 53.3 51.7 46.3

47.3 29.4 31.6 11.9 27.7 28.5 27.5 30.8 27.7 47.5 23,5 35.7 29. I 45.0 29.2 23.4 33.5 21.7 21.1 22,9

Mean S.D.

19.9 5.1

50.4 9.4

29.7 8.9

*5/E ratio is calculated as average of ratios of stro na/epitl etium 3or a[t bl¢~ fields and need not equal ratios for overall stroma/epithelial averages.

number of points allocated to epithelium, or lumen was recorded using the mouse; percentage composition of each field, eqt to volume percentage, ~was obtained. Area~ cinoma, infarction, and inflammation w cluded from analysis, as were areas of cysti! tion greater than 0.5 m m in diameter. Morpt was p e r f o r m e d on an average of 54 fie prostate gland (6 cores per gland, average of per core). Approximately 90 minutes of c time was required for c o m p l e t e s t u d y , prostate.

DATA PROCESSING Data retrieval and analysis were su biostatistician (EJ.D.) using an IBM f puter and a statistical program (Stata ( Station, TX). Significance of differenc mined by the paired t test and the pro correlation coefficient. RESULTS D I F F E R E N C E S B E T W E E N PROS'IATE G L A N D 5

Table I shows the 3 major tissue con epithelium, stroma, and l u m e n - - f o r the UROLOGY~ / OCTOBEa/994 / VOLU,VIe

t2

13

]5 FIGURE 3.

8

g

18

19

2O

__+.....*

C R

L C R

L C R

Percent epithelium (%E, vertical axis) in each of the 20 prostate glands, ranked from lowest (gland I, 11.7 % E) to highest (gland 20, 30.8 %E). For each gland, the %E in each sextant is shown (open circles = posterior, crosses = anterior; L = left, C = center, R = right). Perfect symmetry of %E within a prostate would be depicted by two superimposed horizontal lines.

L C R

AREA OF PROSTATE 'alue listed represents the avtudied from the 6 sextants of n Figure 1. The 20 glands are :ler of the percentage epitheial component was 19.9% of .d tissue, ranging from a low a high of 30.8% in gland 20. :entage epithelium both with o r the glandular lumina (ie, solid prostate tissue vs solid ace); the results were highly as component found was the (column 2), averaging 50.4% nge of 32.2% to 74.4%). The sured, the glandular lumina, of 29.7% of each prostate %). depithelial (S/E) ratio (col-. .nds was 3.9 with a range of :ion was found between preore and S/E ratio. :O5IATE GLANDS

i=~ " {!0

®

. GY

perimposed horizontal lines. Symmetry is best seen in glands 1, 3, 13, 16, and 20, where epithelial percentage is nearly the same in all 6 sextants. Lack of symmetry is demonstrated in gland 7, where the epithelial percentage was found to be considerably higher in the left and right anterior sextants than in the left and right posterior sextants. Another example of asymmetry is seen in gland 15, where the left posterior sextant contains substantially more epithelium than the left anterior sextant, while the right anterior sextant contains more epithelium than the right posterior sextant. However, perusal of Figure 2 indicates no tendency for any sextant of a prostate to have a consistently different epithelial percentage from any other sextant. This gross observation was confirmed by analysis of variance d e m o n s t r a t i n g no significant difference in the means of the sextants. Figure 4 depicts the prostatic sextants subdivided into the three innermost and outermost morphometric fields, respectively representing the periurethral and pericapsular prostate gland. In every sextant, the outer prostate was richer in epithelium than the inner prostate; for each prostate, the average ratio of outer to inner epithelial percentage was 1.25. For all glands taken together, the average epithelial percentage of the outer prostate fields was 21.8% +_ 6.0% S.D. and that of the i n n e r prostate fields was 18.4% _+ 5.6% S.D. The differences are statistically significant (p < 0 . 0 5 , paired t test), although considerable overlap is present.

~verage epithelial percentage (right, left, and center; anteor each of the 20 prostate O glands is represented by a Table I, the glands are num'der of epithelial percentage, : lines appear higher on each tnce overall percentage ep11.7% in gland 1 to 30.8%

[.]5f~-OF LIMIIED SAMPLES TO CHARACTERIZE ENTIRE PROSTATE GLAND

:t symmetry of epithelial disin the appearance of two su-

tn Figure 5, the percentage epithelium (%E) for the entire prostate (average of all 54 fields, 6 sextants) is plotted against the %E found in the outer

OCTOBER1994

/ VOLUME44, NUMBER4

489

O U T E R 3 FIELDS

INNER 3 FIELDS

.5o o .4-

o o

i ii

.3-

=<

.2-

.i

LEFT

POST.

CENTER

ANT.

POST.

RIGHT

ANT.

POST.

ANT,

FIGURE4.

Percentage epithelium (% E) in outer vs inner prostate. Mean %E is plotted on vertical axis; sextants are shown on horizontal axis. Boxes depict 25th to 75th percentiles; horizontal lines within boxes depict median value; brackets depict minimum and maximum values. Outliers (values more than 3 S.D. from mean) are shown as isolated points. Overafl ratio between outer and inner prostate is 1.25 (p < 0.05).

3 fields of a single posterior sextant, right or left side. The coefficient of correlation found between the entire gland composition and the makeup of either of these limited samples was high: 0.75 (left) and 0.68 (right) (p <0.01). When S/E ratios were used in place of %E, the coefficients of correlation were 0.57 (right) and 0.69 (left) (p <0.01). For all fields in the 20 prostates studied, the 2 variables, S/E ratio and %E, were highly correlated (r = -0.66, p <0.01). Since the correlation was higher using %E than S/E ratio, we refer to the former measure throughout this text. COMMENT Benign prostatic hyperplasia is widely considered to be a heterogeneous process. 5,r 9,H We have evaluated this belief in a study differing from previous studies in three important ways: quantitative tissue analysis (morphometry) was used, whole glands were systematically studied, and secondary changes of BPH (ie, inflammation, infarction, and cystic

::E

dilatation) were excluded t found that although major. BPH tissue composition may glands, within individual p] rather symmetric and, in fa basis of limited tissue sample have major implications for linens to characterize vario evaluation of BPH "heterog( Morphometry has been a17 limited number of previous firmed the enucleation studi average BPH gland contains larger than any other. In Bar age BPH gland was 60% su and 28% lumina. In the pn gland was 5 0 . 4 % stroma, ] 29.7% lumina. Bartsch's pat: ing operation for relief of at present patients only a few dates for TURR Thus the d! direct support to the bypath( of symptomatic BPH, the eF creases and the stromal cart Other morphometric astir position obtained from biD with highly symptomatic Weber et al. 3 (11.3%), Shap Siegel et al. 13 (18%). In all served variability was great, ponent making up 5% to 24 the present whole-gland de was 11.7% to 30.8%. Price that the epithelial compone] hyperplastic nodules than v size and number increase wi hyperplastic gland. 5 These d siderable quantitative differc BPH tissue composition. Within the individual gl: ponents were distributed n might be expected, there w ence in distribution of %E

UJ 4 O I--

FIGUR

I--I I--. ._1 03

demot:

g3o

l,-

lotion

£3,,. o

t.u

~ 2 0

z

z tad

o

o

o° o

a

o ~

°

~

o

o

0.68 ........

5O

LEFT

o

r=0.75 - - T -

490

o

10

T . . . . . . . . . . . . "T

20

30

r

g

40

50

itheliu~ 54 fiel 3 field, from k zontal

RIGHT UROLOGY"~/

OCIOB~Rt . . . .

or were there any consistent dif.e anterior and posterior parts of However, a statistically signif%E was noted when the inner (periurethral) was compared n (pericapsular): the outer part /o richer in epithelium than the A similar observation was also al.1 Thus to properly compare pies of various prostates, radial issue (inner vs outer) must be e conveniently accomplished at by applying India ink to one haps much of the within-gland /' is attributable to secondary Lamination, infarction, and cyschanges were intentionally ex:nt analysis, resulting in more :n in the other studies. :he question, "Can limited sam: biopsy specimens, be used to tissue composition?", thereby useful information. To answer issue composition of the outer or left posterior prostate core against whole gland composie results indicate an excellent 1 simulated biopsy specimens mposition (Fig. 5). That such a etween actual biopsy samples ] was suggested previously by kers, 2 although these authors i whole prostates. (We did not lle biopsy specimens in these tse the specimens were not alsecond, because radial orienindicated on any of the cores.) a gland rich in epithelium may %E in the outer 3 m m of a sinh a measure might be useful in to androgen-deprivation there epithelial tissue primarily) ,14 racial staining m e t h o d s , the er differentiated into separate fibrous compartments, the relhich may also vary considernt prostates. 4 Although subdimay not be practical using a :oxylin-eosin stain, as in the have demonstrated the utility tsing immunoenzymatic stainn-elate prostatic smooth musresponsiveness to alpha-blockmerry should also be useful in energic or androgen receptors

!994 / VOLUM~44, NUTviS3ESZ4

in the prostate, pending creation of specific staining methods. Another role of prostatic morphometry might be to quantify the epithelial component in relation to serum levels of PSA. Such a correlation would theoretically be of more value than a PSA density or index based on whole-gland volume, 15,16since PSA is the product of the epithelium, exclusively. A suggestion to that effect is present in the morphomettic data of Weber et al. 3 who showed that during hormonal therapy of BPH serum PSA levels more closely correlate with epithelial volume than with whole-gland volume. Very recently, Lepor and colteagues 17 reported that epithelial volume in the transition zone (determined by transrectal ultrasonography and directed biopsy) does exhibit a statistically significant correlation with serum PSA levels in men with BPH. We chose radical prostatectomy specimens for this study because they were readily available and because they were all sectioned and processed in a uniform fashion. Of course, these data are subject to confirmation by whole-organ studies of noncancerous prostates obtained via autopsy. The simulated biopsy results are subject to confirmation by studies of actual biopsy specimens. Results obtained with a compute>assisted rather than manual morphometric technique would also be of interest. Nevertheless, we are able to conclude the following from the present systematic analysis of BPH tissues: major differences in tissue composition may identify different hyperplastic prostates; BPH heterogeneity within individual prostates is not as pronounced as previously believed, when secondary changes are excluded; and quantitative histologic differences between glands, which are potentially important in directing BPH therapies and normalization of serum PSA levels, may be accurately diagnosed by m o r p h o m e t r y of radially oriented biopsy specimens. Leonard S. Marks, M.D. CuIver Medical Plaza Suite 701 3831 Hughes Avenue Culver City, California 90232

ACKNOWLEDGMENT."lo GlennJ. Gormley, M.D., Ph.D., whose early support and encouragement made this work possible. REFERENCES 1. Bartsch G, Muller HR, Oberholzer M, and R.ohr HP: Light microscopic stereologic analysis of the normal human prostate and of benign prostatic hyperplasia. J Urol 122: 487-491, 1979. 2. Shapiro E, Becich Mt, Hartanto V, and Lepor H: The relative proportion of stromal and epithelial hyperplasia is related

49 ]

to the development of symptomatic benign prostatic hyperplasia. J Urol 147: 1293-1297, 1992. 3. Weber JR OesterlingJE, Peters CA, Partin AW, Chan DW, and Walsh PC: The influence of reversible androgen deprivation on serum prostate specific antigen levels in men with benign prostatic hyperplasia. J Urol 141: 987-992, 1989. 4. Shapiro E, Hartanto V, and Lepor H: Quantifying the smooth muscle content of lhe prostate using double-immunoenzymatic staining and color assisted image analysis. J Urol 147: 1.167-1170, 1992. 5. Price H, McNealJE, and Stamey TA: Evolving patterns of tissue composition in benign prostatic hyperplasia as a function of specimen size. Hum Pathol 21: 578-585, 1990. 6. Shapiro E, Hartanto V, and Lepor H: The response to alpha blockade in benign prostatic hyperplasia is related to the percent area density of prostate smooth muscle. Prostate 21: 297-307, 1992. 7. Franks LM: Benign nodular hyperplasia of the prostate: a review. Ann R Coll Surg Engl 14: 92-106, ] 954. 8. Mostofi FK: Benign hyperplasia of the prostate gland, in Campbell MF, and HarrisonJH (Eds): Urology, 3rd ed, Philadelphia, WB Saunders, 1970, pp 1065-1129. 9. Walsh PC: Benign prostatic hyperplasia, in Walsh PC, Retik AB, Stamey TA, and Vaughan ED (Eds): Campbell's Urology, 6th ed, Philadelphia, WB Saunders, 1992, pp 1010-1015. 10. Fu YS, Cheng L, Huang I, Huang S, Wiesmeier E,

492

Wettstein E and ~Veissman M: D~ cal condyloma and intraepithelial tained by- punch biops~z Anal Quar 1989. 11. McNealj: Pathology of ben: Sight into etiology. Urol Clin Nort 12. Rohr HE and Bartsch G: F perplasia: a stromat disease? N e w morphology. Urology 16: 625-6313. Siegel YI, Zaidel L, Hamme A: Morphometric evaluation of b, Eur Urol 18: 71-73, 1990. 14. Huggins C, and Stevens RA benign hypertrophy of the prostate 1940. 15. Benson MC, Whang IS, Par Olsson CA, and Cooner WH: Pros a means of distinguishing benign prostate cancer. J Urol 147: 815-{ 16. Brawer MK, Aramburu EA( Ellis WJ: The inability of prostate hance the predictive value of prost', agnosis of prostatic carcinoma. J l 17. Lepor H, Wang B, and Shap prostatic epithelial volume and ser (PSA) levels. J Urol 151: 294A, lg~ ,.

UROLOGY ~ / OCrO~ER L99~ / VOLO'ME44,

M

{